Scholarlink

Objective <p>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.</p> Materials and methods <p>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.</p> Results <p>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&#xa0;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&#xa0;mm threshold.</p> Conclusion <p>TCCL on MRI has reasonable pooled sensitivity and specificity to predict ECE. While sensitivity remained moderately high for all TCCL thresholds, specificity at 14&#xa0;mm (cutoff close to 15&#xa0;mm suggested by PI-RADS v1.0) threshold was greater than that at 10&#xa0;mm (cutoff suggested by PIRADS v2.1). This cutoff can be helpful in developing objective likelihood scores for ECE prediction on MRI.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

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

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年6月5日

查看全文
  • Dheeman Futela,
  • Mahima Bhargava,
  • Sharath Rama,
  • Sishir Doddi,
  • Yong Chen,
  • Nikhil H Ramaiya,
  • Sree Harsha Tirumani

摘要

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.