NPV of Biparametric and Multiparametric Prostate MRI: A Comparative Systematic Review and Meta-Analysis.

Abstract

<b>BACKGROUND.</b> Evidence supports comparable PPV between biparametric MRI (bpMRI) and multiparametric MRI (mpMRI). However, concern regarding missed cancers limits wider adoption of bpMRI. <b>OBJECTIVE.</b> The purpose of this study was to compare bpMRI and mpMRI in terms of the NPV for clinically significant prostate cancer. <b>EVIDENCE ACQUISITION.</b> Multiple publication databases, trial registries, and conference proceedings were searched over varying time frames to identify studies reporting comparative results for bpMRI and mpMRI. Information was extracted for negative examinations (PI-RADS category or Likert scale score of 1 or 2), which were classified as true- or false-negative for clinically significant prostate cancer (International Society of Urological Pathology [ISUP] grade group ≥ 2), with a pathologic reference standard (biopsy and/or radical prostatectomy). Risk of bias was assessed using QUADAS-Comparative. Pooled NPVs were calculated using random-effects meta-analysis. <b>EVIDENCE SYNTHESIS.</b> The meta-analysis included 18 studies. Fifteen studies evaluated simulated bpMRI examinations (examinations performed as mpMRI but interpreted with the removal of dynamic contrast-enhancement images); three studies compared parallel arms of patients who underwent bpMRI or mpMRI. No study evaluated patients randomly allocated to undergo bpMRI or mpMRI. The reference standard of three studies included longitudinal follow-up biopsy. Pooled NPV was not significantly different between bpMRI (<i>n</i> = 2857 patients) and mpMRI (<i>n</i> = 2751 patients) overall (92% [95% CI, 89-94%] vs 92% [95% CI, 89-94%]; <i>p</i> = .90) in nine studies after exclusion of studies with a high risk of bias in at least one domain of QUADAS-Comparative (92% [95% CI, 86-95%] vs 92% [95% CI, 95-96%]; <i>p</i> = .83), in three studies of 1.5-T examinations only (89% [95% CI, 78-95%] vs 87% [95% CI, 77-93%]; <i>p</i> = .76), in 12 studies of 3-T examinations only (93% [95% CI, 90-95%] vs 93% [95% CI, 91-95%]; <i>p</i> = .90), in 12 studies of biopsy-naive patients only (92% [95% CI, 88-94%] vs 91% [95% CI, 89-93%]; <i>p</i> = .89), or in three studies of previously biopsied patients only (94% [95% CI, 89-97%] vs 94% [95% CI, 85-98%]; <i>p</i> = .95). <b>CONCLUSION.</b> This study found no evidence of a significant difference between bpMRI and mpMRI in terms of NPV for clinically significant prostate cancer. <b>CLINICAL IMPACT.</b> The results provide further support for bpMRI as an alternative to mpMRI in clinical practice. Future studies should include randomized designs with longitudinal follow-up. <b>TRIAL REGISTRATION.</b> PROSPERO identifier CRD42023491456.