Objective: Body composition measures (BCMs) are an important predictor of nutritional status in patients with cancer. Poor nutritional status is common in ovarian cancer (OC) and is a well-known variable that influences cancer treatment and outcome. We aim to establish the role of BCMs measured by computed tomography (CT) in predicting outcomes in patients with OC.
Methods: We retrospectively searched our institutional database for patients with stage IIIC/IV OC who underwent surgery as primary treatment at Mayo Clinic between 1996 and 2005 and had adequate presurgical CT images available. For each patient, 1 axial CT image at the level of the 3rd lumbar vertebra was evaluated. Adipose and lean tissues were discriminated using commercially available software. Cox models were fit to evaluate the relationship between patient factors and overall survival (OS). Associations were summarized using hazard ratios (HRs) and corresponding 95% CIs.
Results: A total of 82 patients were identified, with a median age of 68.4 years. OS at 1 and 5 years was 84.1% and 24.1%, respectively. Older age (P=.01), stage IV disease (P<.001), and subcutaneous and muscular fat<77.21cm(2) (P<.001) were independently associated with poor OS. Longer hospital stay was independently predicted by albumin≤3g/dL (P=.03), suboptimal surgery (P=.02), and subcutaneous and muscular fat<77.21cm(2) (P<.001). Surgical complications were independently predicted only by albumin≤3g/dL (P<.01).
Conclusions: CT BCMs, as indicators of nutritional status, are independent predictors of longer hospital stay and poor OS in patients with OC.
Published by Elsevier Inc.