Head and neck cancer (HNC) is a prevalent and aggressive form of cancer with high mortality rates and significant implications for nutritional status. Accurate assessment of malnutrition in patients with HNC is crucial for optimizing treatment outcomes and improving survival rates. This study aimed to evaluate the use of ultrasound techniques for predicting nutritional status, malnutrition, and cancer outcomes in patients with HNC. A total of 494 patients with HNC were included in this cross-sectional observational study. Various tools and body composition measurements, including muscle mass and adipose tissue ultrasound evaluations, were implemented. Using regression models, we mainly found that high levels of RF-CSA (rectus femoris cross-sectional area) were associated with a decreased risk of malnutrition (as defined with GLIM criteria (OR = 0.81, 95% CI: 0.68–0.98); as defined with PG-SGA (OR = 0.78, 95% CI: 0.62–0.98)) and sarcopenia (OR = 0.64, 95% CI: 0.49–0.82) after being adjusted for age, sex, and BMI. To predict the importance of muscle mass ultrasound variables on the risk of mortality, a nomogram, a random forest, and decision tree models were conducted. RF-CSA was the most important variable under the random forest model. The obtained C-index for the nomogram was 0.704, and the Brier score was 16.8. With an RF-CSA < 2.7 (AUC of 0.653 (0.59–0.77)) as a split, the decision tree model classified up to 68% of patients as possessing a high probability of survival. According to the cut-off value of 2.7 cm², patients with a low RF-CSA value lower than 2.7 cm² had worse survival rates (p < 0.001). The findings of this study highlight the importance of implementing ultrasound tools, for accurate diagnoses and monitoring of malnutrition in patients with HNC. Adipose tissue ultrasound measurements were only weakly associated with malnutrition and not with sarcopenia, indicating that muscle mass is a more important indicator of overall health and nutritional status. These results have the potential to improve survival rates and quality of life by enabling early intervention and personalized nutritional management.

Background: Malnutrition is highly prevalent in patients with head and neck cancer, with relevant consequences in the treatment results. Methods: Multicenter observational study including 514 patients diagnosed with HNC. The morphofunctional assessment was carried out during the first 2 weeks of radiotherapy treatment. A correlation analysis between nutritional variables and groups of malnutrition, a multivariate logistic regression analysis, and a random forest analysis to select the most relevant variables to predict malnutrition were performed. Results: In total, 51.6% were undernourished (26.3% moderately and 25.3% severely). There was a negative correlation between morphofunctional variables and a positive correlation between hsCRP and well vs. moderate and well vs. severe malnutrition groups. The increase in different bioelectrical and ultrasound parameters was associated with a lower risk of moderate and severe malnutrition when groups with different degrees of malnutrition were compared. To predict the importance of morphofunctional variables on the risk of undernutrition, a nomogram, a random forest, and decision tree models were conducted. For the well vs. moderate, for the well vs. severe, and for the moderate vs. severe malnutrition groups, FFMI (cut-off < 20 kg/m²), BCMI (cut-off < 7.6 kg/m²), and RF-Y-axis (cut-off < 0.94 cm), respectively, were the most crucial variables, showing a greater probability of mortality in the two last comparisons. Conclusions: Malnutrition is very prevalent in HNC patients. Morphofunctional assessment with simple tools such as electrical impedance and muscle ultrasound allows an early nutritional diagnosis with an impact on survival. Therefore, these techniques should be incorporated into the daily clinical attention of patients with HNC.