Radiofrequency ablation of 231 unresectable hepatic tumors: indications, limitations, and complications.

BACKGROUND: Radiofrequency ablation (RFA) is increasingly used for the local destruction of unresectable hepatic malignancies. There is little information on its optimal approach or potential complications.
METHODS: Since late 1997, we have undertaken 91 RFA procedures to ablate 231 unresectable primary or metastatic liver tumors in 84 patients. RFA was performed via celiotomy (n = 39), laparoscopy (n = 27), or a percutaneous approach (n = 25). Patients were followed with spiral computed tomographic (CT) scans at 1 to 2 weeks postprocedure and then every 3 months for 2 years.
RESULTS: Intraoperative ultrasound (IOUS) detected intrahepatic disease not evident on the preoperative scans of 25 of 66 patients (38%) undergoing RFA via celiotomy or laparoscopy. In 38 of 84 patients (45%), RFA was combined with resection or cryosurgical ablation (CSA), or both. RFA was used to treat an average of 2.8 lesions per patient, and the median size of treated lesions was 2 cm (range, 0.3-9 cm). The average hospital stay was 3.6 days overall (1.8 days for percutaneous and laparoscopic cases). Ten patients underwent a second RFA procedure (sequential ablations) and, in one case, a third RFA procedure for large (one patient), progressive (seven patients), and/or recurrent (three patients) lesions. Seven (8%) patients had complications: one skin burn; one postoperative hemorrhage; two simple hepatic abscesses; one hepatic abscess associated with diaphragmatic heat necrosis following sequential percutaneous ablations of a large lesion; one postoperative myocardial infarction; and one liver failure. There were three deaths, one (1%) of which was directly related to the RFA procedure. Three of the complications, including one RFA-related death, occurred after percutaneous RFA. At a median follow-up of 9 months (range, 1-27 months), 15 patients (18%) had recurrences at an RFA site, and 36 patients (43%) remained clinically free of disease.
CONCLUSIONS: Celiotomy or laparoscopic approaches are preferred for RFA because they allow IOUS, which may demonstrate occult hepatic disease. Operative RFA also allows concomitant resection, CSA, or placement of a hepatic artery infusion pump, and isolation of the liver from adjacent organs. Percutaneous RFA should be reserved for patients at high risk for anesthesia, those with recurrent or progressive lesions, and those with smaller lesions sufficiently isolated from adjacent organs. Complications may be minimized when these approaches are applied selectively.