Liver resection for noncolorectal liver metastases: the good, the bad, and the ugly

Background Noncolorectal liver metastases (NONCOLMETs) are composed of a heterogeneous group historically associated with poor outcomes. Advances in systemic therapy and liver interventions have renewed interest in liver resection (LR) for selected patients. However, evidence remains inconsistent, and indications vary across tumor types. Methods A comprehensive MEDLINE and Embase search up to February 2025, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, identified studies assessing the outcomes of liver interventions for NONCOLMETs. The studies were grouped by primary tumor. The extracted and summarized data included overall survival (OS), prognostic determinants, and comparisons with nonsurgical management. Results From 899 screened records, survival after LR varied markedly. The most favorable outcomes were seen in neuroendocrine tumors (median OS, 84–120 months) and gastrointestinal stromal tumors (GISTs) responding to tyrosine kinase inhibitors (TKIs; median OS, 70–90 months). Breast and selected urogynecologic cancers showed a median OS of >36 months. However, esophageal, gastric, and pancreatic cancers showed benefit only in exceptionally selected, chemoresponsive liver-only disease. Across tumor types, specific favorable prognostic factors included oligometastatic and liver-only disease; longer disease-free interval; radiologic or biomarker response to systemic therapy, such as carbohydrate antigen 19–9 decline in pancreatic cancer, estrogen receptor/progesterone receptor positivity in breast cancer, human epidermal growth factor receptor 2 or programmed death-ligand 1 expression in gastroesophageal cancer, and TKI response in GIST; and well-differentiated histology. Comparative and propensity-matched analyses consistently suggested a survival benefit from LR in favorable subsets of patients with NONCOLMETs. Conclusion LR may provide meaningful survival benefit in carefully selected patients with biologically favorable, liver-dominant NONCOLMETs. Multidisciplinary evaluation and tumor-specific selection criteria remain essential. Prospective studies are needed to refine the indications.