Ablation of liver metastasis: is preoperative imaging sufficiently accurate?

The recent introduction of cryotherapy and radiofrequency ablation of liver metastasis has expanded the indications for treatment. As technology has advanced, a percutaneous approach has been developed. Percutaneous treatment, however, requires accurate preoperative imaging. From 1993 to 1999, 179 patients underwent operative exploration for treatment of suspected hepatic metastases from colorectal carcinoma. One hundred seventy-seven patients were staged by preoperative CT, two patients were staged by MRI, and complete data were available in 176. Hepatic tumor count by preoperative imaging was compared to intraoperative tumor count obtained by inspection, palpation, ultrasonographic examination using a 3.5/7.5 MHz T probe, and careful gross sectioning of the resected specimen. Post hoc analysis was performed on 35 CT scans by two radiologists who specialize in abdominal CT. These radiologists were blinded to the intraoperative findings. Their interpretations were compared to the intraoperative counts and to each other. Thirty-four (19%) of 179 patients were deemed untreatable at operation because of unsuspected overwhelming liver involvement in 11 (6%) or extrahepatic metastases in 23 (13%). For the group, CT was accurate in 80 patients (45%), showed more lesions than were found in 16 (9%), and showed fewer metastases than were found in 80 (45%). When the preoperative scan predicted a solitary metastasis, it was correct in 45 (65%) of 69 patients and underestimated disease in 24 (35%). In the post hoc analysis, the mean numbers of lesions reported by the two radiologists did not differ from the mean number of tumors found; however, the radiologists' counts agreed on 16 (59%) and disagreed on 11 (41%) of the scans. The accuracy of CT decreased with increasing numbers of lesions. Regardless of the type of preoperative imaging, intraoperative findings altered the course of the operation in 96 (55%) of 176 patients. Preoperative imaging is not sufficiently accurate to permit adequate percutaneous treatment of hepatic metastases from colorectal carcinoma.