Individual patient-reported symptom items discriminate imaging-graded dysphagia in head and neck cancer patients treated with radiotherapy: secondary analysis of pooled prospective studies.

Abstract

To determine clinically actionable thresholds for the MD Anderson Symptom Inventory-Head and Neck (MDASI-HN) -Swallow and -Choke items by comparing discriminant capacity against imaging-based markers of dysphagia in head and neck cancer (HNC) patients undergoing (chemo)radiotherapy (RT). A retrospective secondary analysis used single-institution prospective registries of HNC patients treated with RT who completed the MDASI-HN and a modified barium swallow (MBS) before, during, or after RT. The DynamicImaging Grade ofSwallowing Toxicity(DIGEST) overall (D), efficiency (E), and safety (S) scores were compared to a priori-defined (≥6) and data-driven thresholds using binomial regression. Diagnostic accuracy was evaluated using sensitivity (SN), specificity (SP), PPV and NPV, using DIGEST D, E, S ≥ 1 (any impairment) and ≥ 2 (moderate-severe) as reference standards. Among 264 patients (mean age 62; 87 % male; 47 % nasopharyngeal cancer; 55 % post-RT), -Swallow ≥ 6 was associated with D ≥ 1 and D ≥ 2 (RRs = 1.8-2.6, p < 0.05), showing high SP (92-95 %) but low SN (23-30 %). -Choke ≥ 6 demonstrated stronger associations for D ≥ 2 (RR = 3.5, p < 0.05), and similar SN/SP tradeoffs. Data-driven thresholds improved accuracy: for D ≥ 1, a combined threshold (-Swallow or -Choke) ≥ 2 yielded SN 77 %, SP 64 %; for D ≥ 2, -Choke ≥ 1 showed SN 85 %, SP 69 %, while -Choke ≥ 2 increased SP to 84 %. For aspiration risk (S ≥ 2), -Choke ≥ 1 had SN 88 %, SP 66 %; -Choke ≥ 2 improved SP to 83 %. MDASI-Swallow and -Choke correlate with imaging-based dysphagia. While ≥ 6 may flag moderate-severe impairment, lower thresholds (≥1-2) offer better sensitivity, supporting early detection. Combined thresholds may enhance screening and guide survivorship care.