Early Experience of Online Adaptive Radiation Therapy of Definitive Head and Neck Cancer Patients

Abstract

<h3>Purpose/Objective(s)</h3> Advent of cone beam computed tomography (CBCT) based online adaptive radiotherapy (OnART) systems has reduced the barriers of adaptation by allowing rapid generation of new plans. However, due to the complexity of head and neck cancer (HNC) treatment planning, there is a paucity of OnART data on HNC patients. We present our early OnART experience in definitive radiation of HNC with the hypothesis that OnART would improve target coverage and reduce dose to organs at risk (OAR). <h3>Materials/Methods</h3> Patients with HNC receiving definitive (chemo)radiation who underwent at least one OnART were enrolled on a prospective registry study between 7/2021 and 1/2022. The same OnART software was used for on-couch adaptation for all patients. The frequency of adaptation was at the discretion of the treating physician. Physicians were given the option of delivering one of two plans during adaptation: the original radiation plan transposed onto the CBCT with adapted contours (scheduled) and a new adapted plan generated from the updated contours (adapted). Paired t-test was used to compare the mean doses between scheduled and adapted plans. <h3>Results</h3> Sixteen patients (12 oropharynx, 2 larynx/hypopharynx, 1 sinus, and 1 salivary gland primary) underwent 38 adaptation sessions (median 2; range 1 – 8). Most patients (14/16) were treated in 33 or 35 standard fractions, with SABR delivered to the other two patients. The mean new plan generation and approval time was 22 minutes (median 21; range 9 – 39), the mean physician time at the console was 26 minutes (median 23; range 14 – 45), and the mean patient time in the vault was 44 minutes (median 42; range 28 – 96). The adapted plan was chosen 89.5% (34/38) of the time. The scheduled plan vs adapted plan mean V100% for high-risk PTVs was 87.8% vs 92.4% (p = 0.07), intermediate-risk PTVs was 86.6% vs 98.2% (p = 0.02), and low-risk PTVs was 96.3% vs 98.1% (p < 0.01). The mean hotspot was also lower with adaptation, 108.7% vs 106.7% (p < 0.01). Doses to OARs saw heterogeneous changes from the scheduled to the adapted plan (Table). Most OARs (9/12) showed a decrease in the mean relative difference in dose with the adapted plans, with the larynx (p = 0.045), spinal cord (p < 0.01), mandible (p < 0.01), and brainstem (p = 0.03) reaching significance. <h3>Conclusion</h3> OnART is feasible for HNC patients. Early results show significant improvement in target coverage and homogeneity. On average, OARs saw a modest decrease in dose.