Supplementary Material for: Measurement of body composition in pancreatic cancer: a systematic review, meta-analysis and recommendations for future study design.

Background/Objectives: Sarcopenia in pancreatic cancer may increase the risk of chemotherapy-related toxicity and post-operative morbidity. This systematic review and meta-analysis aimed to quantify the prevalence of sarcopenia in early stage pancreatic cancer. Methods: Relevant studies were identified using Ovid Medline and Elsevier Embase. Pooled estimates of prevalence rates (percentages) and corresponding 95% confidence interval (CI) were computed using a random-effects model to allow for heterogeneity between studies. Results: The majority of the thirty-three studies (n=5593 patients) included in this meta-analysis utilised computed tomography (CT)-derived measures for body composition assessment in patients undergoing pancreatic resection. Reported prevalence of sarcopenia varied between 14 and 74 %, and the pooled prevalence was 39% (95% CI 38- 40%) Heterogeneity was considerable however (I² 93%) and did not improve significantly when controlling for assessment method, and use of pre-defined cut-offs for sarcopenia, limiting potential to evaluate the true impact of sarcopenia. Conclusion: The ready availability of sequential CT offers a valuable opportunity for body composition assessment, but the quality of assessment and interpretation must improve before the impact of body composition on treatment-related outcomes and survival can be assessed. We suggest recommendations for the assessment of body composition for the design of future studies.