2397 - Impact of Hip Prostheses on Restriction Spectrum Imaging for Prostate Cancer Detection

E. M. Qiao¹, D. Do¹, C. Ollison¹, M. Baxter¹, C. C. Conlin², M. Rojo¹, K. Kallis¹, Y. Song¹, K. Santhosh¹, M. E. Hahn², P. Murphy³, J. Kuperman², R. Rakow-Penner², A. M. Dale⁴, and T. M. Seibert⁵; ¹Department of Radiation Medicine and Applied Sciences, University of California San Diego, La Jolla, CA, ²Department of Radiology, University of California San Diego, La Jolla, CA, ³Department of Radiology, University of California - San Diego, La Jolla, CA, ⁴Halicioglu Data Science Institute, University of California San Diego, La Jolla, CA, ⁵University of California - San Diego, San Diego, CA

Purpose/Objective(s): MRI-based restriction spectrum imaging (RSI) restriction score (RSIrs) is a quantitative imaging biomarker for clinically significant prostate cancer (csPCa, grade group =2). RSIrs can improve csPCa detection for diagnosis and radiotherapy planning. Maximum RSIrs within the prostate performs comparably to subjective PI-RADS for classifying csPCa. Prosthetic hip implants can cause artifacts in MRI, but few data exist on how they affect RSIrs in the prostate. We evaluated the impact of hip prostheses on maximum RSIrs values and the biomarker’s utility for these patients. Materials/

Methods: This is a single-institution, retrospective analysis of 731 patients who underwent MRI with RSI for suspected csPCa from 2018-2023. Student’s t-test compared maximum RSIrs for patients with and without prosthetic hip implants. Multivariable logistic regression models evaluated association between hip implant, RSIrs, and biopsy-proven csPCa, adjusting for demographic and prostate-specific variables. Sensitivity and specificity of csPCa detection at varying RSIrs values were calculated.
Results: Among 731 patients, 46 (6%) had a hip implant, and 346 (47%) had csPCa. The mean of within-prostate-maximum RSIrs for patients with and without hip implants was 788 and 307, respectively (t (degrees of freedom): 2.2 (46), p: 0.04). In patients without hip implants, each 100-unit increase of RSIrs was associated with 20% increased odds of csPCa (odds ratio (OR): 1.20; 95% confidence interval (CI): 1.07-1.40; p <0.01). When including patients with hip implants, RSIrs was no longer significantly associated with csPCa diagnosis (OR: 1.03; 95% CI: 0.99-1.07; p: 0.16). After adjusting for interaction between RSIrs and hip implant, though, RSIrs had similar significant association with csPCa diagnosis as when excluding hip implant patients (OR: 1.21; 95% CI: 1.07-1.40; p <0.01, p_interaction: <0.01). Table 1 shows sensitivity and specificity of RSIrs cutoff values stratified by hip implant status.
Conclusion: Patients with hip implants appear to have higher baseline RSIrs values. Utilizing RSIrs without adjusting for hip prostheses may lead to unnecessary biopsies in patients with hip implants. Additional studies should focus on optimal adjustment of RSIrs in the presence of hip prosthesis. Table 1. Sensitivity and specificity for RSIrs cutoffs

	Hip implant (-)	Hip implant (-)	Hip implant (+)	Hip implant (+)
RSIrs	Sensitivity (%)	Specificity (%)	Sensitivity (%)	Specificity (%)
150	93	13	100	11
200	82	42	89	16
250	66	70	77	57
300	50	83	63	68
350	37	92	55	79