Abstract
BACKGROUND: To assess how inter-fraction bowel cavity (BC) motion volume influences accumulated dose in rectal cancer radiotherapy and to translate these effects into a clinically interpretable risk construct for acute bowel toxicity. METHODS: This retrospective cohort included 28 consecutive rectal cancer patients treated with 50 Gy in 25 fractions on a Halcyon v3.0 linac with daily iterative cone beam CT (iCBCT). Patients with > 5% bladder-volume or > 8% external-contour variation across the course were excluded. For each fraction, per-fraction dose was recalculated on the iCBCT using a scanner-specific HU-relative electron density calibration, rigidly aligned to the planning CT (pCT), deformably registered, and voxel-wise accumulated on pCT. Dosimetric endpoints were PTV D(2%), D(50%), D(98%) and whole-bowel (WB) D(0.03cc), D(150cc). Associations between inter-fraction BC motion volume change (ΔV(BC)) and accumulated dose difference (ΔD) were quantified using Spearman's f and ordinary-least-squares slopes (Gy per 10% BC expansion); same-day effects were estimated with patient fixed-effects models. To account for multiple testing, p-values for the five primary endpoints were adjusted using the Benjamini-Hochberg false discovery rate (FDR). Clinical relevance was assessed with a published logistic normal-tissue complication probability (NTCP) model based on WB V(45), defined as the WB volume receiving ≥ 45 Gy. RESULTS: Accumulated dose analyses showed selective elevation of high-dose metrics with preservation of near-minimum target coverage. Compared with the plan, accumulated PTV D(2%) and D(50%) increased by 0.87 Gy and 0.61 Gy, respectively, whereas PTV D(98%) exhibited no meaningful change. WB D(0.03cc) and D(150cc) increased by 1.21 Gy and 1.18 Gy, respectively. Inter-fraction BC expansion was strongly associated with high-dose escalation: for each 10% increase in BC volume, accumulated WB D(0.03cc) increased by 0.31 Gy (95% CI 0.18-0.44; ρ = 0.78) and PTV D(2%) by 0.29 Gy (0.15-0.37; ρ = 0.66). After FDR adjustment, no endpoint met p < 0.05, but the strongest associations (PTV D(2%), WB D(0.03cc)) remained near-significant and directionally consistent with large, biologically plausible effect sizes. Application of the WB V(45) NTCP model indicated modest but consistent risk increments after accumulation, with NTCP for ≥ grade 2 toxicity increasing from 0.75 to 0.79 (+0.04; p = 0.001) and NTCP for grade 3 toxicity increasing from 0.80 to 0.84 (+0.04; p = 0.002). CONCLUSION: Daily iCBCT-based dose accumulation identifies inter-fraction BC motion volume as a selective driver of high-dose escalation to the PTV and WB while largely preserving near-minimum target coverage. Accumulated WB V(45) may serve as a practical, delivery-based marker for on-treatment surveillance and risk-adapted intervention, complementing plan-based evaluation in rectal cancer radiotherapy.