The use of computed tomography images as a prognostic marker in critically ill cancer patients

Summary Background & aims In the intensive care unit (ICU) setting, body composition is typically estimated through anthropometry, which does not specifically quantify skeletal muscle (SM). Recent findings have shown that computed tomography (CT) is a useful tool to identify low SM in critically ill patients, which in turn is associated with poor clinical outcomes. So, the present study aims at comparing low SM assessed by CT to BMI anthropometric data and its association with outcomes in critically ill patients. Methods Observational study was used, which included >18 year-old patients, with over 72 h of ICU length of stay, who had an abdominal CT at ICU admission. Demographic, body mass index (BMI), hospital outcomes and abdominal CT data (SM and adipose tissue at the 3rd lumbar vertebrae) were collected for analysis. ROC curve optimal stratification analysis for hospital mortality was applied to classify people into low SM (sarcopenic) versus normal SM (non-sarcopenic). A Cox regression was applied to find independent associations between sarcopenia and 30-day survival. Results The study involved 99 patients, 56% male, mean-age of 61.6 years old, BMI 24.19 ± 4.49 kg/m2 ; hospital mortality was 26%. According to BMI, 19.4% of the patients were underweight. However, a poor correlation was observed between BMI and SM index by CT: R2 = 0.39, P