Daily Artificial Intelligence-Assisted Adaptive Radiotherapy on Cone-Beam CT for Cancer of the Head and Neck

Abstract

<h3>Purpose/Objective(s)</h3> Adaptive radiation therapy (ART) provides a method to modify the radiation treatment plan secondary to structural and spatial changes that occur during a treatment course. Specifically, ART is relevant in head and neck cancers as many patients will experience changes of the target volumes and organs-at-risk (OARs) during treatment due to combination of treatment response, weight loss, inflammation and effects on normal tissues. Our institution recently began performing daily ART for select cases of definitive Head and Neck cancers using a commercially available, online adaptive platform with AI-assisted workflows on daily cone-beam computed tomography (CBCT). Herein we report our initial dosimetric experience using this novel technique. <h3>Materials/Methods</h3> Twelve consecutive head and neck cancer patients (nasopharynx, oropharynx, larynx) underwent CT-simulation and planned for adaptive treatments. A reference IMRT plan was created and approved using institutional dose/fractionation schema, target goals and OAR constraints. For treatment, the patient was aligned and CBCT acquired. The simulation CT scan was subsequently registered to the CBCT using deformable registration and a synthetic CT scan was generated. AI-based auto-contour and structure deformation of OARs and Targets on CBCT were reviewed and edited by the treating physician. Two plans were generated including, a CT sim-based plan with deformed structures (scheduled) and a re-optimized plan (adaptive) of which both plans evaluated and the best one approved and delivered. Statistical analyses were performed comparing adaptive versus scheduled plans using a paired student's t-test. <h3>Results</h3> 419 adaptive and scheduled radiotherapy plans were generated of which 82.2% adaptive plans were chosen with an average time per session of 19.4 minutes. Adaptive plans demonstrated superior V95 PTV70, PTV60, PTV56 mean coverage by 0.8 (95% CI 0.5%-1.1%, p < 0.001), 2.4% (95% CI 2%-3%, p < 0.001), and 5.9% (95% CI 5.4%-6.5%, p < 0.001). Compared to scheduled plans, high priority OARs demonstrated a reduction in: mean parotid gland right and left dose by 3.2 Gy (95% CI 1.9-4.4, p < 0.001) and 2.7 Gy (95% CI 1-4.4, p < 0.001), respectively and brainstem max dose was reduced by 5.2% (95% CI 3.7%-6.7%, p < 0.001). <h3>Conclusion</h3> Daily AI-assisted adaptive radiation therapy for head and neck cancer may potentially improve the therapeutic index for different tumor types by improving the target coverage and reducing OAR dose. Prospective clinical evaluation is indicated to further this novel approach.