Automatic PI-RADS assignment by means of formal methods.

Area of Science:

• Radiology
• Medical Imaging
• Oncology

Background:

• Prostate Imaging Reporting and Data System (PI-RADS) version 2 is the standard for prostate MRI.
• PI-RADS scores, assigned by radiologists, indicate the likelihood of clinically significant prostate cancer.
• Accurate PI-RADS scoring is crucial for patient management.

Purpose of the Study:

• To develop an automated methodology for assigning PI-RADS scores to prostate MRI.
• To create a system that can automatically mark MRI scans with their corresponding PI-RADS score.

Main Methods:

• A dataset of 91 patients' DWI ADC MRI scans from two institutions was collected.
• Expert radiologists assigned PI-RADS scores to the dataset.
• A formal model was developed using radiomic features and expert-defined properties.

Main Results:

• The methodology utilizes formal methods and radiomic features.
• Experimental analysis demonstrated a specificity and sensitivity of 1.
• The system achieved very high accuracy in assigning PI-RADS scores.

Conclusions:

• The proposed automated methodology accurately assigns PI-RADS scores to prostate MRI.
• This approach has the potential to improve the efficiency and consistency of prostate cancer assessment.
• Further validation may enhance clinical adoption for improved prostate cancer diagnosis.