Colorectal cancer remains one of the leading causes of cancer-related deaths worldwide [1]. Surgical resection represents the only chance of long-term survival. The goal of surgery should be to resect all metastases with negative histological margins while preserving sufficient functional hepatic parenchyma [2]. Currently, the possibility of improving survival in patients with CRLM is favored by advances in medical therapy, such as systemic chemotherapy, and the improvement in surgical techniques for metastatic disease [3, 4].

The laparoscopic approach is associated with a decrease in intraoperative bleeding, fewer postoperative complications, shorter hospital stay, and decrease in health care costs [5]. However, information on long-term oncological outcomes in laparoscopic hepatectomy (LH) for CRLM is limited [6,7,8,9].

A recently published randomized trial (OSLO-COMET) [10] analyzed short-term and long-term outcomes, demonstrating these advantages in oncologic liver surgery. A recent systematic review published by Haney et al. demonstrated that laparoscopic liver surgery for minor resections shows clear perioperative benefits over open liver surgery with moderate to high certainty of evidence. They also report certain limitations in the long-term results in relation to laparoscopic liver resection [11]. Moreover, the status of surgical resection margins is an additional factor that has been evaluated in many series due to its influence on long-term survival [12]. Currently, there is a little evidence about the results of laparoscopic surgical resections margins when compared to open resections.

The aim of this study is to analyze the oncologic outcomes (surgical margin, overall survival, and disease-free survival) in patients with laparoscopic or open liver resection due to CRLM.

Methods

This is a retrospective analysis of a prospectively maintained database with consecutive patients treated with hepatic resection for CRLM in our center between April 2007 and June 2017. Laparoscopic vs open resection was compared. This study complies with ethical standards and was approved by the Institutional Review Board of Hospital Italiano de Buenos Aires (Protocol number 3900). Patients were divided into two groups (laparoscopic vs open). The patients were selected for laparoscopic versus open resection approach (LR vs OR) based on surgeon preference and experience in minimally invasive liver surgery. Only patients who met our institutional inclusion criteria for minimally invasive approach (initial resectability, tumor size ≤ 5 cm, number of metastases ≤ 3, without bilobar disease, absence of extrahepatic disease) were included. Patients who received concomitant treatment with ablative methods were excluded and patients who died within the first 90 postoperative days were included in the study.

Preoperative evaluations consisted of abdominal and chest computed tomography scans (CT) and/or magnetic resonance imaging (MRI) and clinical biochemistry including carcinoembryonal antigen (CEA). The patients were categorized as resectable after evaluation by a multidisciplinary team comprising surgeons, pathologist, oncologist and radiologist.

Synchronous metastases were defined as cases in which patients received a diagnosis of CRLM at the same time or within 6 months of the diagnosis of the primary tumor. Major hepatectomies were defined as resections of three or more liver segments [13].

Demographic and clinicopathological data were collected from each patient, including sex, age, tumor characteristics, preoperative, perioperative, and postoperative variables. The Clavien-Dindo classification system was used to describe postoperative complications (major morbidity DC ≥ 3) [14]. We also analyzed the oncological outcomes: overall survival (OS) (was calculated from the time of resection to death or last follow-up) and disease-free survival (DFS) (from the time of resection to the presence of evidence compatible with recurrence). R0 resection margin was defined as a disease-free margin ≥ 1 mm and R1 as a disease-free margin < 1 mm [15]. R1 resections were classified as follows: R1 vascular when the tumor was detected at the level of the first or second-order Glissonean pedicle or hepatic veins within their last 4 cm before confluence at the inferior vena cava. The tumor was exposed exclusively along the vessel; in R1 parenchymal, the tumor was exposed along the transection plane. OS and DFS were adjusted for the presence of synchronous disease.

Patient follow-up included serum CEA levels and CT of the chest, abdomen, and pelvis every 3 months, the first 2 years, and every 6 months from the third year after liver resection, according to most recent NCCN guidelines [16].

Surgical technique.

The laparoscopic surgical technique used has already been previously described by our group [17].

Statistical analysis

Statistical analyses were carried out using Stata Software for MAC version 13. The distribution of quantitative variables was analyzed graphically using histogram and probability graphs and analytically using the Shapiro-Wilks test. Variables with normal distribution were expressed as mean and standard deviation (SD) and variables with non-normal distribution as the median and interquartile range (IQR). Categorical data was expressed as absolute frequency and percentage. Comparisons between groups for categorical variables were performed by means of the chi-square test or Fisher’s exact test when appropriate.

The Kaplan–Meier method was applied to calculate mean survival values and to construct survival curves. The log-rank test was used to compare survival among the groups. A Cox regression analysis was performed to adjust to overall survival and disease-free survival in relation to other potential confounding factors (synchronicity disease, size, and the number of metastases, bilaterality, preoperative CEA, location of the primary tumor). Statistical significance was reached at p values lower than 0.05.

The Kaplan–Meier method was applied to calculate mean survival values and to construct survival curves. The log-rank test was used to compare survival among the groups. A Cox regression analysis was performed to adjust to overall survival and disease-free survival in relation to other potential confounding factors (synchronicity disease, size, and the number of metastases, bilaterality, preoperative CEA, location of the primary tumor). Statistical significance was reached at p values lower than 0.05.

A propensity score was performed to adjust the indication bias. The predictive model with the largest area under the curve (AUC of 0.76 (CI: 0.64) and the lowest Akaike information criterion (94.5) was selected. The model was used to predict the probability of laparoscopy and the overlap of the Propensity Scores was tested (PS) between the patients who were approached laparoscopically and those who were not in a graphical way with histograms.

The PS was used to adjust in a logistic regression model the positive margin and the global and disease-free survival, reporting the raw OR/HR and adjusted with their respective 95 CI.

Results

From an initial cohort of 263 patients operated for CRLM, open hepatic resection was performed in 237 patients and laparoscopic resection in 26 patients. Of the 237 conventional resection, only 56 presented the established selection criteria similar to laparoscopic resection (Fig. 1). The demographic characteristic and preoperative variables of the two groups are described in Table 1. The analysis shows a synchronous presentation in 28 patients (50%) with an open resection and 6 patients (23%) with a laparoscopic resection (p = 0.021). The extrahepatic disease was present in 3 patients (11.5%) with LR, with a significant difference (p = 0.029). These three cases were intraoperative findings. Simultaneous resection was performed in 24 patients (42%) with an OR and in 5 patients (19%) with LR (p = 0.037). The major hepatectomy (more than 3 segments) was performed in 11 (19.6%) patients with an open approach and 2 (7.6%) patients with a laparoscopic approach (p = 0.209) (Table 2). Postoperative complications were observed in 9 patients (16%) of the open resection group and 2 patients (7.7%) of the laparoscopic resection group (p = 0.3). There was a statistically significant difference in major morbidity; it occurred in 8 patients (14.2%) with OR and the patients with LR did not present complications (p = 0.042). One patient (1.7%) died in the open resection group and there were no deaths in the laparoscopic resection group (p = 0.493) (Table 3). Postoperative hospital stay was significantly shorter after laparoscopic resection (4.5 days; range, 3–6 vs 7 days; range, 5–12; p = 0.001), demonstrating similar results to studies such as Casting et al. (open resection: 28%—laparoscopic resection: 8%) and Nachmany et al. (open resection: 11.4%—laparoscopic resection: 9.6%) among others [18] [19]. The conversion rate of laparoscopic liver resections was 19.2% (5 patients).

Fig. 1
figure 1

Patients selection: CRLM, colorectal liver metastases

Full size image
Table 1 Demographic and perioperative outcomes
Full size table
Table 2 Intraoperative Variables
Full size table
Table 3 Postoperative variables
Full size table

At final pathologic analysis, the resection margin was found positive (R1) in 5 patients with OR and 2 patients with LR (Table 3). In the crude analysis and adjusted by PS, laparoscopic surgery does not increase the positive margin odds (OR: 0.85, 95% CI: 0.15–4.7, p 0.852, adjusted OR: 1.21, 95% CI: 0.18–8.2, p = 0.842). Of these 7 patients, 4 had positive vascular margins and 3 positive parenchymal margins.

The overall survival rate at 1 and 3 years for the open resection group was 92% and 77% respectively and for the laparoscopic resection group was 96% and 75% respectively (p = 0.739). The risk of death was similar in both groups, even after adjusting for PS (HR 0.72, 95% CI: 0.28, p = 0.506; adjusted HR: 0.42, p = 0.169).

Disease-free survival (DFS) at 1 and 3 years was similar in the two groups (Figs. 2 and 3). The risk of recurrence was similar in both groups even after adjusting for PS (HR: 0.74, 95 CI: 0.40, p = 0.344, adjusted HR: 0.67, CI: 5.33, p = 0.280). When adjusting the OS and DFS rates according to the presence or absence of synchronous disease, the risk of recurrence did not increase (p = 0.94; p < 0.001 respectively) (Figs. 4 and 5). The median follow-up was 30 months (range CI: 12–46).

Fig. 2
figure 2

Overall survival of patients who underwent resection of colorectal liver metastases according to operative approach: Kaplan–Meier curve

Full size image
Fig. 3
figure 3

Disease-free survival of patients who underwent resection of colorectal liver metastases according to operative approach: Kaplan–Meier curve

Full size image
Fig. 4
figure 4

Rate adjustment according to the presence or absence of synchronous disease: Kaplan–Meier curve

Full size image
Fig. 5
figure 5

Rate adjustment according to the presence or absence of synchronous disease: Kaplan–Meier curve

Full size image

Discussion

Colorectal cancer is one of the leading causes of cancer-related death worldwide. Liver metastases occur in 60–70% of cases; of these, 20–30% are synchronic and 30–50% metachronic [20]. Currently, surgical resection is considered the only curative treatment for CRLM, with 5-year survival rates ranging between 30 and 58% [21,22,23,24]. Although the oncological outcomes in the different published series are similar, regardless of the type of approach, with the advent of technology in recent years, liver surgeons acquired more and better management of the laparoscopic approach [25]. Furthermore, several studies demonstrated lower intraoperative blood loss, hospital stay, and morbidity rates with the LR [26, 27][28][26, 27]. Our results are consistent with the literature in terms of less intraoperative blood loss, less hospital stay, and morbidity rates with LR (Tables 2 and 3).

Oncological outcomes are analyzed in different retrospective studies, meta-analysis, and propensity score matching where no significant differences are observed in terms of the type of approach. Slim et al. demonstrated equivalent morbidity rates with the literature and improvements in oncological resection but without demonstrating results in terms of OS and DFS [29]. In our series, there was an OS at 1 and 3 years of 92–77% for OR and 96–75% for LR, respectively. DFS at 1 and 3 years was 63–20% for OR and 73–36% for LR respectively.

Our results are comparable with the literature; Koffron et al. published a series of 300 patients in 2007, demonstrating that laparoscopic liver resection is feasible and safe, and offers better outcomes in terms of blood loss, blood transfusion, hospital stay, and morbidity rates. Due to the presence of patients with different types of tumors and poor follow-up, oncological outcomes were not obtained [30].

Published studies regarding oncological outcomes associated with laparoscopy are scarce and poorly defined. More recent series (multicenter studies and meta-analyses) are limited by their heterogeneity [18, 31, 32]; Castaing et al. published a series of two centers that include 120 patients with CRLM in 2009 where half underwent a laparoscopic approach and showed that the surgical approach had no impact on survival or disease recurrence.

In 2015, Shiffman et al. published a meta-analysis that included 610 patients, where they found no differences in oncological outcomes in relation to the type of approach performed.

Furthermore, Zhang et al. in 2017 performed a systematic review that showed that there are no oncological differences regardless of the type of approach.

Robles et al. published in 2019 a prospective randomized study with 204 patients, where they demonstrated similar oncological results in both types of approaches [33]. In another multicenter study, Nguyen et al. demonstrated that the laparoscopic liver resection for colorectal metastasis is safe, feasible, and oncologically compared to open liver resection when performed by experienced surgeons [34].

In our study, the disease presented synchronously in 28 patients (50%) with OR and in 6 patients (23%) with LR (p = 0.021), in accordance with different studies in the literature [35].

The OS and DFS outcomes were adjusted according to the presence or absence of synchronous disease, without significant differences. Multiple clinicopathologic have been established as important determinants of treatment failure for colorectal liver metastases, including the width of the resection margin. Most authors report that this margin is a significant factor that influences OS and DFS. In the past decades, the risk of resection margin < 10 mm was considered a contraindication for liver resection. Ekberg et al. reported margins ≥ 10 mm necessary to improve long-term survival [36]. Currently, different studies show that disease-free resection margins of 1 to 10 mm present the same oncological results as those margins > 10 mm [15]. In a comparative study with propensity score matching Untereiner et al. demonstrated that the negative resection margin rates were similar in both groups [37]. In our study, the resection limit was 1 mm, resulting in resection margin positive (R1) in 5 patients with OR and 2 patients with LR (8.9% and 7.6% respectively; p = 0.852). Results are comparable with diverse studies published to date; the latest systematic review published in the literature did not show significant differences regarding the resection margins according to the type of approach [38].

Vigano et al. published in 2016 the differences between vascular R1 and parenchymal R1. They demonstrated that R1 parenchymal resection is not adequate for CRLM because it is associated with a higher local recurrence risk and worse survival. R1 vascular resection allows local control of the disease equivalent to R0 resection and adequate survival [39]. In our study, the R1 vascular occurred in 4 patients and R1 parenchymal in 3 patients.

Our study has some limitations; it is a retrospective review of data collected prospectively in a single institution. The selections of patients for the type of approach were not strictly random, and the surgical skill set of each surgeon may have influenced their decisions. Furthermore, overall survival and disease-free survival at 5-year outcomes could not be obtained due to difficulties in patients follow-up.

In our single-center study, the oncological outcomes do not show significant differences in terms of surgical margin, overall survival, and disease-free survival according to the surgical approach used in patients with CRLM. We recommend the use of the laparoscopic approach in selected patients, by specialist surgeons and in centers with sufficient experience.