Abstract 5041
Background
The purpose of this study is to evaluate the body composition including skeletal muscle and fat component as a prognostic factor in advanced pancreatic cancer patients.
Methods
Body composition factor comprised skeletal muscle index (SMI), muscle Hounsfield Unit (HU), subcutaneous fat index (SFI), and visceral fat index (VFI), which were measured at L3 level in abdomino-pelvic computed tomography (CT). region was quantified using the pre-established HU thresholds for skeletal muscle tissue (HU: -30∼150). Visceral fat and subcutaneous fat areas (HU:-150∼-30) were also measured. Continuous variables were dichotomized according to the normal range or the best cut-off values by Contal and O’Quigley method. Kaplan-Meier analysis was applied for relevance between each body composition factor and overall survival (OS).
Results
A total of 414 patients was enrolled. Median age was 66 years old with male predominance (49.8%). Median baseline values were 38.7 cm2/m2 for SMI, 33.4 HU, 32.2 cm2/m2 for SFI, and 30.1 cm2/m2 for VFI. Patients with lower value of the body composition factor has poorer survival (median OS; SMI, 6.9 vs. 10.0 months, P < 0.001; HU, 7.6 vs. 9.3 months, P = 0.0004; SFI, 6.6 vs. 9.8 months, P = 0.0008; VFI, 6.6 vs. 9.8 months, P = 0.0008). A body composition risk score was devised by summing the number of lower value of body composition factors, and patients were divided into two groups (low risk = score 0∼1, high risk = score≥2). Patients with high risk score had shorter OS (6.4 vs. 11.5 months). In multivariate analysis, age, ECOG performance status, WBC, neutrophil-lymphocyte ratio (NLR), carcinoembryonic antigen (CEA) at initial presentation was significant. In addition, body composition risk score was independent prognostic factors (hazard ratio, HR 1.56, 95% confidence interval 1.24-1.95, P < 0.001).
Conclusions
The body composition risk score significantly predicted prognosis in patients of pancreatic cancer. CT data are readily obtainable during routine clinical practice, which makes this approach useful for identifying prognostic biomarkers.
Clinical trial identification
Legal entity responsible for the study
Gangnam Severance Hospital.
Funding
Has not received any funding.
Editorial Acknowledgement
Disclosure
All authors have declared no conflicts of interest.