Liver perfusion chemotherapy for selected patients at a high-risk of liver metastasis after resection of duodenal and ampullary cancers

OBJECTIVE: To evaluate the prognostic benefit of postoperative liver perfusion chemotherapy (LPC) in patients who undergo curative resection of duodenal and ampullary cancers. SUMMARY BACKGROUND DATA: Both nodal involvement and pancreatic invasion are poor prognostic indicators after curative resection of ampullary or duodenal cancers due to high incidences of liver metastasis. Therefore, we have performed postoperative LPC on a number of such "high-risk" patients. METHODS: During the period of 1990 to 2005, 72 consecutive patients successfully underwent curative (R0) resection of duodenal or ampullary carcinomas at our institution, The Osaka Medical Center for Cancer and Cardiovascular Diseases. Of these 72 patients, 38 were found to have positive nodal involvement and/or pancreatic invasion based on an intraoperative inspection, and of these, 28 were deemed to be suitable candidates for intraoperative catheterization: 1 catheter was placed into the gastroduodenal artery; another into the portal vein (group A). Postoperatively, they received an infusion of 5-fluorouracil (5-FU: 125 mg/d) via each of the 2 catheters for a period of 28 continuous days. The remaining 44 patients (group B) did not receive any other adjuvant therapy. The survival rates and patterns of disease failure were compared between these 2 groups and their subgroups. RESULTS: All 72 patients survived the operation, and all 28 patients in group A completed their courses of LPC without showing any significant adverse signs. Postoperative histopathology was later performed to get a more accurate picture regarding the degree of nodal involvement and/or pancreatic invasion: In group A, 21 patients (group A1) proved positive for nodal and/or pancreatic invasion whereas 7 patients (group A2) proved negative; and in group B, 16 patients proved positive (group B1) whereas 28 proved negative (group B2). Although group A displayed higher incidences of nodal involvement and pancreatic invasion, the 5-year survival rates for the 2 groups varied only slightly. The 5-year survival rate was 70% in group A1, 85% in group A2, 35% in group B1, and 92% in group B2, respectively. The difference between B1 and B2 and the difference between A1 and B1 were statistically significant, and these differences were conclusively found to be attributable to the different incidences of liver metastasis. CONCLUSION: Through this research, both nodal involvement and pancreatic invasion were confirmed to be reliable predictors of liver metastasis after curative resection of ampullary and duodenal cancers. Since LPC was proven to be effective in preventing the postoperative development of liver metastasis, it should be more actively performed for patients with a high-risk of liver metastasis.