Utility of 18F-rhPSMA-7.3 PET for Imaging of Primary Prostate Cancer and Preoperative Efficacy in N-Staging of Unfavorable Intermediate- to Very High-Risk Patients Validated by Histopathology.

Abstract

18F-rhPSMA-7.3, the lead compound of a new class of radiohybrid prostate-specific membrane antigen (rhPSMA) ligand, is currently in phase III trials for prostate cancer (PCa) imaging. Here, we describe our experience in primary PCa staging. Methods: We retrospectively identified 279 patients with primary PCa who underwent 18F-rhPSMA-7.3 PET/CT (staging cohort). A subset of patients (83/279) subsequently underwent prostatectomy with lymph node (LN) dissection without prior treatment (efficacy cohort). The distribution of tumor lesions was determined for the staging cohort and stratified by National Comprehensive Cancer Network risk score. Involvement of pelvic LNs was assessed retrospectively by 3 masked independent central readers, and a majority rule was used for analysis. Standard surgical fields were rated on a 5-point scale independently for PET and for morphologic imaging. Results were compared with histopathologic findings on a patient, right-vs.-left, and template basis. Results: For the staging cohort, 18F-rhPSMA-7.3 PET was positive in 275 of 279 (98.6%), 106 of 279 (38.0%), 46 of 279 (16.5%), 65 of 279 (23.3%), and 5 of 279 (1.8%) patients for local, pelvic nodal, extrapelvic nodal, metastatic bone, and visceral metastatic disease, respectively. In the efficacy cohort, LN metastases were present in 24 of 83 patients (29%) and were located in 48 of 420 (11%) resected templates and in 33 of 166 (19.9%) hemipelvic templates in histopathology. The majority vote results showed that patient-level sensitivity, specificity, and accuracy for pelvic nodal metastases were 66.7% (95% CI, 44.7%-83.6%), 96.6% (95% CI, 87.3%-99.4%), and 88.0% (95% CI, 78.5%-93.8%), respectively, for 18F-rhPSMA-7.3 PET and 37.5% (95% CI, 19.6%-59.2%), 91.5% (95% CI, 80.6%-96.8%), and 75.9% (95% CI, 65.0%-84.3%), respectively, for morphologic imaging. 18F-rhPSMA-7.3 showed higher interobserver agreement than morphologic imaging (patient-level Fleiss κ = 0.54 [95% CI, 0.47-0.62] vs. 0.24 [95% CI, 0.17-0.31]). A mean SUV ratio of 6.6 (95% CI, 5.2-8.1) documented a high image contrast between local tumors and adjacent low urinary tracer retention. Conclusion: 18F-rhPSMA-7.3 PET offers diagnostic performance superior to morphologic imaging for primary N-staging of newly diagnosed PCa, shows lower interreader variation, and offers good distinction between primary-tumor activity and bladder background activity. With increasing National Comprehensive Cancer Network risk group, an increasing frequency of extraprostatic tumor lesions was observed.