Baseline CT-derived muscle quality influences post-operative morbidity and five-year survival following pancreatic surgery

Abstract

Purpose: Concerns regarding rapid disease progression have delayed the translation and evaluation of prehabilitation in patients with pancreatic cancer (PC). This study sought to investigate the association between baseline CT-derived muscle indices and post-operative outcome following pancreatic surgery. Methods: Consecutive patients who underwent surgery for suspected pancreatic malignancy at a tertiary, high-volume centre between 2010 and 2013 were identified, and patient status censored on 31st December 2018 to allow a minimum five-year follow up period for all participants. Major post-operative morbidity was defined as per Clavien-Dindo classification (>3). Individual muscle mass and attenuation were measured at the third lumbar vertebrae (L3) using Slice-O-Matic software, and pre-defined BMI- and gender-specific cut-offs applied. Associations between muscle indices and the occurrence of post-operative complications was assessed using binary logistic regression analysis, with results reported as odds ratios (OR) and corresponding 95% confidence intervals. The log-rank (Mantel- Cox) test was used to assess the influence of muscle indices on overall survival. Results: Sarcopenia and low muscle attenuation (MA) were present in 46% and 52% of patients respectively (n=276). Low MA was associated with increased post-operative morbidity (OR 1.6, 95% CI 1.07 -2.19, p=0.029), while both sarcopenia and low MA were associated with reduced overall survival (41 vs 62 months, p=0.036, 29 vs 61 months, p=0.0001 respectively). When patients with pathologically confirmed PC were evaluated separately (n=107), the influence of muscle indices on survival were sustained (sarcopenia 16 vs 26 months, p=0.017, low MA 12 vs 30 months, p= 0.016). Conclusion: Low MA and sarcopenia were prevalent among patients undergoing surgery for suspected pancreatic malignancy. Both factors were associated with increased mortality risk and shorter overall survival, while low muscle attenuation was associated with increased major post-operative morbidity risk. CT-derived muscle assessment could aid patient selection and risk stratification for pancreatic cancer, both by identifying a vulnerable patient cohort in need of pre-operative optimisation, and providing an objective assessment parameter for the evaluation of targeted interventions.