ABSTRACTHead and neck cancer (HNC) patients who underwent percutaneous endoscopic gastrostomy (PEG) present malnutrition risk and speech impairments. Their assessment relies on objective anthropometric/laboratory data. Focusing on HNC PEG patients, our aims were to evaluate: 1) outcome; 2) nutritional status when the patients underwent PEG; and 3) association of nutritional status/outcome, creating a survival predictive model. We evaluated the outcome based on NRS 2002, dietary assessment, body mass index (BMI), mid-upper arm circumference (MUAC), triceps skinfold thickness (TSF), mid-arm muscle circumference (MAMC), albumin, transferrin, and cholesterol on the day of gastrostomy. Using BMI, TSF, MAMC, and laboratory data, a survival predictive model was created. Of the 234 patients (cancer stages III–IV), 149 died, 33 were still PEG-fed, and 36 resumed oral intake (NRS-2002≥3, caloric needs <50% in all). BMI was 12.7–43. 189, 197, and 168 patients displayed, respectively, low MUAC, TSF, and MAMC. 91, 155, and 119 patients displayed low albumin, transferrin, and cholesterol. Albumin, cholesterol, and transferrin were strongly associated with the outcome. A predictive model was created, discriminating between short-term survivors (<4 months) and long-term survivors. HNC patients were malnourished. Using anthropometric/laboratory parameters, a predictive model provides discrimination between patients surviving PEG for <4 months and long-term survivors. Teams taking care of PEG patients may provide special support to potential short-term survivors.