Expanding the Evidence-base for Clinician-Graded Dysphagia Using Dynamic Imaging Grade of Swallowing Toxicity...

The Dynamic Imaging Grade of Swallowing Toxicity version 2 (DIGEST_v2) is a clinician-graded scale for assessing pharyngeal dysphagia in modified barium swallow (MBS) studies, initially validated for head and neck cancer (HNC) patients. Given the investigators' published clinical application more broadly in all oncology populations in a comprehensive cancer center, this study evaluated the validity of DIGEST_v2 across diverse non-HNC oncology populations. A retrospective analysis included 386 patients who underwent MBS studies at MD Anderson Cancer Center (2016-2021). Participants were randomly selected to represent a mix of the most common referrals for MBS (neuro-oncology, endocrine/thyroid, thoracic cancers, mixed cancer etiologies in both inpatient and outpatient care). Two independent raters, clinical and research, assigned DIGEST overall, safety, and efficiency grades. Inter-rater reliability was assessed using weighted kappa (κ_w). Criterion validity was evaluated against the MBSImP pharyngeal total score, and convergent construct validity against the MDADI physical score and the PSS-HN diet score, using Spearman's correlation coefficients (r_s) and Kruskal-Wallis tests. Analyses were stratified by disease and setting subgroups. Inter-rater reliability for the overall sample and subgroups was substantial for DIGEST overall, safety, and efficiency grades (κ_w = 0.68-0.71), with moderate agreement for efficiency in the endocrine/thyroid subgroup (κ_w = 0.55). DIGEST effectively differentiated levels of pharyngeal pathophysiology for the overall sample and subgroups (MBSImP: r_s=0.68-0.86, p < 0.05), demonstrating criterion validity. Associations with MDADI physical (r_s=-0.12 to -0.47) and PSS-HN diet (r_s=-0.23 to -0.49) indicated weak to moderate convergent construct validity. DIGEST demonstrated robust psychometric properties for assessing the severity of pharyngeal dysphagia across diverse non-HNC cancer populations, maintaining stable performance in common subgroups referred for MBS studies. This validation supports the broader use of DIGEST in oncology practice, addressing the need for reliable dysphagia assessment tools in oncology. KEY POINTS: - DIGESTv2 demonstrated substantial inter-rater reliability (κw = 0.68-0.71) between clinicians in real-world practice and research laboratory image grading across neuro-oncology, endocrine/thyroid, and thoracic cancer populations using routine clinical imaging data. - Criterion validity was confirmed by strong correlations with reference measures of swallowing function (rs = 0.68-0.86), validating DIGEST's ability to differentiate dysphagia severity. - DIGESTv2 performed consistently across subgroups commonly referred for videofluoroscopic-based swallowing evaluations, supporting its use as a standardized assessment method in oncology care.