AbstractPurpose Daily online adaptive radiotherapy (oART) for head and neck squamous cell carcinoma (HNSCC) holds the potential to holds the potential to dramatically reduce margins accounting for setup uncertainty. However, careful margin assessment is critical due to the additional time required for real-time contouring and planning, which may lead to consequential intrafractional motion. This study aimed to quantify the intrafraction motion margin (IMM) required for cone-beam computed tomography (CBCT)-guided oART in HNSCC. Methods and Materials This retrospective analysis included 24 patients with HNSCC who were treated with CBCT-guided daily oART as part of a prospective clinical trial. Eligible patients were treated for oropharyngeal, laryngeal, and hypopharyngeal HNSCC. One fraction was assessed per week for each patient. For each of the oART sessions, a pre-planning CBCT and a near post-treatment CBCT, acquired before the last field delivery, were obtained. "Gold standard" (GS) GTV and CTV were retrospectively contoured on both CBCTs without time constraints. Radial margins of 1, 2, and 3 mm plus a craniocaudal margin of 2-mm (CBCT slice thickness) were applied to the planning GS contours to assess geometric overlap with post-treatment GS contours. Dosimetric coverage (V100% ≥95% and V95% ≥99%) of the adapted plan generated with pre-planning (PP)-CBCT GS contour on the post-treatment (PT)-CBCT GS contour was evaluated. Workflow times, including the interval from planning CBCT to final beam delivery, were recorded. Results A total of 166 fractions were evaluated, with an average time of 33 minutes (absolute range, 25-58 minutes) between planning CBCT and delivery completion. Geometric overlap and dosimetric coverage analysis showed that a 1-mm axial and 2-mm superior-inferior margin achieved > 99% geometric overlap for all GS targets and met dosimetric coverage goals: primary gross tumor volume receiving 70 Gy (GTVp-70; 99.7% geometric coverage, V100 = 96.9%), nodal gross tumor volume receiving 70 Gy (GTVn-70; 99.0%, 96.6%), primary clinical target volume receiving 63 Gy (CTV-63; 99.4%, 98.7%), nodal GTV receiving 63 Gy for suspicious nodes >17 mm (GTVn-63; 99.1%, 96.8%), and nodal GTV receiving 56 Gy for suspicious nodes ≤17 mm (GTVn-56; 99.6%, 99.0%). Conclusions Despite the time required for the adaptive process, daily oART achieves robust geometric and dosimetric coverage with minimal IMM. This stability suggests that daily oART for HNSCC can significantly facilitate the use of smaller PTV margin.
Liao et al. (Sun,) studied this question.