J Magn Reson Imaging. 2021 Jul 5. doi: 10.1002/jmri.27803. Online ahead of print.
ABSTRACT
BACKGROUND: Diffusion weighted imaging (DWI) is fundamental for prostate cancer (PCa) detection with MRI; however, limited by susceptibility artifact from hip prosthesis.
PURPOSE: To evaluate image quality and ability to detect PCa with quantitative T2-mapping and DWI in men with hip prosthesis undergoing prostate MRI.
STUDY TYPE: Prospective, cross-sectional study.
POPULATION: Thirty consecutive men with hip replacement (18 unilateral, 12 bilateral) undergoing prostate MRI from 2019 to 2021.
FIELD STRENGTH/SEQUENCE: 3-T; multiparametric MRI (T2W, DCE-MRI, echo-planar [EPI]-DWI), T2-mapping (Carr-Purcell-Meiboom-Gill), FOCUS-EPI-DWI, PROPELLER-DWI.
ASSESSMENT: Five blinded radiologists independently evaluated MRI image quality using a 5-point Likert scale. PI-RADS v2.1 scores were applied in four interpretation strategies: 1) T2W-FSE+DCE-MRI+EPI-DWI, 2) T2W-FSE+DCE-MRI+EPI-DWI+FOCUS-EPI-DWI, 3) T2W-FSE+DCE-MRI+EPI-DWI+PROPELLER-DWI, 4) T2W-FSE+DCE-MRI+EPI-DWI+T2-maps. Five-point confidence scores were recorded.
STATISTICAL ANALYSIS: ANOVA, Kruskal-Wallis with pair-wise comparisons by Wilcoxon sign-rank, and paired t-tests, P < 0.05 was considered significant. Cohen’s Kappa (k) for PI-RADSv2.1 scoring and proportion of correctly classified lesions tabulated for pathology-confirmed cases with 95% confidence intervals (CIs).
RESULTS: For all radiologists, T2-map image quality was significantly higher than EPI-DWI, FOCUS-EPI-DWI, and PROPELLER-DWI and similar (P = 0.146-0.706) or significantly better (for two readers) than T2W-FSE and DCE-MRI. PI-RADS v2.1 agreement improved comparing strategy A (k = 0.46) to strategy B (k = 0.58) to strategy C (k = 0.58) and was highest with strategy D which included T2-maps (k = 1.00). Radiologists’ confidence was significantly highest with strategy D. Strategies B and C had similar confidence (P = 0.051-0.063) both significantly outperforming strategy A. Twelve men with 17 lesions had pathology confirmed diagnoses (13 PCa, 4 benign). Strategy D had the highest proportion of correctly classified lesions (76.5-82.4%) with overlapping 95% confidence intervals.
DATA CONCLUSION: T2-mapping may be a valuable adjunct to prostate MRI in men with hip replacement resulting in improved image quality, higher reader confidence, interobserver agreement, and accuracy in PI-RADS scoring.
LEVEL OF EVIDENCE: 1 TECHNICAL EFFICACY STAGE: 2.
PMID:34223675 | DOI:10.1002/jmri.27803
