Study design, selection of patients and grouping
From January 2007 to December 2016, patients receiving surgery at the department of hepatobiliary surgery at Peking University Cancer Hospital for colorectal cancer with synchronous liver metastases were retrospectively collected and reviewed. Synchronous liver metastasis was defined as the occurrence of liver metastases synchronous with or before the confirmed diagnosis of colorectal cancer. Inclusion criteria: resection of primary colorectal cancer and liver metastases was performed with curative intent, and the patients completed the treatment; the combined extrahepatic metastases were also resected. Combined lung metastases without resection was allowable if controllable under chemotherapy[16, 17]. Exclusion criteria: (1) Recurrence after resection of liver metastases; (2) Non-radical surgery; (3) Failing to finish the resection of both colorectal cancer and liver metastases; (4) Combined with other malignancies. Among patients with initially unresectable colorectal liver metastases, those who did not receive second-stage surgery due to failed conversional chemotherapy after the first-stage surgery were thus excluded.
The treatment strategy was determined based on multidisciplinary discussion attended by colorectal surgeons, hepatic surgeons, medical oncologists, radiation oncologist and radiologists. The informed consent was obtained from the patients before the treatment began. The patients were divided into three groups based on surgical treatment strategies: sequential resection (SeR, sequential staged resection of colorectal cancer and liver metastases, without interval chemotherapy); delayed resection (DeR, staged resection of colorectal cancer and liver metastases, with interval chemotherapy); simultaneous resection (SiR, one-stage resection of primary colorectal cancer and liver metastases simultaneously). Simultaneous resection is more often used in patients that the tumor disease burden is not heavy, or the primary lesion is located in the right colon with heavy tumor burden but can be resected through one incision. Staged resection is mainly used in patients with severe symptoms of the primary lesion, or with heavy tumor burden, or with tumors located in the rectum requiring preoperative radiochemotherapy. After the first surgery, patients with resectable liver metastases were treated with sequential resection or neoadjuvant chemotherapy, and patients with unresectable liver metastases were treated with conversional chemotherapy, then evaluated for the next surgery.
The protocol was approved by the ethics committee of Peking University Cancer Hospital and confirmed to the Declaration of Helsinki. The informed consent was harvested from all patients.
Initial assessment
Before initial treatment, all patients received contrast-enhanced MRI of the liver, contrast-enhanced CT of the abdominopelvic cavity or contrast-enhanced MRI of the pelvic cavity (only for rectal cancer patients) and plain CT scan of the chest. If the primary colorectal cancer had already been resected at other centers, the operation note, pathology report and postoperative complications were kept and recorded in details. Radiological assessment was performed to exclude any signs of residual lesions.
Surgery
The tumor number and position were determined by preoperative imaging, intraoperative ultrasound and palpation. The resected scope of liver was determined based on the tumor number and position. For lesions that were located deep in the liver parenchyma and smaller than 2cm in diameter, combined radiofrequency ablation was selectively performed to avoid excessive loss of liver volume. The liver resection margin was usually larger than 5mm, and the margin of 1mm (R1 margin status) for some lesions was acceptable as long as the chemotherapy was effective. Major liver resection was defined as the resection of 3 hepatic segments and above.
Perioperative chemotherapy
Based on the consensus, initial chemotherapy was usually recommended for patients with asymptomatic primary lesions in our center, except for patients’ refusal. Systemic chemotherapy regimens were oxaliplatin and/or irinotecan based regimens in combination with fluorouracil and leucovorin (folfox/folfiri/folfoxiri). Combined molecular targeted agents were selectively used according to the resectability of liver metastases and clinical risk scores. Assessment was performed after 2 or 4 cycles of initial chemotherapy, and those fit for surgery received surgery as soon as possible. For advanced middle and low rectal cancer (T3/T4 and/or N+), combined radiotherapy was given based on local staging. For staged resection, the chemotherapy regimen given between two surgeries was usually the same as the initial chemotherapy, or folfox/capox/folfiri plus or not molecular targeted agents in the absence of initial chemotherapy. Adjuvant chemotherapy was recommended regularly if the patients’ conditions allowed after surgery.
Perioperative safety
Postoperative complications were assessed by Clavien-Dindo grading system[18]. For staged resection, overall incidence of complications was the sum of incidence of complications after two surgeries, and the highest grade of complications after either surgery was taken as the final grade of complications. Severe complications were defined as those of Clavien-Dindo grade 3 or above.
Postoperative follow-up and survival analysis
Radiological assessment was given within 1 month after resection of both colorectal cancer and liver metastases. Later the patients received reexaminations once every 3 months. Overall survival (OS) was defined as the interval from the start of the initial treatment (surgery or chemotherapy) to the last follow-up or death. Recurrence-free survival (RFS) was defined as the interval from resection of both colorectal cancer and liver metastases to the time of the first recurrence. The time of the last follow-up was December 2018 for all cases.
Statistical analysis
Statistical analysis was performed using SPSS 22.0 software. Continuous variables were described by ranges, and intergroup comparisons were conducted using t-test or U test. Categorical variables were expressed as frequencies or percentages. Intragroup comparisons were performed by using chi-square test. Kaplan-Meier survival analysis was performed, and the survival curves were compared by using the log-rank test. Univariate and multivariate analyses were performed using Cox model to identify the prognostic factors. P<0.05 indicated significant difference.