Background: Computed tomography (CT) has been extensively used in predicting suboptimal cytoreduction (SCR) in advanced ovarian cancer (OC). However, disagreements remain in literatures on the predictive value of CT findings for SCR. This meta-analysis was designed to determine the ability of eight preoperative CT findings to predict SCR in advanced OC.
Materials and methods: A comprehensive literature search was conducted for eligible studies to identify the association between the eight preoperative CT findings and SCR in advanced OC. The predictive performances of preoperative CT findings were expressed in terms of sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR) and diagnostic odds ratio (DOR) with pooled proportion.
Results: A total of 10 studies and 1,614 patients were included in this meta-analysis. Large volume ascites had the highest sensitivity (64%, CI 56-71%), with a PLR of 1.3 (CI 1.1-1.5) and an NLR of 0.73 (0.59-0.90), while lymph node involvement had the highest specificity (89%, CI 79-94%). The highest DOR of 3 (CI 2-4) was achieved in peritoneal involvement and large bowel mesentery involvement. The other CT findings had poorer predictive performance.
Conclusion: Preoperative CT findings have a poor discriminative capacity to predict SCR in advanced OC. Preoperative CT predictors should be used with caution amid clinical decision-making.
Keywords: computed tomography; meta-analysis; ovarian cancer; predict; suboptimal cytoreduction.