Two Simple MRI Metrics Can Improve Prostate Cancer Detection and Avoid Unnecessary Biopsies
At Stanford Medicine, for patients with a prostate-specific antigen (PSA) blood test suggesting an elevated risk for prostate cancer, an MRI scan is often the next step instead of proceeding immediately to ultrasound-guided biopsy. The images are then interpreted by a radiologist, and any lesions found are assigned a score based on a standardized scoring system known as PI-RADS (Prostate Imaging Reporting and Data System). A PI-RADS score of 1-2 indicates a low likelihood of clinically significant prostate cancer. A score of 3 suggests an equivocal (intermediate) risk, and scores 4-5 indicate a higher probability.
Even though the PI-RADS scoring system has been used for over a decade, false positives are still a major issue. In fact, lesions with a PI-RADS score of 3 or higher only have a 35% chance of being clinically significant prostate cancer - this means that up to 65% of patients undergo unnecessary biopsy procedures for findings that turn out to be harmless.
A More Precise Approach
A Stanford Medicine research team led by Pejman Ghanouni, MD, PhD, Associate Professor of Radiology, Mirabela Rusu, PhD, Assistant Professor of Radiology, and Geoffrey Sonn, MD, Associate Professor of Urology and, by courtesy, of Radiology (Body MRI), are tackling this problem head-on. In a recent study published in Radiology, the team showed that by adding two simple measures from the initial MRI scan, apparent diffusion coefficient (ADC) and prostate-specific antigen density (PSAD), doctors can more accurately identify which prostate lesions are likely to contain clinically significant prostate cancer.
The team, including Simon John Christoph Soerensen, MD, lead author of the study and PhD student in Epidemiology and Clinical Research, looked at data from nearly 950 men who had prostate MRIs followed by MRI-ultrasound fusion biopsies. Clinically significant prostate cancer was identified in about 11% of lesions scored PI-RADS 3, 33% of lesions scored PI-RADS 4, and 59% of lesions deemed PI-RADS 5 by a radiologist. But when the two additional metrics, ADC and PSAD were factored in, the accuracy of predicting cancer improved significantly.
Adding thresholds for low ADC and high PSAD, the likelihood of a PI-RADS 3 lesion being significant cancer rose from around 1 in 10 to roughly 1 in 3 (11% to 32%). This increase was also pronounced for PI-RADS 4 lesions, with a jump from 1 in 3 (33%) to approximately 3 in 5 (59%). For those with PI-RADS 5 lesions, the addition of these two new metrics meant that the likelihood of the lesion being significant cancer rose from around 3 in 5 (59%) to nearly 3 in 4 (73%).
Top Left: Pejman Ghanouni, MD, PhD, Top Right: Mirabela Rusu, PhD, Bottom Left: Geoffrey Sonn, MD, Bottom Right: Simon John Christoph Soerensen, MD (PhD Student)
Images of the prostate acquired using T2-weighted MRI (left) and diffusion-weighted imaging (DWI; right).
While generally most PI-RADS 3-5 lesions are recommended for biopsy, the study suggests that by incorporating ADC and PSAD thresholds, some men (as many as 29% of the men in this study) could potentially have continued to be monitored without an invasive biopsy.
This more precise approach could mean fewer unnecessary biopsies, better targeting of suspicious areas, and more personalized care overall. For patients, this translates to fewer invasive procedures, less worry about false alarms, and a clearer path forward when cancer is suspected. For doctors, it offers a practical way to use information they already have to make sharper decisions. Together, that adds up to care that is not only more accurate but also more patient-centered.
For patients at Stanford Medicine, these two metrics are already being reported and used for all MRIs in patients with suspected prostate cancer.