3779 - Time- and Response-Guided Adapt-on-Demand Radiation Therapy for Cervical Cancer: Optimal Timing and Dosimetric Benefit from a Prospective Registry

Purpose/Objective(s): Daily adaptive radiation therapy (ART) for cervical cancer is resource intensive and may not confer incremental benefit over less frequent adapt-on-demand radiation therapy (AOD-RT). However, optimal timing, frequency, and patient selection for AOD-RT remain undefined. Using data from a prospective adaptive registry, we hypothesized that AOD-RT will yield quantifiable dosimetric benefits to patients exhibiting key predictive anatomical features, enabling determination of maximal benefit of AOD-RT with reduced physician and patient burden.

Materials/Methods: We analyzed 176 fractions across 46 patients with stage IB3-IVA cervical cancer receiving standard cisplatin-based chemoradiation who were treated using the AOD-RT workflow. Patients received 23-27 fractions, with cervix planning target volume (PTV) doses ranging from 4250-4875 cGy. Organ-at-risk (OAR) doses from the adaptive plan (ADP) were compared with a scheduled plan (SCH), defined as the original plan overlaid on the daily anatomy. Clinical and tumor characteristics—including stage, BMI, nodal status, and change in cervix clinical target volume (CTV)—were recorded. Fractions were grouped into early (1–10), middle (11–20), and late (=21) treatment phases. Linear mixed-effects regression models were used to identify patient- and time-dependent predictors of dosimetric benefit from adaptation.

Results: Compared with SCH, ADP significantly reduced OAR doses across all evaluated dose levels. Mean bowel bag V40 was 268.1 ± 82.2 cc with ADP versus 299.6 ± 112.5 cc with SCH (p < 0.001), with similar significant improvements observed for bladder and rectum. Target coverage also improved with ADP, with significant gains in CTV and PTV D95 and D99. Time-course analysis demonstrated that the maximum benefit of adaptation occurred during week 3 of treatment, with a reduction in mean cervix CTV of 54.6 cc relative to pre-treatment (Table 1). No significant benefit from adaptation was observed prior to week 3. Reduction in cervix CTV emerged as a key predictor of dosimetric benefit, correlating with greater reductions in OAR toxicity. Patients with reduction in CTV and a Preplan rectum V45 greater than the corresponding cohort median value showed ADP dosimetric benefit during or after week 3 (p = 0.038).

Conclusion: AOD-RT provides superior target coverage and OAR sparing compared with non-adaptive treatment in locally advanced cervical cancer. Dosimetric benefit is time-dependent and maximized when adaptation is initiated in the third week of treatment, particularly among patients demonstrating cervix CTV regression. These findings support a selective, data-driven AOD-RT strategy that optimizes benefit while reducing adaptive workload.