Stopping Before Harm: Adaptive Sandhi Thresholds for Dose and Intervention Control in Medicine

Modern medicine unintentionally harms millions of patients by continuing diagnostic and therapeutic interventions beyond the point at which outcomes are already determined. In medical imaging, radiation dose is routinely escalated to ensure certainty, while in radiation therapy and chemotherapy, intervention intensity often continues despite diminishing clinical benefit. These practices expose patients to avoidable radiation burden, toxicity, tissue damage, and long-term complications—not due to lack of technology, but due to the absence of a formal stopping rule that identifies when intervention has already succeeded. This paper introduces Adaptive Sandhi Thresholds, a mathematically grounded control framework that precisely identifies the pre-irreversible commitment point in medical imaging and therapy. By defining Sandhi as the threshold where diagnostic confidence or therapeutic efficacy becomes locked in—before harm accelerates—the framework enables systems to stop intervention earlier without sacrificing clinical reliability. Compatible with existing CT, X-ray, radiotherapy, and chemotherapy platforms, Adaptive Sandhi Thresholds shift medicine from escalation-based practice to outcome-locked precision control. Given the global rise in imaging volume and aggressive therapies, delaying adoption of such threshold-aware intervention control carries its own ethical cost. This work presents a deployable solution to a known problem—and one that can prevent unnecessary harm at scale. What must be treated is treated. Nothing else is sacrificed.