KEY POINTS
- This single-centre retrospective cohort included 137 patients with M0 HNSCC receiving first-line definitive RT between July 2023 and December 2025. Planned primary-tumor dose was approximately 70 Gy in 32–34 fractions, and 84 patients received concurrent chemotherapy.
- An operational definition informed by the 2023 Asian Working Group for Cachexia criteria classified patients as cachectic when BMI was <21 kg/m² or weight loss exceeded 2%, together with anorexia or CRP >5 mg/L. Handgrip strength was unavailable, so this was not a complete prospective AWGC assessment.
- 72/137 patients (52.6%) met the cachexia definition. Definitive RT discontinuation occurred in 25.0% of cachectic versus 4.6% of non-cachectic patients (p<0.001).
- The association persisted after adjustment for age, ECOG performance status and concurrent chemotherapy: cachexia was associated with approximately sixfold higher odds of treatment discontinuation (OR 6.05, 95% CI 1.63–22.49; p=0.007).
- Sensitivity analyses were consistent. Firth penalized regression produced an adjusted OR 5.16, while additional adjustment for hypopharyngeal tumor site yielded an OR 5.01.
- Among the 21 patients who discontinued RT, 16 stopped in the context of treatment-related toxicity or poor tolerance; 15 of these 16 were cachectic. Patients stopping treatment had completed a median 22 fractions and 45.32 Gy.
- Treatment-related adverse events occurred in 97.2% versus 78.5% of cachectic and non-cachectic patients. However, cachexia was not significantly associated with temporary RT interruptions, acute myelosuppression timing, transfusion requirements or NRS-2002 nutritional risk during treatment.
- Cachectic patients had lower baseline albumin (38.09 vs 40.44 g/L) and prealbumin (0.17 vs 0.20 g/L). Exploratory IL-6 and IFN-γ differences lost significance after multiple-testing correction.
CLINICAL TAKEAWAY
Cachexia assessment before definitive head and neck RT may identify patients at unusually high risk of failing to complete curative treatment—even when ECOG status and conventional nutritional screening do not clearly separate them. These data support earlier recognition and intensified supportive care, but they do not show that treating cachexia itself will reduce treatment discontinuation.