Update and external validation of a multivariable prediction model for tube feeding dependency for at least four weeks during chemoradiotherapy for head and neck cancer.

Abstract

6040 Background: Patients who receive chemoradiation or bioradiation (CRT/BRT) for locally advanced head and neck squamous cell carcinoma (LAHNSCC) often experience high toxicity rates, which may interfere with oral intake, leading to (temporary) tube feeding (TF) dependency. International guidelines recommend gastrostomy insertion when the expected use of TF exceeds four weeks. In this study we aimed to update and externally validate a prediction model to identify patients in need for TF for at least four weeks, meeting the international criteria for prophylactic gastrostomy insertion. Methods: This retrospective multicenter cohort study was performed in four tertiary referral head and neck cancer centers in the Netherlands. The prediction model was developed using data from the University Medical Center Utrecht and the Netherlands Cancer Institute. The model was externally validated in patients from the Maastricht University Medical Center and Radboud University Medical Center. The primary endpoint was TF, initiated during or within 30 days after completion of CRT/BRT, and administered for at least four weeks. Potential predictors were retrieved from patient medical records and radiotherapy dose-volume parameters were calculated. Results: The developmental and validation cohort included 409 and 334 patients respectively. Multivariable analysis showed significant predictive value (p < 0.05) for adjusted diet at start of CRT/BRT, percentage weight change prior to treatment initiation, WHO performance status, tumor-site, nodal stage, mean radiation dose to the contralateral parotid gland, and mean radiation dose to the oral cavity. The area under the receiver operating characteristics curve for the updated model was 0.73 and after external validation 0.64. Positive and negative predictive value at 90% cut off were 80.0% and 48.2% respectively. Conclusions: This externally validated prediction model to estimate TF-dependency for at least four weeks in LAHNSCC patients performs well. This model, which will be presented, can be used in clinical practice to guide personalized decision making on prophylactic gastrostomy insertion.