Combining molecular and functional imaging in metastatic castration-resistant prostate cancer (mCRPC).

141 Background: Recently, plasma tumour DNA (ptDNA) has been identified as a potential early noninvasive biomarker of treatment response in mCRPC patients ( Conteduca, Br J Cancer 2020). In this study, we sought to determine whether pre-treatment ptDNA could accurately reflect metabolic tumor burden in mCRPC and if it could be in combination with functional imaging could provide better prognostication. Methods: Between October 2011 and June 2016, 102 plasma samples from mCRPC patients treated with abiraterone or enzalutamide were collected. Targeted next-generation sequencing was performed to determine baseline ptDNA fraction. Maximum standardized uptake value (SUVmax), total lesion activity (TLA), and metabolic tumour volume (MTV) were calculated on 18F-fluorocholine positron emission tomography/computed tomography. A Weibull multiple regression model was adopted to evaluate the combined impact of clinical, molecular and imaging features on overall survival (OS) and to obtain a prognostic score. Each variable was allotted a “partial score” that depended on the size of the regression coefficient. Total scores ranged between 0 and 5.85 and assigned patients to 3 different risk groups according to 18-months survival probability: group I, >70%; group II 30%-70%; and group III, <30%. We estimated OS probabilities by the exponential model and by the Kaplan-Meier method. Results: We observed a significant association between ptDNA levels dichotomized as below or above median plasma tumor fraction (low ptDNA≤0.188 versus high ptDNA>0.188) and median SUVmax (p<0.0001), MTV (p=0.0005) and TLA (p<0.0001). Patients were randomly divided into a training set (n=68) and a validation set (n=34). In the training cohort, we performed a multivariable analysis showing that visceral metastasis, serum LDH, MTV and ptDNA were independent predictors of OS [HR=2.64, 95%CI 1.32-5.26, p=0.006; HR=3.69, 95%CI 1.98-6.87, p<0.0001; HR=1.91, 95%CI 1.13-3.21, p=0.015; and HR=2.64, 95%CI 1.32-5.26, p=0.003, respectively]. In the training set, median OS was significantly different among the 3 risk groups (risk group I, 29.2 months [95% CI, 18.3 to 37.0 months]; risk group II, 15.9 months [95% CI, 10.6 to 24.0 months]; and risk group III, 8.7 months [95% CI, 6.3 to 15.4 months]; p<0.0001). Similar results were observed in the validation set groups (risk group I, 23.4 months [95% CI, 8.1 to 38.5 months]; risk group II, 13.3 months [95% CI, 3.7 to 18.0 months]; and risk group III, 7.3 months [95% CI, 2.6 to 11.8 months]; p=0.001). Conclusions: Integrating plasma DNA analysis with functional imaging may improve prognostic risk stratification and treatment selection in mCRPC patients. A larger prospective evaluation is now warranted.