Closing the loop in PSMA radiopharmaceutical therapy for prostate cancer: from eligibility to decision-validated treatment

PSMA-targeted radiopharmaceutical therapy (RPT) has demonstrated clinical efficacy, but treatment remains largely protocol-driven after eligibility has been established. We propose a domain-specific operationalization of established dynamic-treatment-regime methodology for PSMA-RPT, centered on three constructs: therapeutic exposure rather than binary target positivity, the clinically consequential lesion, and five treatment-relevant failure states. These constructs are embedded in a select-prescribe-verify-adapt framework. Selection characterizes target phenotype, disease geometry, host reserve, and patient goals; prescription defines the intended therapeutic system, activity, interval, and fallback options; verification determines delivered tumor and normal-organ exposure together with response and toxicity; and adaptation links the integrated patient-lesion-host state to prespecified actions. The framework addresses biological, dosimetric, and temporal gaps between eligibility and longitudinal treatment decisions. Current studies show associations between imaging phenotype, biomarkers, early response, absorbed dose, and outcome, but these associations do not establish that acting on the measurement improves care. We therefore propose a staged pathway: sequential multiple assignment randomized trials to refine candidate measurement-to-action rules, followed by a confirmatory strategy trial of the complete adaptive policy. An illustrative strategy-level estimand uses a hierarchical patient-centered outcome prioritizing survival and progression, followed by symptom and functional preservation and treatment burden. The methodological tools are established; the contribution is their PSMA-RPT-specific operationalization. The complete treatment policy, rather than an isolated biomarker, threshold, imaging finding, or cycle, should be the unit of prospective validation.