Purpose: To investigate the predictors of incidence and duration of xerostomia (XT) based on parotid glands (PG), submandibular glands (SMG), and both glands taken as a whole organ (TG) in head-and-neck cancer patients treated with intensity-modulated radiotherapy.
Methods and materials: A prospective study was initiated in May 2003. Sixty-three head-and-neck patients (44 with nasopharynx cancer) were included in the analysis. Using the dose-volume histogram the PG, SMG, and TG mean doses were calculated. Unstimulated and stimulated salivary flow were measured and XT-related questionnaires were compiled before and at 3, 6, 12, 18, and 24 months after radiotherapy. Salivary gland toxicity was evaluated using the Radiation Therapy Oncology Group scale, and Grade >or=3 toxicity was used as the endpoint. The XT incidence was investigated according to descriptive statistics and univariate and multivariate analysis. The Bonferroni method was used for multiple comparison adjustment.
Results: After a reduced flow at 3 months after radiotherapy, recovery of salivary flow was observed over time. Primary site and salivary gland mean doses and volumes were identified in univariate analysis as prognostic factors. Multivariate analysis confirmed that TG mean dose (p = 0.00066) and pretreatment stimulated salivary flow (p = 0.00420) are independent factors for predicting XT.
Conclusion: The TG mean dose correlates with XT as assessed by Radiation Therapy Oncology Group criteria, salivary output, and XT-related questionnaires. Our results suggest that TG mean dose is a candidate dose constraint for reducing XT, requiring considerably more validation in non-nasopharyngeal cancer patients.