Contralateral Submandibular Gland Sparing in Head and Neck Squamous Cell Carcinoma Is Safe and Improves Patient-Reported Outcomes (PROs)

Abstract

To determine the feasibility and potential benefits of sparing contralateral submandibular gland (cSMG) during definitive intensity modulated radiation therapy (IMRT) for head and neck squamous cell carcinoma (HNSCC). We reviewed treatment plans, mean doses (MDs) to organs at risk, patient-reported outcomes, and tumor control in 273 patients with HNSCC treated with definitive IMRT to the bilateral necks at our institution between 2005 and 2014. Patients were periodically given the validated xerostomia questionnaire (XQ) and head and neck quality of life questionnaire (HNQOL) pretreatment and at follow-up. The majority (73%) of patients had oropharyngeal cancer. The vast majority of patients, 93.0%, harbored stage III/ IV disease, and 93.0% of patients received concurrent chemotherapy. Median follow-up was 22.5 months (1–115 months), and 662 surveys were analyzed. The medians of MDs to oral cavity (OC) was 36 Gy (8-67 Gy), combined parotid glands (comPG) 31 Gy (6-58 Gy), ipsilateral SMG (iSMG) 66 Gy (28-72 Gy), and cSMG 37 Gy (8-70 Gy). On univariate analysis, significant predictors of XQ summary score (SS) included the MDs of OC (r=0.44, P<.01), comPG (r=0.63, P<.01), and cSMG (r=0.22, P=.04). On multivariate analyses, the comPG MD (r=0.42, P=.04) and time from treatment (r=-0.14, P=.01) were statistical predictors of the XQSS, and the cSMG MD (r=0.21, P=.09) was a marginally significant predictor for XQSS. At the 6- and 12-month timepoints, the cSMG MD is a significant predictor for XQSS (r=0.82, P=.01 and r=0.57, P=.04, respectively), eating domain (r=0.90, P=.01 and r=0.68, P=.02, respectively), fasting domain (r=0.66, P=.04 and r=0.52, P=.07, respectively) and HNQOL (r=0.77, P<.01 and r=0.49, P=.02, respectively). Using 39 Gy cSMG MD as a prespecified threshold based on dose-saliva relationships, regression modeling showed that patients receiving <39 Gy had significantly favorable XQSS in both univariate (r=-22.61, P<.01 and r=-15.18, P=.01, respectively) and multivariate (r=-26.50, P<.01 and r=-13.02, P=.03, respectively) analysis at 6 and 12 months. Moreover, patients receiving <39 Gy had improvement (P<.01) in their XQSS over time after IMRT, while patients receiving >39 Gy did not (P=.29) (data will be presented). Sixty-three percent (172/273) of the patients received <39Gy to the cSMG to preserve salivary output. In this group, the median of MD to cSMG was 31Gy (8-39 Gy), and the majority (75%) was oropharyngeal cancer. There were 28 total tumor recurrences with no failures in contralateral level IB, and only 1 patient failed on the contralateral neck, within the 54 Gy isodose line and in initially noninvolved level II. cSMG MD predicts for both patient-reported xerostomia and QOL after IMRT. cSMG sparing did not compromise disease control. We recommend keeping the cSMG MD to no more than 39 Gy if clinically possible.