Incorporating MRI-based parameters into a risk model could cut down on the number of unnecessary biopsies performed in patients with suspected prostate cancer.
Incorporating MRI-derived parameters into a clinical risk model could cut down on the number of unnecessary biopsies performed in patients with suspected prostate cancer, according to a new study. This method could still maintain a high rate of diagnosis of clinically significant cancers.
“Transrectal systematic biopsy remains the standard of care for diagnosing prostate cancer. Use of this biopsy has led to an increased detection of low-grade cancers, which can result in overtreatment,” wrote study authors led by Sherif Mehralivand, MD, of the National Cancer Institute in Bethesda, Maryland. “It would be desirable to reduce the biopsy rate in men who ultimately prove to have benign conditions or low-grade disease.”
As multiparametric MRI and MRI-transrectal ultrasound (TRUS) fusion-guided biopsy have become more common, attempts to standardize reading of MRI have emerged. The investigators hypothesized that incorporating MRI-derived prostate volumes and categories from the Prostate Imaging Reporting and Data System Version 2 (PI-RADS v2) into a clinical risk model could reduce biopsy rates.
They included 400 patients in a development cohort, and 251 patients in a validation cohort. All patients underwent MRI, MRI-TRUS fusion-guided biopsy, and 12-core systematic biopsy. All detected lesions were assigned a category based on PI-RADS v2 guidelines, from 1 to 4, and this, along with MRI-derived prostate volume, was incorporated into a model that included age, ethnicity, and other commonly used variables. The results of the study were published in JAMA Oncology.
In the development cohort, 193 patients (48.3%) had clinically significant prostate cancer; in the validation cohort, 96 patients (38.2%) had clinically significant disease. The risk for clinically significant prostate cancer was inversely associated with prostate volume, and increased with prostate-specific antigen density and PI-RADS v2 category.
Compared with the baseline model, the model incorporating MRI increased the area under the curve (AUC) from 72% to 84% (P < .001) in the development cohort. In the validation cohort, the AUC increased from 64% to 84% (P < .001) with the MRI model.
Both false-positive and true-positive rates were improved with the MRI model. The net reduction in false positives using the MRI model, when compared with performing a biopsy in all patients with positive MRI results, was equivalent to performing 18 fewer unnecessary biopsies per 100 men, and with no increase in the number of clinically significant prostate cancers that would go undiagnosed. Overall, the MRI model could help avoid 38% of biopsies, compared with 6% with the baseline clinical risk model.
“When MRI was incorporated into a prediction model, it exhibited improved model fit and superior diagnostic accuracy, reducing unnecessary biopsies while maintaining a similar level of sensitivity for high-risk cancers,” the authors wrote, adding that it should now be validated prospectively in independent cohorts and centers.