External validation of two MRI-based risk calculators in prostate cancer diagnosis

Anna Lena Petersmann, Sebastiaan Remmers, Tilman Klein, Panagiota Manava, Clemens Huettenbrink, Sascha A. Pahernik, Florian A. Distler*

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Background: The diagnosis of (significant) prostate cancer ((s)PC) is impeded by overdiagnosis and unnecessary biopsy. Risk calculators (RC) have been developed to mitigate these issues. Contemporary RCs integrate clinical characteristics with mpMRI findings. Objective: To validate two of these models—the MRI-ERSPC-RC-3/4 and the risk model of van Leeuwen. Methods: 265 men with clinical suspicion of PC were enrolled. Every patient received a prebiopsy mpMRI, which was reported according to PI-RADS v2.1, followed by MRI/TRUS fusion-biopsy. Cancers with ISUP grade ≥ 2 were classified as sPC. Outcome measurements and statistical analysis: Statistical analysis was performed by comparing discrimination, calibration, and clinical utility Results: There was no significant difference in discrimination between the RCs. The MRI-ERSPC-RC-3/4-RC showed a nearly ideal calibration-slope (0.94; 95% CI 0.68–1.20) than the van Leeuwen model (0.70; 95% CI 0.52–0.88). Within a threshold range up to 9% for a sPC, the MRI-ERSPC-RC-3/4-RC shows a greater net benefit than the van Leeuwen model. From 10 to 15%, the van Leeuwen model showed a higher net benefit compared to the MRI-ERSP-3/4-RC. For a risk threshold of 15%, the van Leeuwen model would avoid 24% vs. 14% compared to the MRI-ERSPC-RC-3/4 model; 6% vs. 5% sPC would be overlooked, respectively. Conclusion: Both risk models supply accurate results and reduce the number of biopsies and basically no sPC were overlooked. The van Leeuwen model suggests a better balance between unnecessary biopsies and overlooked sPC at thresholds range of 10–15%. The MRI-ERSPC-RC-3/4 risk model provides better overall calibration.

Original languageEnglish
Pages (from-to)4109-4116
Number of pages8
JournalWorld Journal of Urology
Volume39
Issue number11
Early online date24 Jun 2021
DOIs
Publication statusPublished - Nov 2021

Bibliographical note

Funding Information:
We thank Monique J. Roobol for critically reviewing the study design and manuscript. We thank the Else Kr?ner-Fresenius-Foundation for sustaining the prostate cancer research group at Paracelsus Medical University Nuremberg.

Publisher Copyright:
© 2021, The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.

Fingerprint

Dive into the research topics of 'External validation of two MRI-based risk calculators in prostate cancer diagnosis'. Together they form a unique fingerprint.

Cite this