Introduction

Since the first reported laparoscopic liver resection (LLR) was performed in 1992 [1], LLR has became increasingly accepted by surgeons. Laparoscopic left lateral sectionectomy has become a standard surgical procedure for treating tumors located in segment 2 or 3 [2,3,4]. Moreover, LLR is considered an acceptable alternative treatment to open liver resection (OLR) for malignant and benign liver tumors [5]. However, the safety and feasibility of laparoscopic anatomical liver resection for centrally located tumors (CLTs; segments 4, 5, and 8) requiring central hepatectomy (CH) or right anterior sectionectomy (RAS) have not been established yet. Although there have been some reports of totally laparoscopic anatomical mesohepatectomy, they were case reports presenting limited outcomes of anatomical liver resection [6, 7]. Other reports did not provide comparative analysis between laparoscopic and open approaches [8, 9]. In this article, we report our experience with totally laparoscopic anatomical liver resection in a single-institution series of patients with CLTs of the liver. The purpose of this study was to assess the feasibility and safety of laparoscopic CH and RAS by comparing short-term perioperative outcomes of LLR and OLR for CLTs.

Methods

Study design and population

We retrospectively assessed the data of all patients who underwent anatomical liver resection for CLT such as CH or RAS at a single institution between April 2011 and March 2016. A total of 132 patients underwent surgical resection for CLTs over a 5-year period. Those patients who were diagnosed with metastatic liver disease requiring simultaneous operation (n = 9), non-anatomical resection (n = 6), and redo-hepatectomy (n = 11) were excluded from the study. The remaining 106 patients were divided into two groups according to the type of operation: LLR (n = 20) and OLR (n = 86). Two cases that were converted to open surgery were included in the LLR group for analysis based on an intention to treat approach. Since there were unbalanced covariates, we used one-to-one propensity score matching (PSM) to ensure that laparoscopic and open groups were comparable. Finally, 40 patients were enrolled in this study: the LLR group (n = 20) and the OLR group (n = 20). This study was approved by the institutional review board of our institution.

Preoperative evaluation

The diagnosis of CLT was based on computed tomography of the abdomen. Patient status was discussed at a weekly multidisciplinary meeting. Attendees included hepatobiliary surgeons, hepatologists, interventional radiologists, and radiation oncologists. Generally, liver resection was considered for patients with preserved liver function of Child–Turcotte–Pugh class A and American society of anesthesiologist grade <III [10]. The decision to use a laparoscopic approach was made by the surgeon. During the period of study, LLR was performed by a single surgeon, and the indications were tumors located more than 5 mm from major vascular or biliary structures.

Surgical procedures

The patient was laid on a table in the French position with the surgeon standing between the patient’s legs. Usually, five trocars were used in this operation as shown in Fig. 1. A 12-mm flexible camera trocar was placed through the umbilical port. Pneumoperitoneum was maintained at 11 mmHg. Intraoperative ultrasonography was routinely performed to identify the course of the right and middle hepatic vein and evaluate the resection margin of the tumor. An advanced bipolar device (LigaSure™, Covidien, Mansfield, MA) and ultrasonic shear (Harmonic, Ethicon Endo Surgery, Inc., Johnson & Johnson Medical SPA, Somerville, NJ and Sonicision™, Covidien) were used. The method of parenchymal transection changed at our institution over this period of time. Either instruments and/or Cavitronal Ultrasonic Surgical Aspirator (CUSA, Valleylab, Boulder, CO) was used for parenchymal transection. All anatomical liver resections were performed using a Glissonian approach [11].

Fig. 1
figure 1

Operative diagram showing the positioning of the patient (‘French’ position) and the surgical team

Full size image

The operation began with dissection of the round and falciform ligament until the suprahepatic inferior vena cava (IVC) was exposed, and the groove between the right hepatic vein and the middle hepatic vein was identified. Dissection between the liver and the right anterior Glissonian pedicle was done via blunt dissection using bipolar forceps (Medtronic, Minneapolis, MN) and an endoscopic suction irrigator. After the right anterior pedicle was exposed, it was not divided, but temporarily clamped with a laparoscopic bulldog clamp (Aesculap, Center Valley, PA) as previously described in temporary inflow control of the Glisson (TICGL) technique, shown in Fig. 2a [12]. The demarcation line for parenchymal dissection was verified and marked using cauterization (Fig. 2b).

Fig. 2
figure 2

Temporary inflow control of Glisson (TICGL) technique in laparoscopic right anterior sectionectomy [12]. a The right anterior Glisson was temporarily controlled using a bulldog clamp. b Demarcation and transection of the right intersectional plane.

c Isolation of the right anterior Glisson using nylon tape and retrieval of the bulldog clamp. d Retraction using nylon tape and division of the right anterior Glisson using an endoscopic staple

Full size image

In the cases of difficult clamping of the right anterior Glisson, the whole right Glissonian pedicle was initially clamped to verify the margin of the plane between the right and left lobe. Since the right Glisson is fully exposed after division of this plane, the right anterior branch can be identified much easier.

For RAS, parenchymal transection was carried out along Cantlie’s line. In cases of excessive bleeding during this stage of parenchymal transection, the Pringle maneuver was applied. After more than 2/3rd of the parenchymal transection was done along this plane, the remainder of the parenchymal transection was carried out along the demarcated line between the right anterior and posterior sections. Transection of the parenchyma was performed until the right anterior pedicle was fully exposed. Subsequently, tape was used to encircle the pedicle. The bulldog clamp was removed, and staples were inserted using a vascular endoscopic stapler (EndoGIA™ Tan, Covidien, Mansfield, MA and Echelon™ White, Ethicon Endo Surgery, Inc., Johnson & Johnson Medical SPA, Somerville, NJ).

It is important to retract remnant Glissonian pedicles with tape caudally and into the left, and to apply the stapler as distally as possible in order to prevent stricture of the remnant structures, as shown in Fig. 2d. The remaining parenchyma was then transected with special attention to the hepatic veins branches, since heavy bleeding might occur at this stage due to branch injury. A specimen was retrieved through the Pfannenstiel incision. The cut surfaces of the liver were checked. Bleeding and bile leakages were controlled meticulously. Subsequently, hemostatic tissue sealant was applied and then a Jackson–Pratt drain was inserted at the site of liver resection.

The process of applying the bulldog clamp throughout parenchymal transection and stapling after the right anterior section pedicle was fully exposed and is a technique referred to as the temporary inflow control of the Glisson (TICGL) technique. This slight twist on the procedure allows the surgeon to: (1) perform parenchymal transection under inflow control, thus decreasing bleeding, (2) easily encircle the Glissonian pedicle after it is fully exposed, and (3) perform stapling safely.

For CH, the general procedures including the TICGL technique are similar to RAS with minor differences. The transection plane was 5 mm to the right of the falciform ligament, and the Glissonian pedicles supplying segment 4 were dissected and divided sequentially along the resection plane. Most of the branches only required double clipping. However, if the branch was thought to be too large, a stapler was used. Following the completion of liver transection, the middle hepatic vein was divided using a vascular endoscopic stapler (EndoGIA™ Tan or Echelon™ White).

Study criteria

The main objective of our study was to compare intraoperative outcomes including surgical time, mean blood loss, transfusion rate, and postoperative outcomes including laboratory tests, complication rate, the comprehensive complication index (CCI), length of hospital stay, and mortality between the LLR and OLR groups. Postoperative morbidity was assessed at 90 postoperative days and graded according to Clavien–Dindo classification [13]. The CCI was obtained using a CCI calculator available online (www.assessurgery.com). Postoperative mortality was defined as death within 90 days after liver resection [14]. Definition and grading of bile leakage and liver failure after hepatectomy were described according to the International Study Group of Liver Surgery guidelines [15]. Major vascular proximity was defined as tumor abutting within 1 cm of the expected transection plane of the right anterior portal vein, and within 5 mm of the inferior vena cava and right or middle major hepatic vein in the case of RAS and of the left hepatic vein in the case of CH [16]. Subcapsular tumor was defined as a lesion located <1 cm from the liver edge [17]. In terms of intrahepatic location, ‘deep’ tumors were defined as those with major vascular proximity or that were not subcapsular. Otherwise, the tumor was considered ‘superficial.’

Statistical analysis

Prior to descriptive statistics, Kolmogorov–Smirnov test and Shapiro–Wilk test were used to evaluate normality since the sample size of the LLR group was below 30 patients. Continuous data were presented as median and range. Categorical data were described in numbers and percentages. Statistical analysis was conducted using independent-samples T test or Mann–Whitney test for continuous values and Chi-squared test or Fisher’s exact test for categorical values, especially when expected cell frequencies were below 5. In order to adjust for the difference in patient characteristics between the two groups and to decrease possible selection bias, 1:1 PSM was applied using multiple logistic regression with the nearest-neighbor matching method. Propensity scores were created using baseline characteristics (age, sex, body mass index [BMI], history of previous abdominal surgery, hepatitis B-related etiology, Child–Turcotte–Pugh [CTP] score, preoperative laboratory test, underlying liver cirrhosis, type of surgical procedure, and the American Society of Anesthesiologists [ASA] score), tumor characteristics (malignancy, size, site, major vascular proximity, and subcapsular tumor). A p value of <0.05 was considered statistically significant. Data handling and analysis were performed using the Statistical Package for Social Science for Windows™ release 22.0 (SPSS Inc., Chicago, IL).

Results

Demographic data and our experience with LLR for CLTs

Baseline characteristics of the two groups of patients before and after PSM are summarized in Table 1 according to surgery type. After PSM, all baseline characteristics were similar between the LLR and OLR groups. Histologic diagnosis and surgical type of the LLR group are listed chronologically in Table 2. Ten cases of laparoscopic CH and RAS were performed, respectively. There were no statistically significant differences in mean surgical time between CH and RAS (CH: 411 min; RAS: 406 min, p = 0.897). However, median surgical time in seven cases of superficial tumor was shorter than it was in 13 cases with deep tumor, which represented a statistically significant difference (329 vs 452 min; p = 0.019). The pathologic diagnosis of these 20 cases was hepatocellular carcinoma in 17 cases and benign tumor in three cases. Nine cases occurred in a background of cirrhotic liver. The open conversion rate was 10%. Patient nos. 9 and 11 were converted to open surgery due to injury of the common bile duct and anatomical distortion, respectively.

Table 1 Patient characteristics before and after propensity score matching
Full size table
Table 2 Chronological data of 20 patients who underwent laparoscopic anatomical liver resection for centrally located tumors
Full size table

Operative details and postoperative outcomes of LLR versus OLR groups after PSM

Table 3 summarizes the operative details and postoperative outcomes of propensity-matched groups. Surgical time was significantly longer in the LLR group (median, 388 min; range, 246–661 min) compared to the OLR group (median, 268 min; range, 98–412 min, p < 0.001, Table 3). There were no significant differences between the LLR and OLR groups with regard to the Pringle maneuver (p = 0.176), clamping time (p = 0.502), blood loss (p = 0.746), and surgical margin (p = 0.198). There was no patient with R1 margin which is mostly defined as below 1 mm width in both groups. Two (10%) and one (5%) cases in the LLR and OLR groups, respectively, required blood transfusion. The median duration of postoperative hospitalization in the LLR group was shorter than the OLR group, but this did not represent a statistically significant difference (LLR group: 8 days, range 5–24 days; OLR group: 10 days, range 5–24 days, p = 0.110). There were no significant differences in mortality rate (30 vs 20%; p = 0.716) or mean CCI (4.5 vs 3.9; p = 0.819) between the LLR and OLR group. There was no 90-day postoperative mortality in either group. As shown in Table 4, there was no patient with extended use of patient-controlled analgesia (PCA) in the LLR group. The median duration of PCA (p = 0.002) and intravenous (IV) opioid (p = 0.035) in the LLR group was significantly shorter than in the OLR group.

Table 3 Operative details and postoperative outcomes of propensity-matched patients who underwent LLR and OLR
Full size table
Table 4 Postoperative analgesia of propensity-matched patients who underwent LLR and OLR
Full size table

Discussion

Although the acceptable indications for LLR have been expanded, the location of tumors remains a major challenge when applying totally anatomical LLR. Currently, consensual indications of LLR with respect to the tumor site are lesions located peripherally in liver segments 2–6 [2]. However, CLTs remain a major obstacle in applying a laparoscopic approach because of the technical difficulties and long operation time. We sought to determine whether this procedure was feasible when performed by experienced surgeons.

Several studies have evaluated the feasibility of laparoscopic approach for CLTs [6,7,8,9]. However, they were not comparative studies between laparoscopic and open approaches. Only one study compared laparoscopic and open liver resection [18]. However, various operation types have been reported according to the location of tumor. Comparatively, we selected a homogenous group of patients who underwent CH or RAS for tumors located in segments 4, 5, and 8.

The median surgical time for LLR was longer than that for OLR, but similar to the results of laparoscopic left lateral sectionectomy [19]. Abdominal closure time was usually shorter in LLR because of the small incision. However, parenchymal transection was slower in LLR compared to OLR due to technical difficulties. In left lateral sectionectomy, the time required for parenchymal transection was usually shorter compared to the duration of the operation. Both RAS and CH are the types of liver resection that require the largest area of parenchymal transection. These two parenchymal transection planes are the main cause of the longer total operation time observed for LLR for CLTs [9]. However, our study only reported a preliminary experience of LLR for CLTs. The cohort consisted of all consecutive RASs and CHs since the beginning of the laparoscopic program at our institution. Based on the results of laparoscopic left lateral sectionectomy, it is hypothesized that accumulation of experience can reduce operation time. Moreover, if the Pringle maneuver is applied more frequently, the operation time may be decreased [20]. In our study, the Pringle maneuver was applied less frequently in the LLR, compared to OLR, but this did not represent a statistically significant difference. Blood loss and transfusion rates may also decrease with further experience. However, due to the complexity of the operation, overall liver-related morbidities such as bile leakage will not be easy to reduce.

Intrahepatic location of CLTs can affect operation time. In our study, the median surgical time of LLR for superficial tumors was significantly shorter than that of deep tumors (321 vs 448 min; p = 0.019). If we adopt stricter indications of LLR for CLTs on the basis of intrahepatic location, LLR operation time may be decreased, to a level comparable to OLR.

The TICGL technique provides an easy, safe, and quick way of performing anatomical liver resection. It allows the surgeon to: (1) perform parenchymal transection under inflow control to decrease blood loss, (2) easily encircle the Glissonian pedicle, which is done after the pedicle is fully exposed, and (3) perform stapling safely [12]. This technique has been used since mid-2012 (patient no. 3 in Table 2) at our center. We analyzed the operation time of 8 patients (no. 2 vs no. 3, 5, 6, 7, 10, 15, and 16 in Table 2) according to use of the TICGL technique without the Pringle maneuver. Before applying TICGL, the operation time of patient no. 2 was 448 min. The mean operative time after applying TICGL was 340 min. By applying the TICGL technique routinely, in addition to accumulation of surgical experiences, RAS and/or CH may be completed within 5 h, which may allow more surgeons to perform this technically demanding operation. Additionally, no patients in the LLR group experienced postoperative bile duct stricture, reflecting the safety of this technique.

One important previously reported benefit of LLR is the shorter hospital stay compared to OLR. This result is in concordance with other reported studies on different types of liver resection [14, 21, 22]. However, our propensity-matched study did not demonstrate that the same results could be duplicated in liver resections of CLTs. The median hospital stay in the LLR group was 8 days, which appears to be rather long. Hospital stay is dependent on the medical system of each country. Because of the low admission cost in Korea, most of the patients who underwent major operations were reluctant to be discharged within 7 postoperative days. In addition, routine postoperative CT was taken on day 7 to confirm possible unexpected complications. More recently, our patients have been discharged on postoperative day 5 without routine postoperative CT. In the LLR group of our study, the mean hospital stay of 14 patients who had no morbidities was 8 days. If our recent clinical pathway had been applied earlier, the median hospital stay of the LLR group would have been decreased.

Another benefit of LLR was that the intensity of postoperative pain was less than OLR. Significant reduction of IV analgesia following surgery is important to achieve enhanced recovery after surgery (ERAS) protocol and to reduce length of hospital stay. In our center, IV PCA was routinely given to patients with major liver resection. And if patients request extended use of PCA, we permitted extension use of PCA for two times. No patients in the LLR group had extended use of PCA, and duration of PCA in the LLR group was significantly shorter than in the OLR group in our study.

This study has several limitations. First, the sample size of the LLR group was smaller than that of the OLR group, which was the main drawback of this study. More studies with a larger sample size are necessary to confirm our results. In addition, our study focused on short-term outcomes in order to evaluate the operative feasibility. Further studies are needed to assess long-term outcomes and to demonstrate the oncological safety of the laparoscopic approach, including overall and disease-free survival in patients with malignant tumors. Moreover, LLR was performed by a single surgeon, whereas OLR was performed by several surgeons. This limitation can act as selection bias of postoperative outcomes.

Conclusion

According to a difficulty scoring system for LLR [23], RAS and CH should be performed by laparoscopic surgeons with extensive experience who are able to successfully perform easy and intermediate difficulty surgeries without much difficulty. Therefore, our findings may not be generalizable to all laparoscopic liver surgeons, but LLR as difficult as RAS and CH can be performed with comparable results to OLR by experienced surgeons.