Advancing Prostate Cancer Imaging for Better Biopsies

Yale experts use machine-learning techniques to improve the diagnosis of prostate cancer

Men have good reason to be concerned about prostate cancer; for them, here in the United States, it’s the second most common (and third most deadly) kind of cancer. But despite how common it is, prostate cancer can be challenging to diagnose, since traditional technology is somewhat imprecise. Yale Medicine urologist Preston Sprenkle, MD, and his research team are working to change that.  

Most men cringe when they consider the process required to obtain the biopsy that allows doctors to identify prostate cancer. Relying on ultrasound guidance to locate the prostate, the current approach requires insertion of a number of biopsy needles, one after the next, to take random samples in the hope of getting one from the area of concern. This approach has been necessary because it’s difficult to see inside the prostate and know where to properly target the biopsy. As a result, the randomness of the biopsy sampling means cancerous lesions are sometimes missed.  

“The prostate gland sits in a difficult location, and it’s a soft organ,” says Dr. Sprenkle. “It rotates and moves all over in the pelvis, making it hard to evaluate for biopsy.”  

Now new techniques are improving this process. Yale was an early adopter of a new technology that blends ultrasound with magnetic resonance (MR) imaging. This algorithm-based tool, called MR-US fusion, creates 3-D images of the prostate that enable doctors to identify suspicious lesions and target them for biopsy.  

This video explains how Dr. Sprenkle and his colleagues, including John Onofrey, PhD, use clinical data and machine learning to improve fusion methods for improved biopsy guidance.  

To learn more, visit yalemedicine.org.