2803 - A Single Patient-Reported Outcome Item Successfully Detects Imaging-Graded Subclinical Swallow Dysfunction in Patients with Head and Neck Cancer Receiving Radiotherapy
Presenter(s)
B. Manduchi1,2, T. C. Jefferson3, A. C. Moreno4, C. E. Barbon5, C. D. Fuller6, and K. A. Hutcheson1; 1MD Anderson Cancer Center, Houston, TX, 2UTHealth Houston, Houston, TX, 3University of Mississippi Medical Center, Jackson, MS, 4Department of Radiation Oncology, University of Texas MD Anderson Cancer Center, Houston, TX, 5Department of Head and Neck Surgery, The University of Texas MD Anderson Cancer Center, Houston, TX, 6Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX
Purpose/Objective(s):
Dysphagia is a major toxicity of radiotherapy (RT) for head and neck cancer (HNC), yet access to objective swallow assessment via modified barium swallow (MBS) studies remains limited in both clinical and research settings. The MD Anderson Dysphagia Inventory (MDADI), the most common self-reported measure of dysphagia-related quality of life in HNC, could help stratify patients to prioritize for MBS referral. Using patterns of single MDADI items to trigger MBS referral, instead of calculation-based composite scores, has potential to simplify and optimize clinical use. This study evaluated the ability of single MDADI item scores to detect MBS-detected dysphagia in HNC patients receiving RT.Materials/Methods: A retrospective analysis was performed using prospective registry data from HNC patients treated with RT from 2021 to 2024 (institutional protocols: PA14-0947, 2024-0022, 2018-0052, 2015-0238). Patients completed same day MDADI and MBS before, during, or after RT. Dysphagia severity was graded for each MBS using DIGEST (D), a CTCAE-aligned 5-point scale. Breiman CART models were constructed considering each of the 20 single patient-reported MDADI items to detect any grade dysphagia (D=1) and moderate-severe dysphagia (D=2). Regression analysis confirmed threshold effects.
Results:
The cohort included 298 HNC patients. Assessments were conducted pre-RT in 111 patients (40%), during RT in 17 (6%), and post-RT in 152 (54%; median 155 days post-RT). Mild (D=1) and moderate-severe (D=2) dysphagia was observed in 28% and 14% of patients, respectively. CART analysis revealed that the single MDADI item P7 (“It takes me longer to eat because of my swallowing problem”) significantly distinguished between D=0 and D=1. The model's performance was confirmed in the validation set (AUC=0.72, 95% CI: 0.59-0.83), with a sensitivity of 0.73 and specificity of 0.70. Regression analysis confirmed the threshold effect (p<0.001). None of the MDADI single items significantly identified D=2 in the D=0-1 vs. D=2 model.Conclusion:
The single MDADI item P7, related to self-reported physical function, effectively detected imaging-graded subclinical swallow dysfunction in ~70% HNC patients before, during or after receiving RT. This simple, clinic-friendly approach could facilitate early dysphagia detection and optimize MBS utilization in radiation oncology practice with a ~30% misclassification rate, future work will combine PRO data with other clinical metrics to potentially improve model performance.