Interreader Agreement with Prostate Imaging Reporting and Data System Version 2 for Prostate Cancer Detection: A Systematic Review and Meta-Analysis

Abstract

Purpose: We evaluated interreader agreement with PI-RADSÒ (Prostate Imaging Reporting and Data System) version 2 for detection of prostate cancer. Materials and Methods: We searched MEDLINEÒ, EmbaseÒ and the Cochrane Library between 2015 and June 3, 2019 to identify original research reporting interreader agreement in the use of PI-RADS version 2. Quality of the retrieved studies was assessed by 2 independent reviewers using the Guidelines for Reporting Reliability and Agreement Studies. Pooled k for PI-RADS version 2 was calculated, and a head-to-head comparison with version 1 was performed for the available studies. Subgroup analysis was performed according to zonal anatomy (peripheral or transitional zone), cutoff value (4 or higher, or 3 or higher) and specific imaging sequences (T2-weighted, diffusion-weighted and dynamic contrast enhanced). Meta-regression analysis was performed to assess the cause of study heterogeneity. Results: A total of 30 studies (4,095 patients) were included. Pooled k of PI-RADS version 2 was 0.61 (95% CI 0.55e0.67). In 4 studies evaluating head-to-head comparisons PI-RADS versions 1 and 2 showed similar pooled k values (0.61, 95% CI 0.33e0.90 vs 0.68, 95% CI 0.57e0.79; p[0.61). Substantial interreader agreement was noted with a cutoff of 4 or higher (k[0.61) and moderate agreement was observed with a cutoff of 3 or higher (k[0.57), peripheral zone (k[0.64), transitional zone (k[0.49) and the 3 magnetic resonance imaging sequences (k 0.42e0.58). Difference in reader experience was the single significant factor affecting study heterogeneity (p[0.01). Conclusions: PI-RADS version 2 provides substantial interreader agreement in overall scoring in patients with suspicious prostate cancer, with a similar level of agreement to version 1.

Performance of Prostate Imaging Reporting and Data System Version 2.1 for Diagnosis of Prostate Cancer: A Systematic Review and Meta‐Analysis

“…11 Considering the relatively low positive predictive value of PI-RADS, 6 the reduction in the number of suspicious lesions (category ≥3) to an acceptable range with PI-RADS v2.1 may potentially increase positive predictive values for diagnosing csPCa and decrease the number of false-positive results. However, because a cutoff of ≥3 still has limitations for determining whether patients require biopsy or not because of the suboptimal specificity and considerable variability in the current PI-RADS, 5,10,11 further dedicated criteria incorporating clinical risk factors and clear definitions of imaging features might be needed to minimize For transition zone lesions, our meta-analysis revealed pooled sensitivity of 90% and pooled specificity of 76%, values that are comparable with those for whole gland evaluation. Considering that PI-RADS v2 showed relatively lower diagnostic performance and lower specificity in the transition zone compared with the peripheral zone, 28,29 the comparable diagnostic performance of PI-RADS v2.1 between the transitional zone and whole gland might indicate improvement in diagnostic performance in the transitional zone with PI-RADS v2.1.…”

“…7 Given the improved clarification of lexicons regarding lesion shape or margin and the new definition of atypical nodules in the transition zone for category 2 lesions, these modifications are expected to improve discrimination between PI-RADS category 1-2 lesions and lesions of category ≥3. 5,11 However, because the total category shift from category 2 to 3 between PI-RADS v2 and v2.1 is limited to a very small proportion of the patient population (0.2-2%), when summing the upgrades and downgrades, 11,27 the overall changes between the two versions might not be significant. Although there are conflicting reports on whether PI-RADS v2.1 provides an improvement in sensitivity and specificity in the diagnosis of csPCa in comparison with PI-RADS v2, 11,12,14,26 the large number of patients (over 1000) included in our meta-analysis should make it useful for evaluating the diagnostic performance of PI-RADS v2.1 in comparison with PI-RADS v2.…”

“…[1][2][3] Although PI-RADS version 2 (v2) provides a highly sensitive tool for detecting PCa, 4 variability across institutions and readers, especially for lesions in the transition zone, remains a problem. 5,6 To address this issue, the updated version of PI-RADS version 2.1 (v2.1) was released in early 2019. 7 To promote a more objective approach, PI-RADS v2.1 contains revisions of some previously unclear lexicons and has added the role of diffusion-weighted imaging (DWI) for transition zone lesions.…”

ESUR/ESUI position paper: developing artificial intelligence for precision diagnosis of prostate cancer using magnetic resonance imaging

“…High levels of expertise also enable the adoption of MRI approaches that avoid contrast medium injections (biparametric MRI; bpMRI) [12][13][14], which can help to increase patient throughput at a lower cost. High reader expertise also minimizes variations [15] in clinically significant cancer yields within the MRI suspicion categories, thus improving the uniformity and reliability of MRI findings for clinical decision-making. Also, expert readers have a lower percentage of uncertain diagnoses.…”

A multifaceted approach to quality in the MRI-directed biopsy pathway for prostate cancer diagnosis

“…As pointed out, the most impactful factors affecting the "rule-in" and "rule-out" performances are different. Secondly, measurements, control, and assurance of the sources of variability should be undertaken, such as compliance with technical standards [30], image quality assessment scores [18], inter-and intra-reader/center variability [31], and cancer detection rates (including underdiagnosis cancer rates). Third, we need to develop performance measures for the main individuals participating in the MRI biopsy pathway including radiologists, urologists, and pathologists (and others), ultimately laying down the foundations for the certification of the main players, once tolerances are known [17].…”