Laparoscopic hand-assisted liver resection for tumours in the left lateral section.

CONTEXT: The role of the laparoscopic left lateral sectionectomy (LLLS) is debatable, and Level-1 data are lacking. AIMS: The aim of the study is to evaluate the feasibility and safety of this approach. SETTINGS AND
DESIGN: This was a retrospective study. SUBJECTS AND METHODS: From 2007 to 2014, patients undergoing LLLS were identified from two institutions. STATISTICAL ANALYSIS USED: Continuous variables were compared between groups with Student's t-test or Mann-Whitney test, as appropriate by type of distribution. Categorical variables were compared with Chi-square or Fisher's exact test, depending on the number of observations.
RESULTS: Thirty-eight patients were included in the study. The mean age was 63.5 + 13 years (range, 31-89), and the mean number of tumours was 1.7 + 1.5. Eleven (29%) patients underwent LLS combined with an additional liver resection (combined resections group). The mean duration of the operation and the mean estimated blood loss were significantly decreased in the LLS group compared to the combined resection group (101 + 71 min vs. 208 + 98 min and 216 + 217 ml vs. 450 + 223 ml;P < 0.05 for both, respectively). The major complications rate was 8% and no mortality occurred.
CONCLUSIONS: In a subset of carefully selected cases, LLLS may provide the benefits of laparoscopy. This does not appear to compromise perioperative morbidity rates. We believe that this approach may serve as a training platform for surgical trainees.

INTRODUCTION

Laparoscopic surgery became a common approach for many abdominal procedures in the 1990s after case studies, followed by several randomised controlled trials, showed their efficacy. Benefits such as a shorter length of hospital stay and recovery, as well as less pain, fuelled the rapid adoption of these minimally invasive procedures.[12] Laparoscopic liver resections were first reported in 1993 and can be performed in different ways which vary in their complexity and difficulty.[1] Three categories of surgical approaches have been defined; pure laparoscopy, hand-assisted laparoscopy and hybrid technique.[1] In the pure laparoscopic procedure, the entire resection is completed through laparoscopic ports, whereas in hand-assisted laparoscopic surgery (HALS), a hand port is electively placed to facilitate the procedure. Finally, the hybrid technique is defined as a procedure which starts as a pure laparoscopic or HALS procedure, but the resection is carried out through a mini-laparotomy incision.[1]

In 2008, a panel of international experts in 'The Louisville Statement' proposed that 'acceptable indications for laparoscopic liver resection are patients with solitary lesions, 5 cm or less, located in lever Segments II–VI' and that 'the laparoscopic approach to left lateral sectionectomy (LLS) should be considered standard practice.'[1] Laparoscopic LLS (LLLS) involves the removal of Segments II and III and thus has encouraged hepatobiliary surgeons to quickly focus on this procedure because of its easily accessible peripheral location. Furthermore, the thin parenchyma along the ligamentum venosum groove, the fact that that hilar biliary dissection is usually not needed, and that the left hepatic vein can be easily controlled with a vascular stapler, makes LLLS a first choice procedure for both benign and malignant tumors.[34] However, the role of the laparoscopic approach to LLS, outside specialised centres, is still debatable, and Level-1 data are lacking.

Therefore, the aim of this study was to evaluate the feasibility and safety of the HALS approach to LLS from reporting our experience in patients with benign and malignant tumours.

SUBJECTS AND METHODS

All patients who underwent laparoscopic liver resection of LLS, from August 2007 to December 2014, were identified from a prospectively maintained surgical database of the Department of Surgery 'A' at Carmel Medical Center and the Division of Surgery at Rabin Medical Center. This study was conducted in accordance with the Declaration of Helsinki and was approved by the Carmel Medical Center Ethics Committee (0020-16-CMC) and the Rabin Medical Center Ethics Committee (0066-16-RMC).

The indications for any liver resection at our institution, either open or laparoscopic, were determined during a weekly multidisciplinary conference. For resection of benign lesions, the surgical indications were the presence of symptoms, a diagnosis of hepatic adenoma or echinococcal cyst and an uncertain diagnosis of cystadenoma. The surgical indications for resection of malignant lesion included peripheral cholangiocarcinoma and hepatocellular carcinoma with child Class A and without a sign of portal hypertension. The indications for metastases were the occurrence of a liver tumour during follow-up or at diagnosis of patients with colorectal and non-colorectal cancer. Patients were not considered for surgical resection if their history included decompensated cirrhosis and cardiac or respiratory failure. The pre-operative workup included blood tests, tumour markers, imaging modalities (computed tomography [CT], positron emission CT and magnetic resonance imaging) and characterisation of the specific tumour (number, location, size and relation to intrahepatic vascular or biliary structures). The patients underwent standard evaluation for major surgery by an anaesthesiologist. All the patients provided informed consent. All procedures were performed under the direction of the same attending surgeon (R.H.)

Surgical technique

LLLS was performed with the patient in a supine 30° reversed Trendelenburg position and with split legs. In brief, three trocars (two 12 mm and one 5 mm) were inserted in the upper midline abdomen, and a hand-assisted device (at the beginning, we used LapDisc, Ethicon, Cincinnati, OH, USA, but later, we switched to GelPort, Applied Medical, CA, USA) was placed in the right abdomen [Figure 1]. We used a supraumbilical cutdown to establish pneumoperitoneum, with a 12-mm port; however, in patients with previous abdominal surgery, we began with a right abdominal incision (7–8 cm). The surgeon dissected the adhesions and then inserted the hand port and inserted a 12-mm supraumbilical trocar into an adhesion-free area. We find such port placements to provide satisfactory ergonomics and allow adequate exposure. The pneumoperitoneum was generated with CO2 at a pressure of 12–15 mmHg, and visual exploration of the liver and the abdominal cavity was conducted with a 30° laparoscope. The abdominal cavity was explored manually, and intra-abdominal sonography of the liver was performed. Ultrasound was used to confirm the extent of tumour, determine the number of lesions, identify potentially hazardous intrahepatic vascular or biliary structures and demarcate surgical tumour resection margins. The round and falciform ligaments were divided from the anterior abdominal wall towards the inferior vena cava, and the left triangular ligament was divided with LigaSure™ (LigaSure 5 mm; Valleylab, Boulder, CO, USA) to free the left lobe. The lateral wall of the left hepatic vein was exposed, and care was taken to avoid injury to the left hepatic and phrenic veins. The posterior surface of Segments II–III was exposed, and the lesser omentum was checked for the presence of an aberrant left hepatic artery. The lesser omentum was divided by LigaSure™. The liver was transected on a line just left of the falciform ligament using LigaSure™ until the portal pedicles of Segments II–III were exposed. The portal pedicle of Segments II–III was divided with Endo GIA™ staplers (vascular cartridge, Endo GIA™, Covidien, Norwalk, CT, USA) applied two or three times. The transaction was continued until the left hepatic vein was reached after which it was stapled with a small amount of surrounding liver tissue. The specimen was extracted through the hand-assisted device. The argon beam coagulator was used for haemostasis. The surgical field was irrigated and checked for bleeding or bile leakage, and residual fluid was removed by suction. An abdominal drain was usually placed through a 5-mm port site, and the pneumoperitoneum was vented. All specimens were sent fresh for pathologic examination to measure the surgical margins.

Figure 1

Trocar location and surgeon's position. Two 12-mm trocars were inserted in the umbilicus (camera) and in the upper midline. Another 5-mm trocar was inserted as indicated. A hand-assisted device was placed in a horizontal incision (7–8 cm)

Statistical analysis

Descriptive and comparative statistical analyses were performed using the Statistical Software for the Social Sciences version 21(IBM Corp., Armonk, NY). Continuous variables were compared between groups with Student's t-test or Mann–Whitney test, as appropriate by type of distribution. Categorical variables were compared with Chi-square or Fisher's exact test, depending on the number of observations. A P < 0.05 was considered statistically significant.

RESULTS

Patients' characteristics

During the study period, 38 consecutive patients who met our study criteria underwent hand-assisted LLLS. Table 1 summarises the characteristics of the patients and their tumours. There were 18 males and 20 females with a mean age of 63.5 ± 13 years (range, 31–89). Previous abdominal surgery had been performed in 16 (42%) patients, including laparoscopic colorectal resection in eight patients, open colorectal resection in eight patients and two of these patients had undergone open hepatectomy. The mean tumour size was 3.6 ± 2.9 cm, and the mean number of tumours was 1.7 ± 1.5.

Table 1

Patients and tumors characteristics of the entire study cohort (n=38)

Characteristics	All patients (n=38)
Gender
M	18 (47%)
F	20 (53%)
Age, years	63.5+13 (31-89)
Previous abdominal surgery	16 (42%)
Indication
benign	13 (34%)
Adenoma	3 (7.9%)
Hemangioma (sclerosing)	2 (5.2%)
Echinococcus	5 (13.2%)
Cystadenoma	2 (5.2%)
Pseudo-tumor	1 (2.6%)
Malignant	25 (66%)
CRLM	22 (58%)
Cholangiocarcinoma	3 (7.9%)
Lesion
Size, cm	3.6+2.86 (0.6-10)
Number	1.7+1.5 (1-6)

Continuous variables are expressed as mean±SD (range); categorical variables are expressed as n (%). CRLM, colorectal liver metastases

Thirteen (34%) patients had benign disease while 25 (66%) had malignant lesions [Table 1]. Regarding the benign lesions, resections were performed with diagnostic intent in eight patients while five patients underwent resection for symptomatic hydatid cysts. All patients with benign disease had single nodule, with a mean diameter of 6.6 ± 3.1 cm (range 2.5–14). The malignant tumours included colorectal liver metastasis (CRLM; 22 patients [88%]) and intrahepatic cholangiocarcinoma (3 patients [12%]). Table 2 summarises the characteristics of the patients with CRLMs (n = 22) and their tumours. Nine of those (41%) were metachronous while 13 (59%) were synchronous. Of these 13 patients, six were treated with the liver- first approach. The mean number of metastases per patient was 2 ± 1.7 (range, 1–6) with a mean diameter of 2.7 ± 2.1 (0.6–9) cm. Neoadjuvant systemic therapy (FOLFOX or FOLFIRI plus bevacizumab or cetuximab) was administered to 14 (64%) patients with CRLM, leading to a partial response in 11 patients and stable disease in three patients.

Table 2

Patients and tumors characteristics of the patients with colorectal liver metastases (n=22)

Characteristics	Patients with CRLM (n=22)
Stage of the primary tumor at presentation
IIIA	3 (13.6%)
IIIB	6 (27.2%)
VI	13 (59.2%)
Liver First approach	6 (27.2%)
CEA, mcg/L	21.4+19 (1.3-68)
PET-CT performed	22 (100%)
Number of metastases	2+1.7 (1-6)
Size, cm	2.75+2.08 (0.6-9)
Neoadjuvant
Yes	14 (64%)
No	8 (36%)

Continuous variables are expressed as mean±SD (range); categorical variables are expressed as n (%). CRLM, colorectal liver metastases; CEA, carcinoembryonic antigen; PET-CT, Positron emission tomography-computed tomography. The classification of stage was defined according to the National Comprehensive Cancer Network definitions.[27]

Intraoperative results

All 38 patients underwent LLS resection, while 11 (29%) patients underwent LLS combined with an additional non-anatomical liver resection; details about the type of liver resections are summarised in Table 3. Associated non-liver-related surgical procedures were performed in 4 (10.4%) of 38 patients (cholecystectomy, right colectomy and anterior resection). The procedure was completed laparoscopically in 37 (97%) patients. An intraoperative adverse event occurred in one patient (2.6%) necessitating conversion to laparotomy due to bleeding from the phrenic vein during liver mobilisation that required a blood transfusion. This event occurred at the beginning of our experience, and it reflects our safe and low threshold for conversion. The mean duration of the operation was 130 ± 61 (45–288) min but was significantly shorter in the LLS group compared to the combined resection group (101 ± 71 vs. 208 ± 98 min; P = 0.0001). The mean estimated blood loss was 270 ± 216 ml, and it was significantly decreased in the LLS group compared to the combined resection group (216 ± 217 vs. 450 ± 223 ml; P = 0.016) with 16% of the patients receiving perioperative blood transfusions.

Table 3

Perioperative outcomes of all patients

Characteristics	n (%)
Left lateral sectionectomy (LLS)	38 (100%)
Combined Non-anatomical resection	11 (29%)
Segment 4	3
Segment 4/8	3
Segment 5	2
Segment 6	2
Segment 7	1
Cholecystectomy	2 (5.2%)
Colectomy	2 (5.2%)
Operative time, min
All patients	130+61 (45-288)
LLS only	101+71 (45-180)
LLS combined resections	208+98 (140-288)
Estimated blood loss, ml
All patients	270+216
LLS only	216+217
LLS combined resections	450+223
Administration of blood transfusion	6 (15.7%)
Conversion to open	1 (2.6%)
Mean length of hospital stay, days	4.36+2.75 (2-17)
Postoperative complications
Overall	8 (21%)
Grade II	5 (13.2%)
Grade III	2 (5.2%)
Grade IV	1 (2.6%)
Bile Leak	2 (5.2%)
Margins
Distance (cm)	1.05+0.85 (0-3.5)
R1 resection (margin distance <1 mm), n (%)	2 (5.2%)

Continuous variables are expressed as mean±SD (range); categorical variables are expressed as n (%)

Post-operative results

The post-operative outcomes are summarised in Table 3. According to the Clavien–Dindo classification, 8 (21%) patients developed at least one post-operative complication. Two patients (5.2%) had a biliary leak from the hepatic transection surface that was resolved without intervention. Grade III complications were recorded in two patients. One was a cirrhotic patient (without portal hypertension) who developed an upper gastrointestinal bleeding that necessitated endoscopic clipping of a visible vessel in a duodenal ulcer. Later, he developed liver failure that required medical treatment and repeated abdominal taps. The other patient developed a non-ST elevation myocardial infarction and underwent cardiac catheterisation. A Grade IV complication was observed in one patient who was diagnosed with partially resolved cerebral stroke, and the CT scan revealed cerebral meningioma that was removed surgically. The mean hospital stay was 4.4 ± 2.7 days (range, 2–17). The mean width of the surgical margins was 1.05 ± 0.85 (range, 0–3.5) cm, while two patients had an R1 resection that was associated with a combined non-anatomic resection of CRLM.

DISCUSSION

While minimally invasive surgery is gaining wide acceptance in many surgical subspecialties,[5] the implementation of laparoscopic liver surgery is relatively limited. A large series (n = 4152 liver resections) from the Memorial Sloan Kettering group[6] demonstrated that the percentage of minimally invasive cases increased from 0% in the first era (1993–1999) to 5% in the third era (2007–2012). This limited acceptance of the laparoscopic approach to liver surgery is also reflected by the early termination of a recent prospective trial owing to slow patient accrual. This trial[7] was initially designed to compare LLLS with open LLS. Therefore, in the current study, we report our experience with the implementation of the laparoscopic approach to LLS as we believe that these data may help disseminate the laparoscopic approach to LLS.

This study mostly demonstrates the safety and feasibility of the laparoscopic approach to LLS. The feasibility was determined by the low conversion rate (2.6%) and the reasonable mean operative time of 130 min. These numbers are comparable to other reports in the literature. The group from Singapore General Hospital[8] demonstrated a similar conversion rate of 9%, and their median operative time was 167 min. Other studies[3910111213] observed operative times for this procedure ranging from 140 to 260 min, with a blood loss of 50–200 ml, whereas in this study, the estimated blood loss was 270 ml (and in the LLS-only group, it was 216 ml). The safety of the procedure was demonstrated by the acceptable overall complication rate and the zero-mortality rate. Others have published morbidity rates in the range of 14%–20%,[81113] whereas in this study, an overall morbidity rate of 21% was recorded. The morbidity in this study mainly consisted of minor complications (63%). Moreover, our mean length of hospital stay is consistent with other studies in the literature that reported a range between 2 and 7 days.[3910111213] It is notable that the safety and feasibility of the procedure did not compromise the early oncological outcomes, as evidenced by the R1 resection rate of 5%, which is in line with other studies that investigated either open or laparoscopic approaches to LLS.[814]

The substantial driving forces for the dissemination of the laparoscopic approach to LLS are the potential patients' benefits and the advantages regarding surgical training. Despite the lack of Level-1 data, previous reports have observed shorter length of hospital stay associated with the laparoscopic approach to LLS.[38151617] Of note, the shorter length of hospital stay did not compromise other perioperative outcomes such as complication rates. The value of a structured training programme to the education of surgical trainees in laparoscopic procedures has been established, and at present, many training programmes, worldwide, are required to include the 'fundamentals of laparoscopic surgery' course in their curriculum.[1819] This educational added value is more pronounced in laparoscopic liver surgery, which is challenging, as it requires dissection of the portal pedicles, transection and mobilisation of the liver parenchyma. We believe that LLLS can serve as a gradual training step for residents and fellows since it requires laparoscopic skills (e.g. mobilisation of the liver, dissection of the inflow/outflow vessels and transection of the parenchyma) that later can be implemented in more complex laparoscopic liver resections. In addition, the laparoscopic approach to LLS has been previously shown to be cost-effective and associated with lower costs compared to the open approach.[20]

As a retrospective cohort analysis that is mainly descriptive, this study has inherent limitations such as selection bias and a limited sample size. Ideally, the question regarding the benefits of the laparoscopic approach to LLS would be answered in a randomised prospective trial. However, given the relatively low incidence, the strongly held preferences of the surgical groups that manage these patients and the public demand, such a trial is less likely to be accomplished. The ORANGE II trial was designed to address this question, but unfortunately, it was stopped prematurely owing to slow accrual.[7] Therefore, it is our premise that the main strength of the current study is that our data may help consolidate the role of the laparoscopic approach to LLS.

CONCLUSIONS

Laparoscopic approach to LLS is feasible and safe and in a subset of carefully selected cases may provide the benefits of laparoscopy such as a relatively short length of hospital stay. This does not appear to compromise the perioperative morbidity rates and the resection margin status. At the same time, we believe that the laparoscopic approach may serve as a training platform for surgical trainees. Prospective multi-institutional registries are needed to validate these data.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.