AUA 2024: Integrating MR and Ultrasound Images for AI-based Prostate Cancer Detection in Transrectal Ultrasound Images: A Comparative Assessment with Clinicians

( The 2024 American Urological Association (AUA) Annual Meeting held in San Antonio, TX was host to a prostate cancer detection and screening moderated poster session. Dr. Hassan Jahanandish presented the results of a comparative analysis of clinician performance versus magnetic resonance imaging (MRI) and ultrasound integration for artificial intelligence-based prostate cancer detection in transrectal ultrasound images.

MRI-guided biopsies have significantly improved the detection of clinically significant prostate cancer. Yet, most biopsies are still performed using b-mode transrectal ultrasound (TRUS) imaging alone. However, TRUS biopsy only detects ~50% of cancer lesions. Recent studies have focused on machine learning approaches for detecting prostate cancer in MR images. This limits the utilization of these models to men who undergo a pre-biopsy prostate MRI. To make machine learning models more accessible to all men undergoing a biopsy, the investigators developed a novel deep learning framework that utilizes biomarkers from both MRI and TRUS images during training to detect prostate cancer foci in TRUS images alone, eliminating the need for MR images in deployment. They evaluated this machine learning model’s performance against four urologists.

This deep learning framework was trained and tested on a dataset of 102 patients (82 training and 20 test cases), with whole-mount pathology as the ground truth. A baseline TRUS-only model was trained using the same dataset with no guidance from the multimodal model. Four urologists, with an average of 5 years of experience reading TRUS prostate images, reviewed the test cohort's TRUS images and manually annotated suspicious lesions without time restrictions.


The multimodal-guided TRUS-only model achieved a sensitivity and specificity of 80% and 70%, respectively. This significantly outperformed the unguided baseline model, which achieved a performance of 54% and 48%, respectively. Furthermore, expert clinicians achieved a lower sensitivity of 35%, with a higher specificity of 92%, compared to this machine learning approach. 


These results demonstrate the effectiveness of a machine learning approach in integrating MRI and TRUS images for prostate cancer detection in TRUS images. The higher sensitivity compared to expert clinicians shows promise for enhancing prostate cancer biopsy diagnosis.

Presented by: Hassan Jahanandish, PhD, Postdoctoral Scholar, Department of Urology, Stanford University, Palo Alto, CA 

Written by: Rashid Sayyid, MD, MSc – Society of Urologic Oncology (SUO) Clinical Fellow at The University of Toronto, @rksayyid on Twitter during the 2024 American Urological Association (AUA) Annual Meeting, San Antonio, TX, Fri, May 3 – Mon, May 6, 2024.