Abstract
242443
Introduction: The phase III VISION trial demonstrated that 177Lu-PSMA-617 (PSMA-RLT) prolonged overall survival in patients with metastatic castration-resistant prostate cancer (mCRPC). Common side effects of PSMA-RLT include xerostomia, bone pain, and myelosuppression. The Brief Pain Inventory (BPI) and Functional Assessment of Cancer Therapy-Prostate (FACT-P) scores have also been previously validated for pain response and health-related quality of life analysis in mCRPC patients. The objective of this project is to evaluate the role of baseline PSMA PET/CT (bPET) in predicting hematologic toxicity and patient-reported outcomes in mCRPC patients undergoing PSMA-RLT.
Methods: Patients enrolled in the Expanded Access Program for PSMA-RLT at UCLA (NCT: 04825652) with available bPET, toxicity data, and patient-reported outcomes (BPI, FACT-P, and Xerostomia Questionnaire) were included in this retrospective analysis. TRAQinform IQ technology (AIQ Solutions-Madison, WI) was used to conduct region of interest (ROI)-based analyses at bPET, and SUVmean and volume (cm3) across all lesion-ROI across the whole-body, the skeleton outside lesion-ROI, and salivary glands were extracted. Cox regression models were used to evaluate the following associations:
• Whole-body tumor volume and SUVmean with time to worsening of FACT-P scores (defined as a decrease of >10 points from baseline) and time to first occurrence of Grade 3 or 4 hematologic toxicity (by CTCAE)
• Skeleton volume and SUVmean with time to worsening of BPI Pain Severity and Pain Interference scores (defined as a 30% increase or at least 2 points from baseline) and time to first occurrence of Grade 3 or 4 hematologic toxicity
A mixed-effect linear regression model was also used to evaluate the associations between salivary gland volume and SUVmean with Xerostomia Questionnaire scores (scores ranged from 11 to 33, with higher scores indicating increased salivary gland toxicity).
Results: 61 mCRPC patients were included. In univariate Cox regression models, skeletal volume and whole-body SUVmean were associated with a longer time to first occurrence of Grade 3 or 4 hematologic toxicity (HR = 0.95, p = 0.021; HR = 0.75, p = 0.004 respectively), while whole-body tumor volume was associated with a shorter time to first occurrence (HR = 1.07, p = 0.001). In multivariate Cox regression models evaluating associations between skeleton volume and SUVmean with time to worsening of BPI Pain Interference scores, the time since diagnosis was associated with a longer time to worsening (HR = 0.85, p = 0.01; HR = 0.86, p = 0.011 respectively). In multivariate Cox regression models evaluating associations between skeleton and whole-body SUVmean with time to first occurrence of Grade 3 or 4 hematologic toxicity, baseline hemoglobin was associated with a longer time to first occurrence (HR = 0.61, p = 0.005; HR = 0.62, p = 0.007 respectively). In a mixed-effect linear regression evaluating associations between salivary gland SUVmean and volume with Xerostomia Questionnaire scores, the number of cycles was associated with higher scores (mean estimate: 0.68, p = 0.002, mean estimate: 0.68, p = 0.001 respectively). In multivariate Cox regression models evaluating associations between whole-body SUVmean with time to worsening of FACT-P scores, whole-body SUVmean was associated with a longer time to worsening (HR = 0.74, p = 0.033), while time since diagnosis was associated with a shorter time to worsening (HR = 1.09, p = 0.047). Table 1 summarizes the results of the multivariate analyses.
Conclusions: In this retrospective analysis of 61 mCRPC patients treated with PSMA-RLT, whole-body SUVmean was associated with a longer time to worsening of FACT-P scores in a multivariate Cox regression model, although bPET parameters were generally not predictive of hematologic toxicity or patient-reported outcomes in other multivariate analyses. Validation in a larger, prospective, multicenter cohort is warranted.