Our objective was to evaluate the clinical application of third lumbar vertebra (L3)-computer tomography (CT)-determined sarcopenia as a marker of muscle mass in cancer inpatients diagnosed with malnutrition according to the Global Leadership Initiative on Malnutrition (GLIM) criteria and to establish its association with 6-month mortality. Methods: This was an observational, prospective study in patients from an inpatient oncology unit. We performed a nutritional assessment according to GLIM criteria, including muscle cross-sectional area at L3 by CT and skeletal muscle index (SMI). Six-month mortality was evaluated. Results: A total of 208 patients were included. The skeletal muscle cross-sectional area at L3 was 136.2 ± 32.5 cm2 in men and 98.1 ± 21.2 cm2 in women. The SMI was 47.4 ± 12.3 cm2/m2 in men and 38.7 ± 8.3 cm2/m2 in women. Sarcopenia (low SMI) was detected in 59.6% of the subjects. Using SMI as a marker of low muscle mass in application of GLIM criteria, we found 183 (87.9%) malnourished patients. There were 104 deaths (50%) at 6 months. The deceased patients had a lower skeletal muscle cross-sectional area (112.9 ± 27.9 vs. 126.1 ± 37.8 cm2; p = 0.003) and a lower SMI (41.3 ± 9.5 vs. 45.7 ± 12.9 cm2/m2; p = 0.006). An increased risk of 6-month mortality was found in malnourished patients according to GLIM criteria using SMI (HR 2.47; 95% confidence interval 1.07–5.68; p = 0.033). Conclusions: Low muscle mass, assessed by L3-CT, was observed to affect more than half of cancer inpatients. The deceased patients at 6 months had a lower skeletal muscle cross-sectional area and SMI. Malnutrition according to GLIM criteria using CT-determined sarcopenia was shown to adequately predict 6-month mortality.