10 mm (PI-RADS v2.1) versus 15 mm (PI-RADS v1.0) tumor capsule contact length in predicting extracapsular extension in prostate cancer: Meta-analysis and systematic review

Abstract

Objective: To determine the diagnostic performance of tumor capsule contact length (TCCL) and to determine its optimal cut-off on prostate MRI to predict extra-capsular extension (ECE) at histopathology of radical prostatectomy specimen in patients with prostate cancer.

Materials and methods: An electronic search of the PUBMED and EMBASE databases was performed until December 31, 2024 to identify studies analyzing TCCL as a predictor of ECE on prostate MRI. Pooled sensitivity and specificity of all studies were plotted in a hierarchical summary receiver operating characteristic plot and the risk of ECE was evaluated using bivariate random-effects meta-analysis. Meta-regression analysis for various TCCL cut-offs was performed.

Results: Twenty-three eligible studies were found. Reported sensitivity (between 47 - 94%) and specificity (between 46 - 88%) showed significant heterogeneity between studies, without evident threshold effect. Pooled sensitivity and specificity of TCCL for predicting ECE was 76.3% and 68.8% respectively with AUC of 0.787 for the summary ROC curve. Meta-regression analysis showed no significant difference in diagnostic performance according to various TCCL cutoffs, measurement methods, or ISUP grade distribution. Studies with 14 mm threshold had similar sensitivity (73%), but greater specificity (74% vs. 70%) and diagnostic odds ratio (8.3 vs. 6.5), when compared to 10 mm threshold.

Conclusion: TCCL on MRI has reasonable pooled sensitivity and specificity to predict ECE. While sensitivity remained moderately high for all TCCL thresholds, specificity at 14 mm (cutoff close to 15 mm suggested by PI-RADS v1.0) threshold was greater than that at 10 mm (cutoff suggested by PIRADS v2.1). This cutoff can be helpful in developing objective likelihood scores for ECE prediction on MRI.

Keywords: Extraprostatic extension; Magnetic resonance imaging; Meta-analysis; Prostate cancer.

Fig. 1

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram showing study selection process for the meta-analysis

Fig. 2

Summary of quality assessment of diagnostic accuracy studies (QUADAS-2) questionnaire results

Fig. 3

Coupled Forest plot of sensitivity and specificity. Horizontal error-bars indicate 95% confidence intervals

Fig. 4

Hierarchical summary receiver operating characteristic curve (hsROC) from the bivariate model for diagnostic accuracy of TCCL to predict extraprostatic extension