Xerostomia and its predictors following parotid-sparing irradiation of head-and-neck cancer.
PURPOSE: To assess long-term xerostomia in patients receiving parotid-sparing radiation therapy (RT) for head-and-neck cancer, and to find the patient and therapy-related factors that affect its severity. PATIENTS AND METHODS: From March 1994 through January 2000, 84 patients received comprehensive bilateral neck RT using conformal and multisegmental intensity-modulated RT (IMRT) aiming to spare the major salivary glands. Before RT and periodically through 2 years after the completion of RT, salivary flow rates from each of the major salivary glands were selectively measured. At the same time intervals, each patient completed an 8-item self-reported xerostomia-specific questionnaire (XQ). To gain a relative measure of the effect of RT on the minor salivary glands, whose output could not be measured, the surfaces of the oral cavity (extending to include the surface of the base of tongue) were outlined in the planning CT scans. The mean doses to the new organ ("oral cavity") were recorded. Forty-eight patients receiving unilateral neck RT were similarly studied and served as a benchmark for comparison. Factors predicting the XQ scores were analyzed using a random-effects model. RESULTS: The XQ was found to be reliable and valid in measuring patient-reported xerostomia. The spared salivary glands which had received moderate doses in the bilateral RT group recovered to their baseline salivary flow rates during the second year after RT, and the spared glands in the unilateral RT group, which had received very low doses, demonstrated increased salivary production beyond their pre-RT levels. The increase in the salivary flow rates during the second year after RT paralleled an improvement in xerostomia in both patient groups. The improvement in xerostomia was faster in the unilateral compared with the bilateral RT group, but the difference narrowed at 2 years. The major salivary gland flow rates had only a weak correlation with the xerostomia scores. Factors found to be independently associated with the xerostomia scores were the pre-RT baseline scores, the time since RT, and the mean doses to the major salivary glands (notably to the submandibular glands) and to the oral cavity. CONCLUSION: An improvement over time in xerostomia, occurring in tandem with rising salivary production from the spared major salivary glands, suggests a long-term clinical benefit from their sparing. The oral cavity mean dose, representing RT effect on the minor salivary glands, was found to be a significant, independent predictor of xerostomia. Thus, in addition to the major salivary glands, sparing the uninvolved oral cavity should be considered as a planning objective to further reduce xerostomia.