Learning curve for magnetic resonance imaging/ultrasound fusion prostate biopsy in detecting prostate cancer using cumulative sum analysis.
Curr Urol
; 17(3): 159-164, 2023 Sep.
Article
en En
| MEDLINE
| ID: mdl-37448610
Background: Targeted magnetic resonance (MR) with ultrasound (US) fusion-guided biopsy has been shown to improve detection of prostate cancer. The implementation of this approach requires integration of skills from radiologists and urologists. Objective methods for assessment of learning curves, such as cumulative sum (CUSUM) analysis, may be helpful in identifying the presence and duration of a learning curve. The aim of this study is to determine the learning curve for MR/US fusion-guided biopsy in detecting clinically significant prostate cancer using CUSUM analysis. Materials and methods: Retrospective analysis was performed in this institutional review board-approved study. Two urologists implemented an MR/US fusion-guided prostate biopsy program between March 2015 and September 2017. The primary outcome measure was cancer detection rate (CDR) stratified by Prostate Imaging Reporting and Data System (PI-RADS) scores assigned on the MR imaging. Cumulative sum analysis quantified actual cancer detection versus a predetermined target satisfactory CDR of MR/US fusion biopsies in a sequential case-by-case basis. For this analysis, satisfactory performance was defined as >80% CDR in patients with PI-RADS 5, >50% in PI-RADS 4, and <20% in PI-RADS 1-3. Results: Complete data were available for MR/US fusion-guided biopsies performed on 107 patients. The CUSUM learning curve analysis demonstrated intermittent underperformance until approximately 50 cases. After this inflection point, there was consistently good performance, evidence that no further learning curve was being encountered. Conclusions: At a new center implementing MR/US fusion-guided prostate biopsy, the learning curve was approximately 50 cases before a consistently high performance for prostate cancer detection.
Texto completo:
1
Colección:
01-internacional
Base de datos:
MEDLINE
Tipo de estudio:
Prognostic_studies
Idioma:
En
Revista:
Curr Urol
Año:
2023
Tipo del documento:
Article
País de afiliación:
Estados Unidos
Pais de publicación:
Estados Unidos