PROBLEMS WITH PROSTATE MRI

The use of multiparametric (MP) MRI has improved the detection of clinically important prostate cancers and has cut down on the detection of indolent cancers.

However, new findings from a study involving 100 consecutive men with prostate cancer who were treated at a major imaging center show that a substantial number of cancers are missed.

About 16% of lesions were missed, and approximately 5% of clinically important prostate cancers (>5 mm, Gleason score > 3+3) were underestimated on MP MRI, according to Baris Turkbey, MD, of the Molecular Imaging Program at the National Cancer Institute in Bethesda, Maryland, and colleagues.

Overall, prostate cancer size was underestimated by at least 30% in eight (8%) of 100 patients.

Their study was published online October 20 in Radiology.

Use of advanced MRI for patients with prostate cancer is being increasingly studied, as it holds promise for detecting potentially aggressive prostate tumors as well as possibly sparing some men from having to undergo invasive biopsies. Last year, the first study of MRI as a primary prostate cancer screening tool reported that its use was "feasible" for the general population.

The authors of the current study note that although MP MRI has been validated in a number of studies, the modality fails to detect some clinically important cancers.

"It is widely quoted that the 'miss rate' on MRI for clinically significant tumors is between 10% and 20%," said Dr Turkbey. "Our study showed comparable results."

The purpose of this study was to illustrate cases of MRI-negative prostate cancer to show that, even on technically good MRI imaging studies from an experienced site, MRI is imperfect, he explained.

"It misses some tumors, and it underestimates others. On balance, however, MRI has consistently been shown to be of a net benefit, as it shows 30% to 40% more clinically significant lesions than random biopsies."

It's very similar to the situation with mammograms. Dr Baris Turkbey

Therefore, MRI is a very useful adjunct in prostate cancer diagnosis, but it is not 100% accurate. "Patients need to understand that ― and that it's very similar to the situation with mammograms," Dr Turkbey said.

High Detection Rate, but False Negatives Occur

In this study, Dr Turkbey and his colleagues sought to identify and characterize false negative, clinically important prostate cancers that were missed during MP MRI assessment in a retrospective, single-center study.

The cohort consisted of men who had undergone MP MRI and subsequent radical prostatectomy, which allowed for a definitive analysis of their cancerous lesions. The men's imaging studies were prospectively interpreted by two genitourinary radiologists who were blinded to the histopathologic findings.

Two other readers, who were not blinded to the histopathologic findings, assigned a Prostate Imaging Reporting and Data System (PI-RADS) version 2 score to each clinically important lesion (0.5 mm; Gleason score, 3+3) that was missed on MP MRI.

A total of 162 clinically important lesions were included in the final analysis. The duration between MRI and prostatectomy was 128 days (median, 120 days; range, 2 - 390 days).

At least one clinically important cancer was detected at initial MRI interpretation in 99 patients (99%). In one individual (aged 59 years; prostate-specific antigen [PSA] level, 3.75 ng/mL; PSA density, 0.045 ng/ml²), all lesions were missed on MRI. At least one clinically important lesion was missed in 26 (26%) of the cohort, and the size of clinically important cancers was underestimated in eight (8%) patients.

Taken together, 136/162 lesions (84.0%) were correctly identified by MRI.

Eight lesions (4.9%) were prospectively identified, but tumor size was underestimated on MRI (Gleason score, 3+4; n = 5 [3.1%]; Gleason score, 4+4; n = 2 [1.2%]; Gleason score, 4+5; n = 1 [0.6%]).

A total of 26 (16%) lesions were missed during the initial interpretation (one per patient). Gleason scores in this patient subset were 3+4 (n = 17 [65%]), 4+3 (n = 1 [4%]), 4+4 (n = 7 [27%]), and 4+5 (n = 1 [4%]).

During the retrospective review, eight (31%) PI-RADS 1 lesions were not detected on MP MRI; 18 (69%) lesions were visible (PI-RADS 2, n = 7 [27%]; PI-RADS 3, n = 6 [23%]; PIRADS 4, n = 5 [20%]). Within this group, five (62%) were Gleason 3+4 lesions, and three (38%) were Gleason 4+4 lesions.

"We think that we can detect about half of the missed lesions and reduce the number of greatly underestimated lesions using computer-aided diagnosis algorithms," Dr Turkbey pointed out. "However, a number of patients will escape detection of their clinically significant prostate cancer using MRI alone. Therefore, we and many other centers continue to advocate for systematic biopsies in addition to targeted biopsies, and we are investigating several aspects of this phenomenon."

Only 16 Patients Underwent MP MRI Before Biopsy: What That Means

Commenting on the study, Hashim U. Ahmed, PhD, BM, chair of urology and consultant urologic surgeon at Imperial College, London, United Kingdom, pointed out that all the patients in this series had biopsy-proven cancer.

"The patients had identifiable cancer already, and as the authors show, 99 of the 100 men had one clinically significant cancer lesion correctly identified by MRI," Dr Ahmed said. "So, for the individual, there is a large probability of MP MRI identifying a significant cancer."

He explained that it is already known that MP MRI has a miss rate for significant disease. In a clinical trial conducted earlier this year by Dr Ahmed and his colleagues, MP MRI administered as a triage test before the first transrectal ultrasound biopsy identified at least one quarter of all patients who could safety avoid undergoing a biopsy. Results of this study also suggested that this technology might improve the detection of clinically significant prostate cancer.

"If one uses template mapping biopsies, as we did in the multicenter, prospective, blinded, validation PROMIS study of 11 expert and nonexpert centers, no cancers of Gleason 4+3 or higher of any volume were missed by MP MRI," he said. "This is already leaps and bounds better than the current transrectal systematic biopsy."

This is already leaps and bounds better than the current transrectal systematic biopsy. Dr Hashim Ahmed

Dr Ahmed did note that the current study did have some methodologic problems.

"Only 16 of 100 were true prebiopsy MP MRI," he told Medscape Medical News. "We have known for over a decade that postbiopsy MRI scans are significantly degraded with biopsy-related inflammation and bleeding for between 3 and 6 months after a biopsy, and in some cases, up to 9 to 12 months. The study also only evaluated the two largest lesions on the report ― if there were significant MRI lesions after this upper limit of two, it seems the authors denoted them as 'negative' on the MP MRI."

In explaining another concerning factor, he said, "We also need to be cognizant of the fact that MP MRI will not be the sole source of cancer localization and extent. Did the biopsies detect significant disease just outside and adjacent to the MRI lesion and in other areas? We don't know this.

"We do know from a number of other studies, and this study confirms this, that MP MRI can underestimate the volume of the lesion on pathology," Dr Ahmed continued. "This information needs to be incorporated into targeted biopsy strategies and into treatment strategies, such as dose-escalation radiotherapy or focal ablative therapy."