Survival analysis between laparoscopic and open hepatectomy for hepatocellular carcinoma: a meta-analysis based on reconstructed time-to-event data.

Abstract

Laparoscopic hepatectomy (LH) has been widely used in the treatment of hepatocellular carcinoma (HCC). It is generally believed that the long-term outcomes of LH are not inferior to open hepatectomy (OH). However, the quality of evidence is low. The purpose of this study was to reconstruct time-to-event data for meta-analysis based on Kaplan-Meier curves from propensity-score matched studies and compare survival rates following LH and OH for hepatocellular carcinoma. All published propensity-score matched studies reported in English that compared LH and OH for hepatocellular carcinoma with Kaplan-Meier curves were screened. Patients' survival information was reconstructed with the aid of a computer vision program. Different models (fixed-effects model for two-stage survival analysis and Cox regression for one-stage survival analysis) were performed for sensitivity analysis. In addition to the primary meta-analysis, two specific subgroup analyses were performed on patients by types of resection, cirrhosis status. Time-to-event data were extracted from 45 propensity-score matched studies (N = 8905). According to the time-to-event data and the reconstructed Kaplan-Meier curves, the cumulative overall survival rate was 49.0% and 50.9% in the LH and OH cohorts, respectively, a log-rank test did not demonstrate statistical significance (p > 0.05). The cumulative recurrence-free survival (RFS) probability was both close to 0.0%. The median RFS time was 49.1 (95% CI 46.1 ~ 51.7) and 44.3 (95% CI 41 ~ 46.1) months. The difference in disease status was statistically significant by the Log-rank test (p < 0.05). Using the random-effects model of two-stage analysis, the minor hepatectomy subgroup (HR = 1.32, 95% CI [1.09, 1.55], I2 = 6.2%, p = 0.383) and the shared fragile model of one-stage analysis (HR = 1.44 95% CI [1.23, 1.69], p < 0.001) suggested that LH could significantly prolong RFS of patients compared with OH. This result was consistent with sensitivity analysis using different models. This study was the first reconstructed time-to-event data based on a high-quality propensity-score matching study to compare the survival outcomes of LH and OH in the treatment of HCC. Results suggested that LH can improve RFS in patients with HCC undergoing minor hepatectomy and may also benefit long-term RFS in overall patients.