Strategy to select nasopharyngeal cancer patients for adaptive radiotherapy.

Background: Nasopharyngeal cancer (NPC) patients experience significant anatomical changes during radiation treatment (RT). Adaptive radiotherapy (ART) can be initiated in response to specific events or at scheduled intervals during treatment, prompting questions about patient selection, timing, and reliable triggers for re-CT and replanning. This study aimed to develop a method for selecting NPC patients for ART before treatment and identify the optimal timing for its implementation. Materials and methods: We retrospectively evaluated NPC patients treated at University Hospitals Leuven between 2016 and 2023, assessing volumetric changes in the parotid glands, air cavities, maxillary sinuses, and body contour at nasopharyngeal and low-neck levels throughout treatment. Structures were contoured on the initial simulation CT, weekly CBCTs, and final treatment day, resulting in eight measurements per parameter. Body contour changes were evaluated at the nasopharyngeal level (odontoid process) and neck level (lower edge of the third cervical vertebra). Measurements included volume, transverse diameters, and radius angles of 50° and 310°. Kaplan-Meier analysis was used for overall survival (OS) and local control (LC), while longitudinal volumetric data were analyzed with linear mixed models. Continuous variables were dichotomized to create a binary variable, for the purpose to define cut-off values for significant predictive variables. Results: Of the 47 NPC patients analyzed, 2- and 5-year loco-regional control rates were 95%, with overall survival rates of 100% and 93%, respectively. Significant reductions in parotid gland volume and body contour were observed, particularly between weeks 3 and 4. Predictors of parotid gland volume reduction included bulky nodal disease and larger initial parotid volume, with thresholds of 15 mm and 56 cc, respectively. Body contour changes during the fourth week correlated with high N-stage (≥ N3), higher initial BMI (>28 kg/m²), bulky lymph nodes (15 mm), higher initial mean planned dose to the ipsilateral parotid gland (> 31 Gy) and larger initial primary tumor Clinical Target Volume (> 93 cc). Induction chemotherapy was significantly associated with fewer sinonasal air cavity changes than concomitant chemoradiotherapy. Conclusion: It is recommended to re-evaluate the RT plan during the period between fractions 15 and 20 of treatment. Patients treated for NCP could be selected for ART based on the following clinical criteria at diagnosis: N3 or higher classification, presence of a bulky lymph node larger than 15 millimeters, initial BMI exceeding 28 kg/m², mean planned dose to the ipsilateral parotid glands greater than 31 Gy, cumulative volume of the parotid glands greater than 56 cc, high-dose CTV of the primary tumor greater than 93 cc, and receiving RT with no prior induction chemotherapy. Validation of these pre-treatment clinical predictors in a large, prospective dataset is essential before clinical usage.