Computed tomographic features of colorectal carcinoma liver metastases predict posthepatectomy patient survival

PURPOSE: The gross appearance of colorectal carcinoma liver metastases reflects the biologic behavior of the tumor, yielding prognostic information. The aims of this retrospective study were to determine whether preoperative computed tomographic features of colorectal carcinoma liver metastases reflect the gross appearance of resected specimens and whether these computed tomographic hepatic features predict survival after hepatectomy. METHODS: Eighty-five patients underwent curative partial hepatectomy for colorectal carcinoma liver metastases. Preoperative computed tomographic features of the largest hepatic deposit were classified by the contour of advancing margin of the tumor into two types: lobular tumors with indentations with an acute angle and nonlobular tumors without such indentations. The correlation between computed tomographic features and 18 other clinicopathologic factors was examined. RESULTS: The overall five-year survival rate was 34.1 percent. Of 85 hepatic tumors examined, 49 were lobular and 36 were nonlobular. Computed tomographic features correlated significantly with gross appearance (P = 0.007). Univariate analysis revealed that computed tomographic features (P < 0.0001), gross appearance (P = 0.0063), size of the largest hepatic deposit (P = 0.0075), age (P = 0.0140), and satellite lesions (P = 0.0443) were significant prognosticators. The five-year survival rates in patients with lobular and nonlobular tumors were 10.4 and 66.1 percent, respectively. By multivariate analysis, computed tomographic features (P < 0.0001) and size of the largest hepatic deposit (P = 0.0419) were independently significant. CONCLUSIONS: Computed tomographic features of colorectal carcinoma liver metastases correlate with their gross appearance. The computed tomographic characterization of liver metastases is the most important independent prognostic factor in patients undergoing curative hepatectomy.