Enhanced computed tomography imaging features predict tumor grade in pseudomyxoma peritonei

摘要

Background: Because few studies have focused on the correlation between computed tomography (CT) signs and tumor grade in pseudomyxoma peritonei (PMP), we evaluated predictive value of abdominal enhanced CT in identifying high- vs. low-grade cases. Methods: In all, 75 patients diagnosed with PMP after surgery were consecutively recruited. The preoperative enhanced CT images were retrospectively analyzed for ascites, hepatic scalloping, omental and peritoneal lesion appearance, intralesional calcification and septa, and peripheral organ involvement. Logistic regression models were applied to analyze the relationship of CT signs with PMP grade. Receiver operating characteristic curves were generated to evaluate the potential utility of CT signs in detecting high-grade PMP. Results: Massive ascites (P=0.017) and peritoneal solid nodules (P<0.001) were more common in high-grade cases. Multivariate logistic regression identified massive ascites odds ratio (OR) =4.389, 95 confidence interval (CI): 1.210-15.921; P=0.025 and peritoneal solid nodules (OR =19.932, 95 CI: 3.560-111.596; P<0.001) as independent predictors of high-grade PMP. For the 55 patients with hepatic scalloping, the maximum thickness of mucin deposition at the hepatic scalloping wave in high-grade PMP was thinner than that in low-grade PMP (P=0.021). Thickness of mucin deposition at the hepatic scalloping wave (OR =0.346, 95 CI: 0.148-0.809; P=0.014) was an independent predictor of high-grade PMP, with a cutoff value of 18.6 mm. Cancer antigen 125 (CA125) combined with CT signs was significantly better at diagnosing high-grade PAD than was CA125 alone in both the overall patients area under the ROC curve (AUC): 0.812 vs. 0.656; P=0.020 and those with hepatic scalloping (AUC: 0.859 vs. 0.600; P=0.007). Conclusions: The CT signs of high-grade PMP significantly differ from those of low-grade PMP, and thus combining CT signs with CA125 may be highly valuable for classifying PMP.