SYSTEM FOR THE DETECTION OF TUMORAL MASSES BASED ON MAGNETIC RESONANCE IMAGING
The computer aided diagnosis (CAD) system allows automatic detection of suspicious areas of prostate cancer starting from magnetic resonance images. It can also predict cancer aggressiveness in order to provide a useful tool not only for radiologists during the reporting phase but also for urologists and/or oncologists in selecting the best treatment option for each subject.
The CAD consists of a graphical interface that allows the simultaneous visualization of different magnetic resonance images that can be scrolled simultaneously thanks to automatic alignment. The user can activate and disable the view of a colorimetric map automatically processed by the CAD. This map assigns at each pixel a color ranging between red (high probability) and blue (low probability) based on the probability of cancer presence. For each suspected area the CAD also provides an indication of the PIRADS and a probability index related to tumor aggressiveness. The user can evaluate these and other quantitative and/or dimensional parameters for each suspicious regions or for new areas manually outlined and, based on this information, the system automatically creates a structured report.
- Support radiologists for prostate cancer detection;
- Management of active surveillance patients, thanks to the tracking of cancer volume over time and related functional parameters evolution;
- Guidance tool for performing targeted biopsies under ultrasound or MRI guidance.
- Significant increase in sensitivity and specificity of radiologists in the identification of prostate cancers;
- Substantial reduction of reporting times;
- Possibility of providing a second opinion to radiologists, especially in small centers where there is no strong expertise for the analysis of magnetic resonance images of the prostate;
- Better selection of patients to be biopsied, to reduce the number of unnecessary biopsies;
- Possibility of using the CAD to guide biopsies, reducing the number biopsy cores and increasing sampling sensitivity for assessment of cancer aggressiveness.