Disorders of energy metabolism is a common phenomenon in cancer patients. Changes in resting energy expenditure (REE) combined with inadequate nutrition support appear to be causes of nutritional depletion in cancer patients. In clinical practice, REE is typically calculated using predictive equations. The aim of this study was to determine the agreement between REE estimated by predictive equations and REE measured by IC in Portuguese cancer patients. Differences in measured REE between patients with different types of digestive cancers were also assessed. REE was measured by indirect calorimetry (IC) in 61 patients with cancer diagnosis (gastric cancer, cholangiocarcinoma, pancreatic cancer, liver cancer and colorectal cancer). Measured REE values were compared with those estimated by equations of Harris-Benedict, Schofield, Ireton-Jones, Mifflin-St.Jeor and Barcellos I and II. Mean Respiratory Quotient (RQ) was 0.77 ± 0.09, which indicates high lipids utilization as substrate. No statistically significant differences between REE or RQ from patients with different cancer types were observed. All equations underestimate REE: Harris-Benedict, mean difference -648 kcal (limits of agreement +627 to -1923 kcal); MifflinSt.Jeor, mean difference -694 kcal (limits of agreement +544 to -193 kcal); Schofield, mean difference -531 kcal (limits of agreement +662 to -1723 kcal); and Ireton-Jones, mean difference -556 kcal (limits of agreement +774 to -1887 kcal). Barcellos I and II showed lower mean difference when compared to measured REE, +59 and + 52 kcal, respectively, although presenting wide limits of agreement, +1542 to -1424 kcal and +1429 to -1326, respectively. Although Barcellos Equations underestimate less and enable more accurate average REE prediction in cancer patients, still present wide limits of agreement and therefore clinically important differences in REE estimation may be found at individual level. Our results support the appropriateness of measuring REE by IC to better adequate the nutrition support to cancer patients. Further research is needed to improve the current knowledge base of energy expenditure in cancer patients, and to improve the accuracy of existing predictive equations.