While parotid-sparing intensity modulated radiotherapy (IMRT) has demonstrated superiority to conventional RT in terms of observer-rated xerostomia, patient-reported outcome measures (PROMs) have only marginally improved. We investigated how sparing all salivary glands affects PROMs.

Patients treated to the bilateral neck with all-gland-sparing IMRT answered xerostomia (XQ) and head-and-neck quality of life (HNQOL) questionnaires. Longitudinal regression was used to assess the relationship between questionnaire scores and mean bilateral parotid gland (bPG), contralateral submandibular gland (cSMG), and oral cavity (OC) doses. Marginal R and Akaike information criterion (AIC) were used for model evaluation.

252 patients completed approximately 600 questionnaires. On univariate analysis, bPG, cSMG, and OC doses significantly correlated with XQ-summary, XQ-eating, and HNQOL-eating scores. On multivariate analysis, bPG and OC doses significantly correlated with XQ-summary, XQ-eating, and HNQOL-eating scores; and cSMG dose with HNQOL-summary. Combining doses to all three structures yielded the highest R for XQ-summary, XQ-rest, XQ-eating, and HNQOL-eating. In the 147 patients who received a mean cSMG dose ≤39Gy, there were no failures in contralateral level IB.

Reducing doses to all salivary glands maximizes PROMs. A cSMG dose constraint of ≤39Gy does not increase failure risk.