RESUMO
BACKGROUND: The ideal treatment approach for colorectal cancer (CRC) with synchronous liver metastases (SCRLM) remains debated. We performed a network meta-analysis (NMA) comparing the 'bowel-first' approach (BFA), simultaneous resection (SIM), and the 'liver-first' approach (LFA). METHODS: A systematic search of comparative studies in CRC with SCRLM was undertaken using the Embase, PubMed, Web of Science, and CENTRAL databases. Outcome measures included postoperative complications, 30- and 90-day mortality, chemotherapy use, treatment completion rate, 3- and 5-year recurrence-free survival, and 3- and 5-year overall survival (OS). Pairwise and network meta-analysis were performed to compare strategies. Heterogeneity was assessed using the Higgins I2 statistic. RESULTS: One prospective and 43 retrospective studies reporting on 10 848 patients were included. Patients undergoing the LFA were more likely to have rectal primaries and a higher metastatic load. The SIM approach resulted in a higher risk of major morbidity and 30-day mortality. Compared to the BFA, the LFA more frequently resulted in failure to complete treatment as planned (34% versus 6%). Pairwise and network meta-analysis showed a similar 5-year OS between LFA and BFA and a more favorable 5-year OS after SIM compared to LFA (odds ratio 0.25-0.90, p = 0.02, I2 = 0%), but not compared to BFA. CONCLUSION: Despite a higher tumor load in LFA compared to BFA patients, survival was similar. A lower rate of treatment completion was observed with LFA. Uncertainty remains substantial due to imprecise estimates of treatment effects. In the absence of prospective trials, treatment of stage IV CRC patients should be individually tailored.
