1. Comparison of baseline data among the three groups
A total of 233 consecutive cases conformed to the inclusion criteria, with 49 cases in the SeR group, 98 cases in the DeR group, and 86 cases in the SiR group. A comparison of the baseline data among the three groups is shown in Table 1. In the SeR group, the proportion of patients with a primary lesion in the rectum was significantly higher than that in the other two groups (P < 0.05). Moreover, the median number of liver metastases in the SiR group was lower than that in the SeR group (P = 0.011). In addition, the three groups of patients showed no significant differences in sex, age, T or N stage of the primary lesion, liver metastatic lesion diameter, initial CEA level, or distribution in one or two lobes. The proportion of combined extrahepatic metastases was low in all groups, with no significant differences among the three groups.
2. Comparison of chemotherapy regimen, surgical strategies and postoperative complications among the three groups
A comparison of the chemotherapy regimens and surgical strategies among the three groups is shown in Table 2. Compared with the DeR group, a higher proportion of patients received initial chemotherapy in the SeR and SiR groups (P < 0.05). It was as high as 83.7% in the SeR group. In addition, the proportion of patients receiving initial chemotherapy with molecularly targeted agents was also higher in the SeR group (P < 0.05).
In the DeR group, the median number of chemotherapy cycles between the two surgeries was 4 (1, 25), and the median interval was 20.7 (10.1, 77.2) weeks. In the SeR group, the median interval was 5.7 (3.4, 14) weeks between the two surgeries. There were no significant differences in the proportion of patients receiving adjuvant chemotherapy among the three groups. Except for the higher proportion of portal occlusion in the SeR group, the three groups showed no significant differences in the use of extensive liver resection or combined radiofrequency ablation, intraoperative blood loss or postoperative resection margin status. A total of 34 patients in the SeR group and 6 patients in the DeR group underwent treatment with the “liver first” strategy. Only 4 patients underwent the PVL procedure, and no ALPPS (associating liver partition and portal vein ligation for staged hepatectomy) procedures were performed in any group.
Regarding postoperative safety, the perioperative mortality rate was 0 in all groups. The incidence of overall complications was 48.8% in the SiR group, higher than that in the DeR group (31.6%, P = 0.017) and the SeR group (40.8%, P = 0.369). The incidence of severe postoperative complications was 26.7% in the SiR group, which was higher than that in the DeR group (11.2%, P = 0.007) and higher but not significantly higher than that in the SeR group (16.3%, P = 0.166). The SeR and DeR groups showed no significant difference in the overall incidence of postoperative complications or the incidence of severe complications. The details of the complications and the Clavien-Dindo grades are shown in Table 2 and Supplementary table 1.
3. Survival analysis
There was no significant difference in the median OS (45 vs. 43 months, P = 0.887) or RFS (9 vs. 8 months, P = 0.714) between patients undergoing simultaneous resection and staged resection (Fig. 1). The 1-year and 3-year survival rates were 93.9% and 50.1% in the SeR group, 94.9% and 64.8% in the DeR group, and 93.0% and 55.2% in the SiR group, respectively. The median OS in the SeR group was lower than that in the DeR group (37 vs. 48 months, P = 0.019), but it was not significantly different from that in the SiR group (37 vs. 43 months, P = 0.378). The 1-year and 3-year recurrence-free survival rates were 22.4% and 18.4% in the SeR group, 43.9% and 24.2% in the DeR group, and 31.4% and 19.6% in the SiR group, respectively. The RFS in the SeR group was also lower than that in the DeR group (6 vs. 10 months, P = 0.033), but the difference was not statistically significant compared with that in the SiR group (6 vs. 8 months, P = 0.275) (Fig. 2).
4. Univariate and multivariate analyses of overall survival
Univariate analysis was performed using the Cox regression model to identify the factors influencing OS (Table 3). The results show that the T and N stage of the primary lesion, diameter of liver metastatic lesions, resection strategies and adjuvant chemotherapy were correlated with OS. Other imbalanced factors in baseline and treatment, such as the position of the primary lesion, number of liver metastases, administration of initial chemotherapy, and use of molecular target drugs, did not affect OS. Cox multivariate analysis indicated that a stage T4 lesion, lymph node-positive primary tumour, tumour size >30 mm and selective sequential resection (relative to delayed resection) were correlated with poor prognosis (Table 4).