HIT-PIRADS: prostate MRI and hybrid intelligence
We investigate how AI image analysis and clinical expertise can support the interpretation of prostate MRI.
How should image evidence and model explanations be presented for expert assessment of prostate MRI?
Method overview
Combining image evidence with clinical interpretation
Complementary MRI channels provide different information about the same region.
Schematic of the research approach; image patterns and measurements are illustrative.
Research overview
Prostate MRI interpretation requires the reader to integrate several image contrasts with the location and appearance of a finding. HIT-PIRADS studies hybrid intelligence for prostate cancer diagnosis and patient management. The research connects automated analysis with the clinician’s interpretation, rather than treating a model score as a complete assessment.
The project examines MRI-based learning and explanations that relate model outputs to radiographic findings. Questions of image quality, localization, and agreement across imaging sequences are relevant to this work. Evaluation must consider how the available imaging and clinical reference information support a particular claim.
Combining image evidence with clinical interpretation
- Analyze complementary MRI contrasts and the correspondence between them.
- Identify image regions associated with a model’s output.
- Study the relationship between model evidence and expert interpretation.
Selected research sources
Research enquiries
Contact the lab