Final Report Summary - EDOCAL (Early detection of cancer using photonic crystal lasers)
Executive summary:
Early detection of cancer saves lives, improves quality of life and reduces health care cost. In Europe there are 3 million new cancer cases every year. According to the World Health Organisation (WHO) the burden of cancer can be reduced by systematic and equitable implementation of evidence based strategies for cancer prevention, early detection and management of patients with cancer. 33 % of the cancer burden could be reduced if detected early and treated adequately. The WHO concludes that early diagnosis of cancers could save the developed world an average of EUR 25 billion per year on medical costs and EUR 50 billion per year on lost productivity.
Early-cancer diagnosis today consists of recognition of some of the early signs by the patient themselves or detection from screenings carried out on specific population groups. If something suspicious is found, the next step is to perform a blood analysis and/or to look with non-invasive imaging techniques. Current imaging techniques are widely used in diagnostic procedures but they cannot give conclusive evidence of cancer. Today, only a biopsy and the subsequent pathologist interpretation can give a definitive diagnosis of cancer and this is not always sufficient. In the EDOCAL project the proof of concept has been achieved for early detection of cancer based on blue laser technology. The technology has been tested and proven for oesophageal cancer detection and has the potential for early detection of many other types of cancer including stomach and colorectal. The technical principle behind the EDOCAL concept is as follows.
In cancer, tissue cells tend to create additional blood vessels to support their growth. The molecule protoporphyrin (PpIX) is generally present in blood vessels and exhibits red fluorescence when excited with blue light in the 375 - 425 nm range. Cancer cells are detected by observing the red fluorescence at a matching excitation wavelength. The wavelength of the laser has to be tuneable as the matching wavelength depends on the person, type of illness and presence of other chemicals. To increase the intensity of the red fluorescence, a porphyrin prodrug can be administered either orally or locally which synthesises PpIX in the cancerous (cells) tissue. By using a tuneable laser system and combination of endogenous fluorescence (as a result of extra blood) and exogenous fluorescence (produced by administration of a porphyrin prodrug), it is possible to accurately distinguish between normal and cancerous tissue.
The results of the EDOCAL project can be summarised as follows:
(a) A tuneable blue laser prototype was defined and the system built for research purposes.
(b) Experiments to detect cancer using shifts in excitation and emission wavelengths were carried out in 70 patients and we obtained a total of 1 557 spectra from 290 sites in the oesophagus. Using the prototype, each site was illuminated with multiple wavelengths followed by registration of the corresponding auto-fluorescence spectrum and tissue biopsy for histological correlation.
(c) Results from the EDOCAL prototype show that it is possible to discriminate early cancer with 80 % sensitivity and 81 % specificity, which meets the performance requirements targeted at the start of the EDOCAL project.
(d) Proof of concept has been achieved for early detection of cancer based on a tuneable blue laser technology.
(e) Preliminary specifications for the first demonstrator have been defined.
(f) The first patent application has been filed and new filings for the latest results are being worked on.
The final goal of the small and medium-sized enterprises (SMEs) is the creation of a device that can be used for first line cancer detection. This device will reduce the need for biopsies, make the diagnosis more reliable, less dependent on human interpretation, make early cancer detection available to more people and give the SMEs access to a new and quickly growing market: reliable, life-saving medical diagnostics. The EDOCAL project combined state-of-the-art laser and imaging technology with leading medical research, expertise and hands on practical medical treatment experience to validate the first prototype and investigate whether or not it not it would be effective in discriminating early cancerous tissue from healthy tissue. The project has succeeded in establishing proof of concept for oesophageal cancer detection. The technique was validated using cell lines in vitro before applying the procedure to patients. It is our intention to investigate the use the same principle method to diagnose and treat other cancer types and to use the results of this project to create a demonstrator that can be tested in the field.
Project context and objectives:
Globally there are 24.6 million people living with cancer. By 2020 there will be over 30 million. Cancer is a leading cause of death worldwide. From a total of 58 million deaths worldwide in 2005, cancer accounted for 7.6 million (or 13 %). Deaths from cancer in the world are projected to continue rising, with an estimated 9 million people dying from cancer in 2015 and 11.4 million dying in 2030.
In Europe there are 3 million new cancer cases per year (38 countries), with 2 million new cases in the EU-25 alone. This represents 340 new cases / 100 000 people per year. One in three men and one in four women will be inflicted with cancer during their life. There is no doubt about the fact that cancer is an increasingly important factor in the global burden of disease with tremendous economic impact.
The United State (US) National Institute of Health (NIH) estimated the overall annual costs for cancer in 2000 at EUR 112 billion in the US alone, with EUR 38 billion for direct medical costs and the rest as a result of lost productivity. These figures can, in their opinion be doubled to cover the rest of the developed world. The total global cost of cancer in 2000 was EUR 150 billion. The figure for direct medical costs alone per patient in Europe is EUR 31 000 and rising.
According to the WHO the burden of cancer can be reduced by systematic and equitable implementation of evidence based strategies for cancer prevention, early detection and management of patients with cancer. Up to 33 % of the cancer burden could be reduced by implementing preventative strategies aimed at reducing the exposure to cancer risks. Another 33 % of the cancer burden could be reduced if detected early and treated adequately. The WHO concludes that, early diagnosis of cancers could save the developed world an average of EUR 25 billion per year on medical costs and USD 50 billion per year on realised productivity. The direct savings on medical treatment alone would be more than EUR 10 000 per patient if early detection was possible.
The purpose of the EDOCAL project is to create a breakthrough tool for early cancer detection by combining state of the art laser and imaging technology with leading medical research (expertise and hands on practical medical treatment experience). We expect that the results from the medical research will enable the SMEs to create a device that can be used for first line cancer detection, making early cancer detection more accessible to more people all over the world. This will create a new mass market for the products created by the SMEs and will ensure for them a position in a new and quickly growing market: Cheap, reliable, medical diagnostics for everyone.
In spite of the huge increase in certain diseases attributed to increased prosperity, growth in product sales (e.g. endoscopes) for treatment of these diseases has not happened at the same rate. Reports by Frost & Sullivan indicate the main reasons for the slow growth to be: lack of new and technically exciting products, the high cost associated with initial purchase of endoscopes and resistance to new technologies. EDOCAL aims to provide the innovative breakthrough, at reasonable costs by combining advanced, proven and low cost telecom and state of the art semiconductor technologies with the latest advances and insights into medical procedures provided by the research and technological development (RTD) partners. EDOCAL also aims to enable the European companies to differentiate themselves from the high-end segment players in Japan and the US through the use of innovative features and will increase the technology lead of Europe compared to the low-cost original equipment manufacturer (OEM) players from China and India.
Early cancer diagnosis today consists of recognition of some of the 'early' signs by the patient themselves or detection from screenings carried out on specific population groups. Typical signs are: lumps, sores, persistent indigestion, persistent coughing and bleeding from the body orifices. If something suspicious is found the next step is to perform a blood analysis and/or to look with non-invasive imaging techniques. Current imaging techniques are widely used in diagnostic procedures but they cannot give conclusive evidence of cancer. Today, only a biopsy and the subsequent pathologists interpretation can give a definitive diagnosis of cancer and even sometimes this is also insufficient. A biopsy is a procedure to remove a piece of tissue or a sample of cells from the body so that it can be analysed in a laboratory. There are five different types of biopsy procedures depending on type of cancer suspected and the location of the suspicious cells.
The main problems with current cancer diagnostic techniques and procedures are:
(1) The lack of early primary detection methods: Early detection depends on the ability to discriminate healthy from (potentially) cancerous tissue, before physical symptoms like sores and ulcers start to appear. In the early stages when the impact of many cancers is still minimal, it is difficult and often impossible to conclusively discriminate healthy from malignant tissue, leading to incorrect / inconclusive results.
(2) High costs due to late interventions: When cancer is not detected in an early stage, complicated surgical interventions, chemotherapy and radiotherapy with limited long term success rates become necessary. The direct and indirect costs are huge when compared to the costs associated with an early diagnosis and the resulting intervention which are sufficient if the cancer has not had a chance to mature and spread. For example, early diagnosis and treatment of gastro intestinal (e.g. Barrett's) cancer has a high success rate via endoscopic intervention. However when detected late, surgery is always necessary and chances of success are low. At this stage there is a mortality rate of 3 - 5 % during surgery and a survival rate of only 20 % after 5 years. As stated earlier, direct savings per patient per year would be approximately EUR 10 000. In the European Union (EU) alone 1.48 million new cancer cases are handled, every year, which adds up to a saving in direct cost of about EUR 15 billion per year due to intervention at an earlier stage.
(3) Sensitivity and specificity of existing imaging techniques are too low. Magnetic resonance imaging (MRI) and computed tomography (CT) scans when used for the types of cancer that are most difficult to detect, have a sensitivity and a specificity of roughly 40 - 70 % and early neoplastic lesions are often too small to be detected by these scans. Sensitivity and specificity of existing endoscopic procedures is roughly 50 - 80 %. This means that there is insufficient contrast between the healthy and cancerous tissue, making it impossible to make a reliable diagnosis. Therefore a biopsy with all of its drawbacks is still the golden standard for conclusive cancer diagnosis.
(4) Non-existence of reliable red flag tools: Typically the sensitivity of existing tools is between 60 - 70 %, which is not reliable enough to be of much use in daily medical practice. Today no online surgical tools exist that can clearly and immediately discriminate healthy from cancerous tissue over a large surface area. Oncologists agree that a wide angled, easy to use, real-time detection tool small enough to be integrated into an endoscope would be an enormous help to them, showing them where to intervene and allowing them to track their progress. Such a tool does not exist today.
(5) The conclusive biopsy lab analysis is slow: Typically is takes about 1 week before results of the biopsy are available and another 1 - 2 weeks before the next treatment step can be taken or appointments scheduled. Furthermore there is the risk that the wrong tissue is sampled for biopsy (sampling error) and that too little malignant tissue or too much healthy tissue is removed from a patient.
The purpose of this project is to create a breakthrough tool for early cancer detection by combining state of the art laser and imaging technology with leading medical research, expertise and hands on experience with application of the tools.
The main focus of the EDOCAL project is to medically validate the use of the first prototype based on the dynamic blue technology for early cancer detection applications. This will be done using standard commercially available endoscopes extended with the dynamic blue laser prototypes.
The SMEs will cooperate with world class medical RTD institutions to fill the gap in medical know how and experience of the SMEs themselves and to enable linking the RTDs leading medical experience with breakthrough optical technologies. The technical principle behind the EDOCAL concept is as follows.
In cancer, tissue cells tend to create additional blood vessels to support their growth. The molecule PpIX is generally present in blood vessels and exhibits red fluorescence when excited with blue light in the 375 - 425 nm range. Cancer cells are detected by observing the red fluorescence at a matching excitation wavelength. The wavelength of the laser has to be tuneable as the matching wavelength depends on the person, type of illness and presence of other chemicals. To increase the intensity of the red fluorescence, a porphyrin prodrug can be administered either orally or locally which synthesises PpIX in the cancerous (cells) tissue. By using a tuneable laser system and combination of endogenous fluorescence (as a result of extra blood) and exogenous fluorescence (produced by administration of a porphyrin prodrug), it is possible to accurately distinguish between normal and cancerous tissue.
This prototype tool will use a tuneable blue laser (laser with multiple blue wavelengths) to selectively excite an accumulated photo sensitiser, called proto-porphyrin in (pre) tissue. This sensitiser is taken by the patient to enhance contrast between healthy and cancerous tissue. Reflected light is filtered using state of the art laser technology. The underlying technology platform will also be applicable for future medical diagnostic solutions such as non-invasive glucose monitoring, but the single focus of the EDOCAL project is to create an early cancer detection tool, the Photonic Cancer Detector (PCD).
The PCD will have the following unique characteristics:
(a) Contact free, high resolution, non-destructive laser based cancer detection imaging system.
(b) Accurate and conclusive detection of many cancers in organic tissue: (targeted: sensitivity above 80 % and specificity above 80 %). Sensitivity is the more important of these two parameters. A sensitivity of 100 % means that the test recognises all people with the cancer as having it. A high sensitivity is vital when early diagnosis and treatment is crucial. A low sensitivity means that many people who have the disease go undetected. A specificity of 100 % means that all healthy people are recognised as being healthy. A high specificity is important when the treatment or diagnosis can be harmful to the patient. A low specificity means that people who do not have the disease are diagnosed (and treated) as having it.
(c) Wide angle and location (1 x 1 mm2) specific detection.
(d) Real time: the difference between healthy and cancerous tissue will instantly light up, providing a real-time diagnostic tool for use during intervention. Surgeons can track progress during surgery.
(e) Portable (size and weight of a mobile phone) and robust solution.
(f) Easy to use by non-trained medical experts.
(g) Economical solution: target price less than EUR 12 500.
(h) Standalone use or endoscopic integration.
We see two different product executions for the PCD.
(1) A standalone tool for early detection / surgical intervention. Target users: medical specialists, GPs, nurses, dentists, research and development (R&D) groups, institutes and universities.
(2) Endoscope integration. Target users: medical specialists.
In this project we focus on the standalone prototype device.
The PCD as standalone tool can be applied to detect up to 10 % of all European cancer cases and can potentially be useful in up to 80 % of all surgical interventions as it provides real-time, wide angle primary detection capabilities. Surgery is still the leading type of cancer treatment. For lung and blood related cancers we expect the PCD to have no use. When integrated in an endoscope it can be used to detect and intervene in about 27.5 % of all cancer cases.
The PCD can also be used as a first screening tool for GPs, trained nurses or dentists. This will provide them with a reliable, and fast, detection tool that does not exist today. Benefits will be instant: conclusive diagnosis of oral cavity and skin cancer (6 % of all cancer cases), resulting in earlier stage intervention, enhancing the success rate of cancer treatment and reducing hospital stays.
In summary, the main objectives of this project are:
(a) To develop a breakthrough contact free, high resolution tool for early cancer detection with a targeted sensitivity above 80 % and specificity above 80 %. The solution will support both wide angle and location specific (1x1mm2) view, and provide real-time information that can be used for red flag support during medical procedures.
(b) To validate the use of the early cancer detection tool in a clinical setting, optimising the original concept where appropriate.
(c) To generate scientifically supported information about possible relationships between the sensitiser complexes and their preference for attaching to specific cancerous areas.
(d) To prepare and gather the initial clinical data to be used in the medical regulatory approval phase, building on our current expertise for medical device certification of red laser applications.
Project results:
The main focus of the EDOCAL project was to medically validate the use of the first prototype based on the dynamic blue technology for the early detection of cancer. To do this four work packages (WPs) were defined as follows:
(1) specification and test systems;
(2) fluorescence spectroscopy in model biological systems and cell lines;
(3) analysis of biopsy and endoscopic mucosal resection (EMR) specimens from patients with BE;
(4) device architecture.
WP3: Specification and test systems
The purpose of this task was to define and build a tuneable fluorescence spectroscopy system. The excitation wavelengths should be tuneable within a wavelength range of 365 - 450 nm. Light was directed to the sample via a bifurcated optical fibre. The fluorescence emission was collected by the same fibre and spectrally resolved using a diode array spectrometer. The emitted signal was analysed using a laptop computer with commercial software. The system was tested using solutions of a photosensitiser, and also using photosensitiser-treated cell lines. The specification for the first system was defined by the RTDs and this first system (called dynamic blue system) was built by the SMEs and delivered to the RTDs for testing.
At NWH the focus was on in vitro fluorescence measurements of porphyrins in model biological systems; the test environment was designed and constructed incorporating the dynamic blue system to investigate the fluorescence in screening both photosensitiser solutions and different cell lines.
The prototype was connected to a shutter by a 400 / 440 fibre and the excitation signal was transported to the sample via a bifurcated fibre. The bifurcated fibre collected and delivered the fluorescence emission signal from the sample to the connected diode array spectrometer (470 - 700 nm) and computer for spectral analysis and dissemination. The distal end of the bifurcated fibre was connected to a probe (6 cm in length) coated in plastic via a subminiature version A (SMA) connector. The fluorescence in photosensitiser solutions and different cell lines was measured, for example, with the probe positioned perpendicular to a monolayer of cells.
Porphyrins have a high fluorescence quantum yield and are intermediate molecules in the biosynthesis of haem. Normally this pathway is kept under tight negative feedback control and porphyrins are not produced in excess. Porphyrins are synthesised from the precursor 5-aminolaevulinic acid (ALA), which when added to cultured cells can bypass the feedback mechanism of haem synthesis, resulting in the overproduction of proto-porphyrin - the penultimate molecule in the pathway. Methyl aminolaevulinic (MAL) acid is an esterfied analogue of ALA. To characterise the test set-up, the photo-sensitiser PpIX was dissolved in dimethyl sulfoxide (DMSO) and the pro-drugs ALA and MAL were added to cultured skin keratinocytes and osophageal adenocarcinoma (CA) cells.
Methodology
Human oesophageal OE19 CA cells were maintained in RPMI 1640 medium containing 10 % foetal calf serum; human HaCaT keratinocytes were maintained in DMEM medium containing 5 % foetal calf serum. Cells were grown at 37 degrees of Celsius in a humidified atmosphere of 95 % air: 5 % carbon dioxide (CO2). For experiments, cells were seeded into 60 mm tissue culture dishes, 24 hours before addition of the pro-drugs ALA and MAL. DMSO containing or not pro-drugs was added to the media and the cells were incubated for 4 or 24 hours before a fluorescence reading was taken.
The stock solution of PpIX (Sigma, Dorset, United Kingdom (UK)) was made by dissolving the powder in DMSO. A working solution of PpIX (5 µM) was prepared and wrapped in tin foil. ALA and MAL (Sigma, Dorset, UK) were prepared in DMSO immediately before use. The concentration of DMSO in culture media was maintained at or below 0.5 % (v / v). Solutions were wrapped in tin foil and all experiments carried out under subdued lighting in a specially adapted photobiology laboratory.
PpIX solutions and intracellular PpIX were excited with blue light (368 - 446.5 nm) and fluorescence emission was detected using a diode array spectrometer.
The test system was capable of delivering and detecting PpIX fluorescence in solution. The fluorescence emission maxima of PpIX in DMSO, as determined by the test system and by a F-2500 fluorescence spectrophotometer (Hitachi, UK) were compared. The results showed that both instruments were able to detect PpIX in solution following excitation with different wavelengths (368, 403, 408.5 416.3 417 and 446.5 nm). There was a difference in the peak wavelengths according to the two different systems and, also, the wavelength shift was not consistent.
The test set-up was also capable of delivering and detecting PpIX fluorescence in different cell lines. Both HaCaT keratinocytes and oesophageal OE19 CA cells effectively converted pro-drug to porphyrin. The porphyrin was identified as proto-porphyrin by high-performance liquid chromatography (HPLC) analysis. Further to this, conversion to PpIX increased over a 24-hour period, which is supported by previous work undertaken by us using HPLC analysis. Differences in peak intensity, as shown by peak height, were observed by different excitation wavelengths (403 and 408.5 nm) in ALA-treated OE19 cells. This result could be due to variations in the output power from each diode. As expected, no PpIX fluorescence was detected in the control OE19 cells but PpIX fluorescence was detected in ALA- and MAL-treated OE19 cells. This result confirms the findings from other instruments and that previously noted in the literature.
The specification for the system to be used at AMC was identical to that for NWH, however the used was different. The AMC focussed on ex-vivo measurements on endoscopic resection specimens for the characterisation of specific tissue (auto)fluorescence. Therefore, a set-up had to be developed to investigate freshly obtained EMR specimens. To closely resemble an in-vivo setting, a fibre-optic probe was specially designed to deliver the excitation signal to the tissue through the working channel of the endoscope and to carry back the fluorescence signal to the connected spectrometer and computer for spectral analysis. The endoscope, probe and spectrometer were fixed on a bench and the EMR specimen was placed in an oblique angle below the tip of the probe, thus mimicking the in vivo endoscopic setting.
Conclusions
The test set-up was tested successfully. The system demonstrated that it was capable of reproducibly delivering and detecting a fluorescent signal derived from either proto-porphyrin IX solutions or pro-drug induced cell lines. The findings were consistent with those found using other instruments and also reported in the literature.
Feedback from NWH on optimisation that was required for the system was discussed with the SMEs and a second improved system was designed and built. Although the system required optimisation, the characterisation of the viability of the source to produce adequate fluorescence signals was successfully achieved and the system could be used to carry out the work of WP4.
The ex vivo set-up was built, tested and optimised, after which preliminary measurements were performed. The excitation signal was adequately delivered to the targeted tissue and relevant fluorescence spectra were obtained. Although limitations to the laser system warranted further optimisation of the excitation hardware, the actual ex-vivo test set-up proved to be adequate to produce ex vivo fluorescence spectra as required for WP5.
The in vivo set-up was developed based on experience with the ex-vivo system and was adapted from bench to bedside for actual in-patient endoscopic fluorescence spectroscopy. In-vivo measurements were performed and proved to achieve relevant fluorescence spectra. The laser prototype needed further optimisation but the characterisation of prerequisites for adequate in-vivo fluorescence spectroscopy was successfully achieved and the system proved to be sufficient for use in the in vivo tests in WP5.
WP4: Fluorescence spectroscopy in model biological systems and cell lines
The aim of WP4 was to study factors that influence differences in fluorescence intensity and determine the optimum excitation conditions to allow differentiation between normal and cancerous cell lines. This work was led by NWH and has two main areas of investigation:
(a) in vitro fluorescence spectroscopy of porphyrins in model biological systems;
(b) investigation of fluorescence induced by photonic crystal laser in cell lines.
Methods
For this work the second version of the dynamic blue system was used, together with the Hitachi U3010 double beam absorbance spectrophotometer and the Hitachi F2500 fluorescence spectrophotometer fitted with a red-sensitive photomultiplier. Matched quartz cuvettes were used throughout, and all work was carried out in a specially adapted photobiology laboratory with ambient light levels below 1 lux. For the spectrophotometer work, excitation wavelengths encompassing the range of the dynamic blue device were used. The optical radiation characterisation was performed using a Bentham spectroradiometer system and the absolute radiant power was measured using an Ophir laser power sensor connected to a Nova II metre.
Key results
The photo-sensitiser environment influenced the absorption peak maximum and shape for both PpIX and Foscan. In terms of fluorescence, 1:75 liposomal solutions of PpIX most closely reflected the data obtained from ALA-treated normal human skin. The fluorescence intensity of photo-sensitisers presented to NW2 - the second dynamic blue iteration - device in different environments was in agreement with data previously published using traditional spectrophotomeric means. The power adjusted measurements of PpIX concentration taken with NW2 over the range 0.1 - 1.0 µM (consistent with concentrations found in human skin following application of ALA) were linear and in line with measurements made using the Hitachi F2500 spectrophotometer. The limits of quantifiable detection (S:N > 10) of PpIX and Foscan fluorescence by NW2 were comparable to the Hitachi F2500 (< 0.1 µM). NW2 was actually more effective at detecting Foscan fluorescence in PBS compared to the Hitachi F2500 (fluorescence not detected). The precision of measurements of photo-sensitisers presented to NW2 in different environments was good as indicated by the coefficient of variation (%CV). The %CV of the fluorescence maxima was generally in the region of 0.03 %. Peak intensity measurement was least precise in solutions that encouraged aggregation of the photo-sensitisers, and most repeatable in liposome suspension at less than 5 %.
In the biologically mimicking solutions of bovine serum albumin (BSA), fetal bovine serum (FCS) and liposomes, tuning the excitation wavelength of PpIX from 400 - 410 nm shifted the emission peak maximum by 0.2 0.5 and 0.3 nm respectively. For the chlorin these shifts were smaller (0.2 0.2 and 0.1 nm respectively). Thus the porphyrin seemed to be more affected by tuning the excitation wavelength than Foscan (Refer to D4.1 and 4.M4). In human skin, fluorescence from endogenous chromophores (auto-fluorescence; AF) shifts by 8.7 nm from 513.9 nm to 522.6 nm when the excitation wavelength was changed from 400 to 410 nm.
Cell-specific differences in fluorescence were observed between cell lines derived from the epidermis (HaCaT), oesophagus (OE19), brain (SHSY5Y) and bladder (HT1197). All cell types were able to effectively convert ALA or MAL to PpIX over time with the more transformed the cell phenotype accumulating more cell-associated PpIX. More porphyrins consistently accumulated in OE19 or HT1197 cells than in HaCaT or SHSY5Y cells. The majority of porphyrin was PpIX, except in OE19 media, where coproporphyrin (4-COOH) was the most dominant after 24 hours. Spectroscopic analysis (I620/I635) using the NW3 the third Dynamic Blue iteration - device illustrated a prominent 620 nm peak in OE19 cell media, which could be 4-COOH. This 620 nm peak was associated with some (OE19, HT1197) but not all (SHSY5Y, HaCaT) tissue cell types. Factors that influence the differences observed between these cell lines were investigated, in particular, PpIX efflux and haem synthesis. The present results do not suggest a major impact of the porphyrin transporter protein within the clinically relevant incubation time used in this study. Distinct differences in fluorescence emission at 620 and 635 nm may thus indicate progression towards a more cancerous cell type. All excitation wavelengths were able to induce adequate fluorescence signals, except the 367 nm light-emitting diode (LED). PpIX fluorescence peak intensity shifts were observed between 4 and 24 hours, with all cell lines; the highest in OE19 and HT1197 cells. These cell lines also accumulated the highest cell associated concentrations of PpIX. Emission wavelength shifts, however, where not detected between excitation wavelengths or between cell lines. Finally, auto-fluorescence from cells could not be observed.
Blue visible light up to a dose of 5 J / cm2 was not in itself phototoxic or photogenotoxic to cell types on its own when compared to baseline (dark) values. Exposure of ALA treated cells to light resulted in a small but significant increase in DNA strand breaks, however no effects on cell survival were subsequently observed 24 hours later in the phototoxicity assay. The doses that were used encompassed and exceeded those that might be delivered to tissue by NW3 in its current form, and changes started to be seen in the gel electrophoresis assay at doses of 80 mJ / cm2. Non-light exposed cells did not exhibit any toxicity. The data confirm the desired specification for a device emitting no more than 1 mW / cm2 or higher. It is also notable that no measurable changes were seen in the phototoxicity assay, (which is particularly vulnerable to damage to the cell membranes) or in the morphology of the cells. Therefore the changes resulting in the small increase in deoxyribonucleic acid (DNA) migration did not manifest as a measurable change in cell viability 24 hours later.
Skin measurements
Methodology
5-aminolaevulinic acid (ALA 20 % w / w) (Mandeville Medicines, UK) was prepared on aluminium Finn chambers on Scanpor (8 mm) (Smart Practice, USA). ALA was applied to the inner surface of the forearm of a healthy volunteer. The application site was covered with Tegaderm film (3M Healthcare, Germany) and a Mepore (M?lnlycke Healthcare, Sweden) dressing for 6 hours prior to removal and any excess cream was wiped off. PpIX fluorescence readings with NW2 were recorded by placing the probe perpendicularly in contact with the skin. Furthermore, additional fluorescence readings were recorded as controls from the normal skin located approximately 10 cm distal to the application site. Each measurement was performed in specially adapted photobiology laboratory with ambient light levels below 1 lux.
Summary of the work
The spectrum for PpIX in organic solvents was typical of that for porphyrin molecules with a Soret band (max 400 - 407 nm) and four subsequent smaller Q bands as shown in the literature with the position of the Soret band, but not the Q band positions, highly dependent upon the solvent used. These spectra represent monomeric solutions. However, in an aqueous solution, the Soret band position shifts to a shorter wavelength alongside the broadening of all absorption bands and a strong decrease in peak intensity. These observations are indicative of a highly aggregated solution. In FCS and BSA solutions, the Soret band shift to a shorter wavelength remains but not to the same extent as PpIX in an aqueous solution, and the 4 Q bands remain more consistent with that of PpIX in an organic solvent which indicates a more monomeric solution than that of PpIX in the aqueous environment. In liposomes, the Soret band and Q bands are in a similar position to that of PpIX in organic solutions along with sharper absorption bands compared to PpIX in FCS or BSA and this is due to further monomerisation of PpIX in this lipid environment. Therefore, the absorption properties of a PpIX solution are highly dependent upon aggregation state. The trends observed in the 25 µM solutions were also observed in the 5 µM, thus trends were not dependent upon PpIX concentration. Generally, peak intensity of the Soret band increased in a positive, linear relationship with increasing PpIX concentration. The Soret band shift to shorter wavelengths at high PpIX concentrations in solutions of BSA and FCS may be due to a mixture of monomeric and aggregated species. PpIX was most soluble in DMSO, followed by BSA, serum and MeOH, but had limited solubility in EtOH and acetone.
The spectrum for Foscan in ethanol was typical of that of chlorin type molecules with a Soret band (max 416 nm) and a strong absorbance in the red region (max 650 nm). In aqueous solutions, unlike that of PpIX, the Soret and Q band position shifts to a longer wavelength which has been previously reported in the literature and is indicative of an aggregated species with this group of molecules. The aggregation of Foscan in PBS was further supported by peak broadening and the strong decrease in peak intensity of all absorption peaks. In FCS, BSA and, in particular, the liposome solutions, monomerisation was evident with a blue shift in Soret and Q band positions alongside peak sharpening when compared to the aqueous solutions. Again, the absorption properties are highly dependent upon the aggregation state of the Foscan molecules.
Conclusions
The environment influenced the absorption peak maximum and shape of both PpIX and Foscan. The PpIX Soret band maximum shifted from 362.5 nm in aqueous buffer to 410.5 nm in liposome suspension (a red shift of 48 nm). For Foscan there was a blue shift of 14 nm from 434 nm to 420 nm. In terms of fluorescence, 1:75 liposomal solutions of PpIX most closely reflected the data obtained from ALA-treated normal human skin. The fluorescence intensity of photo-sensitisers presented to NW2 in different environments was in agreement with data previously published using traditional spectrophotomeric means. The blue shift in PBS, compared to the other conditions where Foscan is more monodispersed, is indicative of the presence, amongst larger aggregates, of H-type (or face-to-face) oligomers.
The power adjusted measurements of PpIX concentration taken with NW2 over the range 0.1 - 1 µM (consistent with concentrations found in human skin following application of ALA) were linear and in line with measurements made using the Hitachi F2500 spectrophotometer. This is encouraging for the possibility of using the laser device for quantification of photo-sensitiser in tissue.
The limits of quantifiable detection (S:N > 10) of PpIX and Foscan fluorescence by NW2 were comparable to the Hitachi F2500 (< 0.1 µM). NW2 was actually more effective at detecting Foscan fluorescence in PBS compared to the Hitachi F2500 (fluorescence not detected). This may be due to a combination of low quantum yield, due to the formation of aggregates, and large inner filter effect due to the concentration of the sample. The precision of measurements of photo-sensitisers presented to NW2 in different environments was good as indicated by the %C V. The %CV of the fluorescence maxima was generally in the region of 0.03 %. Peak intensity measurement was least precise in solutions that encouraged aggregation of the photo-sensitisers, and most repeatable in liposome suspension at less than 5 %.
In the biologically mimicking solutions of BSA, FCS and liposomes, tuning the excitation wavelength of PpIX from 400-410 nm shifted the emission peak maximum by 0.2 0.5 and 0.3 nm respectively. For the chlorin these shifts were smaller (0.2 0.2 and 0.1 nm respectively). Thus the porphyrin seemed to be more affected by tuning the excitation wavelength than Foscan.
In human skin, fluorescence from endogenous chromophores (auto-fluorescence; AF) blue shifted by 8.7 nm from 513.9 nm to 522.6 nm when the excitation wavelength was changed from 400 to 410 nm. The peak maximum of PpIX did not change. As the endogenous fluorescence emission reflects the nature of the fluorophore, the combination of AF peak maxima, intensity and PpIX peak maxima and intensity may be a powerful tool in differentiating tissue staging in cancer and other diseases.
Results from screening the 100 samples
The spectrum for PpIX in organic solvents was typical of that for porphyrin molecules with a Soret band (max 400 - 407 nm) and four subsequent smaller Q bands as shown in the literature with the position of the Soret band, but not the Q band positions, highly dependent upon the solvent used. These spectra represent monomeric solutions. However, in an aqueous solution, the Soret band position shifts to a shorter wavelength alongside the broadening of all absorption bands and a strong decrease in peak intensity. PpIX was most soluble in DMSO, followed by MeOH, but had limited solubility in EtOH and acetone.
The spectrum for Foscan in ethanol was typical of that of chlorin type molecules with a Soret band (max 416 nm) and a strong absorbance in the red region (max 650 nm). In aqueous solutions, unlike that of PpIX, the Soret and Q band position shifts to a longer wavelength which has been previously reported in the literature and is indicative of an aggregated species with this group of molecules. The aggregation of Foscan in PBS was further supported by peak broadening and the strong decrease in peak intensity of all absorption peaks.
The environment influenced the absorption peak maximum and shape of both PpIX and Foscan. The PpIX Soret band maximum red-shifted by 44.5 nm when the solution was changed from aqueous buffer to DMSO. Foscan blue-shifted by 17.5 nm. The solubility of Foscan in buffer was very poor. The blue shift in PBS, compared to the other conditions where Foscan is more monodispersed, is indicative of the presence, amongst larger aggregates, of H-type (or face-to-face) oligomers.
The power adjusted measurements of PpIX concentration taken with NW2 over the range 0.1 - 1.0 µM (consistent with concentrations found in human skin following application of ALA) were linear and in line with measurements made using the Hitachi F2500 spectrophotometer. This is encouraging for the possibility of using the laser device for quantification of photo-sensitiser in tissue.
NW2 was actually more effective at detecting Foscan fluorescence in PBS compared to the Hitachi F2500 (fluorescence not detected). This is likely due to a combination of low quantum yield, due to the formation of aggregates, and large inner filter effect due to the concentration of the sample.
The precision of measurements of photo-sensitisers presented to NW2 in different environments was good as indicated by the %CV. The %CV of the fluorescence maxima was generally in the region of 0.03 %. Peak intensity measurement was least precise in solutions that encouraged aggregation of the photosensitisers (PBS & MeOH in the case of PpIX; PBS in the case of Foscan). PpIX was barely soluble in acetone so these data are not shown. The best solutions for reproducibility were DMSO (PpIX) and EtOH (Foscan).
In the solvents DMSO, MeOH and PBS, tuning the excitation wavelength of PpIX from 400-410 nm shifted the emission peak maximum by 0.0 0.1 and 0.5 nm respectively. For the chlorin the shifts in EtOH & PBS were 0.0 and 0.3 nm respectively. Thus the porphyrin seemed to be more affected by tuning the excitation wavelength than Foscan.
WP5: Analysis of biopsy and EMR specimens from patients with BE
This WP was led by AMC and investigated the following:
(a) In vitro fluorescence microscopy and spectroscopy of biopsy specimens from patients with Barrett's esophagus and high grade dysplasia or early mucosal cancer before and after oral administration of 5-ALA.
(b) 5-ALA induced fluorescence imaging of early Barretts cancers using ex vivo spectroscopy measurements of EMR specimens.
In patients with Barrett's oesophagus (BO), the normal epithelial lining is replaced by columnar lined epithelium, so called intestinal metaplasia. Malignant degeneration may occur in this intestinal metaplasia, through a series of precancerous changes graded on histology: low-grade intraepithelial neoplasia (LGIN), high-grade intraepithelial neoplasia (HGIN) and intramucosal carcinoma or early CA. When detected at this premalignant stage, these dysplastic areas can be curatively treated with minimally invasive techniques, such as endoscopic resection. Therefore, regular endoscopic surveillance is advised in order to detect dysplastic areas at an early stage. However, these areas are usually flat and do not stand out from the non-dysplastic Barretts epithelium. Advanced imaging methods are thus required to increase the detection rate of premalignant lesions in BO. One of the most promising approaches is the use of fluorescence spectroscopy. Different tissue types have distinct natural fluorescence characteristics, which can be detected by exciting the tissue with a certain wavelength of light and measuring the resulting fluorescent light. Also changes within a tissue type can be detected, such as dysplastic mucosa in Barrett's epithelium. Endogenous tissue fluorescence is called auto-fluorescence.
The work focused on the fluorescence characteristics of Barrett's epithelium, either in vivo or assessed in biopsies and endoscopic resection specimens. This work aimed to assess the optimal excitation wavelength(s) to induce (auto)fluorescence, and to study the shift of the excitation and fluorescence spectra which may be associated with changing tissue characteristics. Fluorescence spectroscopy may be used to distinguish non-dysplastic from dysplastic Barrett's tissue in the blue light (ultraviolet / visual) wavelength range. A widely used diagnostic fluorescent agent for in the gastrointestinal tract is 5-aminolevulinic-acid (5-ALA) induced proto-porphyrin-IX (PpIX). The endogenous, photoactive PpIX is an immediate precursor of heme and is converted from 5-ALA intracellularly. This conversion depends on metabolic activity of the cells and hence will be different between different cell types. After administration of excess exogenous 5-ALA, PpIX selectively accumulates in malignant tissue. PpIX induced fluorescence thus can be used to enhance the diagnostic value of fluorescence spectroscopy by increasing the signal (malignant tissue) to noise (background normal tissue) ratio. PpIX will reach peak concentration approximately 4 - 6 hours after oral administration. 5-ALA was used to enhance the discriminating characteristics of the fluorescence induced by the dynamic blue prototype. Fluorescence spectroscopy on biopsies was performed before and after 5-ALA administration.
The other focus of the AMC was in-vivo endoscopic fluorescence spectroscopy. For this purpose, a specially designed optical fibre probe is equipped with 2 arms one is connected to the spectrometer and the other one to the prototype light source box. The whole system, including the computer with the laser driver and spectral analysis software, was mounted on a modified trolley, in order to be able to move closely to the patient and the endoscopist. A small plastic cap was placed on the tip of the endoscope to ensure stable positioning of the endoscope and the probe on the tissue. The probe was subsequently moved through the working channel of the endoscope and placed on the tissue under direct vision. A transparent plastic cap was attached to the tip of the endoscope to ensure a stable position of the endoscope and the probe's tip on the oesophageal tissue.
Patients with and without dysplasia in a BO were treated with an EMR in accordance with current guidelines. In this ex-vivo experiment we tried to closely resemble the in-vivo situation during endoscopy: an optical fiber probe, designed to deliver the excitation light and to collect the emitted fluorescent light was connected to the dynamic blue prototype and a spectrometer and subsequently passed through the working channel of the endoscope onto the EMR specimen. Fluorescence spectroscopy was performed on areas suspicious for dysplasia and on adjacent inconspicuous areas on the specimen. This was immediately followed by histological sampling of a biopsy from the EMR specimen of the corresponding measurement site, to correlate the histological diagnosis to the measured fluorescence spectrum.
Patients with and without early esophageal neoplasia originating from a BO were included in this study. All patients underwent high-resolution white light endoscopy (WLE), auto-fluorescence endoscopy (AFI) and narrow band imaging (NBI). Multiple areas suspicious for dysplasia in the Barrett's segment, were investigated with in-vivo spectroscopy using the Dynamic Blue multi-wavelength prototype and a specially designed optical fiber that was passed through the working channel of the endoscope. A small transparent cap was used on the tip of the endoscope to ensure stable positioning of the probe and the correlation of the corresponding biopsies. Spectroscopy was immediately followed by histological sampling of the corresponding area. Also non-suspicious control areas were investigated and sampled. Auto-fluorescence spectra were recorded, using different excitation wavelengths (365, 395, 405, 410, 417, 445 nm). In addition, white light (WL) and background (bg) spectra were obtained to allow correction of the auto-fluorescence spectra.
After the spectroscopic evaluation, biopsies and EMR specimens were fixed in formalin, embedded in paraffin and routinely cut and stained with haematoxylin and eosin (H&E). Histopathological assessment of the biopsies was performed by an expert GI-pathologist, who recorded the presence of intestinal metaplasia and neoplasia according to the WHO-classification: no-dysplasia, indefinite for dysplasia, LGIN, HGIN or invasive cancer. In case of cancer: depth of tumour invasion, grade of differentiation, presence of lymph-vascular invasion was assessed, as well as the presence of neoplasia at the deep (vertical) resection margins.
Methods: Analysis of the autofluorescence spectra
Analysis of autofluorescence spectra (with white light correction)
Ex-vivo and in-vivo collected fluorescence spectra were analysed using two distinct methods: emission intensity analysis and spectral shape analysis. Fluorescence spectra obtained using a single excitation wavelength and combinations of spectra obtained using different excitation wavelengths were analysed to find the optimal (combination of) wavelength(s).
All auto-fluorescence spectra were corrected by the background signal and normalised with respect to the total intensity of the spectrum (area under the curve).
The intensity of emitted fluorescence as a function of excitation wavelength differs per tissue type and may therefore be used as a discriminating tool. Selected spectra corresponding to the same histopathologic category were averaged. The averaged spectra of non-dysplastic Barrett's (IM) were used as a baseline value. A baseline corrected spectrum was determined by subtracting the baseline spectrum from the spectra of HGIN / CA categories. Intensity ratios were selected according to the results of the baseline corrected spectrum. ROC curves and area under curve (AUC) were used to determine sensitivity and specificity. The statistical relevance of the intensity ratios was determined by two-tailed independent Student's t-test
Spectral shape analysis
The shape of a spectrum depends on the excitation wavelength and the tissue type. Spectral shape analysis was performed by calculating the first derivative and comparing the peak-to-peak deviation. In addition principal component analysis (PCA) was performed which is another method to identify patterns in the spectra to distinguish between non-dysplastic Barrett's (IM) and grouped HGIN and early CA. Intensity and spectral shape analysis results were compared and the more significant one was chosen for further analysis.
Advanced emission intensity analysis
White light correction
Reflectance spectra generated by the white light source were used to correct for the spectral distortion of the auto-fluorescence spectra due to scattering and absorption mainly from hemoglobin absorption. Both white light corrected and non-white light corrected spectra were analysed and compared.
Excitation - emission matrix (EEM)
To further optimise the emission intensity analysis an EEM was calculated for the non-white light corrected and for the white light corrected auto-fluorescence spectra. EEM visualise the intensities in a colour map. Differences in emission wavelengths in relationship to excitation wavelength and tissue type were visualised.
Double intensity ratio
Double intensity ratios of 2 emission spectra from the same location on the tissue obtained from 2 different excitations were used to further optimise the sensitivity and specificity. All combinations of different wavelength excitations were tested.
Results: Ex vivo spectroscopy in Barrett's patients with and without dysplasia.
11 EMRs were performed on seven patients with dysplasia in a BO. All patients underwent an EMR using the EMR-cap technique of both a neoplastic and a non-neoplastic area. The EMR-specimen was immediately retrieved from the patient and freshly pinned onto a paraffin block for direct ex-vivo examination. The dynamic blue prototype was activated (405 nm) and ex-vivo auto-fluorescence spectroscopy and corresponding histology sampling were subsequently performed on a total of 26 mucosal areas in Barretts mucosa in these 11 EMR specimens.
All obtained fluorescence spectra were first corrected by the background signal and then normalised with respect to the total intensity of the fluorescence spectrum (area under the curve). Selected spectra corresponding to the same histopathologic category were averaged. The averaged spectrum of non-dysplastic Barrett (IM) were used as a baseline value. A baseline corrected spectrum was determined by subtracting the baseline spectrum from the spectra of HGIN / EC categories. The resulting graph is deemed the baseline corrected spectrum.
The ex vivo measured auto-fluorescence spectra of normal Barrett's and HGIN / EC show differences in the relative intensity and in the spectral shape.
Results: In vivo spectroscopy in Barretts patients with and without dysplasia; phase 1
Patient and lesion characteristics
In phase 1 and 2 a total of 70 patients were included, in whom a total of 290 areas were investigated with in-vivo auto-fluorescence spectroscopy using the Dynamic Blue prototype. The results from intensity and shape analysis, the intensity ratios show a higher significance in discriminating non-dysplastic Barrett (IM) from grouped HGIN and early CA. The determined sensitivity and specificity however is not adequate for clinical decision making. Therefore advanced analysis was applied.
Conclusions on autofluorescence spectra (without white light correction)
For single excitation with 365nm the sensitivity is 100 % but the specificity (38 %) is too low. The combination of 365nm and 405nm excitation improves the specificity from 38 up to 50 %. Improvements to increase the specificity are however still needed.
The intensity ratios 560 / 640 are higher for all excitation wavelengths compared to the 495 / 560 ratios. At 365nm excitation the difference between non-dysplastic Barrett (IM) and grouped HGIN and early CA is highest. Analysis of combined excitation is done by calculating the double intensity ratio for the emission wavelength 560/640.
Conclusions on autofluorescence spectra (with white light correction)
For single excitation with 365nm the specificity is 100% but the sensitivity is with 67 % still too low. The combination of 365 and 405 nm excitation improves the sensitivity from 67 % up to 100 %.
Conclusions on phase 1
The developed algorithm combines 365 nm and 405 nm excitation in which the double intensity ratio at 560nm and 640 nm of white light corrected auto-fluorescence spectra is used. The 640nm red emission probably reflects the emitted fluorescence of porphyrin.
Results: In vivo spectroscopy in Barretts patients with and without dysplasia; phase 2
In phase 2 the developed algorithm of phase 1 was applied to a new set of data.
Analysis of auto-fluorescence spectra (without white light correction)
Auto-fluorescence spectra of 365nm and 405 nm are analysed with the intensity ratios 560 / 640 and 495 / 560. For single excitation 405 nm has a sensitivity of 75 % and a specificity of 89 % for the emission intensity ratios at 495nm and 560nm to distinguish between non-dysplastic Barrett's (IM) and grouped HGIN and early CA as at 365 nm excitation.
The results of the two tailed independent Students t-test gives a p-value of 0.002 which is a statistical significant difference of the double intensity ratios of 365 and 405 nm and may therefore be used to discriminate between non-dysplastic Barrett (IM) and grouped high-grade HGIN and early CA.
Conclusions on white light corrected auto-fluorescence spectra
For white light corrected auto-fluorescence spectra analysis on the new data set the combination of 365nm and 405nm does improve the sensitivity and specificity compared to single excitation. A sensitivity of 80% and a specificity of 81 % are reached with the double intensity ratios at 495nm and 560nm. The student t-test shows that there is a significant difference between non-dysplastic Barrett (IM) and grouped HGIN and early CA.
Conclusions on phase 2
In phase 2 the optimised method for data analysis developed in phase 1 was applied to a new set of data. The results confirm that the white light corrected auto-fluorescence spectra in combination with 365 and 405 nm excitation gives highest sensitivity and specificity. Contrary, the results in phase 2 suggest that double intensity ratio at 495 and 560 nm increases the specificity compared to the previous mentioned double intensity ratio of 560nm and 640nm.
Conclusions of phase 1
The work at the AMC Amsterdam has significantly expanded compared to the task originally described by performing ex-vivo spectroscopy on 26 sites in 11 EMR specimens from 7 Barrett's patients and 1 557 in vivo spectroscopy measurements on 290 sites in 70 BO patients during real-time endoscopy. A total of 77 patients were investigated in phase 1. In addition, these tests used the first generations dynamic blue prototype with a multi wavelength laser set-up. The first part of WP5 therefore provided much more data and information than originally scheduled. First, data were largely collected in-vivo instead of the artificial in-vitro setting. Second, data were collected from a much larger group of patients than originally planned, thus increasing the statistical power and the extrinsic validity of the project. Third, data were obtained using the dynamic blue prototype which more closely resembles the actual clinical context of the final product. Fourth, the first phase allowed technical, methodological, and statistical investigations that were not only imperative for the execution of WP5 of the EDOCAL project but also allowed technical improvements of the prototype beneficial to the dynamic blue project.
Conclusions on ex-vivo spectroscopy measurements on EMR specimens
A special set-up using the available technology from the dynamic blue project was designed to perform ex-vivo and in-vivo fluorescence spectroscopy measurements on Barretts mucosa. Ex vivo spectroscopy was performed on EMR specimens of patients with and without early neoplasia in BO. This allowed testing of the prototype in a more realistic way than would be possible if only tissue biopsies would have been obtained and provided information that was directly transferable to the further testing of the laser prototype during real-time endoscopy.
Conclusions on in-vivo measurements in Barrett patients
In-vivo auto-fluorescence spectroscopy was performed by passing a specially designed light probe through the working channel of an endoscope onto the mucosa during real-time endoscopy. Spectroscopy sites were selected by using high-resolution WLE, AFI and narrow-band imaging during specially scheduled endoscopy programs. Using the dynamic blue prototype, each site was illuminated with multiple wavelengths followed by registration of the corresponding auto-fluorescence spectrum and tissue biopsy for histological correlation. A total of 1 557 spectra were obtained from 290 sites in 70 patients.
Conclusions on auto-fluorescence spectroscopy with dynamic blue technology
The experiments conducted for D5.1 have shown that auto-fluorescence spectroscopy, using a combination of 365 and 405 nm excitation and double intensity ratios at 495nm and 560nm emission wavelength of white light corrected auto-fluorescence spectra, is able to discriminate early neoplasia (HGIN / EC) with 80 % sensitivity and 81 % specificity from non-dysplastic Barrett epithelium.
We hypothesised that these parameters could be further improved by adding a photosensitiser. Therefore, in D5.2 5-ALA induced fluorescence as a function of different excitation wavelengths will be assessed before and after the administration of 5-ALA. Dynamic Blue test set-up for D5.2
Specification of the test set-up for 5-ALA induced fluorescence experiments
In D5.2 the AMC focuses on investigating 5-ALA induced PpIX fluorescence using a simulated in-vivo model. To closely resemble the in-vivo setting, the CAM model was developed on which freshly obtained BO biopsy specimens and oesophageal CA cells were transplanted.
In D5.2 the same Dynamic Blue prototype set-up was used as during the in-vivo auto-fluorescence measurements in patients with BO. An optical fibre probe was specially designed to deliver the excitation signal to the tissue and to carry back the fluorescence signal to the connected spectrometer and computer for spectral analysis. The specimen was placed below the tip of the probe, thus mimicking the in-vivo endoscopic setting.
Contact details: Ms Gillian Mimnagh
CEO, 2M Engineering LTD, De Run 4352
5503LN Veldhoven
The Netherlands
List of websites: http://www.edocal.eu
Early detection of cancer saves lives, improves quality of life and reduces health care cost. In Europe there are 3 million new cancer cases every year. According to the World Health Organisation (WHO) the burden of cancer can be reduced by systematic and equitable implementation of evidence based strategies for cancer prevention, early detection and management of patients with cancer. 33 % of the cancer burden could be reduced if detected early and treated adequately. The WHO concludes that early diagnosis of cancers could save the developed world an average of EUR 25 billion per year on medical costs and EUR 50 billion per year on lost productivity.
Early-cancer diagnosis today consists of recognition of some of the early signs by the patient themselves or detection from screenings carried out on specific population groups. If something suspicious is found, the next step is to perform a blood analysis and/or to look with non-invasive imaging techniques. Current imaging techniques are widely used in diagnostic procedures but they cannot give conclusive evidence of cancer. Today, only a biopsy and the subsequent pathologist interpretation can give a definitive diagnosis of cancer and this is not always sufficient. In the EDOCAL project the proof of concept has been achieved for early detection of cancer based on blue laser technology. The technology has been tested and proven for oesophageal cancer detection and has the potential for early detection of many other types of cancer including stomach and colorectal. The technical principle behind the EDOCAL concept is as follows.
In cancer, tissue cells tend to create additional blood vessels to support their growth. The molecule protoporphyrin (PpIX) is generally present in blood vessels and exhibits red fluorescence when excited with blue light in the 375 - 425 nm range. Cancer cells are detected by observing the red fluorescence at a matching excitation wavelength. The wavelength of the laser has to be tuneable as the matching wavelength depends on the person, type of illness and presence of other chemicals. To increase the intensity of the red fluorescence, a porphyrin prodrug can be administered either orally or locally which synthesises PpIX in the cancerous (cells) tissue. By using a tuneable laser system and combination of endogenous fluorescence (as a result of extra blood) and exogenous fluorescence (produced by administration of a porphyrin prodrug), it is possible to accurately distinguish between normal and cancerous tissue.
The results of the EDOCAL project can be summarised as follows:
(a) A tuneable blue laser prototype was defined and the system built for research purposes.
(b) Experiments to detect cancer using shifts in excitation and emission wavelengths were carried out in 70 patients and we obtained a total of 1 557 spectra from 290 sites in the oesophagus. Using the prototype, each site was illuminated with multiple wavelengths followed by registration of the corresponding auto-fluorescence spectrum and tissue biopsy for histological correlation.
(c) Results from the EDOCAL prototype show that it is possible to discriminate early cancer with 80 % sensitivity and 81 % specificity, which meets the performance requirements targeted at the start of the EDOCAL project.
(d) Proof of concept has been achieved for early detection of cancer based on a tuneable blue laser technology.
(e) Preliminary specifications for the first demonstrator have been defined.
(f) The first patent application has been filed and new filings for the latest results are being worked on.
The final goal of the small and medium-sized enterprises (SMEs) is the creation of a device that can be used for first line cancer detection. This device will reduce the need for biopsies, make the diagnosis more reliable, less dependent on human interpretation, make early cancer detection available to more people and give the SMEs access to a new and quickly growing market: reliable, life-saving medical diagnostics. The EDOCAL project combined state-of-the-art laser and imaging technology with leading medical research, expertise and hands on practical medical treatment experience to validate the first prototype and investigate whether or not it not it would be effective in discriminating early cancerous tissue from healthy tissue. The project has succeeded in establishing proof of concept for oesophageal cancer detection. The technique was validated using cell lines in vitro before applying the procedure to patients. It is our intention to investigate the use the same principle method to diagnose and treat other cancer types and to use the results of this project to create a demonstrator that can be tested in the field.
Project context and objectives:
Globally there are 24.6 million people living with cancer. By 2020 there will be over 30 million. Cancer is a leading cause of death worldwide. From a total of 58 million deaths worldwide in 2005, cancer accounted for 7.6 million (or 13 %). Deaths from cancer in the world are projected to continue rising, with an estimated 9 million people dying from cancer in 2015 and 11.4 million dying in 2030.
In Europe there are 3 million new cancer cases per year (38 countries), with 2 million new cases in the EU-25 alone. This represents 340 new cases / 100 000 people per year. One in three men and one in four women will be inflicted with cancer during their life. There is no doubt about the fact that cancer is an increasingly important factor in the global burden of disease with tremendous economic impact.
The United State (US) National Institute of Health (NIH) estimated the overall annual costs for cancer in 2000 at EUR 112 billion in the US alone, with EUR 38 billion for direct medical costs and the rest as a result of lost productivity. These figures can, in their opinion be doubled to cover the rest of the developed world. The total global cost of cancer in 2000 was EUR 150 billion. The figure for direct medical costs alone per patient in Europe is EUR 31 000 and rising.
According to the WHO the burden of cancer can be reduced by systematic and equitable implementation of evidence based strategies for cancer prevention, early detection and management of patients with cancer. Up to 33 % of the cancer burden could be reduced by implementing preventative strategies aimed at reducing the exposure to cancer risks. Another 33 % of the cancer burden could be reduced if detected early and treated adequately. The WHO concludes that, early diagnosis of cancers could save the developed world an average of EUR 25 billion per year on medical costs and USD 50 billion per year on realised productivity. The direct savings on medical treatment alone would be more than EUR 10 000 per patient if early detection was possible.
The purpose of the EDOCAL project is to create a breakthrough tool for early cancer detection by combining state of the art laser and imaging technology with leading medical research (expertise and hands on practical medical treatment experience). We expect that the results from the medical research will enable the SMEs to create a device that can be used for first line cancer detection, making early cancer detection more accessible to more people all over the world. This will create a new mass market for the products created by the SMEs and will ensure for them a position in a new and quickly growing market: Cheap, reliable, medical diagnostics for everyone.
In spite of the huge increase in certain diseases attributed to increased prosperity, growth in product sales (e.g. endoscopes) for treatment of these diseases has not happened at the same rate. Reports by Frost & Sullivan indicate the main reasons for the slow growth to be: lack of new and technically exciting products, the high cost associated with initial purchase of endoscopes and resistance to new technologies. EDOCAL aims to provide the innovative breakthrough, at reasonable costs by combining advanced, proven and low cost telecom and state of the art semiconductor technologies with the latest advances and insights into medical procedures provided by the research and technological development (RTD) partners. EDOCAL also aims to enable the European companies to differentiate themselves from the high-end segment players in Japan and the US through the use of innovative features and will increase the technology lead of Europe compared to the low-cost original equipment manufacturer (OEM) players from China and India.
Early cancer diagnosis today consists of recognition of some of the 'early' signs by the patient themselves or detection from screenings carried out on specific population groups. Typical signs are: lumps, sores, persistent indigestion, persistent coughing and bleeding from the body orifices. If something suspicious is found the next step is to perform a blood analysis and/or to look with non-invasive imaging techniques. Current imaging techniques are widely used in diagnostic procedures but they cannot give conclusive evidence of cancer. Today, only a biopsy and the subsequent pathologists interpretation can give a definitive diagnosis of cancer and even sometimes this is also insufficient. A biopsy is a procedure to remove a piece of tissue or a sample of cells from the body so that it can be analysed in a laboratory. There are five different types of biopsy procedures depending on type of cancer suspected and the location of the suspicious cells.
The main problems with current cancer diagnostic techniques and procedures are:
(1) The lack of early primary detection methods: Early detection depends on the ability to discriminate healthy from (potentially) cancerous tissue, before physical symptoms like sores and ulcers start to appear. In the early stages when the impact of many cancers is still minimal, it is difficult and often impossible to conclusively discriminate healthy from malignant tissue, leading to incorrect / inconclusive results.
(2) High costs due to late interventions: When cancer is not detected in an early stage, complicated surgical interventions, chemotherapy and radiotherapy with limited long term success rates become necessary. The direct and indirect costs are huge when compared to the costs associated with an early diagnosis and the resulting intervention which are sufficient if the cancer has not had a chance to mature and spread. For example, early diagnosis and treatment of gastro intestinal (e.g. Barrett's) cancer has a high success rate via endoscopic intervention. However when detected late, surgery is always necessary and chances of success are low. At this stage there is a mortality rate of 3 - 5 % during surgery and a survival rate of only 20 % after 5 years. As stated earlier, direct savings per patient per year would be approximately EUR 10 000. In the European Union (EU) alone 1.48 million new cancer cases are handled, every year, which adds up to a saving in direct cost of about EUR 15 billion per year due to intervention at an earlier stage.
(3) Sensitivity and specificity of existing imaging techniques are too low. Magnetic resonance imaging (MRI) and computed tomography (CT) scans when used for the types of cancer that are most difficult to detect, have a sensitivity and a specificity of roughly 40 - 70 % and early neoplastic lesions are often too small to be detected by these scans. Sensitivity and specificity of existing endoscopic procedures is roughly 50 - 80 %. This means that there is insufficient contrast between the healthy and cancerous tissue, making it impossible to make a reliable diagnosis. Therefore a biopsy with all of its drawbacks is still the golden standard for conclusive cancer diagnosis.
(4) Non-existence of reliable red flag tools: Typically the sensitivity of existing tools is between 60 - 70 %, which is not reliable enough to be of much use in daily medical practice. Today no online surgical tools exist that can clearly and immediately discriminate healthy from cancerous tissue over a large surface area. Oncologists agree that a wide angled, easy to use, real-time detection tool small enough to be integrated into an endoscope would be an enormous help to them, showing them where to intervene and allowing them to track their progress. Such a tool does not exist today.
(5) The conclusive biopsy lab analysis is slow: Typically is takes about 1 week before results of the biopsy are available and another 1 - 2 weeks before the next treatment step can be taken or appointments scheduled. Furthermore there is the risk that the wrong tissue is sampled for biopsy (sampling error) and that too little malignant tissue or too much healthy tissue is removed from a patient.
The purpose of this project is to create a breakthrough tool for early cancer detection by combining state of the art laser and imaging technology with leading medical research, expertise and hands on experience with application of the tools.
The main focus of the EDOCAL project is to medically validate the use of the first prototype based on the dynamic blue technology for early cancer detection applications. This will be done using standard commercially available endoscopes extended with the dynamic blue laser prototypes.
The SMEs will cooperate with world class medical RTD institutions to fill the gap in medical know how and experience of the SMEs themselves and to enable linking the RTDs leading medical experience with breakthrough optical technologies. The technical principle behind the EDOCAL concept is as follows.
In cancer, tissue cells tend to create additional blood vessels to support their growth. The molecule PpIX is generally present in blood vessels and exhibits red fluorescence when excited with blue light in the 375 - 425 nm range. Cancer cells are detected by observing the red fluorescence at a matching excitation wavelength. The wavelength of the laser has to be tuneable as the matching wavelength depends on the person, type of illness and presence of other chemicals. To increase the intensity of the red fluorescence, a porphyrin prodrug can be administered either orally or locally which synthesises PpIX in the cancerous (cells) tissue. By using a tuneable laser system and combination of endogenous fluorescence (as a result of extra blood) and exogenous fluorescence (produced by administration of a porphyrin prodrug), it is possible to accurately distinguish between normal and cancerous tissue.
This prototype tool will use a tuneable blue laser (laser with multiple blue wavelengths) to selectively excite an accumulated photo sensitiser, called proto-porphyrin in (pre) tissue. This sensitiser is taken by the patient to enhance contrast between healthy and cancerous tissue. Reflected light is filtered using state of the art laser technology. The underlying technology platform will also be applicable for future medical diagnostic solutions such as non-invasive glucose monitoring, but the single focus of the EDOCAL project is to create an early cancer detection tool, the Photonic Cancer Detector (PCD).
The PCD will have the following unique characteristics:
(a) Contact free, high resolution, non-destructive laser based cancer detection imaging system.
(b) Accurate and conclusive detection of many cancers in organic tissue: (targeted: sensitivity above 80 % and specificity above 80 %). Sensitivity is the more important of these two parameters. A sensitivity of 100 % means that the test recognises all people with the cancer as having it. A high sensitivity is vital when early diagnosis and treatment is crucial. A low sensitivity means that many people who have the disease go undetected. A specificity of 100 % means that all healthy people are recognised as being healthy. A high specificity is important when the treatment or diagnosis can be harmful to the patient. A low specificity means that people who do not have the disease are diagnosed (and treated) as having it.
(c) Wide angle and location (1 x 1 mm2) specific detection.
(d) Real time: the difference between healthy and cancerous tissue will instantly light up, providing a real-time diagnostic tool for use during intervention. Surgeons can track progress during surgery.
(e) Portable (size and weight of a mobile phone) and robust solution.
(f) Easy to use by non-trained medical experts.
(g) Economical solution: target price less than EUR 12 500.
(h) Standalone use or endoscopic integration.
We see two different product executions for the PCD.
(1) A standalone tool for early detection / surgical intervention. Target users: medical specialists, GPs, nurses, dentists, research and development (R&D) groups, institutes and universities.
(2) Endoscope integration. Target users: medical specialists.
In this project we focus on the standalone prototype device.
The PCD as standalone tool can be applied to detect up to 10 % of all European cancer cases and can potentially be useful in up to 80 % of all surgical interventions as it provides real-time, wide angle primary detection capabilities. Surgery is still the leading type of cancer treatment. For lung and blood related cancers we expect the PCD to have no use. When integrated in an endoscope it can be used to detect and intervene in about 27.5 % of all cancer cases.
The PCD can also be used as a first screening tool for GPs, trained nurses or dentists. This will provide them with a reliable, and fast, detection tool that does not exist today. Benefits will be instant: conclusive diagnosis of oral cavity and skin cancer (6 % of all cancer cases), resulting in earlier stage intervention, enhancing the success rate of cancer treatment and reducing hospital stays.
In summary, the main objectives of this project are:
(a) To develop a breakthrough contact free, high resolution tool for early cancer detection with a targeted sensitivity above 80 % and specificity above 80 %. The solution will support both wide angle and location specific (1x1mm2) view, and provide real-time information that can be used for red flag support during medical procedures.
(b) To validate the use of the early cancer detection tool in a clinical setting, optimising the original concept where appropriate.
(c) To generate scientifically supported information about possible relationships between the sensitiser complexes and their preference for attaching to specific cancerous areas.
(d) To prepare and gather the initial clinical data to be used in the medical regulatory approval phase, building on our current expertise for medical device certification of red laser applications.
Project results:
The main focus of the EDOCAL project was to medically validate the use of the first prototype based on the dynamic blue technology for the early detection of cancer. To do this four work packages (WPs) were defined as follows:
(1) specification and test systems;
(2) fluorescence spectroscopy in model biological systems and cell lines;
(3) analysis of biopsy and endoscopic mucosal resection (EMR) specimens from patients with BE;
(4) device architecture.
WP3: Specification and test systems
The purpose of this task was to define and build a tuneable fluorescence spectroscopy system. The excitation wavelengths should be tuneable within a wavelength range of 365 - 450 nm. Light was directed to the sample via a bifurcated optical fibre. The fluorescence emission was collected by the same fibre and spectrally resolved using a diode array spectrometer. The emitted signal was analysed using a laptop computer with commercial software. The system was tested using solutions of a photosensitiser, and also using photosensitiser-treated cell lines. The specification for the first system was defined by the RTDs and this first system (called dynamic blue system) was built by the SMEs and delivered to the RTDs for testing.
At NWH the focus was on in vitro fluorescence measurements of porphyrins in model biological systems; the test environment was designed and constructed incorporating the dynamic blue system to investigate the fluorescence in screening both photosensitiser solutions and different cell lines.
The prototype was connected to a shutter by a 400 / 440 fibre and the excitation signal was transported to the sample via a bifurcated fibre. The bifurcated fibre collected and delivered the fluorescence emission signal from the sample to the connected diode array spectrometer (470 - 700 nm) and computer for spectral analysis and dissemination. The distal end of the bifurcated fibre was connected to a probe (6 cm in length) coated in plastic via a subminiature version A (SMA) connector. The fluorescence in photosensitiser solutions and different cell lines was measured, for example, with the probe positioned perpendicular to a monolayer of cells.
Porphyrins have a high fluorescence quantum yield and are intermediate molecules in the biosynthesis of haem. Normally this pathway is kept under tight negative feedback control and porphyrins are not produced in excess. Porphyrins are synthesised from the precursor 5-aminolaevulinic acid (ALA), which when added to cultured cells can bypass the feedback mechanism of haem synthesis, resulting in the overproduction of proto-porphyrin - the penultimate molecule in the pathway. Methyl aminolaevulinic (MAL) acid is an esterfied analogue of ALA. To characterise the test set-up, the photo-sensitiser PpIX was dissolved in dimethyl sulfoxide (DMSO) and the pro-drugs ALA and MAL were added to cultured skin keratinocytes and osophageal adenocarcinoma (CA) cells.
Methodology
Human oesophageal OE19 CA cells were maintained in RPMI 1640 medium containing 10 % foetal calf serum; human HaCaT keratinocytes were maintained in DMEM medium containing 5 % foetal calf serum. Cells were grown at 37 degrees of Celsius in a humidified atmosphere of 95 % air: 5 % carbon dioxide (CO2). For experiments, cells were seeded into 60 mm tissue culture dishes, 24 hours before addition of the pro-drugs ALA and MAL. DMSO containing or not pro-drugs was added to the media and the cells were incubated for 4 or 24 hours before a fluorescence reading was taken.
The stock solution of PpIX (Sigma, Dorset, United Kingdom (UK)) was made by dissolving the powder in DMSO. A working solution of PpIX (5 µM) was prepared and wrapped in tin foil. ALA and MAL (Sigma, Dorset, UK) were prepared in DMSO immediately before use. The concentration of DMSO in culture media was maintained at or below 0.5 % (v / v). Solutions were wrapped in tin foil and all experiments carried out under subdued lighting in a specially adapted photobiology laboratory.
PpIX solutions and intracellular PpIX were excited with blue light (368 - 446.5 nm) and fluorescence emission was detected using a diode array spectrometer.
The test system was capable of delivering and detecting PpIX fluorescence in solution. The fluorescence emission maxima of PpIX in DMSO, as determined by the test system and by a F-2500 fluorescence spectrophotometer (Hitachi, UK) were compared. The results showed that both instruments were able to detect PpIX in solution following excitation with different wavelengths (368, 403, 408.5 416.3 417 and 446.5 nm). There was a difference in the peak wavelengths according to the two different systems and, also, the wavelength shift was not consistent.
The test set-up was also capable of delivering and detecting PpIX fluorescence in different cell lines. Both HaCaT keratinocytes and oesophageal OE19 CA cells effectively converted pro-drug to porphyrin. The porphyrin was identified as proto-porphyrin by high-performance liquid chromatography (HPLC) analysis. Further to this, conversion to PpIX increased over a 24-hour period, which is supported by previous work undertaken by us using HPLC analysis. Differences in peak intensity, as shown by peak height, were observed by different excitation wavelengths (403 and 408.5 nm) in ALA-treated OE19 cells. This result could be due to variations in the output power from each diode. As expected, no PpIX fluorescence was detected in the control OE19 cells but PpIX fluorescence was detected in ALA- and MAL-treated OE19 cells. This result confirms the findings from other instruments and that previously noted in the literature.
The specification for the system to be used at AMC was identical to that for NWH, however the used was different. The AMC focussed on ex-vivo measurements on endoscopic resection specimens for the characterisation of specific tissue (auto)fluorescence. Therefore, a set-up had to be developed to investigate freshly obtained EMR specimens. To closely resemble an in-vivo setting, a fibre-optic probe was specially designed to deliver the excitation signal to the tissue through the working channel of the endoscope and to carry back the fluorescence signal to the connected spectrometer and computer for spectral analysis. The endoscope, probe and spectrometer were fixed on a bench and the EMR specimen was placed in an oblique angle below the tip of the probe, thus mimicking the in vivo endoscopic setting.
Conclusions
The test set-up was tested successfully. The system demonstrated that it was capable of reproducibly delivering and detecting a fluorescent signal derived from either proto-porphyrin IX solutions or pro-drug induced cell lines. The findings were consistent with those found using other instruments and also reported in the literature.
Feedback from NWH on optimisation that was required for the system was discussed with the SMEs and a second improved system was designed and built. Although the system required optimisation, the characterisation of the viability of the source to produce adequate fluorescence signals was successfully achieved and the system could be used to carry out the work of WP4.
The ex vivo set-up was built, tested and optimised, after which preliminary measurements were performed. The excitation signal was adequately delivered to the targeted tissue and relevant fluorescence spectra were obtained. Although limitations to the laser system warranted further optimisation of the excitation hardware, the actual ex-vivo test set-up proved to be adequate to produce ex vivo fluorescence spectra as required for WP5.
The in vivo set-up was developed based on experience with the ex-vivo system and was adapted from bench to bedside for actual in-patient endoscopic fluorescence spectroscopy. In-vivo measurements were performed and proved to achieve relevant fluorescence spectra. The laser prototype needed further optimisation but the characterisation of prerequisites for adequate in-vivo fluorescence spectroscopy was successfully achieved and the system proved to be sufficient for use in the in vivo tests in WP5.
WP4: Fluorescence spectroscopy in model biological systems and cell lines
The aim of WP4 was to study factors that influence differences in fluorescence intensity and determine the optimum excitation conditions to allow differentiation between normal and cancerous cell lines. This work was led by NWH and has two main areas of investigation:
(a) in vitro fluorescence spectroscopy of porphyrins in model biological systems;
(b) investigation of fluorescence induced by photonic crystal laser in cell lines.
Methods
For this work the second version of the dynamic blue system was used, together with the Hitachi U3010 double beam absorbance spectrophotometer and the Hitachi F2500 fluorescence spectrophotometer fitted with a red-sensitive photomultiplier. Matched quartz cuvettes were used throughout, and all work was carried out in a specially adapted photobiology laboratory with ambient light levels below 1 lux. For the spectrophotometer work, excitation wavelengths encompassing the range of the dynamic blue device were used. The optical radiation characterisation was performed using a Bentham spectroradiometer system and the absolute radiant power was measured using an Ophir laser power sensor connected to a Nova II metre.
Key results
The photo-sensitiser environment influenced the absorption peak maximum and shape for both PpIX and Foscan. In terms of fluorescence, 1:75 liposomal solutions of PpIX most closely reflected the data obtained from ALA-treated normal human skin. The fluorescence intensity of photo-sensitisers presented to NW2 - the second dynamic blue iteration - device in different environments was in agreement with data previously published using traditional spectrophotomeric means. The power adjusted measurements of PpIX concentration taken with NW2 over the range 0.1 - 1.0 µM (consistent with concentrations found in human skin following application of ALA) were linear and in line with measurements made using the Hitachi F2500 spectrophotometer. The limits of quantifiable detection (S:N > 10) of PpIX and Foscan fluorescence by NW2 were comparable to the Hitachi F2500 (< 0.1 µM). NW2 was actually more effective at detecting Foscan fluorescence in PBS compared to the Hitachi F2500 (fluorescence not detected). The precision of measurements of photo-sensitisers presented to NW2 in different environments was good as indicated by the coefficient of variation (%CV). The %CV of the fluorescence maxima was generally in the region of 0.03 %. Peak intensity measurement was least precise in solutions that encouraged aggregation of the photo-sensitisers, and most repeatable in liposome suspension at less than 5 %.
In the biologically mimicking solutions of bovine serum albumin (BSA), fetal bovine serum (FCS) and liposomes, tuning the excitation wavelength of PpIX from 400 - 410 nm shifted the emission peak maximum by 0.2 0.5 and 0.3 nm respectively. For the chlorin these shifts were smaller (0.2 0.2 and 0.1 nm respectively). Thus the porphyrin seemed to be more affected by tuning the excitation wavelength than Foscan (Refer to D4.1 and 4.M4). In human skin, fluorescence from endogenous chromophores (auto-fluorescence; AF) shifts by 8.7 nm from 513.9 nm to 522.6 nm when the excitation wavelength was changed from 400 to 410 nm.
Cell-specific differences in fluorescence were observed between cell lines derived from the epidermis (HaCaT), oesophagus (OE19), brain (SHSY5Y) and bladder (HT1197). All cell types were able to effectively convert ALA or MAL to PpIX over time with the more transformed the cell phenotype accumulating more cell-associated PpIX. More porphyrins consistently accumulated in OE19 or HT1197 cells than in HaCaT or SHSY5Y cells. The majority of porphyrin was PpIX, except in OE19 media, where coproporphyrin (4-COOH) was the most dominant after 24 hours. Spectroscopic analysis (I620/I635) using the NW3 the third Dynamic Blue iteration - device illustrated a prominent 620 nm peak in OE19 cell media, which could be 4-COOH. This 620 nm peak was associated with some (OE19, HT1197) but not all (SHSY5Y, HaCaT) tissue cell types. Factors that influence the differences observed between these cell lines were investigated, in particular, PpIX efflux and haem synthesis. The present results do not suggest a major impact of the porphyrin transporter protein within the clinically relevant incubation time used in this study. Distinct differences in fluorescence emission at 620 and 635 nm may thus indicate progression towards a more cancerous cell type. All excitation wavelengths were able to induce adequate fluorescence signals, except the 367 nm light-emitting diode (LED). PpIX fluorescence peak intensity shifts were observed between 4 and 24 hours, with all cell lines; the highest in OE19 and HT1197 cells. These cell lines also accumulated the highest cell associated concentrations of PpIX. Emission wavelength shifts, however, where not detected between excitation wavelengths or between cell lines. Finally, auto-fluorescence from cells could not be observed.
Blue visible light up to a dose of 5 J / cm2 was not in itself phototoxic or photogenotoxic to cell types on its own when compared to baseline (dark) values. Exposure of ALA treated cells to light resulted in a small but significant increase in DNA strand breaks, however no effects on cell survival were subsequently observed 24 hours later in the phototoxicity assay. The doses that were used encompassed and exceeded those that might be delivered to tissue by NW3 in its current form, and changes started to be seen in the gel electrophoresis assay at doses of 80 mJ / cm2. Non-light exposed cells did not exhibit any toxicity. The data confirm the desired specification for a device emitting no more than 1 mW / cm2 or higher. It is also notable that no measurable changes were seen in the phototoxicity assay, (which is particularly vulnerable to damage to the cell membranes) or in the morphology of the cells. Therefore the changes resulting in the small increase in deoxyribonucleic acid (DNA) migration did not manifest as a measurable change in cell viability 24 hours later.
Skin measurements
Methodology
5-aminolaevulinic acid (ALA 20 % w / w) (Mandeville Medicines, UK) was prepared on aluminium Finn chambers on Scanpor (8 mm) (Smart Practice, USA). ALA was applied to the inner surface of the forearm of a healthy volunteer. The application site was covered with Tegaderm film (3M Healthcare, Germany) and a Mepore (M?lnlycke Healthcare, Sweden) dressing for 6 hours prior to removal and any excess cream was wiped off. PpIX fluorescence readings with NW2 were recorded by placing the probe perpendicularly in contact with the skin. Furthermore, additional fluorescence readings were recorded as controls from the normal skin located approximately 10 cm distal to the application site. Each measurement was performed in specially adapted photobiology laboratory with ambient light levels below 1 lux.
Summary of the work
The spectrum for PpIX in organic solvents was typical of that for porphyrin molecules with a Soret band (max 400 - 407 nm) and four subsequent smaller Q bands as shown in the literature with the position of the Soret band, but not the Q band positions, highly dependent upon the solvent used. These spectra represent monomeric solutions. However, in an aqueous solution, the Soret band position shifts to a shorter wavelength alongside the broadening of all absorption bands and a strong decrease in peak intensity. These observations are indicative of a highly aggregated solution. In FCS and BSA solutions, the Soret band shift to a shorter wavelength remains but not to the same extent as PpIX in an aqueous solution, and the 4 Q bands remain more consistent with that of PpIX in an organic solvent which indicates a more monomeric solution than that of PpIX in the aqueous environment. In liposomes, the Soret band and Q bands are in a similar position to that of PpIX in organic solutions along with sharper absorption bands compared to PpIX in FCS or BSA and this is due to further monomerisation of PpIX in this lipid environment. Therefore, the absorption properties of a PpIX solution are highly dependent upon aggregation state. The trends observed in the 25 µM solutions were also observed in the 5 µM, thus trends were not dependent upon PpIX concentration. Generally, peak intensity of the Soret band increased in a positive, linear relationship with increasing PpIX concentration. The Soret band shift to shorter wavelengths at high PpIX concentrations in solutions of BSA and FCS may be due to a mixture of monomeric and aggregated species. PpIX was most soluble in DMSO, followed by BSA, serum and MeOH, but had limited solubility in EtOH and acetone.
The spectrum for Foscan in ethanol was typical of that of chlorin type molecules with a Soret band (max 416 nm) and a strong absorbance in the red region (max 650 nm). In aqueous solutions, unlike that of PpIX, the Soret and Q band position shifts to a longer wavelength which has been previously reported in the literature and is indicative of an aggregated species with this group of molecules. The aggregation of Foscan in PBS was further supported by peak broadening and the strong decrease in peak intensity of all absorption peaks. In FCS, BSA and, in particular, the liposome solutions, monomerisation was evident with a blue shift in Soret and Q band positions alongside peak sharpening when compared to the aqueous solutions. Again, the absorption properties are highly dependent upon the aggregation state of the Foscan molecules.
Conclusions
The environment influenced the absorption peak maximum and shape of both PpIX and Foscan. The PpIX Soret band maximum shifted from 362.5 nm in aqueous buffer to 410.5 nm in liposome suspension (a red shift of 48 nm). For Foscan there was a blue shift of 14 nm from 434 nm to 420 nm. In terms of fluorescence, 1:75 liposomal solutions of PpIX most closely reflected the data obtained from ALA-treated normal human skin. The fluorescence intensity of photo-sensitisers presented to NW2 in different environments was in agreement with data previously published using traditional spectrophotomeric means. The blue shift in PBS, compared to the other conditions where Foscan is more monodispersed, is indicative of the presence, amongst larger aggregates, of H-type (or face-to-face) oligomers.
The power adjusted measurements of PpIX concentration taken with NW2 over the range 0.1 - 1 µM (consistent with concentrations found in human skin following application of ALA) were linear and in line with measurements made using the Hitachi F2500 spectrophotometer. This is encouraging for the possibility of using the laser device for quantification of photo-sensitiser in tissue.
The limits of quantifiable detection (S:N > 10) of PpIX and Foscan fluorescence by NW2 were comparable to the Hitachi F2500 (< 0.1 µM). NW2 was actually more effective at detecting Foscan fluorescence in PBS compared to the Hitachi F2500 (fluorescence not detected). This may be due to a combination of low quantum yield, due to the formation of aggregates, and large inner filter effect due to the concentration of the sample. The precision of measurements of photo-sensitisers presented to NW2 in different environments was good as indicated by the %C V. The %CV of the fluorescence maxima was generally in the region of 0.03 %. Peak intensity measurement was least precise in solutions that encouraged aggregation of the photo-sensitisers, and most repeatable in liposome suspension at less than 5 %.
In the biologically mimicking solutions of BSA, FCS and liposomes, tuning the excitation wavelength of PpIX from 400-410 nm shifted the emission peak maximum by 0.2 0.5 and 0.3 nm respectively. For the chlorin these shifts were smaller (0.2 0.2 and 0.1 nm respectively). Thus the porphyrin seemed to be more affected by tuning the excitation wavelength than Foscan.
In human skin, fluorescence from endogenous chromophores (auto-fluorescence; AF) blue shifted by 8.7 nm from 513.9 nm to 522.6 nm when the excitation wavelength was changed from 400 to 410 nm. The peak maximum of PpIX did not change. As the endogenous fluorescence emission reflects the nature of the fluorophore, the combination of AF peak maxima, intensity and PpIX peak maxima and intensity may be a powerful tool in differentiating tissue staging in cancer and other diseases.
Results from screening the 100 samples
The spectrum for PpIX in organic solvents was typical of that for porphyrin molecules with a Soret band (max 400 - 407 nm) and four subsequent smaller Q bands as shown in the literature with the position of the Soret band, but not the Q band positions, highly dependent upon the solvent used. These spectra represent monomeric solutions. However, in an aqueous solution, the Soret band position shifts to a shorter wavelength alongside the broadening of all absorption bands and a strong decrease in peak intensity. PpIX was most soluble in DMSO, followed by MeOH, but had limited solubility in EtOH and acetone.
The spectrum for Foscan in ethanol was typical of that of chlorin type molecules with a Soret band (max 416 nm) and a strong absorbance in the red region (max 650 nm). In aqueous solutions, unlike that of PpIX, the Soret and Q band position shifts to a longer wavelength which has been previously reported in the literature and is indicative of an aggregated species with this group of molecules. The aggregation of Foscan in PBS was further supported by peak broadening and the strong decrease in peak intensity of all absorption peaks.
The environment influenced the absorption peak maximum and shape of both PpIX and Foscan. The PpIX Soret band maximum red-shifted by 44.5 nm when the solution was changed from aqueous buffer to DMSO. Foscan blue-shifted by 17.5 nm. The solubility of Foscan in buffer was very poor. The blue shift in PBS, compared to the other conditions where Foscan is more monodispersed, is indicative of the presence, amongst larger aggregates, of H-type (or face-to-face) oligomers.
The power adjusted measurements of PpIX concentration taken with NW2 over the range 0.1 - 1.0 µM (consistent with concentrations found in human skin following application of ALA) were linear and in line with measurements made using the Hitachi F2500 spectrophotometer. This is encouraging for the possibility of using the laser device for quantification of photo-sensitiser in tissue.
NW2 was actually more effective at detecting Foscan fluorescence in PBS compared to the Hitachi F2500 (fluorescence not detected). This is likely due to a combination of low quantum yield, due to the formation of aggregates, and large inner filter effect due to the concentration of the sample.
The precision of measurements of photo-sensitisers presented to NW2 in different environments was good as indicated by the %CV. The %CV of the fluorescence maxima was generally in the region of 0.03 %. Peak intensity measurement was least precise in solutions that encouraged aggregation of the photosensitisers (PBS & MeOH in the case of PpIX; PBS in the case of Foscan). PpIX was barely soluble in acetone so these data are not shown. The best solutions for reproducibility were DMSO (PpIX) and EtOH (Foscan).
In the solvents DMSO, MeOH and PBS, tuning the excitation wavelength of PpIX from 400-410 nm shifted the emission peak maximum by 0.0 0.1 and 0.5 nm respectively. For the chlorin the shifts in EtOH & PBS were 0.0 and 0.3 nm respectively. Thus the porphyrin seemed to be more affected by tuning the excitation wavelength than Foscan.
WP5: Analysis of biopsy and EMR specimens from patients with BE
This WP was led by AMC and investigated the following:
(a) In vitro fluorescence microscopy and spectroscopy of biopsy specimens from patients with Barrett's esophagus and high grade dysplasia or early mucosal cancer before and after oral administration of 5-ALA.
(b) 5-ALA induced fluorescence imaging of early Barretts cancers using ex vivo spectroscopy measurements of EMR specimens.
In patients with Barrett's oesophagus (BO), the normal epithelial lining is replaced by columnar lined epithelium, so called intestinal metaplasia. Malignant degeneration may occur in this intestinal metaplasia, through a series of precancerous changes graded on histology: low-grade intraepithelial neoplasia (LGIN), high-grade intraepithelial neoplasia (HGIN) and intramucosal carcinoma or early CA. When detected at this premalignant stage, these dysplastic areas can be curatively treated with minimally invasive techniques, such as endoscopic resection. Therefore, regular endoscopic surveillance is advised in order to detect dysplastic areas at an early stage. However, these areas are usually flat and do not stand out from the non-dysplastic Barretts epithelium. Advanced imaging methods are thus required to increase the detection rate of premalignant lesions in BO. One of the most promising approaches is the use of fluorescence spectroscopy. Different tissue types have distinct natural fluorescence characteristics, which can be detected by exciting the tissue with a certain wavelength of light and measuring the resulting fluorescent light. Also changes within a tissue type can be detected, such as dysplastic mucosa in Barrett's epithelium. Endogenous tissue fluorescence is called auto-fluorescence.
The work focused on the fluorescence characteristics of Barrett's epithelium, either in vivo or assessed in biopsies and endoscopic resection specimens. This work aimed to assess the optimal excitation wavelength(s) to induce (auto)fluorescence, and to study the shift of the excitation and fluorescence spectra which may be associated with changing tissue characteristics. Fluorescence spectroscopy may be used to distinguish non-dysplastic from dysplastic Barrett's tissue in the blue light (ultraviolet / visual) wavelength range. A widely used diagnostic fluorescent agent for in the gastrointestinal tract is 5-aminolevulinic-acid (5-ALA) induced proto-porphyrin-IX (PpIX). The endogenous, photoactive PpIX is an immediate precursor of heme and is converted from 5-ALA intracellularly. This conversion depends on metabolic activity of the cells and hence will be different between different cell types. After administration of excess exogenous 5-ALA, PpIX selectively accumulates in malignant tissue. PpIX induced fluorescence thus can be used to enhance the diagnostic value of fluorescence spectroscopy by increasing the signal (malignant tissue) to noise (background normal tissue) ratio. PpIX will reach peak concentration approximately 4 - 6 hours after oral administration. 5-ALA was used to enhance the discriminating characteristics of the fluorescence induced by the dynamic blue prototype. Fluorescence spectroscopy on biopsies was performed before and after 5-ALA administration.
The other focus of the AMC was in-vivo endoscopic fluorescence spectroscopy. For this purpose, a specially designed optical fibre probe is equipped with 2 arms one is connected to the spectrometer and the other one to the prototype light source box. The whole system, including the computer with the laser driver and spectral analysis software, was mounted on a modified trolley, in order to be able to move closely to the patient and the endoscopist. A small plastic cap was placed on the tip of the endoscope to ensure stable positioning of the endoscope and the probe on the tissue. The probe was subsequently moved through the working channel of the endoscope and placed on the tissue under direct vision. A transparent plastic cap was attached to the tip of the endoscope to ensure a stable position of the endoscope and the probe's tip on the oesophageal tissue.
Patients with and without dysplasia in a BO were treated with an EMR in accordance with current guidelines. In this ex-vivo experiment we tried to closely resemble the in-vivo situation during endoscopy: an optical fiber probe, designed to deliver the excitation light and to collect the emitted fluorescent light was connected to the dynamic blue prototype and a spectrometer and subsequently passed through the working channel of the endoscope onto the EMR specimen. Fluorescence spectroscopy was performed on areas suspicious for dysplasia and on adjacent inconspicuous areas on the specimen. This was immediately followed by histological sampling of a biopsy from the EMR specimen of the corresponding measurement site, to correlate the histological diagnosis to the measured fluorescence spectrum.
Patients with and without early esophageal neoplasia originating from a BO were included in this study. All patients underwent high-resolution white light endoscopy (WLE), auto-fluorescence endoscopy (AFI) and narrow band imaging (NBI). Multiple areas suspicious for dysplasia in the Barrett's segment, were investigated with in-vivo spectroscopy using the Dynamic Blue multi-wavelength prototype and a specially designed optical fiber that was passed through the working channel of the endoscope. A small transparent cap was used on the tip of the endoscope to ensure stable positioning of the probe and the correlation of the corresponding biopsies. Spectroscopy was immediately followed by histological sampling of the corresponding area. Also non-suspicious control areas were investigated and sampled. Auto-fluorescence spectra were recorded, using different excitation wavelengths (365, 395, 405, 410, 417, 445 nm). In addition, white light (WL) and background (bg) spectra were obtained to allow correction of the auto-fluorescence spectra.
After the spectroscopic evaluation, biopsies and EMR specimens were fixed in formalin, embedded in paraffin and routinely cut and stained with haematoxylin and eosin (H&E). Histopathological assessment of the biopsies was performed by an expert GI-pathologist, who recorded the presence of intestinal metaplasia and neoplasia according to the WHO-classification: no-dysplasia, indefinite for dysplasia, LGIN, HGIN or invasive cancer. In case of cancer: depth of tumour invasion, grade of differentiation, presence of lymph-vascular invasion was assessed, as well as the presence of neoplasia at the deep (vertical) resection margins.
Methods: Analysis of the autofluorescence spectra
Analysis of autofluorescence spectra (with white light correction)
Ex-vivo and in-vivo collected fluorescence spectra were analysed using two distinct methods: emission intensity analysis and spectral shape analysis. Fluorescence spectra obtained using a single excitation wavelength and combinations of spectra obtained using different excitation wavelengths were analysed to find the optimal (combination of) wavelength(s).
All auto-fluorescence spectra were corrected by the background signal and normalised with respect to the total intensity of the spectrum (area under the curve).
The intensity of emitted fluorescence as a function of excitation wavelength differs per tissue type and may therefore be used as a discriminating tool. Selected spectra corresponding to the same histopathologic category were averaged. The averaged spectra of non-dysplastic Barrett's (IM) were used as a baseline value. A baseline corrected spectrum was determined by subtracting the baseline spectrum from the spectra of HGIN / CA categories. Intensity ratios were selected according to the results of the baseline corrected spectrum. ROC curves and area under curve (AUC) were used to determine sensitivity and specificity. The statistical relevance of the intensity ratios was determined by two-tailed independent Student's t-test
Spectral shape analysis
The shape of a spectrum depends on the excitation wavelength and the tissue type. Spectral shape analysis was performed by calculating the first derivative and comparing the peak-to-peak deviation. In addition principal component analysis (PCA) was performed which is another method to identify patterns in the spectra to distinguish between non-dysplastic Barrett's (IM) and grouped HGIN and early CA. Intensity and spectral shape analysis results were compared and the more significant one was chosen for further analysis.
Advanced emission intensity analysis
White light correction
Reflectance spectra generated by the white light source were used to correct for the spectral distortion of the auto-fluorescence spectra due to scattering and absorption mainly from hemoglobin absorption. Both white light corrected and non-white light corrected spectra were analysed and compared.
Excitation - emission matrix (EEM)
To further optimise the emission intensity analysis an EEM was calculated for the non-white light corrected and for the white light corrected auto-fluorescence spectra. EEM visualise the intensities in a colour map. Differences in emission wavelengths in relationship to excitation wavelength and tissue type were visualised.
Double intensity ratio
Double intensity ratios of 2 emission spectra from the same location on the tissue obtained from 2 different excitations were used to further optimise the sensitivity and specificity. All combinations of different wavelength excitations were tested.
Results: Ex vivo spectroscopy in Barrett's patients with and without dysplasia.
11 EMRs were performed on seven patients with dysplasia in a BO. All patients underwent an EMR using the EMR-cap technique of both a neoplastic and a non-neoplastic area. The EMR-specimen was immediately retrieved from the patient and freshly pinned onto a paraffin block for direct ex-vivo examination. The dynamic blue prototype was activated (405 nm) and ex-vivo auto-fluorescence spectroscopy and corresponding histology sampling were subsequently performed on a total of 26 mucosal areas in Barretts mucosa in these 11 EMR specimens.
All obtained fluorescence spectra were first corrected by the background signal and then normalised with respect to the total intensity of the fluorescence spectrum (area under the curve). Selected spectra corresponding to the same histopathologic category were averaged. The averaged spectrum of non-dysplastic Barrett (IM) were used as a baseline value. A baseline corrected spectrum was determined by subtracting the baseline spectrum from the spectra of HGIN / EC categories. The resulting graph is deemed the baseline corrected spectrum.
The ex vivo measured auto-fluorescence spectra of normal Barrett's and HGIN / EC show differences in the relative intensity and in the spectral shape.
Results: In vivo spectroscopy in Barretts patients with and without dysplasia; phase 1
Patient and lesion characteristics
In phase 1 and 2 a total of 70 patients were included, in whom a total of 290 areas were investigated with in-vivo auto-fluorescence spectroscopy using the Dynamic Blue prototype. The results from intensity and shape analysis, the intensity ratios show a higher significance in discriminating non-dysplastic Barrett (IM) from grouped HGIN and early CA. The determined sensitivity and specificity however is not adequate for clinical decision making. Therefore advanced analysis was applied.
Conclusions on autofluorescence spectra (without white light correction)
For single excitation with 365nm the sensitivity is 100 % but the specificity (38 %) is too low. The combination of 365nm and 405nm excitation improves the specificity from 38 up to 50 %. Improvements to increase the specificity are however still needed.
The intensity ratios 560 / 640 are higher for all excitation wavelengths compared to the 495 / 560 ratios. At 365nm excitation the difference between non-dysplastic Barrett (IM) and grouped HGIN and early CA is highest. Analysis of combined excitation is done by calculating the double intensity ratio for the emission wavelength 560/640.
Conclusions on autofluorescence spectra (with white light correction)
For single excitation with 365nm the specificity is 100% but the sensitivity is with 67 % still too low. The combination of 365 and 405 nm excitation improves the sensitivity from 67 % up to 100 %.
Conclusions on phase 1
The developed algorithm combines 365 nm and 405 nm excitation in which the double intensity ratio at 560nm and 640 nm of white light corrected auto-fluorescence spectra is used. The 640nm red emission probably reflects the emitted fluorescence of porphyrin.
Results: In vivo spectroscopy in Barretts patients with and without dysplasia; phase 2
In phase 2 the developed algorithm of phase 1 was applied to a new set of data.
Analysis of auto-fluorescence spectra (without white light correction)
Auto-fluorescence spectra of 365nm and 405 nm are analysed with the intensity ratios 560 / 640 and 495 / 560. For single excitation 405 nm has a sensitivity of 75 % and a specificity of 89 % for the emission intensity ratios at 495nm and 560nm to distinguish between non-dysplastic Barrett's (IM) and grouped HGIN and early CA as at 365 nm excitation.
The results of the two tailed independent Students t-test gives a p-value of 0.002 which is a statistical significant difference of the double intensity ratios of 365 and 405 nm and may therefore be used to discriminate between non-dysplastic Barrett (IM) and grouped high-grade HGIN and early CA.
Conclusions on white light corrected auto-fluorescence spectra
For white light corrected auto-fluorescence spectra analysis on the new data set the combination of 365nm and 405nm does improve the sensitivity and specificity compared to single excitation. A sensitivity of 80% and a specificity of 81 % are reached with the double intensity ratios at 495nm and 560nm. The student t-test shows that there is a significant difference between non-dysplastic Barrett (IM) and grouped HGIN and early CA.
Conclusions on phase 2
In phase 2 the optimised method for data analysis developed in phase 1 was applied to a new set of data. The results confirm that the white light corrected auto-fluorescence spectra in combination with 365 and 405 nm excitation gives highest sensitivity and specificity. Contrary, the results in phase 2 suggest that double intensity ratio at 495 and 560 nm increases the specificity compared to the previous mentioned double intensity ratio of 560nm and 640nm.
Conclusions of phase 1
The work at the AMC Amsterdam has significantly expanded compared to the task originally described by performing ex-vivo spectroscopy on 26 sites in 11 EMR specimens from 7 Barrett's patients and 1 557 in vivo spectroscopy measurements on 290 sites in 70 BO patients during real-time endoscopy. A total of 77 patients were investigated in phase 1. In addition, these tests used the first generations dynamic blue prototype with a multi wavelength laser set-up. The first part of WP5 therefore provided much more data and information than originally scheduled. First, data were largely collected in-vivo instead of the artificial in-vitro setting. Second, data were collected from a much larger group of patients than originally planned, thus increasing the statistical power and the extrinsic validity of the project. Third, data were obtained using the dynamic blue prototype which more closely resembles the actual clinical context of the final product. Fourth, the first phase allowed technical, methodological, and statistical investigations that were not only imperative for the execution of WP5 of the EDOCAL project but also allowed technical improvements of the prototype beneficial to the dynamic blue project.
Conclusions on ex-vivo spectroscopy measurements on EMR specimens
A special set-up using the available technology from the dynamic blue project was designed to perform ex-vivo and in-vivo fluorescence spectroscopy measurements on Barretts mucosa. Ex vivo spectroscopy was performed on EMR specimens of patients with and without early neoplasia in BO. This allowed testing of the prototype in a more realistic way than would be possible if only tissue biopsies would have been obtained and provided information that was directly transferable to the further testing of the laser prototype during real-time endoscopy.
Conclusions on in-vivo measurements in Barrett patients
In-vivo auto-fluorescence spectroscopy was performed by passing a specially designed light probe through the working channel of an endoscope onto the mucosa during real-time endoscopy. Spectroscopy sites were selected by using high-resolution WLE, AFI and narrow-band imaging during specially scheduled endoscopy programs. Using the dynamic blue prototype, each site was illuminated with multiple wavelengths followed by registration of the corresponding auto-fluorescence spectrum and tissue biopsy for histological correlation. A total of 1 557 spectra were obtained from 290 sites in 70 patients.
Conclusions on auto-fluorescence spectroscopy with dynamic blue technology
The experiments conducted for D5.1 have shown that auto-fluorescence spectroscopy, using a combination of 365 and 405 nm excitation and double intensity ratios at 495nm and 560nm emission wavelength of white light corrected auto-fluorescence spectra, is able to discriminate early neoplasia (HGIN / EC) with 80 % sensitivity and 81 % specificity from non-dysplastic Barrett epithelium.
We hypothesised that these parameters could be further improved by adding a photosensitiser. Therefore, in D5.2 5-ALA induced fluorescence as a function of different excitation wavelengths will be assessed before and after the administration of 5-ALA. Dynamic Blue test set-up for D5.2
Specification of the test set-up for 5-ALA induced fluorescence experiments
In D5.2 the AMC focuses on investigating 5-ALA induced PpIX fluorescence using a simulated in-vivo model. To closely resemble the in-vivo setting, the CAM model was developed on which freshly obtained BO biopsy specimens and oesophageal CA cells were transplanted.
In D5.2 the same Dynamic Blue prototype set-up was used as during the in-vivo auto-fluorescence measurements in patients with BO. An optical fibre probe was specially designed to deliver the excitation signal to the tissue and to carry back the fluorescence signal to the connected spectrometer and computer for spectral analysis. The specimen was placed below the tip of the probe, thus mimicking the in-vivo endoscopic setting.
Contact details: Ms Gillian Mimnagh
CEO, 2M Engineering LTD, De Run 4352
5503LN Veldhoven
The Netherlands
List of websites: http://www.edocal.eu
