Cause of residual organ-at-risk dose differences

Identify whether the residual organ-at-risk dose offset between uncertainty-guided digital twin online adaptive proton therapy plans and physician-approved offline replans is attributable to reuse of the offline replan objective template, use of patient-specific robust margins, or another component of the planning framework.

Background

The online adaptive plans achieved target coverage and spinal-cord and brainstem tolerance comparable to the offline replans, but exhibited residual increases in organ-at-risk doses, including spinal cord, brainstem, oral cavity, and parotid doses. The differences were concentrated in a small number of patients and structures near steep dose gradients.

The study reused each patient's offline-replan objective template and applied the patient-specific uncertainty-guided robust margins to the online plans. Because these factors were not varied independently, the authors could not determine which one produced the observed organ-at-risk dose offset. Disentangling their contributions remains an unresolved methodological question for improving the framework.