Prognostic indicators lose their value with repeated resection of colorectal liver metastases.

Background: The liver is the most common site of colorectal liver metastases (CRLM) and surgical resection improves overall survival in selected patients. Here, we investigate outcomes and relevant prognostic factors after repeated hepatic resections for CRLM.

Methods: From a prospective database, 578 patients who underwent 788 resections of colorectal liver metastases were included into this study. In total, 169 patients underwent a second and 41 patients had a third operation due to intrahepatic metastatic recurrence. Univariate and multivariate analyses were performed to determine prognostic risk factors.

Results: 5-year overall survival was 36.7% (95% CI: 30.2%; 43.2%) and 10-year survival was 20.3% (95% CI: 7.6%; 33.0%) in patients undergoing single resection. In patients undergoing a second or third resection, 5- and 10-year survival rates were 56.6% (95% CI: 45.0%; 68.2%) and 21.9% (95% CI: 6.8%; 37.0%) or 53.2% (95% CI: 32.4%; 74.0%) and 25.4%, respectively. In patients undergoing single resection, established markers (number, size and pattern of CRLM [p = 0.030/0.015/<0.001], R-status [p = 0.001], surgical/medical complications [p = 0.001/0.008], CEA-level [p = 0.001] and Fong-Score [p = 0.02]) were significantly associated with survival. In patients undergoing three resections, the only predictive markers were pT-stage of the primary tumor in univariate analysis (p = 0.013) and metachronous metastasis and medical complications in multivariate analysis (p = 0.001/0.025). The Fong-Score had no predictive value in patients undergoing two (p = 0.08) or three (p = 0.7) resections.

Conclusion: Established prognostic indicators are not applicable in patients undergoing repeated CRLM resection. In a highly-selected group of patients, repeated hepatic resections can be performed safely with favorable long-term outcomes.