The accuracy of multiparametric MRI in men with negative biopsy and elevated PSA level—Can it rule out clinically significant prostate cancer?

Abstract

PurposeTo assess the performance of multiparametric magnetic resonance imaging (mp-MRI) in patients with previous negative transrectal ultrasound (TRUS) guided prostate biopsy.Materials and methodsFifty-four patients with at least 1 previous negative TRUS prostate biopsy underwent mp-MRI in the form of T2-weighted, diffusion-weighted, and dynamic contrast-enhanced imaging. This was followed by transperineal template systematic prostate biopsies. Analysis was done based on 2 sectors per prostate, right and left (108 sectors out of 54 prostates). mp-MRI was scored using an ordinal scale 1 to 5 based on the suspicion of the presence of clinically significant disease. We used 6 different definitions for clinically significant disease and tested the performance of mp-MRI at each single definition.ResultsMedian age was 64 (range, 39–75), median PSA level was 10 (range, 2–23), and median number of biopsies was 45 (range, 21–137). Cancer of any volume and any grade was detected in 34 of 54 (63%) patients. mp-MRI accuracy at detection of clinically significant cancer using University College London (UCL) definition 2 (any Gleason score of 4 or maximum cancer core length of ≥4 mm or both) showed sensitivity of 76%, specificity of 42%, positive predictive value of 38%, and negative predictive value of 79%. For a different definition of significant tumor (UCL definition 1; dominant Gleason score 4 or maximum cancer core length ≥6 mm or both), the sensitivity was 90%, specificity 42%, positive predictive value 26%, and negative predictive value 95%.Conclusionsmp-MRI showed good performance at both detection and ruling out clinically significant disease, according to the definition used. mp-MRI can then be used as a triage test in the population with persistently elevated or rising PSA levels to select patients that can avoid unnecessary prostate biopsy.

Multiparametric magnetic resonance imaging outperforms the Prostate Cancer Prevention Trial risk calculator in predicting clinically significant prostate cancer

“…Conversely, 3T MP-MRI with endorectal coil for the identification of suspicious prostate lesions and subsequent MRI/TRUS fusion-guided prostate biopsy has exhibited high levels of sensitivity and negative predictive values (>90%) in detecting clinically significant disease. 8,18 In the current study, we observed an overall CDR of 64.6% which is comparable to the 54% to 63% range reported in the literature. 11,19,20 We observed an AUC of 0.769 (95% CI, 0.703-0.834) for lesion suspicion level on MP-MRI in predicting high-grade CaP, which is 10 percentage points better than the PCPTHG AUC of 0.676 (95% CI, 0.592-0.751; P 5 .091).…”

“…However, it has been demonstrated that MP-MRI of the prostate exhibits a high negative predictive value of >90%. 18 Moreover, we previously validated the MRI/TRUS fusion-guided biopsy platform in a community setting, attesting to the reproducibility of results. 11 Furthermore, our images, which were obtained using 3T MRI with a 16-channel cardiac coil placed on the anterior pelvis and an endorectal coil, were of superior quality (for MRI parameters, see online supporting information).…”

Evidence-based guideline recommendations on multiparametric magnetic resonance imaging in the diagnosis of prostate cancer: A Cancer Care Ontario clinical practice guideline

“…A total of 8663 original studies in English were identified and 15 were analyzed. [2][3][4][6][7][8][23][24][25][26][27][28][29][30][31] The overall quality varied from low to moderate. Fourteen conference abstracts will require followup in future updated versions.…”

“…• In three studies (n=570) 8,27,31 with a prevalence of CSPC of 18-34%, the ranges of sensitivity, specificity, positive predictive value, and negative predictive value of mpMRI to detect CSPC were 68-100%, 41-91%, 29-87%, and 79-100%, respectively.…”

“…• Seven eligible studies 4,8,[27][28][29][30][31] addressed the second objective. The quality of evidence ranged from poor to moderate.…”