In addition to provider based assessment, e.g. Common Terminology Criteria for Adverse Events (CTCAE) grades, there is an increasing use of patient-reported outcome (PRO) questionnaires, such as the University of Michigan Xerostomia Questionnaire (XQ), the Sydney Swallow Questionnaire (SSQ), and the MD Anderson Dysphagia Inventory (MDADI). The aim of this study was to evaluate the performance of different measurements for xerostomia and dysphagia compared against an objective value, in this case, weight loss (WL) at 6 months after RT. We hypothesized that PROs would have a higher correlation to WL than provider-based assessments with the CTCAE scale. Eligible cases included patients treated from January 2014 to February 2019 with completed XQ, SSQ, MDADI, CTCAE xerostomia, and CTCAE dysphagia scores at 6 months post-treatment. Patients were also required to have their weight recorded at baseline (before or within the first week of the start of treatment) and at 6 months after completion of RT. Patients’ demographic and clinical features were also included in the model, including age at start of treatment, gender, race, smoking status, attending physician, tumor stage, tumor site, chemotherapy, human papillomavirus association, and use of enteral support. WL was calculated by the percent change at 6 months compared to baseline and then binarized into high (WL>5%) and low (WL≤5%). Logistic regression was performed to assess the association between WL and variables. Six classification models that adjusted patients’ demographics, clinical features, xerostomia (XQ and CTCAE xerostomia), and dysphagia (SSQ, MDADI, and CTCAE dysphagia) were developed. The area under the receiver operating curve (AUC) was used to evaluate the performance of xerostomia and dysphagia measurements (Table). All data analysis was performed using STATA v.15. Of 103 patients who met the inclusion criteria, 78 (76%) were in the high WL group. Median XQ score was 31/80; median SSQ and MDADI were 348.9/1700 and 89/100, respectively. 40 (38.9%) patients had CTCAE xerostomia grade≥2, and 17 (16.5%) were classified as CTCAE dysphagia grade≥2. Among the six models predicting WL at 6 months after RT (Table), the model using XQ and SSQ yielded the best performance (AUC = 0.60). The prediction of WL at 6 months post RT was improved using patient-reported xerostomia and dysphagia as measured by XQ and SSQ scores. Future healthcare providers may consider PROs to not only better measure and monitor symptoms, but to direct therapeutic interventions and predict for treatment effects.Abstract 3379; Table 1Model performance with different xerostomia and dysphagia measurementsXerostomiaDysphagiaAUCXQSSQ0.60XQMDADI0.56XQCTCAE dysphagia0.55CTCAE xerostomiaSSQ0.59CTCAE xerostomiaMDADI0.53CTCAE xerostomiaCTCAE dysphagia0.52 Open table in a new tab