MRI-Guided Triage Could Reduce Overdiagnosis in Localized Prostate Cancer

Use of multiparametric MRI could allow a significant portion of men with elevated PSA levels to avoid undergoing a biopsy without missing clinically significant prostate cancers.

The use of multiparametric magnetic resonance imaging (MP-MRI) could allow a significant portion of men with elevated prostate-specific antigen (PSA) levels to avoid undergoing a biopsy without missing clinically significant prostate cancers, according to results of a new study presented at the 2016 American Society of Clinical Oncology (ASCO) Annual Meeting, held June 3–7 in Chicago (abstract 5000).

“Men with elevated PSA or other risk factors are currently undergoing transrectal ultrasound-guided biopsies, which have a number of errors,” said Hashim Uddin Ahmed, PhD, of University College London in the United Kingdom, who presented the study. This method tends to overdiagnose clinically indolent lesions, miss clinically significant lesions, and incorrectly classify important cancers as unimportant.

The PROMIS study tested wither a MP-MRI prior to biopsy might eliminate some of those issues. It included a total of 576 men at 11 centers in the United Kingdom who underwent the MRI prior to both a template prostate mapping biopsy (used as reference test) and a transrectal ultrasound-guided (TRUS) biopsy. The MRI yielded a LIKERT score of 1 to 5, with 3 or higher being scored as a positive test. The radiologists and pathologists were blinded to results of the biopsies and MRI results.

TRUS biopsy had a sensitivity of 48%, a specificity of 96%, a positive predictive value (PPV) of 90%, and a negative predictive value (NPV) of 74%. MRI, meanwhile, had a higher sensitivity of 93%, a specificity of 41%, a PPV of 51%, and a NPV of 89%. Ahmed said there was a “strong correlation” between MRI score and the presence or absence of clinically significant disease, with scores of 4 or 5 highly likely to be harboring significant cancer.

The MRI’s sensitivity was significantly better than TRUS biopsy, with an odds ratio (OR) of 0.06 (95% CI, 0.02–0.12; P < .0001). The same was true of NPV, with an OR of 0.34 (95% CI, 0.21–0.55; P < .0001). The biopsy was significantly better than MRI with regard to specificity and PPV, but Ahmed said “if we were to use MRI as a triage test to identify men who could avoid a biopsy, then we need a high sensitivity and a high NPV.”

TRUS biopsy alone would have missed 119 clinically significant prostate cancers in this cohort, compared with 17 with MRI used as a triage test followed by biopsy. In total, 27% of men in the study could have avoided a biopsy safely, Ahmed said. There would also have been less overdiagnosis of insignificant cancer (16% with TRUS biopsy alone vs 11% with MRI-guided triage).

“The high sensitivity and NPV of MP-MRI justify its use as a triage test to identify those men who might avoid a primary biopsy,” Ahmed concluded. “However, the low specificity and PPV of prostate MP-MRI indicate that men should still undergo a biopsy if they have a suspicious MRI.”

The discussant for the session, Eric A. Klein, MD, of the Cleveland Clinic in Ohio, agreed that the data suggest using MRI would do a better job of not overdetecting low-grade cancers and a better job of not missing biologically significant cancers, “which is really what we want an effective screening program to do.” He noted that prospective data on this method will still be needed in order to determine if MRI-guided triage is a cost-effective method for screening for prostate cancer.