MRI-based risk calculator may curb unnecessary prostate cancer biopsies

NEW YORK (Reuters Health) – A risk calculator based on multiparametric MRI (mpMRI) may reduce the number of unnecessary biopsies in patients with prostate cancer while still detecting clinically significant disease, researchers suggest.

“There is growing evidence that mpMRI, in conjunction with targeted biopsies, can detect clinically significant prostate cancer much better than the currently deployed so-called systematic biopsy,” Drs. Baris Turkbey and Sherif Mehralivand of the U.S. National Cancer Institute said in a joint email. “However, in most studies, mpMRI is considered on an individual basis.”

“In this study, we sought to evaluate prostate mpMRI as a biomarker in combination with other well-known predictors like age, ethnicity, prior biopsy status, digital rectal examination and prostate volume,” they told Reuters Health.

“To achieve this, we created a baseline model with all variables except mpMRI and a second model including mpMRI-based PI-RADS (prostate imaging reporting and data system) scores.”

The development cohort was made up of 400 patients from a single institution, and the validation cohort included 251 patients from two independent institutions. Mean age at biopsy was about 64 in both cohorts; more than 80% of the participants were white. Median PSA scores ranged from 6 to 6.6.

As reported online February 22 in JAMA Oncology, 48.3% of the development cohort and 38.2% of the validation cohort had clinically significant prostate cancer, defined as a Gleason score greater than or equal to 3 + 4.

“The diagnostic accuracy improved significantly when PI-RADS scores were added, indicating an additional benefit of mpMRI (area under the curve increased from 72% to 84% in the development cohort),” Drs. Turkbey and Mehralivand noted.


In the validation cohort, mpMRI improved the AUC from 64% to 84%, compared with the baseline model.

At a risk threshold of 20%, the mpMRI model had a significantly lower false-positive rate than the baseline model (46% vs. 92%), with only a small reduction in the true-positive rate (89% vs. 99%).

“Further analysis showed that due to the net reduction of false positives by applying this model,” Drs. Turkbey and Mehralivand said, “a significant number of biopsies could be avoided.”

Specifically, 18 fewer biopsies per 100 men could have been performed, with no increase in the number of patients with missed clinically significant prostate cancers.

“Interestingly, the model showed robustness and the results were similar although the workflows in all three institutions were different,” Drs. Turkbey and Mehralivand noted. “We therefore think that our model can be used by other centers that are using mpMRI and PI-RADS version 2 to determine which patients might benefit from a targeted biopsy and which won’t.”

“Furthermore,” they concluded, “we think that our results will change the view of the scientific community on multiparametric prostate MRI as an additional biomarker among others, instead of a singular diagnostic modality.”

Dr. S. Adam Ramin, medical director of Urology Cancer Specialists in Los Angeles called the study “important and promising.”

“This predictive model can be used to counsel patients about the need for biopsy, reduce false negatives and prevent the need to perform unnecessary prostate biopsies,” he told Reuters Health.

However, he said by email, “more studies will be needed to further validate results. Also, the predictive model is somewhat confusing to use clinically and, therefore, may not be readily adopted.”

“Newer testing techniques available for the clinic can also predict which patients with elevated PSA will need a prostate biopsy,” he added. “These (include) newer blood tests and urine tests that, in combination with PSA, can aid urologists and their patients in determining (whether or not) to proceed with prostate biopsy.”


JAMA Oncol 2018.


This article was from Reuters and was legally licensed through the NewsCred publisher network. Please direct all licensing questions to