Prostate‐specific membrane antigen positron emission tomography/computed tomography imaging as a precision diagnostic at prostate cancer recurrence after radical prostatectomy: Modeling long‐term survival

Abstract Background Prostate‐specific membrane antigen positron emission tomography/computed tomography (PSMA‐PET/CT) is affecting the management of patients with prostate cancer with biochemical recurrence after radical prostatectomy. The long‐term outcomes of tailoring salvage treatment on the basis of PSMA‐PET/CT status remain to be determined. Methods A decision‐analytic model was developed to project incremental life‐years of strategies that allocate treatments at biochemical recurrence after radical prostatectomy on the basis of PSMA‐PET/CT status (PSMA‐metastatic vs. PSMA‐nonmetastatic). Modeled treatments are local/regional (radiation) or systemic (hormone therapy and doublet therapy), administered immediately or delayed. PSMA‐metastatic status was assumed to lead to treatment intensification, whereas PSMA‐nonmetastatic status would lead to deintensification. To project survival, data on progression to metastasis from a clinical cohort were combined with registry data on postmetastasis survival. Because of the lack of data on long‐term treatment benefits by PSMA status, survival was projected by varying the hazard ratio (HR) for disease‐specific death among PSMA‐metastatic versus PSMA‐nonmetastatic patients under delayed or local/regional regimens (HR1) and under immediate systemic regimens (HR2). Results Mean life‐years are projected to be 15.5 under the non–PSMA‐tailored strategy, and mean incremental life‐years range from 0.38 to 0.81 depending on HR1 and HR2. A greater benefit is projected when PSMA‐metastatic status is more adverse under salvage regimens that do not include systemic agents. Conclusions This decision‐analytic modeling study projects that PSMA‐PET/CT–guided management at biochemical recurrence after radical prostatectomy yields a modest survival benefit under the specified model inputs and assumptions regarding treatment distributions. These findings may complement emerging data on the corresponding economic costs and health‐related quality of life.