Periodic Reporting for period 2 - PCAVISION (Prostate Cancer Diagnosis, Localisation and Characterisation using Ultrasound)
Okres sprawozdawczy: 2023-04-01 do 2024-12-31
Prostate cancer (PCa), is a type of cancer with the highest incidence (19%) and second mortality rate (8%) in western men. In 2018 in Europe, 140,000 male patients were newly diagnosed with PCa and over 100,000 patients have died from Prostate cancer. Numbers are rising due to the ageing EU population.
The present medical diagnosis procedures are largely based on the execution of multi-biopsy procedure(s) on the patient. Each multi-biopsy procedure is invasive, painful for the patient and incurs significant risk (10%-20%) of adverse health events such as sepsis (8%). Screening programs - comparable to breast cancer screening for woman - are not possible based on biopsy. In recent years, in a few EU countries, an image-based diagnosis method has been added to the PCa diagnosis pathway: Magnetic Resonance Imaging (MRI). A dedicated MRI imaging protocol offers an additional diagnosis tool for detecting and characterising PCa tumour growth.
However, MRI based PCa diagnosis has drawbacks: scarce patient access (no public healthcare coverage in multiple EU countries) and long waiting times of 35 days median, increased national healthcare cost and dependent on availability of highly trained radiologist. The gap between the need for image based diagnosis and the availability is rapidly widening.
PCaVision is a novel PCa imaging diagnostic method relying on modern ultrasound imaging equipment and advanced signal processing algorithms. Advantages are: equipment costs 96% less and there is not need for a specialised radiologist; the urologist performs the diagnosis. Hence PCaVision makes broad scale access for patients possible, improves PCa healthcare and reduces national healthcare cost.
With PCaVision the gap can be closed for the benefit of all EU patients.
The present medical diagnosis procedures are largely based on the execution of multi-biopsy procedure(s) on the patient. Each multi-biopsy procedure is invasive, painful for the patient and incurs significant risk (10%-20%) of adverse health events such as sepsis (8%). Screening programs - comparable to breast cancer screening for woman - are not possible based on biopsy. In recent years, in a few EU countries, an image-based diagnosis method has been added to the PCa diagnosis pathway: Magnetic Resonance Imaging (MRI). A dedicated MRI imaging protocol offers an additional diagnosis tool for detecting and characterising PCa tumour growth.
However, MRI based PCa diagnosis has drawbacks: scarce patient access (no public healthcare coverage in multiple EU countries) and long waiting times of 35 days median, increased national healthcare cost and dependent on availability of highly trained radiologist. The gap between the need for image based diagnosis and the availability is rapidly widening.
PCaVision is a novel PCa imaging diagnostic method relying on modern ultrasound imaging equipment and advanced signal processing algorithms. Advantages are: equipment costs 96% less and there is not need for a specialised radiologist; the urologist performs the diagnosis. Hence PCaVision makes broad scale access for patients possible, improves PCa healthcare and reduces national healthcare cost.
With PCaVision the gap can be closed for the benefit of all EU patients.
The activities performed are:
Creation of Ground Truth data
To date we have included 400 patients in the Research Clinical trial which is part of the project. For each patient, a multi-model ultrasound scan was made. From this cohort, around 220 patients were diagnosed (via the to the existing standard of care) with a form of PCa requiring a Radical Prostatectomy. The prostates of these patients we sliced, digitised and annotated for PCa by a team of pathologists. The resulting annotated full pathology is our Ground Truth. Via a complex reconstruction and registration process we were able to map the pathology Ground Truth onto the 3D ultrasound scan data. The latter required development of multiple dedicated software tools.
Development and validation of the algorithms
A complex pipeline of various algorithms were developed. The processing pipeline spans pre-processing, feature extraction and classification. The end-result is determining areas of interest in the 3D scan space where presence of PCa is expected. Classification is performed using classical ML and deep learning techniques. On the voxel level we have achieved an ROC-AUC higher then 90%,
Development of the PCaVision application
The algorithm is encapsulated in an extensive and complex software application. This software application (among other) automatically communicates with the ultrasound scanner, executes the algorithm function and generates all output file for the Graphical User Interface. Multiple error detection and logging functions are present lifting the reliability up to a Class IIB medical device level. The PCaVision application also interfaces with fusion based Targeted Biopsy systems.
Validation
An extensive scientific validation process has commenced. This process includes various usability studies, a reader study and a multi-clinic 500 patient head to head comparison agains the current clinical pathway based on MRI. The results of the validation is expected in the next reporting period.
Creation of Ground Truth data
To date we have included 400 patients in the Research Clinical trial which is part of the project. For each patient, a multi-model ultrasound scan was made. From this cohort, around 220 patients were diagnosed (via the to the existing standard of care) with a form of PCa requiring a Radical Prostatectomy. The prostates of these patients we sliced, digitised and annotated for PCa by a team of pathologists. The resulting annotated full pathology is our Ground Truth. Via a complex reconstruction and registration process we were able to map the pathology Ground Truth onto the 3D ultrasound scan data. The latter required development of multiple dedicated software tools.
Development and validation of the algorithms
A complex pipeline of various algorithms were developed. The processing pipeline spans pre-processing, feature extraction and classification. The end-result is determining areas of interest in the 3D scan space where presence of PCa is expected. Classification is performed using classical ML and deep learning techniques. On the voxel level we have achieved an ROC-AUC higher then 90%,
Development of the PCaVision application
The algorithm is encapsulated in an extensive and complex software application. This software application (among other) automatically communicates with the ultrasound scanner, executes the algorithm function and generates all output file for the Graphical User Interface. Multiple error detection and logging functions are present lifting the reliability up to a Class IIB medical device level. The PCaVision application also interfaces with fusion based Targeted Biopsy systems.
Validation
An extensive scientific validation process has commenced. This process includes various usability studies, a reader study and a multi-clinic 500 patient head to head comparison agains the current clinical pathway based on MRI. The results of the validation is expected in the next reporting period.
The current demand for image based PCa diagnosis supersedes the available MRI capacity. This results in waiting times ranging form weeks to multiple months.
The first key impact of the availability of PCaVision is that 100,000+ more of patients gain rapid access to state-of-the-art image based PCa diagnosis. This seriously reduced the waiting time anxiety for patients and expedites the execution of the clinical treatment.
The second impact is a cost saving. Projection indicate that in the EU only the utilisation of PCaVision will save the European healthcare over EUR 100M within the first five year after introduction.
Finally, the EU member states are exploring the feasibility and added value of PCa screening for men. Similar to Breast cancer screening for women.
The introduction of a PCa screening is projected to require over 500K additional image based diagnosis procedures. PCaVision can play an important role in making screening feasible.
The first key impact of the availability of PCaVision is that 100,000+ more of patients gain rapid access to state-of-the-art image based PCa diagnosis. This seriously reduced the waiting time anxiety for patients and expedites the execution of the clinical treatment.
The second impact is a cost saving. Projection indicate that in the EU only the utilisation of PCaVision will save the European healthcare over EUR 100M within the first five year after introduction.
Finally, the EU member states are exploring the feasibility and added value of PCa screening for men. Similar to Breast cancer screening for women.
The introduction of a PCa screening is projected to require over 500K additional image based diagnosis procedures. PCaVision can play an important role in making screening feasible.