Final Report Summary - LCAOS (A Nanoscale Artificial Nose to easily detect Volatile Biomarkers at Early stages of Lung Cancer and Related Genetic Mutations)
Executive Summary:
The LCAOS project has enabled the detection of lung cancer (LC) volatile biomarkers present in exhaled breath as well as in headspace of LC tissues/cells to screen and identify high risk groups for LC, and to monitor the therapy provided to people affected by LC. This goal was obtained by a strong multi-national and interdisciplinary consortium, composed of leading material scientists with additional proven ability to realize novel, prototype nanoscale devices, algorithmic groups, device developing and manufacturing SMEs and a company from the diagnostics field together with clinical institutions. Indeed, the LCAOS team have successfully developed, studied and implemented the following main action items: (i) developing molecularly-modified cross-selective Silicon nanowire Field Effect Transistors (Si NW FETs) and chemiresistive films of molecularly-modified gold nanoparticles in conjugation with pattern recognition methods - This new tool was termed Nano Artificial NOSE (NA-NOSE); (ii) testing the feasibility of the developed devices for sensing volatile biomarkers of LC from in-vitro tissues (including histologic/genetic sub-categories of LC) and from exhaled breath; (iii) developing an improved understanding of the signal transduction of the various volatile biomarkers, aided by sophisticated pattern recognition methods; and by (iv) developing performing clinical-related studies to assess LC conditions in actual patients and tissues and in the presence of real-world confounding signals.
The developed ‘NA-NOSE’ tool (15 ×15 × 5 cm; 0.8 kg; 5-10 min for analysis time) was examined both in lab-track and in clinical-track with more than 600 volunteers in several centres worldwide (Sheba Tal-Hashomer Hospital; Liverpool Lung Cancer Centre; Denver Colorado Lung Cancer Centre; Fox Chase Cancer Centre; Mayo Clinics; The First Affiliated Hospital of Anhui Medical University), with 30% of the samples being blinded. The blind analysis of the breath samples exhibited discrimination between early stages (I&II) and advances stages (III&IV) of lung cancer (81% average accuracy), between benign and malignant tumours (83% average accuracy), between SCLC and NSCLC (79% average accuracy), and between adenocarcinoma and squamous cell carcinoma (80% average accuracy). Many confounding factors (e.g. geographical location, age, gender, ethnics, smoking habit, etc.) were successfully neutralized using advanced hardware and software algorithms. In a series of in-vitro studies, the LCAOS partners have shown the feasibility to sense the secondary metabolic expressions for fine genetic alterations such as mutation in KRAS, P53, EGFR and EML4-ALK from the cells headspace (in a petri dish). Some of the alterations were associated with general carcinogenesis while the other part predicts response to targeted therapy.
The LCAOS achievements have indicated for the potential to reduce cancer mortality, by enabling widespread, trustworthy screening, especially suitable for high-risk populations. The ‘NA-NOSE’ was found to be suitable for use outside of specialist settings and would significantly reduce costs in health budgets. In addition, the ‘NA-NOSE’ was found to be suitable for immediate diagnosis of fresh LC tissues in operating rooms, where a dichotomic diagnosis is crucial to guide surgeons. The easy-to-use (no medical specialists required) ‘NA-NOSE’ technology detects cancer based on a change in the blood chemistry and/or metabolic activity (which is reflected in the chemical composition of the exhaled breath and cell/tissue headspace) rather than by tumor imaging, thus permitting earliest cancer detection i.e. before a tumor of detectable size has formed. The effectiveness of the ‘NA-NOSE’ in detecting LC volatile biomarkers specifically and selectively was experimentally and clinically proven to provide a launch pad for identifying other types of cancer from simple analysis of clinical samples, including breast, colon, and prostate cancers.
Project Context and Objectives:
Lung cancer is the most lethal cancer accounting for 28% of cancer deaths globally. In Europe there are 383,900 incident cases every year, resulting in 341,800 deaths. The average lifetime cost of lung cancer patients in Europe ranges between €46,000 and €61,000 per patient. Detecting lung cancer in its early stages, while still localized, can be expected to increase the 5-year-survival rate by three to four times. Unfortunately, diagnostic tests currently available, such as bronchoscope biopsy, pulmonary puncture and CT scan, occasionally miss tumours, which is costly. They are unsuitable for widespread screening because they are not efficient in terms of time, and are not free of complications. On the other hand, cancerous cells can be distinguished from non-cancerous cells on the basis of intracellular or extracellular biomarkers. Detection methods based on specific recognition of intracellular biomarkers (e.g. DNA/RNA/Proteins) require previous knowledge of specific mutations and copy number alterations in DNA/RNA or of changes in the regulation of protein expression inside the cells, assessed by expression and methylation profiling. Similarly, detection methods based on specific recognition of extracellular biomarkers such as histopathology, bioimaging, and antibody arrays require prior knowledge of biomarkers on cell surfaces. Antibody arrays can provide an effective but complex approach to cancer detection, diagnosis and prognosis. Nevertheless, there is no single biomarker or a combination of biomarkers with sufficient sensitivity and specificity to differentiate between normal, cancerous, and metastatic cell types.
An alternative approach to diagnosing cancer is emerging, which relies on volatile biomarkers emitted from cell membranes. The principle behind this approach is focused on the cell membrane, which consists primarily of amphipathic phospholipids, carbohydrates and many integral membrane proteins that are distinct for different cell types. Since tumor growth is accompanied by gene changes that may lead to oxidative stress, a peroxidation of the cell membrane species causes volatile biomarkers to be emitted. Some of these biomarkers appear in distinctively different mixture compositions, depending on whether a cell is healthy or cancerous. The remainder of the emitted biomarkers appear only in cancerous cells, and not in cells in a healthy state. Of particular significance to this approach is the fact that each type of cancer has its own unique pattern of volatile biomarkers, so that the presence of one cancer would not screen other cancer types. These volatile biomarkers can be detected either directly from the headspace of the cancer cells (i.e. the mixture of volatile biomarkers trapped above the cancer cells in a sealed vessel) or via the exhaled breath. The principle in the latter case is that cancer-related changes in blood chemistry are reflected in measurable changes to the breath through exchange via the lung.
The rationale for diagnosing cancer by means of volatile biomarkers has led to several important findings through spectrometry and/or spectroscopy-based techniques. To date, the use of these techniques has been impeded by the need for expensive equipment, high levels of expertise required to operate such instruments, the speed required for sampling and analysis, and the need for preconcentration techniques. Taking these considerations into account, the current proposal aims to harness nanotechnology, biomedical engineering, pulmonary medicine, medical oncology, and computation strategies in a multidisciplinary effort to develop and test a highly sensitive, inexpensive, and fast-response non-invasive Nanoscale Artificial Nose (NA-NOSE) for the detection of pre-neoplastic volatile biomarkers that indicate an increased risk of lung cancer (e.g. bronchial gene methylation, sputum atypia) and the presence of lung cancer. Further to these analyses, the proposed project investigates the bronchial epithelial histology (tissue) for its dysplastic appearance, early gene methylation changes, and genome-wide sequencing. In-vitro studies with lung cancer cells and tissues are performed, focusing on the differences between subtypes of lung cancer, in controlled conditions. The NA-NOSE’s ability to detect gene abnormalities in the lung cancer cells also be assessed.
The major objectives of the LCAOS project were divided between “strategic objectives”, “scientific objectives”, “technological objectives” and “policy objectives”.
STRATEGIC OBJECTIVES:
LCAOS’s extensive collaborative effort aims to make a significant contribution to the goals of the Strategic Research Agendas (SRA) of the relevant European Technology Platforms: both for Nanomedicine (ETPN) and for Smart System Integrations (EPoSS). The main strategic objectives of this project included (but are not confined to):
• Performing ambitious, but relatively short-term research within the Consortium of eight world-class R&D groups, well beyond conventional approaches, to realize a highly novel ‘NA-NOSE’ that has the potential for highly sensitive detection of cancer biomarkers in a simple, inexpensive and non-invasive manner.
• Establishing interaction and communication between the different disciplines involved in LCAOS, and between academia and industry. By sharing expertise and knowledge (including via an exchange of personnel) between academia, clinical institutions, SMEs, and diagnostics companies, the newly integrated ‘NA-NOSE’ was produced for clinical-track applications by the end of the project. The industrial participants, in addition to participating in the research, play a role in shaping the academic research through their feedback to the approaches developed. The academic partners provide access to the industrial partners to test clinical samples and advanced characterization methods.
• Facilitating the swift transfer of clinical and academic research results, as well as hardware and software directly to SMEs/companies specialized in commercializing devices and methods. SMEs is a crucial role in achieving this objective, which also includes validation and benchmarking of the novel ‘NA-NOSE’ at clinical sites, and performing comparisons with other diagnostic and screening tools.
• Given the multi-disciplinary nature of the LCAOS Consortium, the partners are able to provide multi-disciplinary training to researchers and students. The partners exploit this opportunity by exchanging personnel for extended periods of time, as well as organizing workshops and 2 mini-conferences (summer-school) throughout the four-year effort.
• To make most effective use of resources, a lab-track and a clinical-track run in parallel. In the lab-track, a prototype ‘NA-NOSE’ is developed and tested. In the clinical-track a more professionalized version of the ‘NA-NOSE’ is tested in clinical environments to carry out screening of high-risk population, monitoring of therapy progress and use of the device to direct surgery.
SCIENTIFIC OBJECTIVES
To support the strategic objectives and to enable the novel ‘NA-NOSE’ to identify and detect volatile biomarkers in real confounding environments, the partners have determined the following scientific objectives:
• To develop arrays of differently functionalized Si NW FETs that form a novel ‘NA-NOSE’ able to detect and identify volatile biomarkers specific in their mixture for LC as well as other types of cancer and even sub-types of LC (.
• To design and apply reliable pattern recognition models able to detect, with the minimum number of failures possible, the presence of cancerous volatile biomarkers in real samples, which exhibit a high chemical complexity.
• To test the feasibility of the developed ‘NA-NOSE’ to sense volatile biomarkers of LC from in-vitro tissues (including histologic/genetic sub-categories of LC) and from human exhaled breath.
• To test the feasibility of the developed ‘NA-NOSE’ to sense precancerous conditions, invasive LC (including for its subcategories) and cross-correlation between the volatile biomarker signatures and other known biomarkers (e.g. EGFR mutation).
• To test the feasibility of the ‘NA-NOSE’ to monitor the treatment/therapy success of LC.
• To reach a thorough understanding of the signal transduction mechanism of the various volatile biomarkers and metabolites, aided by sophisticated statistical and pattern recognition methods.
• To develop improved systems that will enable the ‘NA-NOSE’ to clearly distinguish the targeted biomarkers from environmental clutter, using methylation, expression profiling, and genome-wide sequencing in order to correlate LC’s metabolite signature with genetic aberrations in the related pathways that will improve treatment strategy.
• To perform clinical studies to assess LC conditions in actual patients and actual tissues. One of the aims in the initial clinical studies will be to explore the effectiveness of the important sensing signals in the presence of real-world confounding signals.
• To reach a thorough understanding of the molecular biology and pathways of the multistep process of lung carcinogenesis by combining the molecular transcriptomic and genomic data obtained in biopsies at all stages of lung squamous carcinogenesis, and with isolated organic compounds detected in exhaled breath or headspace of in-vitro or tissue samples.
• To further sub-categorize the profiles of volatile biomarkers for each of the sub-types of LC, where each profile will express a different level of tumor activity and aberrancy. The data emerging from this project will provide a very powerful metabolite, clinical and genetic dataset, which will require high-level bioinformatic analytical techniques. The medical-related studies will permit us to obtain the necessary feedback required to bioengineer a viable technology for the detection and identification of LC volatile biomarkers.
TECHNOLOGICAL OBJECTIVES
The realization of a highly-efficient and cost-effective ‘NA-NOSE’ for the detection and identification of volatile biomarkers related to (sub-types of) LC from breath samples or headspace of LC cells/tissues is carried out at two industry/SME sites, aiding with the expertise ofone of the SME partners in the field of medical diagnostics. These industrial partners take the ‘NA-NOSE’ prototype design to a higher level of miniaturization and will refine and automate features such as data acquisition and data interpretation and software integration, making the hand-held device simple-to-use, even be used during surgery. To fulfill these highly ambitious goals, the following technological objectives were set, to be achieved chiefly within the tasks performed by the industrial partners:
• Integration of an array of cross-reactive molecule-terminated Si NW FETs and molecule-terminated nanoparticle (MNPs) chemiresistive films in a ‘NA-NOSE’, with proven ability to detect and identify volatile biomarkers of LC, into a VLSI (~20mm×15mm) chip, preferably using VLSI fabrication techniques currently available.
• Design and prototype realization of a microfluidics flow-cell, able to be used as a host for the Si NW FET- and MNP-based (VLSI) chip, and as an exposure cell for breath and tissue/cells headspace samples. In this context, detailed objectives include the design and construction of: (i) a microfluidic cover to serve as an interface between the environments and directs the samples onto the Si NWs array; (ii) a flow-cell package to interface the flow cell with the ’NA-NOSE’; and (iii) a customized holder to provide both electrical and fluidic connections by means of press-fit connections.
• Design and construction of a prototype ‘NA-NOSE’ integrating the Si NW FET- and MNP-based (VLSI) chip (VLSI) chip and microfluidics system. This prototype is aimed to be portable and stand-alone for its use in clinical studies. The system is equipped with a microcontroller for fluidics control and data acquisition. Either an external computer is connected via USB, or an embedded computer will be used for data analysis, storage and user interface. This allows the same software platform (MultiSens) to be used during algorithm development and testing. MultiSens will be extended with an easy-to-use interface, to allow typical personnel to operate the system efficiently in a clinical environment.
• Integration of the prototypes into the MultiSens software packages in such a way that allow the other groups to embed and test their algorithms with the prototypes efficiently. The structure of MultiSens allows different algorithms (feature extraction, filtering, classification, and drift and cross interference compensation) to be tested and combined during the laboratory and clinical phases.
• Development of a system for testing and training the ‘NA-NOSE’ using artificial breath, which allows characterizing of the system in a defined and reproducible way. This has a great impact on progress of sensor development, since the measurement of the sensitivity (and cross-sensitivity) and selectivity of the sensor is readily available.
POLICY OBJECTIVES
Today, European research groups and companies/SMEs compete well in the field of (nano)biomedical devices to detect diseases including cancer biomarkers. European knowledge and expertise has led many technological developments for non-invasive prediction, diagnosis, monitoring and prognosis tools. For European leadership to be maintained, substantial R&D efforts with adequate funding, such as the LCAOS proposal, will be required. LCAOS aims to enable the European Technology Platforms for Nanomedicine (ETPN) and Smart System Integrations (EPoSS) to advance the novel concepts for early detection and identification of cancer biomarkers from breath samples and tissue/cells headspace, and with the help of the strong, multi-disciplinary consortium, to take a leadership role in establishing the next generation of easy-to-use, highly-sensitive screening, diagnosis, monitoring, and prognosis tools. LCAOS planning includes making the novel device attractive to biomedical markets. European device manufacturers shall get benefit from participation in LCAOS to launch highly innovative ‘NA-NOSE’ products. For these equipment manufacturers, LCAOS provides an early opportunity to commercialise devices with high potential, in the fast-growing area of non-invasive, real-time, and inexpensive biomedical tools. The participation of a leading SME in the diagnostics field will guarantee the benchmarking of the novel ‘NA-NOSE’ technology and advance deployment.
Project Results:
1. Fabrication of Si NWs
The fabrication of Si NWs with controlled diameter, length, and electronic properties is essential to the NA-NOSE that can be used for cancer detection. In the LCAOS project, the MPL partner presented a few pathways to grow and to improve Si NWs based sensing application performance and reliability though modified NWs chemical composition and dopant profiles. The fabrication of such NWs requires a state of the art growth conditions with superior control over growth parameters such as precursors and carrier gases type flow rates and partial pressures, growth temperature and so on. A true breakthrough over the LCAOS project was achieved mainly by setting up a new plasma enhanced chemical vapor deposition (PECVD) machine that enables the above mentioned requirements and controlling the overall growth parameters. In CVD, the growth of the Si NWs follows the VLS principle, where the temperature during the CVD process is chosen to be above the eutectic temperature and below the melting temperatures of gold and silicon. Electronic properties of the NWs can be conveniently tuned by adding doping gases during the Si NW growth. For example, we grow p-type Si NW by introducing B2H6 with silan gas (4 sccm SiH4 /0.1 sccm B2H6), while with similar conditions, n-type was grown by using PH3 as a carrier gas for the phosphine atoms (4 sccm SiH4 /0.1 sccm PH3). However, Si NWs were realized also by RIE technique. Here the doping structure of the SiNW can be controlled by the composition of the used substrate materials. Many different structures could be realized, including the fabrication of diodes at the scale of individual nanowires. The electronic properties of the SiNW have an important addtional role: tuning the sub-threshold senstivity of the FETs based NA-NOSE. Anticipated success on increased sensitivity limit for cancer detection was accomplished though tuning the sub-threshold window of the Si NWs and achieving high on/off ratio of the intrinsic Si NWs. The structure of the Si NWs is an important feature for the NA-NOSE since it determines the resolution of the signal to noise ratio for low concentration detection. Therefore, the structure of the Si NWs after VLS process has been analyzed intensively by means of high resolution transmission electron microscopy (TEM) and electron backscattered diffraction (EBSD). The results show a bi-crystalline structure, with presence of axial twins crossing the entire length of the NW from bottom to tip. These twin defects can be visualized when observing the NW across the [1-10] zone axis, and correspond to a 180º rotation along the (11-1) plane, which is perpendicular to the [112] NW growth axis. The sensitivity of the FET based Si NWs as a channel of the NA-NOSE, was increased by replacing the native oxide of the Si NW with variety of molecules (organic layer) to the surface of each Si NW. PN junctions were fabricated to explore the surface quality and conductivity of the Si NW. Finally, a high sensitivity at the ppm level was achieved after integrating the Si NW into the NA-NOSE device.
2. Fabrication of Si NW FET sensors
The goal of the LCAOS project is to develop a diagnostic tool based on SiNW array FET sensors for non-invasive diagnosis of lung cancer. In order to achieve this aim the Technion optimized the fabrication method for the devices, including the development of a new process for the deposition of the SiNWs on the substrate, called spray coating. The spray coating process is based on a simple principle of equilibrium between the shear force and the surface tension of the droplet containing the SiNWs when it comes in contact with the substrate. This process allowed us to control the number and orientation of the SINWs on the device, thus eliminating variances between the devices due to differences in the NW orientation. In addition, the density of the NWs on the surface was optimized for increased sensitivity of the device, as well as for optimal electrical properties. The next step was the investigation of the various electrical properties (features) extracted from the characteristic Ids-Vgs curve, such as the threshold voltage (Vth) and the charge carrier (hole) mobility (µh), in addition to the value of Ids at various Vgs. Many of these features are independent, resulting from different charge transfer mechanisms, and can therefore be considered as independent virtual sensors. The following step was determining which features are more relevant for VOC sensing, and which region of the Ids-Vgs curve has the greater importance for this specific sensing application (the region which contains the majority of VOC sensitive features). We found that Vth, µh, and the currents at the sub-threshold and linear region were the most important, and hold the highest potential for the differentiation between breath samples.
In addition, the Technion partner investigated the surface functionalization of the devices, and molecularly modified the surface of the Si NWs with a variety of organic molecules. These molecules were Silane based molecules, which varied either in chain length, or in functional group. We tested molecules with varying polarity and bulkiness, as well as several modification methods. The success of the functionalization of the sensors was tested via X-ray photoelectron spectroscopy (XPS). The changes in the molecular functionalization affected the affinity of the different sensors to the various VOCs, and we wanted to test which is the most appropriate surface modifications for the sensing of Lung cancer related VOCs. For that purpose, we exposed the sensors to a mixture of VOCs, as a way to simulate the human breath. These VOCs included compounds which were previously found to relate to Lung cancer.
After testing the sensors' ability to respond to simulations, the Technion partner exposed them to real breath samples of Lung cancer patients who had both early and late Lung cancer, and also to control subjects. The control group is a heterogeneous group that included completely healthy people, as well as people who had non-cancerous Lung conditions. These samples were sent to the Technion partner from both TAU and LIV partners. In the same time, the Technion also worked on adjusting the devices for the microfluidic chamber developed in MICRONIT in order to integrate these devices in the stand alone NA-NOSE. The Technion had to change the contact pads so they could be mounted on the chip more easily and so the flow of the sample had the best orientation with the Si NWs for sensing. In addition, we tested the sensing ability of these devices in the MICRONIT chip, to verify that it was not affected by the small sample volume or the change in the fabrication parameters. Finally, the Technion tested the complete stand-alone device and made sure it is ready to be used for research in clinics (see more details below).
3. Hosting Si NW FETs in Microfluidic Exposure Cells
Microfluidic exposure cells (size 22 x 30 mm2) were developed by Micronit to bring gaseous samples (i.e. exhaled breath) into contact with arrays of eight sensor chips. The sensor chips provided by Technion were attached to the microfluidic chip made from glass, forming both a fluidic and electrical interface between the sensors and the instrument. The internal volume of the fully assembled exposure cell is as low as a couple of microliters with very little internal dead volume thereby offering very quick exchange of the gaseous samples. Furthermore, a holder was designed with the capacity for connecting two exposure cells simultaneously which can be replaced very easily when required. A combined electrical and fluidic interface was developed, providing an easy and reliable means for swapping exposure cells when integrated with the stand-alone instrument.
4. Assembly of Stand-alone NA-NOSE
The JLM partner has designed and developed the NA-NOSE electronics and the related full demonstration instrument. Initial development stage consisted on design of interfacing and readout electronics for preliminary laboratory tests. Two different electronic interfaces were designed for interfacing Si NW FET and GNP sensors. A final version of the electronics for the demonstrator was designed based on the preliminary designs, including both interfaces in a single design, with optimized performance. The following stage in the project involved the design of the breath sampling subsystem and the sample transport subsystem. The demonstrator prototype featured the first version of these subsystems alongside with the electronics. The clinical versions of the instruments improved some aspects of the breath sampling subsystem, in particular the interface to the glass breath buffering tube. In the design process, 3D printed parts, fluidics, sample transport. The full instrument featured a novel integrated breath sampling subsystem alongside with the electronics, including a custom designed flow sensor to identify the phases of patient breathing. A sampling protocol was designed in coordination with TAU and Technion partners. A development grade software tool (research interface) was developed using Labview 2012®, which allowed limited co-development with Technion (who also holds a Labview 2012® license). The clinical grade software interface was developed using Delphi® and is optimized for touch-screen operation. High accuracy, fast readout electronics was demonstrated to match the accuracy of larger, expensive desktop instrumentation. Besides, the extremely compact construction of the breath-analysis demonstrator (NA-NOSE) enabled the development and study of point-of-care applications. In particular the whole design and construction of the innovative breath sampling subsystem has been a major outcome of the project.
5. Pattern Recognition Methods for the NA-NOSE
The UCM was in charge of the mathematical analysis and modeling of the results provided by diverse analytical equipment. The goal of this task was the design of models (software, in the end) to solve numerous problems such as: (1) to locate and assign specific volatile organic compounds to determined patterns to assess the viability of the approach; (2) to find disease-specific compounds (biomarkers) in breath that may aid in diagnosis; and (3) to classify and/or identify LC patients as well as staging them. To carry out these tasks, the UCM employed two main algorithmic blocks: feature selection (FS) and artificial neural networks (ANNs). Many of the databases that were provided contained great amounts of variables, reason why FS was crucial. On the other hand, ANNs were employed to achieve estimative and classificatory models, as they are a reliable set of non-linear mathematical tools that can excel in numerous applications. The combination of these kinds of algorithms gave the NA-NOSE better results than other classic models used in this field in the past. The UCM group has proven that the information gathered from breath analysis is clearly linked to the patient’s status, and can be used to reach early diagnosing tools.
Using the UCM algorithms, the LCAOS partners have been able to set a proof-of-concept that implies the viability of LC diagnosis through breath analysis, as specific disease-related patterns have been found in the databases offered by the analytical equipment. One of the main findings was that different volatile organic compounds (VOCs), comparable to potential LC biomarkers present in breath, provide specific patterns after interacting with cross-reactive sensor arrays. Furthermore, these patterns are also quantity dependent enabling not only identifying VOCs, but quantifying them. Additionally, mixtures of VOCs also were successfully identified employing ANNs. The models designed and optimized consist of supervised algorithms based on multilayer perceptrons. To achieve the best results, the parameters of the ANNs were optimized (some heuristically and others through experimental designs). In these models the input information was attained from the signals from the analytical equipment and the output provides information regarding the type and amount of VOCs. These types of algorithms are able to recognize biomarker profiles which characterize LC and their performance improve and are more reliable with bigger sets of data.
The UCM work has led to the development of a novel and complex software package that employs ANNs to locate the most important features or independent variables for a further estimation or classification. This software has been successfully employed in this project, and will clearly be taken advantage of in the future to continue fighting against LC and other types of cancers and illnesses.
To sum up this part of the work, one of the most important results of LCAOS is that the intelligent algorithms developed in this project guarantee that there are mathematical relations between biomarkers present in the human breath and LC. This relation is very important not only to reach early diagnosis of LC but also to deeply research about chemicals which are required to fight against it. In addition, this knowledge that has been achieved is the starting point for the research of other illnesses. UCM has enabled the last step of the LCAOS project: interpreting and putting to use the databases from the different analytical equipment to achieve early LC diagnosis through breath. The reliability, calculating power, and efficiency of ANNs have shown to be useful traits in the analysis of the intricate databases that are originated from breath analysis.
6. NA-NOSE and Volatolomics of Lung Cancer
Volatolomic analysis by the developed NA-NOSE as well as with complementary mass spectrometry techniques was first carried out by in vitro analysis of lung cell lines with defined mutations, whereby discrimination of individual mutations was possible. Analysis of the NA-NOSE responses allowed significant discrimination between LC and immortal bronchial epithelium (IBE) cells with 96% sensitivity, 86% specificity and 93% accuracy, as well as between the two major LC subtypes: small cell lung cancer (SCLC) and non-SCLC (NSCLC) with 96% accuracy, 100% sensitivity and 75% specificity in leave-one-out cross-validation. The volatolomic headspace of the two subtypes of NSCLC: adenocarcinoma and squamous cell carcinoma yielded 86% sensitivity, 100% specificity and 90% accuracy. SVM analysis of the compounds detected by GC-MS found 5 VOCs to show significant difference between IBE and LC cell lines and between the different histologies in different compositions with high accuracies as well. Histologic classification of lung cancer tumors has a major impact on the therapeutic approach (e.g. surgery or chemotherapy). However, it is reasonable to expect that the clusters of a wider study population would be less defined, and some overlap could occur, thus an extended study would be necessary to confirm the criteria for distinguishing between LC sub-histologies based on the findings presented in this study.
Due to the growing interest in lung cancer target therapy and the need for targeting specific genetic mutation related to the carcinogenesis process, the LCAOS partners (mainly, LIV, TAU, and Technion) looked for specific volatolomic signatures of LC genetic mutations through headspace analysis. Results of this pilot study provided first evidence for the existence of measurable VOC profiles for the most important genetic mutations associated with LC targeted therapy (EGFRmut, KRASmut and EML4-ALK fusion). A group of cell lines representing oncogenes that were wild type (wt) to the tested three mutation of interest served as control group. The three studied oncogenes are usually mutually exclusive and, hence, are considered distinct genetic subtypes of LC. Statistical analysis of the collective signals from the NA-NOSE yielded a comprehensive and highly accurate volatolomic fingerprint assay based on a set of three primary tests and an additional set of six supplementary tests that could correctly identify the oncogenes of the studied cell lines in a blind test with 76% accuracy in global leave-one-out cross-validation of the 9-test assay. Complementary analysis with GC-MS identified five VOCs (triethylamine, toluene, styrene, benzaldehyde and decanal) which could be associated with each of the genetic mutations of interest. While the concept of this study was obtained through in vitro studies, these results gave indications to assume that similar volatolomic pattern could be detected directly from the breath.
A more tractable system, than LC cell lines, to study the effect of specific cancer driver mutations on volatile release was provided by human bronchial epithelial cells (HBEC) that have been genetically manipulated. HBEC cells, immortalized through expression of telomerase and activation of CDK4, are genetically stable compared to cancer cell-lines, with minimal aneuploidy. This provided a parental cell-line in which the effect of additional genetic lesions can be investigated. In the derived cell-line HBEC-3KT53 TP53 has been knocked down by RNAi knockdown, whereas HBEC-3KTR cell-line expresses mutant KRASG12V. In a fourth cell-line, HBEC-3KTR53, both genetic alterations were induced by simultaneous transfection. In this part of the research, the developed NA-NOSE, together with Discriminant Factor Analysis (DFA), demonstrated the ability to discriminate metabolic differences due to individual mutations in lung cancer cells based on volatolomic signature. The classification accuracy of the DFA models obtained for any of the binary comparison were in the range of 77% to 93% in leave-one-out cross-validation and in the range of 64% to 89% in leave-third-out. Sensitivities and accuracies calculated by receiver operating characteristic (ROC) analysis were very high. Wilcoxon P values of the canonical variables (CVs) were below 0.04. Complementary chemical analysis by GC-MS has reported 20 VOCs from the families of alkanes, methylated alkanes and alkenes, benzene derivatives, ketones, aldehydes and alcohols to show significant discrimination between different genetic mutations. The three genetically altered HBEC cell lines and the vector only control (HBEC-3KT) allowed us to study the effect of individual mutations on the VOC signature of lung cancer precursors and potentially model the earliest stages of lung tumorigenesis, since these lesions may be present in those at risk of subsequently developing lung cancer. TP53 and KRAS mutations are common mutations in LC, sometimes occur together, and have been found early in the progression of lung tumorigenesis, thus there is a superior need for early, non-invasive and accurate identification of these LC genetic mutations.
To sum up, in-vitro experiments utilizing a model of early lung cancer progression have provided evidence that volatolomic profiles can distinguish between cells with limited, defined mutations. This supports both the potential ability of breath analysis to detect very early genetic changes (e.g. which may improve risk-based early detection) and very specific genetic changes (e.g. for characterization of patients for targeted molecular therapies).
7. NA-NOSE and Breath Analysis of Lung Cancer
In order to leverage the benefits of breath analysis in the characterisation of lung cancer, the LCAOS project has undertaken molecular (expression and methylation) and genetic (mutation) analysis of lung cancers and other clinical samples likely to form the basis of early detection or diagnosis. This analysis includes a consideration of the relationship between tumour mutation and lung cancer risk. We were able to demonstrate that molecular alterations in lung cancers were not closely associated with the epidemiological risk score, which offers hope that breath analysis would detect lung cancer independent of pre-conceived lung cancer risk – potentially opening up screening to those low risk individuals currently excluded. Furthermore molecular alterations were broadly similar in early cancers detected in asymptomatic individuals by CT screening, supporting a molecular basis for implementation of breath testing in the screening setting.
Given the opportunity to target a well-characterized, high-risk population, recruitment of breath samples by LIV partners has focused on controls from the Liverpool Lung Project population cohort. A total to 168 subjects (each providing replicate samples for parallel analysis by GC-MS and e-nose) was recruited, including pre-diagnosed lung cancer and a number of patients that were diagnosed with lung cancer subsequent to breath sampling. Whilst the diagnostic potential of breath analysis in symptomatic, high risk individuals is self-evident, in order to fully utilize breath analysis in an early detection or screening setting it is important to understand the relationship between breath analysis and other factors known to influence lung cancer risk. Currently epidemiology-based risk markers are the best way of identifying individuals that most benefit from a CT screening approach, which is an important consideration for the cost-benefit analysis of early lung cancer detection programs. In order to address the potential interaction of breath volatolomic signatures with these risk factors, we have included within the LCAOS recruitment a cohort of subjects form the Liverpool Lung Project (LLP), for whom lung cancer risk factors and potential confounding factors have been extensively documented. Breath samples from LLP participants collected by ULIV partners have been analysed at Technion for VOC content (by GC-MS) and by 2 alternative NA-NOSEs, with these large datasets jointly analyzed by UCM and LIV. It is of note, that breath analysis apparently reflects historical smoking history independent of recent smoking activity (which is a common confounding factor in VOC analysis); this may in part be due to a volatolomic signature of smoking-induced lung damage, which has been proposed as a more important risk factor than smoke exposure alone. Together with breath analysis of cases at ULIV and case/control cohorts across multiple sites, the LLP cohort samples also contribute to studies of breath-based lung cancer diagnosis; in particular several samples were collected from asymptomatic subjects who were subsequently diagnosed with lung cancer, providing a basis for investigation of early detection.
For each breath test subject we have collated epidemiological and clinical data with follow-up and stored this information on a secure database. For analysis two sets of data were extracted. The first consisted of a detailed epidemiological (subject) data and technical (sample) data in order to perform in depth analysis of risk factor and confounder variables in relation to VOC patterns. The second dataset combined LIV data with subject and sample data across multiple recruitment sites. In addition to breath samples, additional clinical samples were collected, including blood, plasma, sputum, buccal scrapes and tumour tissue.
DNA sequencing studies were designed to address the relationship between molecular alterations, lung cancer risk and volatolomic signatures arising from metabolic perturbations. An initial cohort representing tumours from patients with very low or very high levels of epidemiological risk was prepared for genome sequencing; additional samples were prepared from lung cancer cases recruited for breath samples and from lung cancers detected in asymptomatic subjects by CT screening. Genome and exome-wide sequencing was successfully performed on all selected tumour samples. Dysplasia samples were not sequenced, as none were found in breath test individuals. Additional lung cancer DNA sequence data was downloaded from public repositories to complement that prepared in house. DNA sequence data was produced from lung cancer cell-line DNA prepared by RMC/TAU partners). This was analysed alongside more extensive cell-line sequence, expression and methylation data available from public repositories. Expression data (in the form of microRNA microarray analysis) and DNA methylation analysis were performed on samples. In addition, other sample types were used as part of studies to: demonstrate a novel risk stratification technique in buccal samples; investigate the use of plasma microRNA for early detection of lung cancer; explore the promising area of plasma cell-free DNA biomarkers (so-called liquid biopsies), for both DNA methylation and DNA mutation. The sequencing data for the LCAOS project thus includes whole genome sequencing of carefully selected lung tumours and exome sequencing for both lung tumors and specific lung cancer cell lines (for which additional VOC analysis was performed in WP6). The data was analyzed in the context of published genomics datasets such as are available from The Cancer Genome Atlas (TCGA), International Cancer Genome Consortium (ICGC) and the Cancer Cell Line Encyclopaedia (CCLE) and aggregative datasets such as the Catalogue of Somatic Mutations in Cancer (COSMIC). Of particular relevance to LCAOS are genomic markers of lung tumour metabolic context since the lung tumour is likely to be a major contributor to the altered breath volatile signature in a cancer patient (albeit alongside the influence of tumour burden on systemic metabolism). Within the external datasets, we have addressed whether specific metabolic pathways are altered at the transcriptional, epigenetic or sequence level in lung cell lines, in tumour subclasses (e.g. adenocarcinoma compared with squamous cell carcinoma), or between tumour and normal lung tissue.
The relationship between epidemiology-based Risk Scores and lung cancer mutations is poorly understood. Although more cancers arise in those people with high epidemiology-based Risk Scores, some people with low Risk Scores also get cancer. At the molecular level lung cancer is not a single disease and it is important to consider both how the molecular genetic diversity of lung cancer might be influenced by the risk factors that are likely to form the basis of early detection and how this molecular genetic diversity might be reflected in the breath volatolomic signatures. To this end we have undertaken a range of molecular studies on lung cancers themselves and on other patient samples that might, together with breath, form the basis of lung cancer detection or characterisation. DNA mutation studies of lung tumours and matched normal DNA have been performed on LLP cases, for which the full range of risk factors, have been documented, with exemplar cases selected from both ends of the lung cancer risk spectrum. In parallel, a larger set of DNA sequencing data from publically available databases (e.g. The Cancer Genome Atlas, TCGA) was also analyzed in relation to their (more limited) risk marker profiles. These large DNA mutation datasets have been augmented with data from LCAOS cases (for which breath VOC profiles are available) and a panel of asymptomatic CT-screen detected cancers (representing the earliest stages of lung cancer). In addition gene expression and methylation studies have been performed, focusing on the cellular pathways implicated in cellular metabolism that contributes to VOC production. Importantly, we were able to demonstrate that tumours arising in patients considered being either high risk (targeted by screening programmes) or low risk harbour the same molecular alterations. This indicates that, should breath analysis be able to distinguish cancers on the basis of such molecular alterations, it would be equally sensitive for those low risk patients currently excluded from screening on the basis of their epidemiological risk profile.
Complementary to the efforts to identify high-risk people for LC, the TAU partner was responsible for executing the translational studies with the developed NA-NOSE device. In the first series of studies, the focus was on the diagnosis and classification of lung cancer. TAU collected exhaled breath samples from lung cancer patients and their matched control in order to evaluate the potential of exhaled breath analysis in lung cancer detection and distinction between early and advanced disease stages. Following the breath collection, together with the related clinical samples, the Technion team has analyzed the exhale breath of a total of 174 patients participated in this study, and Inter-group analysis of 80 lung cancer patients (64 advanced stage) and 31 matched controls showed a significant discrimination between disease and control. The results indicated that breath analysis discriminated early-stage lung cancer from matched controls. Moreover, it discriminated early stage patients from advanced disease patients. On top of that, the LCAOS partners compared breath samples from 72 patients with benign and malignant pulmonary nodules (PNs). For the specific identification of VOCs a complementary chemical analysis using GC-MS was performed. All patients underwent vast clinical evaluation including bronchoscopy, wedge resection and/or lobectomy for final diagnosis. LC patients were histologically classified into non-small cell LC (NSCLC) and small cell lung cancer (SCLC) and into early stage and advance stage of the disease. Nodule size of all malignancies was determined. Sensitivity, specificity and accuracy of the binary comparisons varied between 78% and 92% indicating that breath test might have future potential for the management of patients with PNs. More specifically, the results obtained in this study indicate that breath testing could be used to discriminate between benign and malignant PNs detected by CT, reduce the false-positive rate, minimize the risk of morbidity related to invasive diagnostic procedures and reduce the costs of LC screening. This innovative method may pose as an important non-invasive tool for lung cancer early detection, thus promoting better prognosis and therapeutic possibilities for patients.
In the second series of studies, the TAU team examined the use of breath sampling in monitoring response to anti-cancerous treatment of advanced lung cancer patients. Repeated exhaled breath samples were collected from patients with advanced lung cancer before and under systemic therapy. VOCs profiles were determined by GC-MS and nanomaterial-based array of sensor and correlated with response to therapy, assessed by CT scans as Complete Response (CR), Partial Response (PR), Stable Disease (SD), or Progressive Disease (PD). One hundred forty three breath samples were collected from 39 patients with stage III/IV lung cancer. GC-MS anaylsis identified 3 VOCs as significantly indicating PR/SD samples. One of them was also significantly discriminated between PR/SD and PD. Further, the NA-NOSE signals were able to alarm per a change in tumor response across therapy, i.e. indicating lack of further response to therapy, or developement of resistance to therapy. Therefore, breath analysis, using the NA-NOSE technology, may serve as a serogate marker for response to systemic therapy in lung cancer. Such a monitoring tool can provide the physicians with a quick and simple way to identify patient's response to anti-cancerous treatment in shorter intervals than currently available by CT scans, and so to allow a better decision-making regarding the chosen treatment. In a complementary phase, the Technion partner examined the exhaled VOCs of LC-induced long-term changes in pre-surgery and post-surgery patients. Breath samples were analyzed using two complimentary approaches. The first approach was chemical analysis of breath VOCs that aims to detect significant compounds that show statistically difference between pre-surgery and post-surgery patients using SPME/GC-MS. The second approach is based on constructing a volatolomic signature of breath VOCs by the NA-NOSE. Five VOCs were tentatively identified by GC-MS to have a significant difference between pre- and post-surgery LC patients with P value<0.013. These compounds are from the families of methylated alkenes, ketones, aromatic compounds and methylated alkanes. All five compounds were significantly reduced in post-surgery LC patients compared to pre-surgery patients indicating that their origin stem directly from the affected lung area that was resected. However, this identification is tentative and requires calibration standards for verification and a clinical study with a considerably larger number of patients is required to confirm these results. Leave-one-out cross validation of DFA analysis yielded high sensitivity and specificity of 100% and 80% respectively for the comparison of LC patients (pre-surgery) versus benign stated. For the comparison of LC patients (post-surgery) versus benign and for the comparison of benign before and after surgery the calculated sensitivity, specificity and accuracy were much lower. For the discrimination of LC before and after surgery accuracy, sensitivity and specificity varied between 75%-83%. These results imply that the volatolomic patterns may stem from the responses of the sensors to the local tumor markers, thus classification success reduced significantly once the tumor has been resected. The somewhat better separation of LC versus benign (pre-surgery) compared to LC before and after surgery supports the assumption that VOC levels would remain stable shortly after surgery. The appropriate choose of nanoparticles and recognition layer may identify LC-tumor related breath patterns, during short-term post-surgery follow-up. In the future, breath test might be utilized to complement the current diagnostic methods and to increase sensitivity of LC prediction prior to surgery, making swifter treatment decision possible.
8. Understanding Biochemical Pathways of LC Volatolomics
The IONICON partner supporting the clinical studies by providing the hardware, training and know-how for real-time breath analysis. IONICON has supplied a real-time trace gas analyser (PTR-MS) with a special inlet system for breath analysis (BET sampler) to project partner TAU. After the training of the researchers on the general use of the analyser and operating procedures for breath analysis, several clinical studies on human breath as well as cancer cell lines have been performed at TAU.
With the help of IONICON, TAU assessed the VOCs profile derived from lung cancer cell lines as well as the differences in glucose metabolism on the volatile signature in lung cancer patients through an oral glucose tolerance test (OGTT) using the PTRMS platform. This cohort included forty participants (22 control participants whom are at high risk for lung cancer, 18 study participants whom have active, naïve lung cancer). Pre-OGTT and Post-OGTT blood glucose levels and exhaled breath samples were measured with a lay period of 90 minutes. We revealed 14 masses which enable us to distinguish between the two groups. We concluded that breath analysis could discriminate between the high-risk and LC group. Furthermore, it demonstrated that glucose metabolism leaves a unique VOC pattern in the LC group. These findings may assist in the development of a non-invasive screening method for lung cancer. From the in vitro studies using the PTR-MS, the LCAOS partners identified and quantified masses/ charge (m/e’s) which are exhaled from cell’s headspace (ppt/ppb levels). These finding revealed differences in the volatile signature between LC subtypes which have a different mitochondrial capacity, and particularly, a volatile signature attributed to glycolysis. Additional differences in the breath of cancer patients can be revealed using an oral glucose tolerance tests, exploiting the Warburg effect, a metabolic anomaly of cancer cells. In addition, the in vitro studies in which we profile the VOCs pattern of A549, H2030, H358 and H322 cell lines using the PTR-MS detected numerous masses per charge (m/z’s) which their headspace concentration changed significantly after glycolysis inhibition. The concentration of the following masses m/z’s: 47, 60, 83, 101, 111, 113, 129, 157 decreased after glycolysis inhibition. In contrast, the concentration of m/z 129 in the presence of metabolic inhibitors was reduced in H358 & H322 cell lines while in A549 & H2030 cell lines it was enhanced.
Potential Impact:
Although lung cancer is the most lethal cancer worldwide (with more than 1M deaths every year), the impact of the current proposal is generic and affects every type of cancer, as all cancers share common metabolic and genomic pathways which are the essence of the LCAOS Project. LCAOS technology combines early detection of cancer with personalized treatment guidance and monitoring treatment outcome. Altogether, this project may contribute significantly to a better outcome for millions of cancer patients who are at risk or suffering from this devastating disease.
It is a well-known that early-stage diagnosis of lung cancer and related genetic mutations leads to higher survival rates. Current diagnostic techniques are characterized by one or combination of the following: invasive, expensive, not sufficiently accurate, and/or require the patient to change daily routine - all of which reduce compliance of the population to undergo screening tests. For example, the five-year survival rate in lung cancer is less than 10-15%, causing 1.2 million deaths each year. The main reason behind these figures is late-stage diagnosis, which usually happens at the symptomatic stage, when recovery is less likely even if treatment is given. An expert-panel hosted by IARC (International Agency for Cancer Prevention) December 4-6, 2013 has emphasized that lung cancer is likely to remain a condition of high global health importance for the foreseeable future (at least for the next 30 years) unless effective control measures will be implemented. The recommended algorithm for screening for lung cancer through low dose CT has not adopted by the payers or by the governmental healthcare systems, because it has been criticized by low specificity rate, i.e. high degree of false positivity. In few cases, tests for diagnosis and follow-up treatments of lung cancer have entered the market, but the translation of research findings into clinically useful tests seemed to be lagging. This is perhaps not surprising given the technical, financial, regulatory, and social challenges linked to the discovery, development, validation, and incorporation of genetic-related tests into clinical practice. To explore those challenges and to find ways to overcome them, the National Cancer Policy Forum held the conference “Developing Biomarker-Based Tools for Cancer Screening, Diagnosis and Treatment: The State of the Science, Evaluation, Implementation, and Economics” in Washington, D.C. from March 20 to 22, 2012. The conclusions and recommendations of this conference were taken into account during the LCAOS project.
By regular monitoring of patients bearing pre-malignant lesions it would be possible to identify those progressing towards high-risk conditions (such as high-grade dysplasia) or cancer. This would allow target management (e.g. targeted therapy or surgical management) only of those patients that otherwise would progress to malignant disease. In addition, monitoring of the treatment results would allow evaluation of the efficacy of the scheduled treatment plan. Regular surveillance of patients having undergone treatment (surgery, endoscopic resection, chemotherapy or other) would allow timely identification of those having the progression or relapse of the disease (e.g. metastatic disease following radical surgical management). All the above would result in higher survival rates and create savings for the healthcare organisations due to early detection of life-threatening conditions as well as reduction of unnecessary procedures (i.e. invasive removal of pre-malignant lesions by surgery). Thanks to micro- and nanotechnology, the NA-NOSE, could contain enough hard-wired intelligence and robustness to be used and to deliver a multitude of data/analysis to the practitioner. Furthermore, the use of nanoelectronics will improve the sensitivity of the sensors. New advances in microfluidic technologies have great potential to realise a fully-integrated device that directly delivers full data for a medical screening from a single sample.
The impact of this project is not only early detection, but rather also in directing targeted therapy, as well as monitoring treatment success. The project will allow a full monograph of the sum metabolomics signature of a specific patient. Precise diagnosis allows a more efficient therapy and avoids unnecessary treatment as in the case of in-efficient therapy. The applicability and user-friendliness of the technologies are at the core of the project. By engaging in a constructive dialogue with interest groups and early adopters (e.g. clinicians, patients, caregivers, technology developers, policy makers, etc.), the project ensures that the impact on society is matched with society’s unmet needs and wishes. To do so, the consortium has and will continue to organize various outreach actions, including:
• Knowledge dissemination in the research community through talks and presentations.
• Internal/external communication through different media to publicise the project results (newspaper, radio, TV).
• Science and Society interactions, with the public as well as with industrial partners (open lab workshops, school visits, and general study seminars).
In order to gather the feedback from future potential users on important issues such as acceptance, end-user value and business opportunities, we addressed these questions in focus groups, interviews and specific co-design sessions. Specific business development elements performed within the project aimed to explore the market factors influencing the possible commercialization of the end-product of LCAOS.
1. Scientific Impact
The LCAOS approach is totally different from conventional cancer diagnosis and follow-up methods. It detects cancer based on a change of the blood chemistry and metabolic activity and not on the basis of tumor imaging or pathogenetic morphological changes. Such an operation is comparably simple, and the results can be interpreted automatically using integrated software. Only positively-tested patients will require conventional, unpleasant and expensive imaging diagnostics (e.g. bronchoscopy, CT, MRI, etc.) to validate their condition and locate their tumor before decision on the treatment. The benefits of early stage detection and treatment are anticipated to significantly increase curability rates and lower the healthcare expenditure for societies. It also allows overcoming the heterogeneity of cancer as it exposed to the systemic circulation, which, probably, represents the total disease burden and not only the biopsy’ site.
The LCAOS project combines novel and improved sensing capabilities with statistical and pattern recognition methods. Achieving the goals of this proposal improve the scientific basis for obtaining enhanced performance from sensors and artificial intelligent arrays, especially when used to target low concentrations of compounds. This would provide clear improvement over the developments of artificial sensory systems reported so far, and, will facilitate further progress in the chosen areas of application. Indeed, LCAOS developments in sensor technology could provide invaluable real-world experience and feedback to assist development of such systems for other targeted applications such as security/safety applications and quality control of food and beverage products, which could include a broad range of problems in human health. In addition to the above, the proposed research will address important problems associated with achieving chemical control of the electrical properties of nanomaterials and devices. Arguably, any advances in this area will be of positive benefit to our overall understanding of this pervasive and this critical technology in Europe, and world-wide.
The collected data of each patient has been stored and kept available for retrieval and reanalysis. Within a few years of usage, a great amount of information (daily results, clinical information, lifestyle and other relevant data) will be collected and stored in a secured manner. This database will serve as an enormous potential for making new medical discoveries based on statistical analyses and data mining, allowing comparison of changes within a person’s life, and comparison of large groups of people having similar conditions. These abilities should provide additional insights on the general studies of disease initiation, progress, treatment efficacy enabling future medical breakthroughs.
2. Technological Impact
The LCAO’s strong European research partnerships and networks across national boundaries was necessary to acquire complementary skills and to respond to global economic trends, and to allow the European SMEs and large industries to maintain, and even gain market share, in a market dominated by strong competitors. In this context, the LCAOS project enabled the participating SMEs to collaborate, using their combined expertise in producing working demonstrators of unique, cancer detection and follow-up devices including software and procedures. For that purpose, the participating institutes provided the required materials, test samples, devices, mathematical and software novelties, based on research and process development. The development of a fast, low cost, and transparent-for-user tool though the development of the next generation of health monitoring system, based on systemic miniaturization and integration of heterogeneous microelectronics, nanoelectronics, microfluidics, and bio/chemical principles into a smart system may enable a new approach to improved detection and monitoring large populations with only simple actions on behalf of the patient, thus improving the compliance and reducing overall costs. The main technological development steps that were taken to achieve the targeted demonstrators were: 1) innovative miniaturized VOCs collection unit; 2) new multi-transducer, heterogamous sensors array, where the different sensors are complementary to each the other; 3) new effective data analysis, based on pattern recognition and statistical methods; and 4) data treatment and transmission and database preparation to the practitioner. An increase in sensitivity and selectivity of the LCAOS appeared to depend on improved microfluidics and the measuring setup, which will be developed by flexible SMEs.
3. Strategic Impact
Several strategic impacts have been reached through the LCAOS project. These include, but not confined to:
(a) Nanotechnology has become a multidisciplinary key technology. Today, everything is ‘nano’ in life-sciences and semiconductor technology, because this is the size of molecular structures, cells, DNA and the circuit paths in a computer’s CPU. Thus, it is timely to improve the existing diagnostics, e.g. for cancer screening, by taking advantage of progress in nanotechnology.
(b) The LCAOS project generated decisive knowledge for new applications in the field of early cancer detection and follow-up during treatment processes, by performing research located at the crossroads between different technologies and disciplines. The competitiveness of established European industries is largely dependent on their capacity to integrate knowledge and new technologies. LCAOS support the ability of the participating large industries, SMEs and hospitals to innovate through by advancing cancer screening strategies and methods, and will thus ultimately improve low-invasive treatments and therapies.
(c) The proposed LCAOS research has significant potential to reduce (lung) cancer mortality within the next decade through improved widespread screening and improved compliance, due mainly to being able to get early warnings of cancer at its earliest stages. Furthermore, the proposed approach has the potential to detect even pre-cancerous lesions, and would be suitable for high-risk populations who are not well-diagnosed by conventional techniques, (e.g. x-rays, computer tomography, etc.) that usually detect cancer at later stages, and which are, in addition, unpleasant (low patients’ compliance) and easily accompanied by complications. The LCAOS-based VOC (or breath) analysis can be used outside of specialist settings and could considerably reduce screening costs, both through the low cost of the proposed cancer test, and through earlier diagnosis and, hence, more cost-effective cancer treatments/therapies. The success of the LCAOS project can serve as a launch pad for similar approaches to be used in the diagnosis of other diseases (e.g. diabetes, lung cancer, colorectal cancer, pulmonary hypertension, heart disease, kidney disease, etc.) via VOCs. Ultimately, these developments would match society’s need for rapid and early screening of diseases, and will successfully initiate early therapeutic approaches and rapid treatment.
(d) Since the LCAOS approach has been implemented in easy-to-use devices, very limited human expertise is required for cancer detection and follow-up, in contrast to current screening- or detection-methods based on e.g. imaging techniques. Only positively-tested patients shall be subjected to the conventional imaging-based diagnostics to validate the results and locate their tumor.
(e) A number of graduate students and postdoctoral researchers were deeply involved in the project. Regular contacts between participating laboratories with broad-range expertise provided excellent opportunities for training Ph.D students/post-docs in integrated interdisciplinary techniques (e.g. materials science, nanoelectronics, medicine, nanotechnology based sensors, breath collection and analysis, algorithms developments, etc.). The multidisciplinary nature of the team further increased the gain for young scientists. The annual consortium conferences, the regular WP/sub-project meetings, and international symposia of the research field involving scientists from different disciplines provided forums for the young scientists to present their data/contributions, and interact with other scientists of the consortium and beyond. This networking together with wet-lab work and professional mentoring by the consortium seniors provided the students with versatile training and the post-doctoral fellows to support their career development. Also, in specific subprojects, short-term visits of a few weeks between different consortium partners are made possible for the young scientists to facilitate mobility and transfer of knowledge.
4. Policy and Socio-economic impacts
The LCAOS project contributed to European socio-economic policies to improve living conditions, safety, and the use of vital and economic resources. Although the incidence of lung cancer in age-standardized figures is declining, the total number of new cases is expected to remain stable in foreseeable future (at least the coming 30 years) due to aging and increasing population. In parallel to development of new treatment methods (effective chemotherapy in EGFR-positive lung cancer cases), the recent guidelines are requiring more extensive and expensive investigations and therapies for lung cancer, in particularly when diagnosed at advanced stages. In addition, the recognition of driver mutations improved treatment efficacy and elongated survival. Therefore, it is expected that the total number of patients with lung cancer will increase significantly, because they live longer. This results in higher impact upon healthcare budgets throughout Europe for management of the disease. The NA-NOSE result in earlier detection of the disease when endoscopic management or surgery is curative, therefore this constitutes a significant saving for both private and public healthcare expenditures in Europe. Although the project targeted specific pathologies to which the recorded data needs to be associated to, the NA-NOSE application is to be used in different scenarios that set different performance requirements (sensitivity, specificity, positive and negative prediction rates) as well as interpretability and robustness of the results. The different medical-related scenarios include, but not confined to:
• Screening purpose as a single-time test to be repeated periodically. This can be considered a low-cost simple diagnostic test applied to a large population in organized screening programs. Its main aim is to detect potential patients, who would benefit from more accurate follow-up investigations by a specialist, in particular – endoscopic assessment. As such, especially the sensitivity (detecting actual problem cases) of the method is of importance. Specificity (minimizing false alarms) is important too, but may be prioritised lower, depending on how expensive e.g. further examinations are. The disease prevalence is low, which sets some requirements to the classifier properties as well as datasets to use for training. This is also important that such approach allows avoiding unnecessary endoscopies, and therefore allowing spending the healthcare budgets more efficient.
• Monitoring of patients with revealed high-risk conditions. The current guidelines (e.g. ESMO, NCCN, and others) are recommending annual CT screening for individuals, who are at risk for lung cancer, mostly smokers who are older than 50 years old. With the current recommendation, which consequence with a very high false positive rate, numerous patients will undergo unnecessary invasive investigation. NA-NOSE is an ideal approach for monitoring purposes since the frequency of the measurements can be substantially higher than of endoscopic investigations because of the non-invasive test nature. Specificity is even more important here, as there is a high false positive rate in the current algorithms.
• Surveillance of patients after management procedures, e.g. after surgical or endoscopic removal of cancerous lesions. Substantial proportions of patients following curative surgery or endoscopic removal of the cancer are relapsing within the first five-year period. This could be either local relapse or distal metastasis; the latter cannot be identified with routine endoscopic follow-up. Timely management of the relapsing patients is related to more successful treatment and improved survival. The Scaliger Charter "Rationale of Oncological Follow‐Up after Resection for Cancer" accepted during the International Congress of Lung Cancer in Verona, June 19-22, 2013 has called for the necessity to put all the patients after curative surgery for lung cancer under surveillance programs. NA-NOSE shall be considered excellent solution for such surveillance. Since this population is really high-risk, sensitivity requirements can be strengthened further, and specificity more relaxed.
• Monitoring of efficacy of medication and dose titration. In this case we want to follow over time what has happened as consequence of certain interventions, hence a longitudinal analysis is needed to analyse a patient state over time. The appropriate approach here would be to use a continuous ‘disease burden’ index that evolves over time and show improvement (or relapse) of the disease. In this case we use probabilities, rather than dichotomous yes/no-disease classifications. This could be applied e.g. in patients after surgery or definitive therapies (e.g. chemo- radiation treatment. The importance of continuous monitoring is also in the era of the advanced disease, and particularly in the evolving era of targeted therapies, because tumour develops resistance to the targeted therapies. Early recognition of resistance will derive treatment modification which may contribute to a better therapy in an earlier stage, before the disease re-grow.
The potential for using the NA-NOSE technology in each of the above scenarios for other diseases and conditions (e.g. other types of cancer, neurodegenerative diseases, diabetes, heart diseases and their complications, etc.) contributing to further, significant cost savings. Further use of the NA-NOSE could be envisioned in the applications below:
• The proposed technology could enable early detection/control/screening of infectious and toxic agents/contaminations throughout food chain and/or from producer to consumer.
• The new technology has great potential to open the way for a new industrial scale manufacturing of inexpensive disposable sensors for medical and environmental applications. Indeed, commercialization of the new technology will require establishing new industries besides those which currently exist, including SMEs that engage in developing and manufacturing new instrumentation for both medical and environmental monitoring technologies. The increased number of industries and SMEs will increase competition in this area, which will lead to increased cost savings, and, ultimately, will create significant opportunities for increased employment in both the Europe and its Associated States. Based on the Organization for Economic Co-operation and Development (OECD) Innovation Report April 2011, it is estimated that there are >15000 companies with 20-100 employees. These companies employ >1 Million persons in total
5. Economic Impact
The diagnostic platform developed in LCAOS addresses market demands; it provides improved and high resolution screening and diagnostic, faster response times, increased availability to PoC and at-home testing at low cost. Thus, it is highly competitive and can have a huge economic potential to gain large market share among the screening techniques developed within the multiplexed diagnostics market. More details are listed below.
The estimated market – According to the estimates, in 2014, about 1.9 M people are expected to die of cancer in EU and the US, over 5,000 people per day. Cancer is the second most common cause of death in the US, exceeded only by heart disease, accounting for nearly 1 of every 4 deaths. The relatively new personalized medicine approach, relying on identification of disease-specific genetics, promotes personalized treatment improving survival rates among lung cancer patients, though currently the costs of obtaining genetic information as wide screening approach are too high. Personalized Medicine diagnostics market is expected to be more than €25 Billion by 2018. Diabetes management test and cancer management test are the leading market in this segment. To enhance precision through collection of broader data, the Multiplexed diagnostics and screening branch is gaining high acceptance in clinical research, especially with respect to high volume screening of cancer diseases. Achieving improved screening techniques, and an effective cost reduction to the medical system due to early diagnosis are only part of the factors pushing the growth of this market, estimated to reach almost €5 Billion in 2015. Recent technological advances in sensors, integration of heterogeneous technologies, continuous addition of novel biomarkers and an intensive effort to clinically validate these tests enable the rapid growth of this market. Some of the key players contributing to the health diagnostics market include Genzyme Corporation, GE Healthcare, Roche Diagnostics, Quest Diagnostics Inc., Tecan Group Ltd. and SIEMENS Healthcare Diagnostics, Inc.
Sustainability of health care system – It is a well-known fact that early diagnosis of cancer enables simpler treatment, at significantly lower price and better prognosis for the patient. Nevertheless, high-capacity screening programs must be cost-effective to the overall health system to allow its economical sustainability. In parallel to providing high quality treatment through introduction of new technologies, the healthcare systems need to find ways to become more efficient due to the continuously rising life expectancy. Unfortunately, lung cancer screening has not yet gained EU governments support and currently is not implemented, partially due to the associated high economic burden and the limited ability of early-stage diagnosis. Today there are more than 40,000 different in-vitro diagnostic products available, providing information to doctors and patients on a huge range of conditions; yet screening devices cost remarkably little - total expenditure was equivalent to €21 per person per year. This is less than what most people spend on the mobile phone bill per month. By comparison, healthcare expenditure on pharmaceuticals is more than €450 per head of population per year. NA-NOSE enables fast, relatively inexpensive and minimally invasive in-vitro screening and diagnostics of (lung) cancer and will make an important contribution towards addressing this problem.
Personalized medicine based on disease-related genetic mutations will provide more value for money not only because of improved drug effectiveness and reduced toxicity, but also through faster favorable outcome and shorter treatment periods. Studies on the cost-effectiveness of personalized medicines are still on the way but promising results are available. The introduction of a companion diagnostic strategy in advanced non-small cell lung cancer reduced overall treatment costs by more than € 800 compared to current treatment.
All technical components, including the NA-NOSE sensing chips, were developed with the aim of future cost-effective line-production, so that the NA-NOSE could be distributed at an affordable price. Thus, wide-spread implementation of the NA-NOSE to screen lung cancer would be more cost-effective than the conventional screening methods, contributing to the economic sustainability of the healthcare system. In addition, NA-NOSE is extremely easy-to-use and would not require trained medical personnel or specialist doctors. Hence, the initial wide-population screening may be done by the patient himself, reducing the number of referrals to specialist clinics and expensive medical consultations by specialists. Finally, earlier diagnosis and targeted therapy would allow more effective and less costly intervention, reducing hospitalization times and additional costs through co-morbidities and side-effects of rigorous medical intervention.
6. Dissemination and Exploitations Impact
The LCAOS communication and dissemination plan was developed in the early stages of the project to organize the overall project communication and dissemination activities and guarantees the dissemination of its results and assures that the project main findings are spread beyond the consortium to the target groups and to the wider public.
(a) Participation and Presentation at Scientific Conferences and Events
The LCAOS partner has disseminated the knowledge through the main dissemination activities:
• Participation at conferences and events: this activity was particularly important when the first results and recommendations were available. Project partners attended sector related events / conferences / workshops, where they met the target groups, other stakeholders, public authorities and scientific community and, also, raised awareness on the project objectives and results. The main on-going and final results were presented at scientific forums. As the members of the consortium come from different disciplines, they disseminated project results to diverse relevant scientific forums.
(b) Organize workshops and training sessions: During the LCAOS Project, there were three thematic workshops, practical courses and/or summer schools to advance transfer of knowledge, training, networking and career development among the young researchers within the consortium. These included: (a) the “1st LCAOS Workshop” held in Wallerfangen, Germany, in coordination with the International Society of Olfaction and Chemical Sensing (ISOCS) which organised in a short course on breath analysis the day before (13-14 June 2013); (b) The “2d LCAOS Workshop” held in Palma de Majorca, Spain (September 18-19, 2014); and (c) “The Breath Analysis Conference” held in Haifa, Israel (February 23-25). Additionally, the LCAOS partners set up the LCAOS website and the data exchange platform, which allowed easy data exchange between partners. The website has been setup and used as quick dissemination tool. Graphic design as well as flyer design was also performed. On top of these activities, three demonstration events have been performed with the LCAOS demonstrator, a preliminary one at Technion in December 2013, at the second LCAOS workshop in Mallorca (September 2014), and at the final closing meeting at Technion, (March 2015).
(c) Publication of scientific papers and magazines: Scientific excellence and technical skills available within the consortium or built through the LCAOS project were used for the preparation of scientific content for the benefit of the scientific communities and institutions. Most submissions for publication were concentrated during the last half of the project. Examples of journals focused on LCAOS related issues include: Nano Letters, ACS Nano, British Journal of Cancer, International Journal of Cancer, Advanced Materials, Nanomedicine, Sensors and Actuators B, Biosensors and Bioelectronics, IEEE Sensor, Cancer Research, Journal of Breath Research, etc.
The LCAOS consortium was fully committed in ensuring the sustainability and maximum possible exploitation of the project results. All the participating organisations and SMEs were well-placed to successfully carry out such activities, thanks to the wide networks of contacts and collaborating entities that each partner has, as well as their involvement in major international forums. These activities run in parallel and logically follow dissemination and will aim at properly transferring knowledge to the targeted sectors in order to gain support for possible implementation.
In order to transform our NA-NOSE technology into a marketable screening device we must take several steps: (i) prior art review to clarify the “freedom to operate”(FTO) status to potential investors; (ii) clinical evaluation (sensitivity and specificity) of a prototype device; (iii) completing the IP portfolio that covers the technology; and (iv) conducting detailed market research to identify the need for our technology, the optimal placement in the healthcare market, and possible competitors. In our case, the market research also involves planning our strategy towards Investigational Device Exemption (IDE) and from there towards the pivotal clinical trial leading to FDA/CE approval process. After completing these 3 stages, we can proceed to the technology transfer to a mature health care company (see below), or as an intermediate step, seek a grant of the EU, together with an industrial partner, that will afterwards fully take over the funding.
The Technion’s Technology Transfer unit, which, according to the CA, will lead the commercialization during the 4 years after the end of the LCAOS project, supports and guides the LCAOS partners in establishing an IP portfolio of their innovative ideas that were developed during academic research, covering all the costs of patent submission and search. We have already submitted more than 10 patent applications to various patent offices (including Europe and US) to protect the technology and clinical applications of the lung malignancy breath test. These efforts will continue during the prototype setup and testing. IP activity will also include reviewing prior art to clarify the “freedom to operate” status to investors. We will constantly update our review on prior arts and we intend to submit additional patent applications describing in details the features and novelties of the new system and to provide empirical support for our claims.
The LCAOS partners will explore two possible avenues to marketing. Our preferred option is to interest a major multinational company with a healthcare product line, or a well-established company that caters exclusively to the healthcare and medical diagnostics market. For both options, we would offer obtaining an option on our technology and to continue funding the project towards FAD/CE approval through the R&D departments of such companies. Alternatively, the EU offers several funding mechanisms for collaborative industry-academia projects, ranging from ~$3M to ~$6 M. Such type of funding may be obtained following the current grant and can lead from the tested prototype to a final product that is fully assembled in a manufacturing environment – a pre-requisite for initializing the pivotal trial for acquiring FDA or CE approval. Funding for these projects comes mostly from the government (80-90%) while the rest is covered by the industry; this means that a large industry can adopt the project with relative ease and later on take it to the last phase of actual product development. Several companies have already expressed initial interest, on condition that a prototype will be tested successfully, after visits to our laboratory and demonstration of the new technology. These include:
• Alpha Szenszor: a provider of advanced sensing technologies (Spain & France) for a wide range of life science applications. We are having advanced discussions with the company’s CEO, Steve Lerner, on placement and utility of the breath testing technology in the LC screening market (see attached formal letter).
• Siemens: a multinational company offering a wide range of electronics products, including conventional and innovative medical products such as clinical information technology systems; in-vitro diagnostics equipment; and wide range of imaging equipment (e.g. CT, MRI, PET, PET/CT). Siemens has expressed specific interest in combining our breath testing technology with LDCT screening for future population-based LC screening.
• Abbott: a multinational company that markets health care products. Its broad range of medical tests and diagnostic systems are used worldwide to diagnose and monitor diseases as well as assess other important indicators of general health. Abbott Europe and US have expressed interest in combining our breath testing technology with their screening devices.
Commercialization of the IP aspects will be carried out using appropriate method for the different target audiences (e.g. scientific community, clinical educators, general public). This will start by market research in the form of interviews with an established network of clinicians to define user requirements (usability, costs, integration in to the clinic workflow, etc.). The NA-NOSE prototype that is protected with IP portfolio will be presented to a wider group of clinicians to receive ‘market feedback’. This step is essential for planning of our strategy towards IDE and from there towards clinical trial leading to FDA/CE approval (after the funding period). In parallel, we will devote major effort to the research of the possibly unfolding markets especially in the US, Europe and Japan as described above. We will conduct this research together with an external partner (subcontractor), which will provide information on the nature of our specific edge and advantages with respect to other products in the market and recommend the best strategy to penetrate it. Through the entire project, and especially after clinical results will become available, we intend to initiate stronger contacts with potential interested companies and investors, as described above. For this purpose an executive summary and a presentation will be prepared as soon as the first clinical results will be available, which will be constantly updated throughout the project. Through the grant period, novel aspects that would be considered as inventions will be disclosed and submitted as patent applications. The cost and maintenance of this is covered by the Technion’s Technology Transfer unit. At the end of this process, LuMaSense will do market research together with an external consultant in the last third of the project. A worldwide force or scientific specialists (Technical Directors, not sales force) would be involved in introducing and disseminating the new test to clinicians and clinical lab staff and help to arrange these collaborations.
List of Websites:
Public website address: http://LCAOS.eu
Relevant Contact details:
Dr. Jan Mitrovics
Executive Director / Geschäftsführer
JLM Innovation GmbH
Vor dem Kreuzberg 17, 72070 Tübingen, Germany
Tel: +49 (7071) 5667730 Fax: +49 (7071) 5667731
http://www.jlm-innovation.de
Registergerichtgericht Stuttgart, HRB 735296
The LCAOS project has enabled the detection of lung cancer (LC) volatile biomarkers present in exhaled breath as well as in headspace of LC tissues/cells to screen and identify high risk groups for LC, and to monitor the therapy provided to people affected by LC. This goal was obtained by a strong multi-national and interdisciplinary consortium, composed of leading material scientists with additional proven ability to realize novel, prototype nanoscale devices, algorithmic groups, device developing and manufacturing SMEs and a company from the diagnostics field together with clinical institutions. Indeed, the LCAOS team have successfully developed, studied and implemented the following main action items: (i) developing molecularly-modified cross-selective Silicon nanowire Field Effect Transistors (Si NW FETs) and chemiresistive films of molecularly-modified gold nanoparticles in conjugation with pattern recognition methods - This new tool was termed Nano Artificial NOSE (NA-NOSE); (ii) testing the feasibility of the developed devices for sensing volatile biomarkers of LC from in-vitro tissues (including histologic/genetic sub-categories of LC) and from exhaled breath; (iii) developing an improved understanding of the signal transduction of the various volatile biomarkers, aided by sophisticated pattern recognition methods; and by (iv) developing performing clinical-related studies to assess LC conditions in actual patients and tissues and in the presence of real-world confounding signals.
The developed ‘NA-NOSE’ tool (15 ×15 × 5 cm; 0.8 kg; 5-10 min for analysis time) was examined both in lab-track and in clinical-track with more than 600 volunteers in several centres worldwide (Sheba Tal-Hashomer Hospital; Liverpool Lung Cancer Centre; Denver Colorado Lung Cancer Centre; Fox Chase Cancer Centre; Mayo Clinics; The First Affiliated Hospital of Anhui Medical University), with 30% of the samples being blinded. The blind analysis of the breath samples exhibited discrimination between early stages (I&II) and advances stages (III&IV) of lung cancer (81% average accuracy), between benign and malignant tumours (83% average accuracy), between SCLC and NSCLC (79% average accuracy), and between adenocarcinoma and squamous cell carcinoma (80% average accuracy). Many confounding factors (e.g. geographical location, age, gender, ethnics, smoking habit, etc.) were successfully neutralized using advanced hardware and software algorithms. In a series of in-vitro studies, the LCAOS partners have shown the feasibility to sense the secondary metabolic expressions for fine genetic alterations such as mutation in KRAS, P53, EGFR and EML4-ALK from the cells headspace (in a petri dish). Some of the alterations were associated with general carcinogenesis while the other part predicts response to targeted therapy.
The LCAOS achievements have indicated for the potential to reduce cancer mortality, by enabling widespread, trustworthy screening, especially suitable for high-risk populations. The ‘NA-NOSE’ was found to be suitable for use outside of specialist settings and would significantly reduce costs in health budgets. In addition, the ‘NA-NOSE’ was found to be suitable for immediate diagnosis of fresh LC tissues in operating rooms, where a dichotomic diagnosis is crucial to guide surgeons. The easy-to-use (no medical specialists required) ‘NA-NOSE’ technology detects cancer based on a change in the blood chemistry and/or metabolic activity (which is reflected in the chemical composition of the exhaled breath and cell/tissue headspace) rather than by tumor imaging, thus permitting earliest cancer detection i.e. before a tumor of detectable size has formed. The effectiveness of the ‘NA-NOSE’ in detecting LC volatile biomarkers specifically and selectively was experimentally and clinically proven to provide a launch pad for identifying other types of cancer from simple analysis of clinical samples, including breast, colon, and prostate cancers.
Project Context and Objectives:
Lung cancer is the most lethal cancer accounting for 28% of cancer deaths globally. In Europe there are 383,900 incident cases every year, resulting in 341,800 deaths. The average lifetime cost of lung cancer patients in Europe ranges between €46,000 and €61,000 per patient. Detecting lung cancer in its early stages, while still localized, can be expected to increase the 5-year-survival rate by three to four times. Unfortunately, diagnostic tests currently available, such as bronchoscope biopsy, pulmonary puncture and CT scan, occasionally miss tumours, which is costly. They are unsuitable for widespread screening because they are not efficient in terms of time, and are not free of complications. On the other hand, cancerous cells can be distinguished from non-cancerous cells on the basis of intracellular or extracellular biomarkers. Detection methods based on specific recognition of intracellular biomarkers (e.g. DNA/RNA/Proteins) require previous knowledge of specific mutations and copy number alterations in DNA/RNA or of changes in the regulation of protein expression inside the cells, assessed by expression and methylation profiling. Similarly, detection methods based on specific recognition of extracellular biomarkers such as histopathology, bioimaging, and antibody arrays require prior knowledge of biomarkers on cell surfaces. Antibody arrays can provide an effective but complex approach to cancer detection, diagnosis and prognosis. Nevertheless, there is no single biomarker or a combination of biomarkers with sufficient sensitivity and specificity to differentiate between normal, cancerous, and metastatic cell types.
An alternative approach to diagnosing cancer is emerging, which relies on volatile biomarkers emitted from cell membranes. The principle behind this approach is focused on the cell membrane, which consists primarily of amphipathic phospholipids, carbohydrates and many integral membrane proteins that are distinct for different cell types. Since tumor growth is accompanied by gene changes that may lead to oxidative stress, a peroxidation of the cell membrane species causes volatile biomarkers to be emitted. Some of these biomarkers appear in distinctively different mixture compositions, depending on whether a cell is healthy or cancerous. The remainder of the emitted biomarkers appear only in cancerous cells, and not in cells in a healthy state. Of particular significance to this approach is the fact that each type of cancer has its own unique pattern of volatile biomarkers, so that the presence of one cancer would not screen other cancer types. These volatile biomarkers can be detected either directly from the headspace of the cancer cells (i.e. the mixture of volatile biomarkers trapped above the cancer cells in a sealed vessel) or via the exhaled breath. The principle in the latter case is that cancer-related changes in blood chemistry are reflected in measurable changes to the breath through exchange via the lung.
The rationale for diagnosing cancer by means of volatile biomarkers has led to several important findings through spectrometry and/or spectroscopy-based techniques. To date, the use of these techniques has been impeded by the need for expensive equipment, high levels of expertise required to operate such instruments, the speed required for sampling and analysis, and the need for preconcentration techniques. Taking these considerations into account, the current proposal aims to harness nanotechnology, biomedical engineering, pulmonary medicine, medical oncology, and computation strategies in a multidisciplinary effort to develop and test a highly sensitive, inexpensive, and fast-response non-invasive Nanoscale Artificial Nose (NA-NOSE) for the detection of pre-neoplastic volatile biomarkers that indicate an increased risk of lung cancer (e.g. bronchial gene methylation, sputum atypia) and the presence of lung cancer. Further to these analyses, the proposed project investigates the bronchial epithelial histology (tissue) for its dysplastic appearance, early gene methylation changes, and genome-wide sequencing. In-vitro studies with lung cancer cells and tissues are performed, focusing on the differences between subtypes of lung cancer, in controlled conditions. The NA-NOSE’s ability to detect gene abnormalities in the lung cancer cells also be assessed.
The major objectives of the LCAOS project were divided between “strategic objectives”, “scientific objectives”, “technological objectives” and “policy objectives”.
STRATEGIC OBJECTIVES:
LCAOS’s extensive collaborative effort aims to make a significant contribution to the goals of the Strategic Research Agendas (SRA) of the relevant European Technology Platforms: both for Nanomedicine (ETPN) and for Smart System Integrations (EPoSS). The main strategic objectives of this project included (but are not confined to):
• Performing ambitious, but relatively short-term research within the Consortium of eight world-class R&D groups, well beyond conventional approaches, to realize a highly novel ‘NA-NOSE’ that has the potential for highly sensitive detection of cancer biomarkers in a simple, inexpensive and non-invasive manner.
• Establishing interaction and communication between the different disciplines involved in LCAOS, and between academia and industry. By sharing expertise and knowledge (including via an exchange of personnel) between academia, clinical institutions, SMEs, and diagnostics companies, the newly integrated ‘NA-NOSE’ was produced for clinical-track applications by the end of the project. The industrial participants, in addition to participating in the research, play a role in shaping the academic research through their feedback to the approaches developed. The academic partners provide access to the industrial partners to test clinical samples and advanced characterization methods.
• Facilitating the swift transfer of clinical and academic research results, as well as hardware and software directly to SMEs/companies specialized in commercializing devices and methods. SMEs is a crucial role in achieving this objective, which also includes validation and benchmarking of the novel ‘NA-NOSE’ at clinical sites, and performing comparisons with other diagnostic and screening tools.
• Given the multi-disciplinary nature of the LCAOS Consortium, the partners are able to provide multi-disciplinary training to researchers and students. The partners exploit this opportunity by exchanging personnel for extended periods of time, as well as organizing workshops and 2 mini-conferences (summer-school) throughout the four-year effort.
• To make most effective use of resources, a lab-track and a clinical-track run in parallel. In the lab-track, a prototype ‘NA-NOSE’ is developed and tested. In the clinical-track a more professionalized version of the ‘NA-NOSE’ is tested in clinical environments to carry out screening of high-risk population, monitoring of therapy progress and use of the device to direct surgery.
SCIENTIFIC OBJECTIVES
To support the strategic objectives and to enable the novel ‘NA-NOSE’ to identify and detect volatile biomarkers in real confounding environments, the partners have determined the following scientific objectives:
• To develop arrays of differently functionalized Si NW FETs that form a novel ‘NA-NOSE’ able to detect and identify volatile biomarkers specific in their mixture for LC as well as other types of cancer and even sub-types of LC (.
• To design and apply reliable pattern recognition models able to detect, with the minimum number of failures possible, the presence of cancerous volatile biomarkers in real samples, which exhibit a high chemical complexity.
• To test the feasibility of the developed ‘NA-NOSE’ to sense volatile biomarkers of LC from in-vitro tissues (including histologic/genetic sub-categories of LC) and from human exhaled breath.
• To test the feasibility of the developed ‘NA-NOSE’ to sense precancerous conditions, invasive LC (including for its subcategories) and cross-correlation between the volatile biomarker signatures and other known biomarkers (e.g. EGFR mutation).
• To test the feasibility of the ‘NA-NOSE’ to monitor the treatment/therapy success of LC.
• To reach a thorough understanding of the signal transduction mechanism of the various volatile biomarkers and metabolites, aided by sophisticated statistical and pattern recognition methods.
• To develop improved systems that will enable the ‘NA-NOSE’ to clearly distinguish the targeted biomarkers from environmental clutter, using methylation, expression profiling, and genome-wide sequencing in order to correlate LC’s metabolite signature with genetic aberrations in the related pathways that will improve treatment strategy.
• To perform clinical studies to assess LC conditions in actual patients and actual tissues. One of the aims in the initial clinical studies will be to explore the effectiveness of the important sensing signals in the presence of real-world confounding signals.
• To reach a thorough understanding of the molecular biology and pathways of the multistep process of lung carcinogenesis by combining the molecular transcriptomic and genomic data obtained in biopsies at all stages of lung squamous carcinogenesis, and with isolated organic compounds detected in exhaled breath or headspace of in-vitro or tissue samples.
• To further sub-categorize the profiles of volatile biomarkers for each of the sub-types of LC, where each profile will express a different level of tumor activity and aberrancy. The data emerging from this project will provide a very powerful metabolite, clinical and genetic dataset, which will require high-level bioinformatic analytical techniques. The medical-related studies will permit us to obtain the necessary feedback required to bioengineer a viable technology for the detection and identification of LC volatile biomarkers.
TECHNOLOGICAL OBJECTIVES
The realization of a highly-efficient and cost-effective ‘NA-NOSE’ for the detection and identification of volatile biomarkers related to (sub-types of) LC from breath samples or headspace of LC cells/tissues is carried out at two industry/SME sites, aiding with the expertise ofone of the SME partners in the field of medical diagnostics. These industrial partners take the ‘NA-NOSE’ prototype design to a higher level of miniaturization and will refine and automate features such as data acquisition and data interpretation and software integration, making the hand-held device simple-to-use, even be used during surgery. To fulfill these highly ambitious goals, the following technological objectives were set, to be achieved chiefly within the tasks performed by the industrial partners:
• Integration of an array of cross-reactive molecule-terminated Si NW FETs and molecule-terminated nanoparticle (MNPs) chemiresistive films in a ‘NA-NOSE’, with proven ability to detect and identify volatile biomarkers of LC, into a VLSI (~20mm×15mm) chip, preferably using VLSI fabrication techniques currently available.
• Design and prototype realization of a microfluidics flow-cell, able to be used as a host for the Si NW FET- and MNP-based (VLSI) chip, and as an exposure cell for breath and tissue/cells headspace samples. In this context, detailed objectives include the design and construction of: (i) a microfluidic cover to serve as an interface between the environments and directs the samples onto the Si NWs array; (ii) a flow-cell package to interface the flow cell with the ’NA-NOSE’; and (iii) a customized holder to provide both electrical and fluidic connections by means of press-fit connections.
• Design and construction of a prototype ‘NA-NOSE’ integrating the Si NW FET- and MNP-based (VLSI) chip (VLSI) chip and microfluidics system. This prototype is aimed to be portable and stand-alone for its use in clinical studies. The system is equipped with a microcontroller for fluidics control and data acquisition. Either an external computer is connected via USB, or an embedded computer will be used for data analysis, storage and user interface. This allows the same software platform (MultiSens) to be used during algorithm development and testing. MultiSens will be extended with an easy-to-use interface, to allow typical personnel to operate the system efficiently in a clinical environment.
• Integration of the prototypes into the MultiSens software packages in such a way that allow the other groups to embed and test their algorithms with the prototypes efficiently. The structure of MultiSens allows different algorithms (feature extraction, filtering, classification, and drift and cross interference compensation) to be tested and combined during the laboratory and clinical phases.
• Development of a system for testing and training the ‘NA-NOSE’ using artificial breath, which allows characterizing of the system in a defined and reproducible way. This has a great impact on progress of sensor development, since the measurement of the sensitivity (and cross-sensitivity) and selectivity of the sensor is readily available.
POLICY OBJECTIVES
Today, European research groups and companies/SMEs compete well in the field of (nano)biomedical devices to detect diseases including cancer biomarkers. European knowledge and expertise has led many technological developments for non-invasive prediction, diagnosis, monitoring and prognosis tools. For European leadership to be maintained, substantial R&D efforts with adequate funding, such as the LCAOS proposal, will be required. LCAOS aims to enable the European Technology Platforms for Nanomedicine (ETPN) and Smart System Integrations (EPoSS) to advance the novel concepts for early detection and identification of cancer biomarkers from breath samples and tissue/cells headspace, and with the help of the strong, multi-disciplinary consortium, to take a leadership role in establishing the next generation of easy-to-use, highly-sensitive screening, diagnosis, monitoring, and prognosis tools. LCAOS planning includes making the novel device attractive to biomedical markets. European device manufacturers shall get benefit from participation in LCAOS to launch highly innovative ‘NA-NOSE’ products. For these equipment manufacturers, LCAOS provides an early opportunity to commercialise devices with high potential, in the fast-growing area of non-invasive, real-time, and inexpensive biomedical tools. The participation of a leading SME in the diagnostics field will guarantee the benchmarking of the novel ‘NA-NOSE’ technology and advance deployment.
Project Results:
1. Fabrication of Si NWs
The fabrication of Si NWs with controlled diameter, length, and electronic properties is essential to the NA-NOSE that can be used for cancer detection. In the LCAOS project, the MPL partner presented a few pathways to grow and to improve Si NWs based sensing application performance and reliability though modified NWs chemical composition and dopant profiles. The fabrication of such NWs requires a state of the art growth conditions with superior control over growth parameters such as precursors and carrier gases type flow rates and partial pressures, growth temperature and so on. A true breakthrough over the LCAOS project was achieved mainly by setting up a new plasma enhanced chemical vapor deposition (PECVD) machine that enables the above mentioned requirements and controlling the overall growth parameters. In CVD, the growth of the Si NWs follows the VLS principle, where the temperature during the CVD process is chosen to be above the eutectic temperature and below the melting temperatures of gold and silicon. Electronic properties of the NWs can be conveniently tuned by adding doping gases during the Si NW growth. For example, we grow p-type Si NW by introducing B2H6 with silan gas (4 sccm SiH4 /0.1 sccm B2H6), while with similar conditions, n-type was grown by using PH3 as a carrier gas for the phosphine atoms (4 sccm SiH4 /0.1 sccm PH3). However, Si NWs were realized also by RIE technique. Here the doping structure of the SiNW can be controlled by the composition of the used substrate materials. Many different structures could be realized, including the fabrication of diodes at the scale of individual nanowires. The electronic properties of the SiNW have an important addtional role: tuning the sub-threshold senstivity of the FETs based NA-NOSE. Anticipated success on increased sensitivity limit for cancer detection was accomplished though tuning the sub-threshold window of the Si NWs and achieving high on/off ratio of the intrinsic Si NWs. The structure of the Si NWs is an important feature for the NA-NOSE since it determines the resolution of the signal to noise ratio for low concentration detection. Therefore, the structure of the Si NWs after VLS process has been analyzed intensively by means of high resolution transmission electron microscopy (TEM) and electron backscattered diffraction (EBSD). The results show a bi-crystalline structure, with presence of axial twins crossing the entire length of the NW from bottom to tip. These twin defects can be visualized when observing the NW across the [1-10] zone axis, and correspond to a 180º rotation along the (11-1) plane, which is perpendicular to the [112] NW growth axis. The sensitivity of the FET based Si NWs as a channel of the NA-NOSE, was increased by replacing the native oxide of the Si NW with variety of molecules (organic layer) to the surface of each Si NW. PN junctions were fabricated to explore the surface quality and conductivity of the Si NW. Finally, a high sensitivity at the ppm level was achieved after integrating the Si NW into the NA-NOSE device.
2. Fabrication of Si NW FET sensors
The goal of the LCAOS project is to develop a diagnostic tool based on SiNW array FET sensors for non-invasive diagnosis of lung cancer. In order to achieve this aim the Technion optimized the fabrication method for the devices, including the development of a new process for the deposition of the SiNWs on the substrate, called spray coating. The spray coating process is based on a simple principle of equilibrium between the shear force and the surface tension of the droplet containing the SiNWs when it comes in contact with the substrate. This process allowed us to control the number and orientation of the SINWs on the device, thus eliminating variances between the devices due to differences in the NW orientation. In addition, the density of the NWs on the surface was optimized for increased sensitivity of the device, as well as for optimal electrical properties. The next step was the investigation of the various electrical properties (features) extracted from the characteristic Ids-Vgs curve, such as the threshold voltage (Vth) and the charge carrier (hole) mobility (µh), in addition to the value of Ids at various Vgs. Many of these features are independent, resulting from different charge transfer mechanisms, and can therefore be considered as independent virtual sensors. The following step was determining which features are more relevant for VOC sensing, and which region of the Ids-Vgs curve has the greater importance for this specific sensing application (the region which contains the majority of VOC sensitive features). We found that Vth, µh, and the currents at the sub-threshold and linear region were the most important, and hold the highest potential for the differentiation between breath samples.
In addition, the Technion partner investigated the surface functionalization of the devices, and molecularly modified the surface of the Si NWs with a variety of organic molecules. These molecules were Silane based molecules, which varied either in chain length, or in functional group. We tested molecules with varying polarity and bulkiness, as well as several modification methods. The success of the functionalization of the sensors was tested via X-ray photoelectron spectroscopy (XPS). The changes in the molecular functionalization affected the affinity of the different sensors to the various VOCs, and we wanted to test which is the most appropriate surface modifications for the sensing of Lung cancer related VOCs. For that purpose, we exposed the sensors to a mixture of VOCs, as a way to simulate the human breath. These VOCs included compounds which were previously found to relate to Lung cancer.
After testing the sensors' ability to respond to simulations, the Technion partner exposed them to real breath samples of Lung cancer patients who had both early and late Lung cancer, and also to control subjects. The control group is a heterogeneous group that included completely healthy people, as well as people who had non-cancerous Lung conditions. These samples were sent to the Technion partner from both TAU and LIV partners. In the same time, the Technion also worked on adjusting the devices for the microfluidic chamber developed in MICRONIT in order to integrate these devices in the stand alone NA-NOSE. The Technion had to change the contact pads so they could be mounted on the chip more easily and so the flow of the sample had the best orientation with the Si NWs for sensing. In addition, we tested the sensing ability of these devices in the MICRONIT chip, to verify that it was not affected by the small sample volume or the change in the fabrication parameters. Finally, the Technion tested the complete stand-alone device and made sure it is ready to be used for research in clinics (see more details below).
3. Hosting Si NW FETs in Microfluidic Exposure Cells
Microfluidic exposure cells (size 22 x 30 mm2) were developed by Micronit to bring gaseous samples (i.e. exhaled breath) into contact with arrays of eight sensor chips. The sensor chips provided by Technion were attached to the microfluidic chip made from glass, forming both a fluidic and electrical interface between the sensors and the instrument. The internal volume of the fully assembled exposure cell is as low as a couple of microliters with very little internal dead volume thereby offering very quick exchange of the gaseous samples. Furthermore, a holder was designed with the capacity for connecting two exposure cells simultaneously which can be replaced very easily when required. A combined electrical and fluidic interface was developed, providing an easy and reliable means for swapping exposure cells when integrated with the stand-alone instrument.
4. Assembly of Stand-alone NA-NOSE
The JLM partner has designed and developed the NA-NOSE electronics and the related full demonstration instrument. Initial development stage consisted on design of interfacing and readout electronics for preliminary laboratory tests. Two different electronic interfaces were designed for interfacing Si NW FET and GNP sensors. A final version of the electronics for the demonstrator was designed based on the preliminary designs, including both interfaces in a single design, with optimized performance. The following stage in the project involved the design of the breath sampling subsystem and the sample transport subsystem. The demonstrator prototype featured the first version of these subsystems alongside with the electronics. The clinical versions of the instruments improved some aspects of the breath sampling subsystem, in particular the interface to the glass breath buffering tube. In the design process, 3D printed parts, fluidics, sample transport. The full instrument featured a novel integrated breath sampling subsystem alongside with the electronics, including a custom designed flow sensor to identify the phases of patient breathing. A sampling protocol was designed in coordination with TAU and Technion partners. A development grade software tool (research interface) was developed using Labview 2012®, which allowed limited co-development with Technion (who also holds a Labview 2012® license). The clinical grade software interface was developed using Delphi® and is optimized for touch-screen operation. High accuracy, fast readout electronics was demonstrated to match the accuracy of larger, expensive desktop instrumentation. Besides, the extremely compact construction of the breath-analysis demonstrator (NA-NOSE) enabled the development and study of point-of-care applications. In particular the whole design and construction of the innovative breath sampling subsystem has been a major outcome of the project.
5. Pattern Recognition Methods for the NA-NOSE
The UCM was in charge of the mathematical analysis and modeling of the results provided by diverse analytical equipment. The goal of this task was the design of models (software, in the end) to solve numerous problems such as: (1) to locate and assign specific volatile organic compounds to determined patterns to assess the viability of the approach; (2) to find disease-specific compounds (biomarkers) in breath that may aid in diagnosis; and (3) to classify and/or identify LC patients as well as staging them. To carry out these tasks, the UCM employed two main algorithmic blocks: feature selection (FS) and artificial neural networks (ANNs). Many of the databases that were provided contained great amounts of variables, reason why FS was crucial. On the other hand, ANNs were employed to achieve estimative and classificatory models, as they are a reliable set of non-linear mathematical tools that can excel in numerous applications. The combination of these kinds of algorithms gave the NA-NOSE better results than other classic models used in this field in the past. The UCM group has proven that the information gathered from breath analysis is clearly linked to the patient’s status, and can be used to reach early diagnosing tools.
Using the UCM algorithms, the LCAOS partners have been able to set a proof-of-concept that implies the viability of LC diagnosis through breath analysis, as specific disease-related patterns have been found in the databases offered by the analytical equipment. One of the main findings was that different volatile organic compounds (VOCs), comparable to potential LC biomarkers present in breath, provide specific patterns after interacting with cross-reactive sensor arrays. Furthermore, these patterns are also quantity dependent enabling not only identifying VOCs, but quantifying them. Additionally, mixtures of VOCs also were successfully identified employing ANNs. The models designed and optimized consist of supervised algorithms based on multilayer perceptrons. To achieve the best results, the parameters of the ANNs were optimized (some heuristically and others through experimental designs). In these models the input information was attained from the signals from the analytical equipment and the output provides information regarding the type and amount of VOCs. These types of algorithms are able to recognize biomarker profiles which characterize LC and their performance improve and are more reliable with bigger sets of data.
The UCM work has led to the development of a novel and complex software package that employs ANNs to locate the most important features or independent variables for a further estimation or classification. This software has been successfully employed in this project, and will clearly be taken advantage of in the future to continue fighting against LC and other types of cancers and illnesses.
To sum up this part of the work, one of the most important results of LCAOS is that the intelligent algorithms developed in this project guarantee that there are mathematical relations between biomarkers present in the human breath and LC. This relation is very important not only to reach early diagnosis of LC but also to deeply research about chemicals which are required to fight against it. In addition, this knowledge that has been achieved is the starting point for the research of other illnesses. UCM has enabled the last step of the LCAOS project: interpreting and putting to use the databases from the different analytical equipment to achieve early LC diagnosis through breath. The reliability, calculating power, and efficiency of ANNs have shown to be useful traits in the analysis of the intricate databases that are originated from breath analysis.
6. NA-NOSE and Volatolomics of Lung Cancer
Volatolomic analysis by the developed NA-NOSE as well as with complementary mass spectrometry techniques was first carried out by in vitro analysis of lung cell lines with defined mutations, whereby discrimination of individual mutations was possible. Analysis of the NA-NOSE responses allowed significant discrimination between LC and immortal bronchial epithelium (IBE) cells with 96% sensitivity, 86% specificity and 93% accuracy, as well as between the two major LC subtypes: small cell lung cancer (SCLC) and non-SCLC (NSCLC) with 96% accuracy, 100% sensitivity and 75% specificity in leave-one-out cross-validation. The volatolomic headspace of the two subtypes of NSCLC: adenocarcinoma and squamous cell carcinoma yielded 86% sensitivity, 100% specificity and 90% accuracy. SVM analysis of the compounds detected by GC-MS found 5 VOCs to show significant difference between IBE and LC cell lines and between the different histologies in different compositions with high accuracies as well. Histologic classification of lung cancer tumors has a major impact on the therapeutic approach (e.g. surgery or chemotherapy). However, it is reasonable to expect that the clusters of a wider study population would be less defined, and some overlap could occur, thus an extended study would be necessary to confirm the criteria for distinguishing between LC sub-histologies based on the findings presented in this study.
Due to the growing interest in lung cancer target therapy and the need for targeting specific genetic mutation related to the carcinogenesis process, the LCAOS partners (mainly, LIV, TAU, and Technion) looked for specific volatolomic signatures of LC genetic mutations through headspace analysis. Results of this pilot study provided first evidence for the existence of measurable VOC profiles for the most important genetic mutations associated with LC targeted therapy (EGFRmut, KRASmut and EML4-ALK fusion). A group of cell lines representing oncogenes that were wild type (wt) to the tested three mutation of interest served as control group. The three studied oncogenes are usually mutually exclusive and, hence, are considered distinct genetic subtypes of LC. Statistical analysis of the collective signals from the NA-NOSE yielded a comprehensive and highly accurate volatolomic fingerprint assay based on a set of three primary tests and an additional set of six supplementary tests that could correctly identify the oncogenes of the studied cell lines in a blind test with 76% accuracy in global leave-one-out cross-validation of the 9-test assay. Complementary analysis with GC-MS identified five VOCs (triethylamine, toluene, styrene, benzaldehyde and decanal) which could be associated with each of the genetic mutations of interest. While the concept of this study was obtained through in vitro studies, these results gave indications to assume that similar volatolomic pattern could be detected directly from the breath.
A more tractable system, than LC cell lines, to study the effect of specific cancer driver mutations on volatile release was provided by human bronchial epithelial cells (HBEC) that have been genetically manipulated. HBEC cells, immortalized through expression of telomerase and activation of CDK4, are genetically stable compared to cancer cell-lines, with minimal aneuploidy. This provided a parental cell-line in which the effect of additional genetic lesions can be investigated. In the derived cell-line HBEC-3KT53 TP53 has been knocked down by RNAi knockdown, whereas HBEC-3KTR cell-line expresses mutant KRASG12V. In a fourth cell-line, HBEC-3KTR53, both genetic alterations were induced by simultaneous transfection. In this part of the research, the developed NA-NOSE, together with Discriminant Factor Analysis (DFA), demonstrated the ability to discriminate metabolic differences due to individual mutations in lung cancer cells based on volatolomic signature. The classification accuracy of the DFA models obtained for any of the binary comparison were in the range of 77% to 93% in leave-one-out cross-validation and in the range of 64% to 89% in leave-third-out. Sensitivities and accuracies calculated by receiver operating characteristic (ROC) analysis were very high. Wilcoxon P values of the canonical variables (CVs) were below 0.04. Complementary chemical analysis by GC-MS has reported 20 VOCs from the families of alkanes, methylated alkanes and alkenes, benzene derivatives, ketones, aldehydes and alcohols to show significant discrimination between different genetic mutations. The three genetically altered HBEC cell lines and the vector only control (HBEC-3KT) allowed us to study the effect of individual mutations on the VOC signature of lung cancer precursors and potentially model the earliest stages of lung tumorigenesis, since these lesions may be present in those at risk of subsequently developing lung cancer. TP53 and KRAS mutations are common mutations in LC, sometimes occur together, and have been found early in the progression of lung tumorigenesis, thus there is a superior need for early, non-invasive and accurate identification of these LC genetic mutations.
To sum up, in-vitro experiments utilizing a model of early lung cancer progression have provided evidence that volatolomic profiles can distinguish between cells with limited, defined mutations. This supports both the potential ability of breath analysis to detect very early genetic changes (e.g. which may improve risk-based early detection) and very specific genetic changes (e.g. for characterization of patients for targeted molecular therapies).
7. NA-NOSE and Breath Analysis of Lung Cancer
In order to leverage the benefits of breath analysis in the characterisation of lung cancer, the LCAOS project has undertaken molecular (expression and methylation) and genetic (mutation) analysis of lung cancers and other clinical samples likely to form the basis of early detection or diagnosis. This analysis includes a consideration of the relationship between tumour mutation and lung cancer risk. We were able to demonstrate that molecular alterations in lung cancers were not closely associated with the epidemiological risk score, which offers hope that breath analysis would detect lung cancer independent of pre-conceived lung cancer risk – potentially opening up screening to those low risk individuals currently excluded. Furthermore molecular alterations were broadly similar in early cancers detected in asymptomatic individuals by CT screening, supporting a molecular basis for implementation of breath testing in the screening setting.
Given the opportunity to target a well-characterized, high-risk population, recruitment of breath samples by LIV partners has focused on controls from the Liverpool Lung Project population cohort. A total to 168 subjects (each providing replicate samples for parallel analysis by GC-MS and e-nose) was recruited, including pre-diagnosed lung cancer and a number of patients that were diagnosed with lung cancer subsequent to breath sampling. Whilst the diagnostic potential of breath analysis in symptomatic, high risk individuals is self-evident, in order to fully utilize breath analysis in an early detection or screening setting it is important to understand the relationship between breath analysis and other factors known to influence lung cancer risk. Currently epidemiology-based risk markers are the best way of identifying individuals that most benefit from a CT screening approach, which is an important consideration for the cost-benefit analysis of early lung cancer detection programs. In order to address the potential interaction of breath volatolomic signatures with these risk factors, we have included within the LCAOS recruitment a cohort of subjects form the Liverpool Lung Project (LLP), for whom lung cancer risk factors and potential confounding factors have been extensively documented. Breath samples from LLP participants collected by ULIV partners have been analysed at Technion for VOC content (by GC-MS) and by 2 alternative NA-NOSEs, with these large datasets jointly analyzed by UCM and LIV. It is of note, that breath analysis apparently reflects historical smoking history independent of recent smoking activity (which is a common confounding factor in VOC analysis); this may in part be due to a volatolomic signature of smoking-induced lung damage, which has been proposed as a more important risk factor than smoke exposure alone. Together with breath analysis of cases at ULIV and case/control cohorts across multiple sites, the LLP cohort samples also contribute to studies of breath-based lung cancer diagnosis; in particular several samples were collected from asymptomatic subjects who were subsequently diagnosed with lung cancer, providing a basis for investigation of early detection.
For each breath test subject we have collated epidemiological and clinical data with follow-up and stored this information on a secure database. For analysis two sets of data were extracted. The first consisted of a detailed epidemiological (subject) data and technical (sample) data in order to perform in depth analysis of risk factor and confounder variables in relation to VOC patterns. The second dataset combined LIV data with subject and sample data across multiple recruitment sites. In addition to breath samples, additional clinical samples were collected, including blood, plasma, sputum, buccal scrapes and tumour tissue.
DNA sequencing studies were designed to address the relationship between molecular alterations, lung cancer risk and volatolomic signatures arising from metabolic perturbations. An initial cohort representing tumours from patients with very low or very high levels of epidemiological risk was prepared for genome sequencing; additional samples were prepared from lung cancer cases recruited for breath samples and from lung cancers detected in asymptomatic subjects by CT screening. Genome and exome-wide sequencing was successfully performed on all selected tumour samples. Dysplasia samples were not sequenced, as none were found in breath test individuals. Additional lung cancer DNA sequence data was downloaded from public repositories to complement that prepared in house. DNA sequence data was produced from lung cancer cell-line DNA prepared by RMC/TAU partners). This was analysed alongside more extensive cell-line sequence, expression and methylation data available from public repositories. Expression data (in the form of microRNA microarray analysis) and DNA methylation analysis were performed on samples. In addition, other sample types were used as part of studies to: demonstrate a novel risk stratification technique in buccal samples; investigate the use of plasma microRNA for early detection of lung cancer; explore the promising area of plasma cell-free DNA biomarkers (so-called liquid biopsies), for both DNA methylation and DNA mutation. The sequencing data for the LCAOS project thus includes whole genome sequencing of carefully selected lung tumours and exome sequencing for both lung tumors and specific lung cancer cell lines (for which additional VOC analysis was performed in WP6). The data was analyzed in the context of published genomics datasets such as are available from The Cancer Genome Atlas (TCGA), International Cancer Genome Consortium (ICGC) and the Cancer Cell Line Encyclopaedia (CCLE) and aggregative datasets such as the Catalogue of Somatic Mutations in Cancer (COSMIC). Of particular relevance to LCAOS are genomic markers of lung tumour metabolic context since the lung tumour is likely to be a major contributor to the altered breath volatile signature in a cancer patient (albeit alongside the influence of tumour burden on systemic metabolism). Within the external datasets, we have addressed whether specific metabolic pathways are altered at the transcriptional, epigenetic or sequence level in lung cell lines, in tumour subclasses (e.g. adenocarcinoma compared with squamous cell carcinoma), or between tumour and normal lung tissue.
The relationship between epidemiology-based Risk Scores and lung cancer mutations is poorly understood. Although more cancers arise in those people with high epidemiology-based Risk Scores, some people with low Risk Scores also get cancer. At the molecular level lung cancer is not a single disease and it is important to consider both how the molecular genetic diversity of lung cancer might be influenced by the risk factors that are likely to form the basis of early detection and how this molecular genetic diversity might be reflected in the breath volatolomic signatures. To this end we have undertaken a range of molecular studies on lung cancers themselves and on other patient samples that might, together with breath, form the basis of lung cancer detection or characterisation. DNA mutation studies of lung tumours and matched normal DNA have been performed on LLP cases, for which the full range of risk factors, have been documented, with exemplar cases selected from both ends of the lung cancer risk spectrum. In parallel, a larger set of DNA sequencing data from publically available databases (e.g. The Cancer Genome Atlas, TCGA) was also analyzed in relation to their (more limited) risk marker profiles. These large DNA mutation datasets have been augmented with data from LCAOS cases (for which breath VOC profiles are available) and a panel of asymptomatic CT-screen detected cancers (representing the earliest stages of lung cancer). In addition gene expression and methylation studies have been performed, focusing on the cellular pathways implicated in cellular metabolism that contributes to VOC production. Importantly, we were able to demonstrate that tumours arising in patients considered being either high risk (targeted by screening programmes) or low risk harbour the same molecular alterations. This indicates that, should breath analysis be able to distinguish cancers on the basis of such molecular alterations, it would be equally sensitive for those low risk patients currently excluded from screening on the basis of their epidemiological risk profile.
Complementary to the efforts to identify high-risk people for LC, the TAU partner was responsible for executing the translational studies with the developed NA-NOSE device. In the first series of studies, the focus was on the diagnosis and classification of lung cancer. TAU collected exhaled breath samples from lung cancer patients and their matched control in order to evaluate the potential of exhaled breath analysis in lung cancer detection and distinction between early and advanced disease stages. Following the breath collection, together with the related clinical samples, the Technion team has analyzed the exhale breath of a total of 174 patients participated in this study, and Inter-group analysis of 80 lung cancer patients (64 advanced stage) and 31 matched controls showed a significant discrimination between disease and control. The results indicated that breath analysis discriminated early-stage lung cancer from matched controls. Moreover, it discriminated early stage patients from advanced disease patients. On top of that, the LCAOS partners compared breath samples from 72 patients with benign and malignant pulmonary nodules (PNs). For the specific identification of VOCs a complementary chemical analysis using GC-MS was performed. All patients underwent vast clinical evaluation including bronchoscopy, wedge resection and/or lobectomy for final diagnosis. LC patients were histologically classified into non-small cell LC (NSCLC) and small cell lung cancer (SCLC) and into early stage and advance stage of the disease. Nodule size of all malignancies was determined. Sensitivity, specificity and accuracy of the binary comparisons varied between 78% and 92% indicating that breath test might have future potential for the management of patients with PNs. More specifically, the results obtained in this study indicate that breath testing could be used to discriminate between benign and malignant PNs detected by CT, reduce the false-positive rate, minimize the risk of morbidity related to invasive diagnostic procedures and reduce the costs of LC screening. This innovative method may pose as an important non-invasive tool for lung cancer early detection, thus promoting better prognosis and therapeutic possibilities for patients.
In the second series of studies, the TAU team examined the use of breath sampling in monitoring response to anti-cancerous treatment of advanced lung cancer patients. Repeated exhaled breath samples were collected from patients with advanced lung cancer before and under systemic therapy. VOCs profiles were determined by GC-MS and nanomaterial-based array of sensor and correlated with response to therapy, assessed by CT scans as Complete Response (CR), Partial Response (PR), Stable Disease (SD), or Progressive Disease (PD). One hundred forty three breath samples were collected from 39 patients with stage III/IV lung cancer. GC-MS anaylsis identified 3 VOCs as significantly indicating PR/SD samples. One of them was also significantly discriminated between PR/SD and PD. Further, the NA-NOSE signals were able to alarm per a change in tumor response across therapy, i.e. indicating lack of further response to therapy, or developement of resistance to therapy. Therefore, breath analysis, using the NA-NOSE technology, may serve as a serogate marker for response to systemic therapy in lung cancer. Such a monitoring tool can provide the physicians with a quick and simple way to identify patient's response to anti-cancerous treatment in shorter intervals than currently available by CT scans, and so to allow a better decision-making regarding the chosen treatment. In a complementary phase, the Technion partner examined the exhaled VOCs of LC-induced long-term changes in pre-surgery and post-surgery patients. Breath samples were analyzed using two complimentary approaches. The first approach was chemical analysis of breath VOCs that aims to detect significant compounds that show statistically difference between pre-surgery and post-surgery patients using SPME/GC-MS. The second approach is based on constructing a volatolomic signature of breath VOCs by the NA-NOSE. Five VOCs were tentatively identified by GC-MS to have a significant difference between pre- and post-surgery LC patients with P value<0.013. These compounds are from the families of methylated alkenes, ketones, aromatic compounds and methylated alkanes. All five compounds were significantly reduced in post-surgery LC patients compared to pre-surgery patients indicating that their origin stem directly from the affected lung area that was resected. However, this identification is tentative and requires calibration standards for verification and a clinical study with a considerably larger number of patients is required to confirm these results. Leave-one-out cross validation of DFA analysis yielded high sensitivity and specificity of 100% and 80% respectively for the comparison of LC patients (pre-surgery) versus benign stated. For the comparison of LC patients (post-surgery) versus benign and for the comparison of benign before and after surgery the calculated sensitivity, specificity and accuracy were much lower. For the discrimination of LC before and after surgery accuracy, sensitivity and specificity varied between 75%-83%. These results imply that the volatolomic patterns may stem from the responses of the sensors to the local tumor markers, thus classification success reduced significantly once the tumor has been resected. The somewhat better separation of LC versus benign (pre-surgery) compared to LC before and after surgery supports the assumption that VOC levels would remain stable shortly after surgery. The appropriate choose of nanoparticles and recognition layer may identify LC-tumor related breath patterns, during short-term post-surgery follow-up. In the future, breath test might be utilized to complement the current diagnostic methods and to increase sensitivity of LC prediction prior to surgery, making swifter treatment decision possible.
8. Understanding Biochemical Pathways of LC Volatolomics
The IONICON partner supporting the clinical studies by providing the hardware, training and know-how for real-time breath analysis. IONICON has supplied a real-time trace gas analyser (PTR-MS) with a special inlet system for breath analysis (BET sampler) to project partner TAU. After the training of the researchers on the general use of the analyser and operating procedures for breath analysis, several clinical studies on human breath as well as cancer cell lines have been performed at TAU.
With the help of IONICON, TAU assessed the VOCs profile derived from lung cancer cell lines as well as the differences in glucose metabolism on the volatile signature in lung cancer patients through an oral glucose tolerance test (OGTT) using the PTRMS platform. This cohort included forty participants (22 control participants whom are at high risk for lung cancer, 18 study participants whom have active, naïve lung cancer). Pre-OGTT and Post-OGTT blood glucose levels and exhaled breath samples were measured with a lay period of 90 minutes. We revealed 14 masses which enable us to distinguish between the two groups. We concluded that breath analysis could discriminate between the high-risk and LC group. Furthermore, it demonstrated that glucose metabolism leaves a unique VOC pattern in the LC group. These findings may assist in the development of a non-invasive screening method for lung cancer. From the in vitro studies using the PTR-MS, the LCAOS partners identified and quantified masses/ charge (m/e’s) which are exhaled from cell’s headspace (ppt/ppb levels). These finding revealed differences in the volatile signature between LC subtypes which have a different mitochondrial capacity, and particularly, a volatile signature attributed to glycolysis. Additional differences in the breath of cancer patients can be revealed using an oral glucose tolerance tests, exploiting the Warburg effect, a metabolic anomaly of cancer cells. In addition, the in vitro studies in which we profile the VOCs pattern of A549, H2030, H358 and H322 cell lines using the PTR-MS detected numerous masses per charge (m/z’s) which their headspace concentration changed significantly after glycolysis inhibition. The concentration of the following masses m/z’s: 47, 60, 83, 101, 111, 113, 129, 157 decreased after glycolysis inhibition. In contrast, the concentration of m/z 129 in the presence of metabolic inhibitors was reduced in H358 & H322 cell lines while in A549 & H2030 cell lines it was enhanced.
Potential Impact:
Although lung cancer is the most lethal cancer worldwide (with more than 1M deaths every year), the impact of the current proposal is generic and affects every type of cancer, as all cancers share common metabolic and genomic pathways which are the essence of the LCAOS Project. LCAOS technology combines early detection of cancer with personalized treatment guidance and monitoring treatment outcome. Altogether, this project may contribute significantly to a better outcome for millions of cancer patients who are at risk or suffering from this devastating disease.
It is a well-known that early-stage diagnosis of lung cancer and related genetic mutations leads to higher survival rates. Current diagnostic techniques are characterized by one or combination of the following: invasive, expensive, not sufficiently accurate, and/or require the patient to change daily routine - all of which reduce compliance of the population to undergo screening tests. For example, the five-year survival rate in lung cancer is less than 10-15%, causing 1.2 million deaths each year. The main reason behind these figures is late-stage diagnosis, which usually happens at the symptomatic stage, when recovery is less likely even if treatment is given. An expert-panel hosted by IARC (International Agency for Cancer Prevention) December 4-6, 2013 has emphasized that lung cancer is likely to remain a condition of high global health importance for the foreseeable future (at least for the next 30 years) unless effective control measures will be implemented. The recommended algorithm for screening for lung cancer through low dose CT has not adopted by the payers or by the governmental healthcare systems, because it has been criticized by low specificity rate, i.e. high degree of false positivity. In few cases, tests for diagnosis and follow-up treatments of lung cancer have entered the market, but the translation of research findings into clinically useful tests seemed to be lagging. This is perhaps not surprising given the technical, financial, regulatory, and social challenges linked to the discovery, development, validation, and incorporation of genetic-related tests into clinical practice. To explore those challenges and to find ways to overcome them, the National Cancer Policy Forum held the conference “Developing Biomarker-Based Tools for Cancer Screening, Diagnosis and Treatment: The State of the Science, Evaluation, Implementation, and Economics” in Washington, D.C. from March 20 to 22, 2012. The conclusions and recommendations of this conference were taken into account during the LCAOS project.
By regular monitoring of patients bearing pre-malignant lesions it would be possible to identify those progressing towards high-risk conditions (such as high-grade dysplasia) or cancer. This would allow target management (e.g. targeted therapy or surgical management) only of those patients that otherwise would progress to malignant disease. In addition, monitoring of the treatment results would allow evaluation of the efficacy of the scheduled treatment plan. Regular surveillance of patients having undergone treatment (surgery, endoscopic resection, chemotherapy or other) would allow timely identification of those having the progression or relapse of the disease (e.g. metastatic disease following radical surgical management). All the above would result in higher survival rates and create savings for the healthcare organisations due to early detection of life-threatening conditions as well as reduction of unnecessary procedures (i.e. invasive removal of pre-malignant lesions by surgery). Thanks to micro- and nanotechnology, the NA-NOSE, could contain enough hard-wired intelligence and robustness to be used and to deliver a multitude of data/analysis to the practitioner. Furthermore, the use of nanoelectronics will improve the sensitivity of the sensors. New advances in microfluidic technologies have great potential to realise a fully-integrated device that directly delivers full data for a medical screening from a single sample.
The impact of this project is not only early detection, but rather also in directing targeted therapy, as well as monitoring treatment success. The project will allow a full monograph of the sum metabolomics signature of a specific patient. Precise diagnosis allows a more efficient therapy and avoids unnecessary treatment as in the case of in-efficient therapy. The applicability and user-friendliness of the technologies are at the core of the project. By engaging in a constructive dialogue with interest groups and early adopters (e.g. clinicians, patients, caregivers, technology developers, policy makers, etc.), the project ensures that the impact on society is matched with society’s unmet needs and wishes. To do so, the consortium has and will continue to organize various outreach actions, including:
• Knowledge dissemination in the research community through talks and presentations.
• Internal/external communication through different media to publicise the project results (newspaper, radio, TV).
• Science and Society interactions, with the public as well as with industrial partners (open lab workshops, school visits, and general study seminars).
In order to gather the feedback from future potential users on important issues such as acceptance, end-user value and business opportunities, we addressed these questions in focus groups, interviews and specific co-design sessions. Specific business development elements performed within the project aimed to explore the market factors influencing the possible commercialization of the end-product of LCAOS.
1. Scientific Impact
The LCAOS approach is totally different from conventional cancer diagnosis and follow-up methods. It detects cancer based on a change of the blood chemistry and metabolic activity and not on the basis of tumor imaging or pathogenetic morphological changes. Such an operation is comparably simple, and the results can be interpreted automatically using integrated software. Only positively-tested patients will require conventional, unpleasant and expensive imaging diagnostics (e.g. bronchoscopy, CT, MRI, etc.) to validate their condition and locate their tumor before decision on the treatment. The benefits of early stage detection and treatment are anticipated to significantly increase curability rates and lower the healthcare expenditure for societies. It also allows overcoming the heterogeneity of cancer as it exposed to the systemic circulation, which, probably, represents the total disease burden and not only the biopsy’ site.
The LCAOS project combines novel and improved sensing capabilities with statistical and pattern recognition methods. Achieving the goals of this proposal improve the scientific basis for obtaining enhanced performance from sensors and artificial intelligent arrays, especially when used to target low concentrations of compounds. This would provide clear improvement over the developments of artificial sensory systems reported so far, and, will facilitate further progress in the chosen areas of application. Indeed, LCAOS developments in sensor technology could provide invaluable real-world experience and feedback to assist development of such systems for other targeted applications such as security/safety applications and quality control of food and beverage products, which could include a broad range of problems in human health. In addition to the above, the proposed research will address important problems associated with achieving chemical control of the electrical properties of nanomaterials and devices. Arguably, any advances in this area will be of positive benefit to our overall understanding of this pervasive and this critical technology in Europe, and world-wide.
The collected data of each patient has been stored and kept available for retrieval and reanalysis. Within a few years of usage, a great amount of information (daily results, clinical information, lifestyle and other relevant data) will be collected and stored in a secured manner. This database will serve as an enormous potential for making new medical discoveries based on statistical analyses and data mining, allowing comparison of changes within a person’s life, and comparison of large groups of people having similar conditions. These abilities should provide additional insights on the general studies of disease initiation, progress, treatment efficacy enabling future medical breakthroughs.
2. Technological Impact
The LCAO’s strong European research partnerships and networks across national boundaries was necessary to acquire complementary skills and to respond to global economic trends, and to allow the European SMEs and large industries to maintain, and even gain market share, in a market dominated by strong competitors. In this context, the LCAOS project enabled the participating SMEs to collaborate, using their combined expertise in producing working demonstrators of unique, cancer detection and follow-up devices including software and procedures. For that purpose, the participating institutes provided the required materials, test samples, devices, mathematical and software novelties, based on research and process development. The development of a fast, low cost, and transparent-for-user tool though the development of the next generation of health monitoring system, based on systemic miniaturization and integration of heterogeneous microelectronics, nanoelectronics, microfluidics, and bio/chemical principles into a smart system may enable a new approach to improved detection and monitoring large populations with only simple actions on behalf of the patient, thus improving the compliance and reducing overall costs. The main technological development steps that were taken to achieve the targeted demonstrators were: 1) innovative miniaturized VOCs collection unit; 2) new multi-transducer, heterogamous sensors array, where the different sensors are complementary to each the other; 3) new effective data analysis, based on pattern recognition and statistical methods; and 4) data treatment and transmission and database preparation to the practitioner. An increase in sensitivity and selectivity of the LCAOS appeared to depend on improved microfluidics and the measuring setup, which will be developed by flexible SMEs.
3. Strategic Impact
Several strategic impacts have been reached through the LCAOS project. These include, but not confined to:
(a) Nanotechnology has become a multidisciplinary key technology. Today, everything is ‘nano’ in life-sciences and semiconductor technology, because this is the size of molecular structures, cells, DNA and the circuit paths in a computer’s CPU. Thus, it is timely to improve the existing diagnostics, e.g. for cancer screening, by taking advantage of progress in nanotechnology.
(b) The LCAOS project generated decisive knowledge for new applications in the field of early cancer detection and follow-up during treatment processes, by performing research located at the crossroads between different technologies and disciplines. The competitiveness of established European industries is largely dependent on their capacity to integrate knowledge and new technologies. LCAOS support the ability of the participating large industries, SMEs and hospitals to innovate through by advancing cancer screening strategies and methods, and will thus ultimately improve low-invasive treatments and therapies.
(c) The proposed LCAOS research has significant potential to reduce (lung) cancer mortality within the next decade through improved widespread screening and improved compliance, due mainly to being able to get early warnings of cancer at its earliest stages. Furthermore, the proposed approach has the potential to detect even pre-cancerous lesions, and would be suitable for high-risk populations who are not well-diagnosed by conventional techniques, (e.g. x-rays, computer tomography, etc.) that usually detect cancer at later stages, and which are, in addition, unpleasant (low patients’ compliance) and easily accompanied by complications. The LCAOS-based VOC (or breath) analysis can be used outside of specialist settings and could considerably reduce screening costs, both through the low cost of the proposed cancer test, and through earlier diagnosis and, hence, more cost-effective cancer treatments/therapies. The success of the LCAOS project can serve as a launch pad for similar approaches to be used in the diagnosis of other diseases (e.g. diabetes, lung cancer, colorectal cancer, pulmonary hypertension, heart disease, kidney disease, etc.) via VOCs. Ultimately, these developments would match society’s need for rapid and early screening of diseases, and will successfully initiate early therapeutic approaches and rapid treatment.
(d) Since the LCAOS approach has been implemented in easy-to-use devices, very limited human expertise is required for cancer detection and follow-up, in contrast to current screening- or detection-methods based on e.g. imaging techniques. Only positively-tested patients shall be subjected to the conventional imaging-based diagnostics to validate the results and locate their tumor.
(e) A number of graduate students and postdoctoral researchers were deeply involved in the project. Regular contacts between participating laboratories with broad-range expertise provided excellent opportunities for training Ph.D students/post-docs in integrated interdisciplinary techniques (e.g. materials science, nanoelectronics, medicine, nanotechnology based sensors, breath collection and analysis, algorithms developments, etc.). The multidisciplinary nature of the team further increased the gain for young scientists. The annual consortium conferences, the regular WP/sub-project meetings, and international symposia of the research field involving scientists from different disciplines provided forums for the young scientists to present their data/contributions, and interact with other scientists of the consortium and beyond. This networking together with wet-lab work and professional mentoring by the consortium seniors provided the students with versatile training and the post-doctoral fellows to support their career development. Also, in specific subprojects, short-term visits of a few weeks between different consortium partners are made possible for the young scientists to facilitate mobility and transfer of knowledge.
4. Policy and Socio-economic impacts
The LCAOS project contributed to European socio-economic policies to improve living conditions, safety, and the use of vital and economic resources. Although the incidence of lung cancer in age-standardized figures is declining, the total number of new cases is expected to remain stable in foreseeable future (at least the coming 30 years) due to aging and increasing population. In parallel to development of new treatment methods (effective chemotherapy in EGFR-positive lung cancer cases), the recent guidelines are requiring more extensive and expensive investigations and therapies for lung cancer, in particularly when diagnosed at advanced stages. In addition, the recognition of driver mutations improved treatment efficacy and elongated survival. Therefore, it is expected that the total number of patients with lung cancer will increase significantly, because they live longer. This results in higher impact upon healthcare budgets throughout Europe for management of the disease. The NA-NOSE result in earlier detection of the disease when endoscopic management or surgery is curative, therefore this constitutes a significant saving for both private and public healthcare expenditures in Europe. Although the project targeted specific pathologies to which the recorded data needs to be associated to, the NA-NOSE application is to be used in different scenarios that set different performance requirements (sensitivity, specificity, positive and negative prediction rates) as well as interpretability and robustness of the results. The different medical-related scenarios include, but not confined to:
• Screening purpose as a single-time test to be repeated periodically. This can be considered a low-cost simple diagnostic test applied to a large population in organized screening programs. Its main aim is to detect potential patients, who would benefit from more accurate follow-up investigations by a specialist, in particular – endoscopic assessment. As such, especially the sensitivity (detecting actual problem cases) of the method is of importance. Specificity (minimizing false alarms) is important too, but may be prioritised lower, depending on how expensive e.g. further examinations are. The disease prevalence is low, which sets some requirements to the classifier properties as well as datasets to use for training. This is also important that such approach allows avoiding unnecessary endoscopies, and therefore allowing spending the healthcare budgets more efficient.
• Monitoring of patients with revealed high-risk conditions. The current guidelines (e.g. ESMO, NCCN, and others) are recommending annual CT screening for individuals, who are at risk for lung cancer, mostly smokers who are older than 50 years old. With the current recommendation, which consequence with a very high false positive rate, numerous patients will undergo unnecessary invasive investigation. NA-NOSE is an ideal approach for monitoring purposes since the frequency of the measurements can be substantially higher than of endoscopic investigations because of the non-invasive test nature. Specificity is even more important here, as there is a high false positive rate in the current algorithms.
• Surveillance of patients after management procedures, e.g. after surgical or endoscopic removal of cancerous lesions. Substantial proportions of patients following curative surgery or endoscopic removal of the cancer are relapsing within the first five-year period. This could be either local relapse or distal metastasis; the latter cannot be identified with routine endoscopic follow-up. Timely management of the relapsing patients is related to more successful treatment and improved survival. The Scaliger Charter "Rationale of Oncological Follow‐Up after Resection for Cancer" accepted during the International Congress of Lung Cancer in Verona, June 19-22, 2013 has called for the necessity to put all the patients after curative surgery for lung cancer under surveillance programs. NA-NOSE shall be considered excellent solution for such surveillance. Since this population is really high-risk, sensitivity requirements can be strengthened further, and specificity more relaxed.
• Monitoring of efficacy of medication and dose titration. In this case we want to follow over time what has happened as consequence of certain interventions, hence a longitudinal analysis is needed to analyse a patient state over time. The appropriate approach here would be to use a continuous ‘disease burden’ index that evolves over time and show improvement (or relapse) of the disease. In this case we use probabilities, rather than dichotomous yes/no-disease classifications. This could be applied e.g. in patients after surgery or definitive therapies (e.g. chemo- radiation treatment. The importance of continuous monitoring is also in the era of the advanced disease, and particularly in the evolving era of targeted therapies, because tumour develops resistance to the targeted therapies. Early recognition of resistance will derive treatment modification which may contribute to a better therapy in an earlier stage, before the disease re-grow.
The potential for using the NA-NOSE technology in each of the above scenarios for other diseases and conditions (e.g. other types of cancer, neurodegenerative diseases, diabetes, heart diseases and their complications, etc.) contributing to further, significant cost savings. Further use of the NA-NOSE could be envisioned in the applications below:
• The proposed technology could enable early detection/control/screening of infectious and toxic agents/contaminations throughout food chain and/or from producer to consumer.
• The new technology has great potential to open the way for a new industrial scale manufacturing of inexpensive disposable sensors for medical and environmental applications. Indeed, commercialization of the new technology will require establishing new industries besides those which currently exist, including SMEs that engage in developing and manufacturing new instrumentation for both medical and environmental monitoring technologies. The increased number of industries and SMEs will increase competition in this area, which will lead to increased cost savings, and, ultimately, will create significant opportunities for increased employment in both the Europe and its Associated States. Based on the Organization for Economic Co-operation and Development (OECD) Innovation Report April 2011, it is estimated that there are >15000 companies with 20-100 employees. These companies employ >1 Million persons in total
5. Economic Impact
The diagnostic platform developed in LCAOS addresses market demands; it provides improved and high resolution screening and diagnostic, faster response times, increased availability to PoC and at-home testing at low cost. Thus, it is highly competitive and can have a huge economic potential to gain large market share among the screening techniques developed within the multiplexed diagnostics market. More details are listed below.
The estimated market – According to the estimates, in 2014, about 1.9 M people are expected to die of cancer in EU and the US, over 5,000 people per day. Cancer is the second most common cause of death in the US, exceeded only by heart disease, accounting for nearly 1 of every 4 deaths. The relatively new personalized medicine approach, relying on identification of disease-specific genetics, promotes personalized treatment improving survival rates among lung cancer patients, though currently the costs of obtaining genetic information as wide screening approach are too high. Personalized Medicine diagnostics market is expected to be more than €25 Billion by 2018. Diabetes management test and cancer management test are the leading market in this segment. To enhance precision through collection of broader data, the Multiplexed diagnostics and screening branch is gaining high acceptance in clinical research, especially with respect to high volume screening of cancer diseases. Achieving improved screening techniques, and an effective cost reduction to the medical system due to early diagnosis are only part of the factors pushing the growth of this market, estimated to reach almost €5 Billion in 2015. Recent technological advances in sensors, integration of heterogeneous technologies, continuous addition of novel biomarkers and an intensive effort to clinically validate these tests enable the rapid growth of this market. Some of the key players contributing to the health diagnostics market include Genzyme Corporation, GE Healthcare, Roche Diagnostics, Quest Diagnostics Inc., Tecan Group Ltd. and SIEMENS Healthcare Diagnostics, Inc.
Sustainability of health care system – It is a well-known fact that early diagnosis of cancer enables simpler treatment, at significantly lower price and better prognosis for the patient. Nevertheless, high-capacity screening programs must be cost-effective to the overall health system to allow its economical sustainability. In parallel to providing high quality treatment through introduction of new technologies, the healthcare systems need to find ways to become more efficient due to the continuously rising life expectancy. Unfortunately, lung cancer screening has not yet gained EU governments support and currently is not implemented, partially due to the associated high economic burden and the limited ability of early-stage diagnosis. Today there are more than 40,000 different in-vitro diagnostic products available, providing information to doctors and patients on a huge range of conditions; yet screening devices cost remarkably little - total expenditure was equivalent to €21 per person per year. This is less than what most people spend on the mobile phone bill per month. By comparison, healthcare expenditure on pharmaceuticals is more than €450 per head of population per year. NA-NOSE enables fast, relatively inexpensive and minimally invasive in-vitro screening and diagnostics of (lung) cancer and will make an important contribution towards addressing this problem.
Personalized medicine based on disease-related genetic mutations will provide more value for money not only because of improved drug effectiveness and reduced toxicity, but also through faster favorable outcome and shorter treatment periods. Studies on the cost-effectiveness of personalized medicines are still on the way but promising results are available. The introduction of a companion diagnostic strategy in advanced non-small cell lung cancer reduced overall treatment costs by more than € 800 compared to current treatment.
All technical components, including the NA-NOSE sensing chips, were developed with the aim of future cost-effective line-production, so that the NA-NOSE could be distributed at an affordable price. Thus, wide-spread implementation of the NA-NOSE to screen lung cancer would be more cost-effective than the conventional screening methods, contributing to the economic sustainability of the healthcare system. In addition, NA-NOSE is extremely easy-to-use and would not require trained medical personnel or specialist doctors. Hence, the initial wide-population screening may be done by the patient himself, reducing the number of referrals to specialist clinics and expensive medical consultations by specialists. Finally, earlier diagnosis and targeted therapy would allow more effective and less costly intervention, reducing hospitalization times and additional costs through co-morbidities and side-effects of rigorous medical intervention.
6. Dissemination and Exploitations Impact
The LCAOS communication and dissemination plan was developed in the early stages of the project to organize the overall project communication and dissemination activities and guarantees the dissemination of its results and assures that the project main findings are spread beyond the consortium to the target groups and to the wider public.
(a) Participation and Presentation at Scientific Conferences and Events
The LCAOS partner has disseminated the knowledge through the main dissemination activities:
• Participation at conferences and events: this activity was particularly important when the first results and recommendations were available. Project partners attended sector related events / conferences / workshops, where they met the target groups, other stakeholders, public authorities and scientific community and, also, raised awareness on the project objectives and results. The main on-going and final results were presented at scientific forums. As the members of the consortium come from different disciplines, they disseminated project results to diverse relevant scientific forums.
(b) Organize workshops and training sessions: During the LCAOS Project, there were three thematic workshops, practical courses and/or summer schools to advance transfer of knowledge, training, networking and career development among the young researchers within the consortium. These included: (a) the “1st LCAOS Workshop” held in Wallerfangen, Germany, in coordination with the International Society of Olfaction and Chemical Sensing (ISOCS) which organised in a short course on breath analysis the day before (13-14 June 2013); (b) The “2d LCAOS Workshop” held in Palma de Majorca, Spain (September 18-19, 2014); and (c) “The Breath Analysis Conference” held in Haifa, Israel (February 23-25). Additionally, the LCAOS partners set up the LCAOS website and the data exchange platform, which allowed easy data exchange between partners. The website has been setup and used as quick dissemination tool. Graphic design as well as flyer design was also performed. On top of these activities, three demonstration events have been performed with the LCAOS demonstrator, a preliminary one at Technion in December 2013, at the second LCAOS workshop in Mallorca (September 2014), and at the final closing meeting at Technion, (March 2015).
(c) Publication of scientific papers and magazines: Scientific excellence and technical skills available within the consortium or built through the LCAOS project were used for the preparation of scientific content for the benefit of the scientific communities and institutions. Most submissions for publication were concentrated during the last half of the project. Examples of journals focused on LCAOS related issues include: Nano Letters, ACS Nano, British Journal of Cancer, International Journal of Cancer, Advanced Materials, Nanomedicine, Sensors and Actuators B, Biosensors and Bioelectronics, IEEE Sensor, Cancer Research, Journal of Breath Research, etc.
The LCAOS consortium was fully committed in ensuring the sustainability and maximum possible exploitation of the project results. All the participating organisations and SMEs were well-placed to successfully carry out such activities, thanks to the wide networks of contacts and collaborating entities that each partner has, as well as their involvement in major international forums. These activities run in parallel and logically follow dissemination and will aim at properly transferring knowledge to the targeted sectors in order to gain support for possible implementation.
In order to transform our NA-NOSE technology into a marketable screening device we must take several steps: (i) prior art review to clarify the “freedom to operate”(FTO) status to potential investors; (ii) clinical evaluation (sensitivity and specificity) of a prototype device; (iii) completing the IP portfolio that covers the technology; and (iv) conducting detailed market research to identify the need for our technology, the optimal placement in the healthcare market, and possible competitors. In our case, the market research also involves planning our strategy towards Investigational Device Exemption (IDE) and from there towards the pivotal clinical trial leading to FDA/CE approval process. After completing these 3 stages, we can proceed to the technology transfer to a mature health care company (see below), or as an intermediate step, seek a grant of the EU, together with an industrial partner, that will afterwards fully take over the funding.
The Technion’s Technology Transfer unit, which, according to the CA, will lead the commercialization during the 4 years after the end of the LCAOS project, supports and guides the LCAOS partners in establishing an IP portfolio of their innovative ideas that were developed during academic research, covering all the costs of patent submission and search. We have already submitted more than 10 patent applications to various patent offices (including Europe and US) to protect the technology and clinical applications of the lung malignancy breath test. These efforts will continue during the prototype setup and testing. IP activity will also include reviewing prior art to clarify the “freedom to operate” status to investors. We will constantly update our review on prior arts and we intend to submit additional patent applications describing in details the features and novelties of the new system and to provide empirical support for our claims.
The LCAOS partners will explore two possible avenues to marketing. Our preferred option is to interest a major multinational company with a healthcare product line, or a well-established company that caters exclusively to the healthcare and medical diagnostics market. For both options, we would offer obtaining an option on our technology and to continue funding the project towards FAD/CE approval through the R&D departments of such companies. Alternatively, the EU offers several funding mechanisms for collaborative industry-academia projects, ranging from ~$3M to ~$6 M. Such type of funding may be obtained following the current grant and can lead from the tested prototype to a final product that is fully assembled in a manufacturing environment – a pre-requisite for initializing the pivotal trial for acquiring FDA or CE approval. Funding for these projects comes mostly from the government (80-90%) while the rest is covered by the industry; this means that a large industry can adopt the project with relative ease and later on take it to the last phase of actual product development. Several companies have already expressed initial interest, on condition that a prototype will be tested successfully, after visits to our laboratory and demonstration of the new technology. These include:
• Alpha Szenszor: a provider of advanced sensing technologies (Spain & France) for a wide range of life science applications. We are having advanced discussions with the company’s CEO, Steve Lerner, on placement and utility of the breath testing technology in the LC screening market (see attached formal letter).
• Siemens: a multinational company offering a wide range of electronics products, including conventional and innovative medical products such as clinical information technology systems; in-vitro diagnostics equipment; and wide range of imaging equipment (e.g. CT, MRI, PET, PET/CT). Siemens has expressed specific interest in combining our breath testing technology with LDCT screening for future population-based LC screening.
• Abbott: a multinational company that markets health care products. Its broad range of medical tests and diagnostic systems are used worldwide to diagnose and monitor diseases as well as assess other important indicators of general health. Abbott Europe and US have expressed interest in combining our breath testing technology with their screening devices.
Commercialization of the IP aspects will be carried out using appropriate method for the different target audiences (e.g. scientific community, clinical educators, general public). This will start by market research in the form of interviews with an established network of clinicians to define user requirements (usability, costs, integration in to the clinic workflow, etc.). The NA-NOSE prototype that is protected with IP portfolio will be presented to a wider group of clinicians to receive ‘market feedback’. This step is essential for planning of our strategy towards IDE and from there towards clinical trial leading to FDA/CE approval (after the funding period). In parallel, we will devote major effort to the research of the possibly unfolding markets especially in the US, Europe and Japan as described above. We will conduct this research together with an external partner (subcontractor), which will provide information on the nature of our specific edge and advantages with respect to other products in the market and recommend the best strategy to penetrate it. Through the entire project, and especially after clinical results will become available, we intend to initiate stronger contacts with potential interested companies and investors, as described above. For this purpose an executive summary and a presentation will be prepared as soon as the first clinical results will be available, which will be constantly updated throughout the project. Through the grant period, novel aspects that would be considered as inventions will be disclosed and submitted as patent applications. The cost and maintenance of this is covered by the Technion’s Technology Transfer unit. At the end of this process, LuMaSense will do market research together with an external consultant in the last third of the project. A worldwide force or scientific specialists (Technical Directors, not sales force) would be involved in introducing and disseminating the new test to clinicians and clinical lab staff and help to arrange these collaborations.
List of Websites:
Public website address: http://LCAOS.eu
Relevant Contact details:
Dr. Jan Mitrovics
Executive Director / Geschäftsführer
JLM Innovation GmbH
Vor dem Kreuzberg 17, 72070 Tübingen, Germany
Tel: +49 (7071) 5667730 Fax: +49 (7071) 5667731
http://www.jlm-innovation.de
Registergerichtgericht Stuttgart, HRB 735296