Low skeletal muscle radiation attenuation and visceral adiposity are associated with overall survival and surgical site infections in patients with pancreatic cancer

Abstract

Background: Cancer cachexia and skeletal muscle wasting are related to poor survival. In this study, quantitative body composition measurements using computed tomography (CT) were investigated in relation to survival, post-operative complications, and surgical site infections in surgical patients with cancer of the head of the pancreas.

Methods: A prospective cohort of 199 patients with cancer of the head of the pancreas was analysed by CT imaging at the L3 level to determine (i) muscle radiation attenuation (average Hounsfield units of total L3 skeletal muscle); (ii) visceral adipose tissue area; (iii) subcutaneous adipose tissue area; (iv) intermuscular adipose tissue area; and (v) skeletal muscle area. Sex-specific cut-offs were determined at the lower tertile for muscle radiation attenuation and skeletal muscle area and the higher tertile for adipose tissues. These variables of body composition were related to overall survival, severe post-operative complications (Dindo-Clavien ≥ 3), and surgical site infections (wounds inspected daily by an independent trial nurse) using Cox-regression analysis and multivariable logistic regression analysis, respectively.

Results: Low muscle radiation attenuation was associated with shorter survival in comparison with moderate and high muscle radiation attenuation [median survival 10.8 (95% CI: 8.8-12.8) vs. 17.4 (95% CI: 14.7-20.1), and 18.5 (95% CI: 9.2-27.8) months, respectively; P < 0.008]. Patient subgroups with high muscle radiation attenuation combined with either low visceral adipose tissue or age <70 years had longer survival than other subgroups (P = 0.011 and P = 0.001, respectively). Muscle radiation attenuation was inversely correlated with intermuscular adipose tissue (r_p = -0.697, P < 0.001). High visceral adipose tissue was associated with an increased surgical site infection rate, OR: 2.4 (95% CI: 1.1-5.3; P = 0.027).

Conclusions: Low muscle radiation attenuation was associated with reduced survival, and high visceral adiposity was associated with an increase in surgical site infections. The strong correlation between muscle radiation attenuation and intermuscular adipose tissue suggests the presence of ectopic fat in muscle, warranting further investigation. CT image analysis could be implemented in pre-operative risk assessment to assist in treatment decision-making.

Keywords: Body composition; Computed tomography; Pancreatic cancer; Radiation attenuation; Surgical site infection; Visceral adipose tissue.

Figure 1

Survival is related to different risk categories in patients with cancer of the head of the pancreas. (A) Kaplan–Meier estimate (univariable analysis): patients with low muscle radiation attenuation had a significantly lower survival than patients with moderate or high radiation attenuation (log‐rank test, P = 0.002). (B) Kaplan–Meier estimate (univariable analysis): patients with both high muscle radiation attenuation and low visceral adipose tissue index had significantly higher survival than other categories (log‐rank test, P = 0.011). (C) Kaplan–Meier estimate (univariable analysis): patients with both high muscle radiation attenuation and low age had a significantly higher survival than other categories, while patients with low muscle radiation attenuation and high age had significantly lower survival (log‐rank test, P = 0.001). RA, radiation attenuation; VAT, visceral adipose tissue index.

Figure 2

Association between computed tomography scan measurements at the third lumbar vertebra and potential confounders of patients with cancer of the head of the pancreas at diagnosis with survival using Cox‐regression analysis. Values are displayed as hazard ratio and 95% confidence interval. Sex, age, body mass index (BMI), and variables that generated a P‐value of <0.1 in univariable analysis were entered in the multivariable analysis. C‐statistic = 0.66 for multivariable analysis. SAT, subcutaneous adipose tissue; VAT, visceral adipose tissue. *P < 0.05.

Figure 3

Association between computed tomography scan measurements and potential confounders with post‐operative complications using logistic regression analysis. Values are displayed as odds ratio and 95% confidence interval. Sex, age, body mass index (BMI), and variables that generated a P‐value of <0.1 in a univariable analysis were entered in the multivariable analysis. C‐statistic = 0.72 for multivariable analysis. SAT, subcutaneous adipose tissue; VAT, visceral adipose tissue. *P < 0.05.

Figure 4

Association between computed tomography scan measurements and potential confounders with surgical site infections using logistic regression analysis. Values are displayed as odds ratio and 95% confidence interval. Sex, age, body mass index (BMI), and variables that generated a P‐value of <0.1 in univariable analysis were entered in the multivariable analysis. C‐statistic = 0.62 for multivariable analysis. SAT, subcutaneous adipose tissue; VAT, visceral adipose tissue. *P < 0.05.