LAUSR

Objective To retrospectively evaluate the safety and efficacy of percutaneous radiofrequency ablation (RFA) in patients with metachronous hepatic metastases arising from pancreatic adenocarcinoma who had previously received curative surgery. Materials and Methods Between 2002 and 2017, percutaneous RFA was performed on 94 metachronous hepatic metastases (median diameter, 1.5 cm) arising from pancreatic cancer in 60 patients (mean age, 60.5 years). Patients were included if they had fewer than five metastases, a maximum tumor diameter of ≤ 5 cm, and disease confined to the liver or stable extrahepatic disease. For comparisons during the same period, we included 66 patients who received chemotherapy only and met the same eligibility criteria described. Results Technical success was achieved in all hepatic metastasis without any procedure-related mortality. During follow-up, local tumor progression of treated lesions was observed in 38.3% of the tumors. Overall median survival and 3-year survival rates were 12 months and 0%, respectively from initial RFA, and 14.7 months and 2.1%, respectively from the first diagnosis of liver metastasis. Multivariate analysis showed that a large tumor diameter of > 1.5 cm, a late TNM stage (≥ IIB) before curative surgery, a time from surgery to recurrence of < 1 year, and the presence of extrahepatic metastasis, were all prognostic of reduced overall survival after RFA. Median overall (12 months vs. 9.1 months, p = 0.094) and progression-free survival (5 months vs. 3.3 months, p = 0.068) were higher in the RFA group than in the chemotherapy group with borderline statistical difference. Conclusion RFA is safe and may offer successful local tumor control in patients with metachronous hepatic metastases arising from pancreatic adenocarcinoma. Patients with a small diameter tumor, early TNM stage before curative surgery, late hepatic recurrence, and liver-only metastasis benefit most from RFA treatment. RFA provided better survival outcomes than chemotherapy for this specific group with borderline statistical difference.