CT-based muscle and adipose measurements predict prognosis in patients with digestive system malignancy

Abstract

The role of skeletal muscle and adipose tissue in the progression of cancer has been gradually discussed, but it needs further exploration. The objective of this study was to provide an in-depth analysis of skeletal muscle and fat in digestive malignancies and to construct novel predictors for clinical management. This is a retrospective study that includes data from Cancer Center, the First Hospital of Jilin University. Basic characteristic information was analyzed by T tests. Correlation matrices were drawn to explore the relationship between CT-related indicators and other indicators. Cox risk regression analyses were performed to analyze the association between the overall survivals (OS) and various types of indicators. A new indicator body composition score (BCS) was then created and a time-dependent receiver operating characteristic curve was plotted to analyze the efficacy of the BCS. Finally, a nomogram was produced to develop a scored-CT system based on BCS and other indicators. C-index and calibration curve analyses were performed to validate the predictive accuracy of the scored-CT system. A total of 575 participants were enrolled in the study. Cox risk regression model revealed that VFD, L3 SMI and VFA/SFA were associated with prognosis of cancer patients. After adjustment, BCS index based on CT was significantly associated with prognosis, both in all study population and in subgroup analysis according to tumor types (all study population: HR 2.036, P < 0.001; colorectal cancer: HR 2.693, P < 0.001; hepatocellular carcinoma: HR 4.863, P < 0.001; esophageal cancer: HR 4.431, P = 0.008; pancreatic cancer: HR 1.905, P = 0.016; biliary system malignancies: HR 23.829, P = 0.035). The scored-CT system was constructed according to tumor type, stage, KPS, PG-SGA and BCS index, and it was of great predictive validity. This study identified VFD, L3 SMI and VFA/SFA associated with digestive malignancies outcomes. BCS was created and the scored-CT system was established to predict the OS of cancer patients.

Keywords: Adipose tissue; Body composition analysis; Computed tomography; Digestive system cancer; Skeletal muscle.

Figure 1

The flow diagram of participants enrolled. ECOG, Eastern Cooperative Oncology Group performance status.

Figure 2

Correlation matrix separately for males and females between CT-related indicators and other variables.

Figure 3

The time-dependent ROC curve of body composition score in predicting overall survivals. *In the graph, the horizontal coordinate time t is measured in months.

Figure 4

The scored-CT system by nomogram used to predict the clinical outcome of patients with digestive system malignancy. Tumor type: 1 = Gastric cancer; 2 = Colorectal cancer; 3 = Hepatocellular carcinoma; 4 = Esophageal cancer; 5 = Pancreatic cancer; 6 = Biliary malignant tumor. Stage: 1 = I; 2 = II; 3 = III; 4 = IV.

Figure 5

Calibration curves of study participants for the nomogram predictions of the 1-, 2-, 3-and 5-year overall survival. (A) Calibration curves of nomogram predictions of the 1-year overall survival; (B) Calibration curves of nomogram predictions of the 2-year overall survival; (C) Calibration curves of nomogram predictions of the 3-year overall survival; (D) Calibration curves of nomogram predictions of the 5-year overall survival.