Laparoscopic left hemihepatectomy is suitable as a first step in pure laparoscopic major hepatectomy

Abstract Aim As a procedure, major laparoscopic liver resection (LLR) remains in the exploration phase. Previous studies have assessed major LLR en bloc, including hepatectomies of varying complexities; however, the number of segments alone does not convey the complexity of a resection. This study aimed to assess operative outcomes of LLR procedures with more than one sectionectomy, and to identify the best procedure as a first step when learning to carry out major LLR in order to make LLR a safer, more widely used procedure. Methods We carried out a retrospective review of the operative outcomes of 120 consecutive patients who underwent pure LLR with more than one sectionectomy. Operative outcomes were compared according to the complexity classification recently published, and the learning curve for each LLR procedure was assessed and compared. Results Operative outcomes, including operative time, blood loss, and the comprehensive complication index, were significantly stratified according to complexity. There were significant differences in operative outcomes among the medium complexity procedures. The operative time for left hemihepatectomy was the shortest, and the amount of blood loss was the lowest among the medium complexity LLR. Operative times for left hemihepatectomy shortened significantly with time and experience (r = −0.639), and the slope of the learning curve was steeper than for right hemihepatectomy and right posterior sectionectomy. Conclusion Left hemihepatectomy is suitable as a first step in pure laparoscopic major hepatectomy and, given its safety and rapid learning curve for surgeons, it could become the gold standard procedure.


| INTRODUCTION
Laparoscopic liver resection (LLR) surgeries are still being developed, meaning they are not as widely used as open liver surgeries, despite their clinical benefits. These benefits include less blood loss, less pain and analgesic requirement, shorter hospital stays, and improved cosmetic results. [1][2][3][4][5][6][7] International Consensus Conferences on LLR have been held to evaluate the status of laparoscopic liver surgery, and these conferences have provided recommendations to aid in the future development of this procedure. 8,9 During the second conference, the major type of LLR that was discussed was an innovative procedure still in the exploration phase with incompletely defined risks; therefore, extending the clinical indications for this LLR should be considered carefully. 9 In several previous studies, various procedures were assessed and discussed, particularly the learning curve, en bloc as a major LLR. 3,[10][11][12][13][14][15] The complexity of each major hepatectomy procedure differs widely, and the number of segments alone does not convey the complexity of a resection. To this end, Lee et al 16,17 recently reported that the complexity of open liver resections should not be classified based on whether the resection is "major or minor," but instead based on the extent of the liver resection.
There are no reports comparing the learning curve for each different type of major LLR procedure. Here, we aimed to investigate which of these major LLR procedures should be the first step for surgeons when starting to carry out major LLR.

| Data source and study population
The prospective database of patients treated with pure LLR with more than one sectionectomy at our institution was retrospectively reviewed. All patients provided informed consent for the procedure.
This study was approved by our institutional review board.
We have been carrying out LLR procedures at our institution since May 1997, and approximately 600 patients have undergone LLR as of October 2017. At first, the laparoscopic peripheral wedge resection procedure was introduced, and the extent of resection was extended in a stepwise method. Laparoscopic left lateral sectionectomy (LLLS) was the first sectionectomy adopted in April 2003. Right hemihepatectomy was the first major LLR introduced in September 2009.
The clinical records of the consecutive 120 patients who underwent a pure LLR with more than one sectionectomy were reviewed to extract the following information for analysis: patient characteristics, tumor characteristics, operative procedures, and operative outcomes. First, the operative outcomes were compared according to Lee's complexity classification system for hepatectomy. 16,17 The classification divides open liver resection procedures into three complexity groups: low, medium, and high. In this system, a left lateral sectionectomy is classified as low complexity. Left hemihepatectomy, right hemihepatectomy, and posterior sectionectomy are considered medium complexity, whereas right anterior sectionectomy and central bisectionectomy are considered high complexity. Next, the outcomes of each LLR procedure classified as medium complexity were compared to evaluate the learning curve, and to explore which procedure was suitable as the first step in major LLR. This analysis was carried out after also excluding cases involving associated procedures (e.g. stoma closure, colectomy, or radiofrequency ablation) and/or multiple hepatectomies. In this study, an expert was defined as a surgeon who had experience in carrying out more than 60 LLR, either as a surgeon or as an assistant surgeon. 10,11 In the present study, three experts and six non-experts carried out the LLR.

| Definitions
Extent of liver resection was classified according to the Brisbane 2000 terminology. 18 Postoperative morbidity and mortality were defined as any complication or death within 90 days, respectively.
Complications were graded according to the Clavien-Dindo classification system and were scored by the comprehensive complication index (CCI ® ). 19

| Surgical technique
Our basic surgical LLR techniques were carried out as previously described. 21,22 The Glissonian approach or individual hilar dissection for the hilar approach was chosen on a case-by-case basis. The liver parenchyma was transected along the demarcation line and the main hepatic vein. After April 2012, the intermittent Pringle maneuver was routinely used except for LLLS. Additionally, low central venous pressure, the reverse Trendelenburg position, low airway pressure, and low tidal volume contributed to less bleeding from the hepatic vein. 5,23,24

| Statistical analysis
Continuous data were expressed as median values with the associated interquartile ranges. Categorical data were expressed as counts, with the associated percentile values calculated. The Kruskal-Wallis test was used to compare continuous data, and the chi-squared test was used for categorical data. P-value <0.05 was considered statistically significant. All statistical analyses were carried out using JMP statistical software (version 9.0.0; SAS Institute, Cary, NC, USA). Table 1. Rate of malignant lesions was 88.3%, median tumor size was 35 mm (24-61 mm), and liver cirrhosis was observed in 15 patients (12.5%).

Patient and tumor characteristics are shown in
Operative methods and outcomes are summarized in Table 2. According to Lee's complexity classification, 46 patients (39.7%) were considered as low, 61 (52.6%) as medium, and nine (7.7%) as high complexity cases. Operative outcomes for the patients who underwent LLR without associated procedures according to Lee's complexity classification are shown in Table 3   Laparoscopic left lateral sectionectomy has been acknowledged as the standard procedure. 22,[26][27][28] In this study, the first 14 cases showed a rapid learning curve, that later plateaued, regardless of whether the procedures were carried out by not only experts but also by nonexpert surgeons. These results suggested that the LLLS procedure was well standardized, and the results aligned with the current view that LLLS should be the standard procedure. Additionally, its learning curve was steeper than that of medium-complexity LLR.  The CCI ® was developed to overcome an underestimation of the true overall postoperative morbidity. 19,20 In this study, the CCI ® was used to evaluate postoperative morbidity, and low values were obtained throughout our study period. These findings suggest that the learning curve did not influence morbidity. The reason for this could be that LLLS was adopted as the first LLR with more than one sectionectomy, which was a relatively easier procedure; therefore, there were no major complications. Additionally, it could be because major LLR was adopted only after adequate experience was acquired with minor LLR. The indications for hepatectomy were extended in a stepwise method to ensure patient safety.
In the present study, the surgeries of two patients in the left hemihepatectomy group were converted to laparoscopic-assisted procedures. The first patient had massive bleeding from the portal vein caused by the Pringle maneuver itself, and mini-laparotomy was   Regarding the operative techniques, there are two methods of hilar dissection, the Glissonian approach and individual dissection. In the present study, the operative outcomes of the two methods could not be compared because of differences in patient backgrounds.
Both methods are generally acceptable, and neither is superior in terms of safety and reproducibility. Therefore, the operative method is left to the surgeon's discretion and preference. 30 Spiegel's lobe resection is another concern. Spiegel's lobe resection requires dissection from the IVC, and division of the Glissonian vessels in Spiegel's lobe and the larger parenchymal transection area. Therefore, the risk of bleeding and bile leakage might increase and prolong the operative time.
This study had several limitations. First, this report includes only the experiences gathered from a single, specialized institution. Additionally, the number of cases in our study group was limited and, consequently, the learning curve for the LLR apart from four procedures could not be evaluated. Therefore, validations, multicenter comparison, learning curve analyses, or comparisons among surgeons with varied experiences are required using independent data sets.
Despite these limitations, we believe that a pure laparoscopic left hemihepatectomy without Spiegel's lobe resection is suitable as a first step for pure laparoscopic major hepatectomy and, given its safety and rapid learning curve for surgeons, could become the gold standard procedure.

DISCLOSURE
The protocol for this research project has been approved by a suit- Author Contribution: All authors were involved in the conception, design, interpretation of data, surgery, and acquisition of data. YH and TT analysed the data.