Optimising Diagnostic Pathways
Prostate cancer
The use case addresses the diagnostic pathway in prostate cancer, specifically biopsy decision after an uncertain MRI finding.
Avoiding Unnecessary Interventions
The Challenge: Managing Diagnostic Uncertainty
Prostate cancer is the most common cancer in men in Europe and a major cause of death.
Yet, latent cases are also frequently detected at autopsy in up to a third or more of older men who died without any evidence of the disease during their lifetimes. Detection of such non-progressive cases is unnecessary and should be avoided.
Overdiagnosis Risk
Frequent detection of non-progressive cases leads to unnecessary curative treatments and diagnostic interventions.
Intermediate Findings
MRI findings with a PIRADS score of 3 pose a clinical challenge, suggesting a low but uncertain risk of significant cancer.
Magnetic resonance imaging (MRI) is among the most important advances in prostate cancer detection in the past decades. MRI is excellent in detecting clinically significant cases, but equally important is its capacity to reduce diagnosis of low-grade cases that do not require curative treatment.
MRI findings are evaluated using PIRADS scores and an intermediate finding (PIRADS score 3), suggesting a low risk of clinically significant cancer poses a clinical challenge.
In this use case. we aim to develop methods to optimise decision about which men with intermediate MRI findings should be biopsied.
AI tools will be developed based on clinical and radiology data to predict the outcome of biopsy and facilitate decision-making.
AI-Driven Decision Support
The Use Case Objectives
The prostate cancer use case involves collecting data from six data nodes on men with a suspicion for prostate with uncertain or intermediate MRI results. The final diagnosisis the main outcome, with men classified as free of prostate cancer (biopsies not needed), those with well-differentiated, clinically insignificant cancers (where biopsies should be avoided), and clinically significant cancers requiring active treatment (where biopsies are needed).
- Develop an AI model to predict final diagnosis and assist in biopsy decision-making.
- Minimize overdiagnosis of low-grade cases and avoid benign biopsies.
- Ensure clinically significant cases are not missed.
- Utilize multimodal data including digitized MRI, PSA levels, and genetic information.
Data on pre-biopsy characteristics will be collected, including age, family history, and previous biopsies, as well as PSA level and digitized MRI data. Genetic information will also be utilised where available. These will be utilised in developing an AI model to assist in predicting the final diagnosis, aiming at minimising over diagnosis of low-grade cases, missed clinically significant cases, as well as benign biopsies.
Preserving Quality of Life
Expected Impact
The prostate cancer use case aims to provide tools for the diagnostic pathway of prostate cancer that would save resources by avoiding unnecessary diagnostic interventions and detection of clinically insignificant cancers that need to undergo active surveillance with repeated consultations, MRI examinations and prostate biopsies. This would also minimize the quality-of-life detriment due to diagnosis of clinically insignificant cases.
For researchers
Developing and evaluating research processes and tools that enable international collaborative research using the European Health Data Space.
For healthcare systems
Saving resources by avoiding unnecessary active surveillance, repeated consultations, and invasive biopsy procedures for non-progressive cases.
For patients
Minimizing the quality-of-life detriment associated with the diagnosis and management of clinically insignificant cases.
A SUSTAINABLE RESEARCH LEGACY
In the process, we will develop and evaluate research processes, procedures and tools that enable international collaborative research using the European Health Data Space.