Association of tumor burden with the eligibility of upfront therapy in patients with castration-naive prostate cancer: A multicenter retrospective study.
107 Background: Although several large-scale studies reported a robust benefit of upfront therapy for mHNPC, not all patients with mHNPC experienced mCRPC progression. As there is a concern for overtreatment in patients with low-risk/volume disease, we aimed to identify potential candidates for upfront therapy in patients with metastatic hormone-naïve prostate cancer (mHNPC). Methods: We retrospectively evaluated 679 patients with mHNPC who were initially treated with conventional androgen deprivation therapy. We defined the patients with progression to metastatic castration-resistant prostate cancer (mCRPC) as potential candidates for upfront therapy. To estimate the suitable candidate for upfront therapy, we retrospectively compared mCRPC progression rate, mCRPC-free survival, and overall survival (OS) after castration-resistance (OS-CR) between the low- and high-volume disease groups. Furthermore, we tried to develop a novel prediction model using deep learning algorithm. Results: The number of patients with mCRPC progression (potential candidates for upfront therapy) was 119 (52%) and 319 (71%) in the low- and high-volume disease groups. The mCRPC progression rate and CRPC-free survival were significantly worse in high-volume disease group ( P < 0.01), but no difference was found for OS-CR. Multivariate Cox regression analysis showed no significant association between tumor volume and OS-CR (HR 1.14, P=0.52). The deep learning model showed not high accuracy of mCRPC prediction (AUC 0.66). Conclusions: Approximately a half of patients with low-volume disease had progression to mCRPC. As the OS-CR in the patients with low-volume disease showed poor prognosis as well as those with high-volume disease, upfront therapy may be needed for a half of patients with low-volume disease.