Open and laparoscopic resection of hepatocellular adenoma: trends over 23 years at a specialist hepatobiliary unit

Abstract

Background

Hepatocellular adenoma (HCA) is a rare benign liver epithelial tumour that can require surgery. This retrospective study reports a 23-year experience of open and laparoscopic resections for HCA.

Methods

Patients with a histological diagnosis of HCA were included in this analysis. Surgical resection was performed in all symptomatic patients and in those with lesions measuring > 5 cm.

Results

Between 1989 and 2012, 62 patients, 59 of whom were female, underwent surgery for HCA (26 by open surgery and 36 by laparoscopic surgery). Overall, 96.6% of female patients had a history of contraceptive use; 54.8% of patients presented with abdominal pain and 11.2% with haemorrhage; the remaining patients were asymptomatic. Patients who underwent laparoscopy had smaller lesions (mean ± standard deviation diameter: 68.3 ± 35.2 mm versus 91.9 ± 42.5 mm; P = 0.022). Operatively, laparoscopic and open liver resection did not differ except in the number of pedicle clamps, which was significantly lower in the laparoscopic group (27.8% versus 57.7% of patients; P = 0.008). Postoperative variables did not differ between the groups. Mortality was nil. Two surgical specimens were classified as HCA/borderline hepatocellular carcinoma. At the 3-year follow-up, all patients were alive with no recurrence of HCA.

Conclusions

Open and laparoscopic liver resections are both safe and feasible approaches for the surgical management of HCA. However, laparoscopic liver resections may be limited by lesion size and location and require advanced surgical skills.

Introduction

Although rare, hepatocellular adenoma (HCA) is the most common benign epithelial tumour of the liver.¹ First described by Edmondson in 1958,² the incidence of HCA has progressively increased since the 1970s to reach approximately three cases per million population per year, of which 90% occur in women.³ This trend may be explained by the wide use of oral oestrogen-based contraceptives among young women, combined with the increasing quality and number of liver imaging techniques.¹ Recently, distinct subtypes of HCA, based on molecular alterations, have been identified.⁴ These types are: (i) H-HCA, in which inactivating mutations of the gene are encoded by hepatocyte nuclear factor-1α (HNF-1α); (ii) β-HCA, which shows a β-catenin-activating mutation; (iii) I-HCA, which represents inflammatory HCA, and (iv) U-HCA, which represents unclassified HCA.⁴ Either H-HCA or I-HCA is found in 80% of patients, whereas β-HCA is found in approximately 10–15%. The I-HCA subtype accounts for 40–50% of all adenomas and has been demonstrated to be associated with alcohol abuse and obesity.⁵ Recently, new risk factors for HCA have emerged, including metabolic syndrome,⁵ use of clomiphene, methyltestosterone or danazol, Klinefelter's syndrome, types Ia, Ib and III glycogen storage disease,⁶ and familial adenomatous polyposis.¹ Hepatocellular adenoma is usually a solitary tumour of variable size, but in 10–24% of patients it presents with multiple nodules; the presence of more than 10 nodules is defined as liver adenomatosis.^7–9 Symptoms can be related to tumour size or to specific complications. Size-related symptoms are usually non-specific, such as right upper quadrant tenderness, dyspepsia or abdominal mass presentation. A specific and potentially life-threatening complication of HCA is tumour rupture with haemorrhage, observed in 20–40% of patients.^10–12 Malignant transformation affects up to 10% of patients and does not distinguish between isolated HCA and liver adenomatosis.¹

Diagnosis is based on multi-phase spiral computed tomography (CT) and multi-phase contrast-enhanced magnetic resonance imaging (MRI), but a definitive diagnosis is possible only after a histopathological examination.^13,14 Biopsy can be useful in subtyping HCA, but is not always recommended for diagnosis because it can induce bleeding and tumour dissemination.^15,16 Moreover, biopsy cannot completely rule out the occurrence of malignant transformation.¹⁷ Although there has been no consensus on the management of HCA, surgical resection is generally indicated for tumours larger than 5 cm, in symptomatic patients, or when there is strong concern regarding malignant transformation.^18,19 For liver adenomatosis, liver transplantation is the only cure because of the frequent bilobar distribution of the lesions.²⁰

The application of laparoscopic liver surgery in benign diseases was first reported in 1991.²¹ Since then, many authors have presented series of patients submitted to laparoscopic resection for benign liver tumours.²² Although there has been a lack of international guidelines, the laparoscopic approach has been used mainly in patients with peripheral lesions located in the left lateral (segments II and III) and anterior (segments IVb, V and VI) liver, which have required limited resections.^23–25 However, major laparoscopic liver resections (more than three segments) have also been reported with successful outcomes when performed by experienced surgeons in specialized centres.^23,26–28

The few studies in the literature that have reported the laparoscopic management of liver adenomas have mostly described heterogeneous populations of patients presenting with different benign liver diseases requiring surgery.^22,23,29 The aim of the present study was to report the 23-year experience of a specialist hepatobiliary unit in the surgical resection of HCA, managed by both open and laparoscopic approaches. The indications, outcomes, complications and advantages of both techniques are discussed.

Materials and methods

This was a retrospective study of a prospectively maintained database on HCAs. Demographic, preoperative, operative and postoperative data for patients diagnosed with HCA and submitted to surgery between 1989 and 2012 were retrieved and analysed after institutional board approval.

Patients were included in the study if the diagnosis of HCA had been confirmed by a histopathological analysis of the surgical specimen. All resected HCAs were classified using common histological and immunohistochemical criteria.³⁰

Surgical resection was performed in all symptomatic patients and in those with lesions measuring >5 cm. Patients with polyadenomatosis (≥10 nodules) were excluded.

Prior to surgery, all patients underwent imaging including abdominal ultrasound, multi-phase spiral CT and multi-phase contrast-enhanced MRI. Liver biopsy was performed selectively when preoperative imaging was unclear. Each patient was individually evaluated in a multidisciplinary meeting during which the best surgical approach and technique were collegially discussed and chosen.

The surgeries were performed using either an open or a laparoscopic approach on a patient-by-patient basis. The main clinical criteria used to select the surgical technique were lesion size and location (i.e. proximity to major vascular structures) and, consequently, the extent of hepatic tissue (i.e. segments) that required to be resected. Surgical decisions were based on imaging studies. Both elective and emergency settings were included. Emergency surgery was performed when tumour rupture was identified, based on clinical symptoms and radiological findings. The surgeries included all types of liver resection (e.g. major resection, bisegmentectomy, segmentectomy, wedge resection), performed according to the location and size of the tumour. The Pringle manoeuvre was not used as a routine procedure.

The same team has previously described laparoscopic liver resection performed using the standard nomenclature.^22,31 Major hepatectomy was defined as resection of three or more liver segments, whereas minor hepatectomy included segmentectomy, bisegmentectomy and wedge procedures.^22,32 For open liver resection, a right subcostal incision, with or without an upper midline extension, was performed. Crush and clamp techniques, oversewing of biliary ducts, monopolar and bipolar cautery, and, more recently, the Cavitron ultrasonic surgical aspirator (CUSA; Tyco Healthcare, Covidien, Inc., Mansfield, MA, USA) and vascular stapler were used. For laparoscopic liver resections, liver parenchyma transection was performed using a combination of the LigaSure (Valleylab, Covidien, Inc.), the Harmonic Scalpel (Ethicon EndoSurgery, Inc., Cincinnati, OH, USA), the Endo GIA vascular stapler (US Surgical Corp., Norwalk, CT, USA), and bipolar cautery. In all patients, prior to surgical transection, intraoperative liver ultrasound was performed to guide and limit the resection margins.

After liver resection, patients were included in a 1-year and 3-year follow-up protocol. The following variables were analysed: patient demographics [e.g. age, sex, body mass index (BMI), medications taken]; imaging studies; laboratory tests; biopsies; surgical technique (i.e. open or laparoscopic); surgical complications (e.g. duration, conversion, blood loss); tumour characteristics (e.g. number, size, location), and short- and longterm postoperative variables (e.g. Clavien–Dindo classification, duration of hospital stay, incisional hernia, recurrence after resection). Moreover, when available, histological slides were re-reviewed for this analysis and HCAs were classified as H-HCA, β-HCA, I-HCA, U-HCA or HCA/borderline hepatocellular carcinoma (HCC) according to the widely recognized system of molecular classification.^4,5

According to data distribution, parametric (Fisher's exact test) and non-parametric (Mann–Whitney U-test) statistics were used to describe the study population and to compare patients submitted to laparoscopic versus open liver resection for HCA. Regression analyses were performed to control the effects of covariables on clinical outcomes. A P-value of ≤0.05 was considered to indicate statistical significance. IBM spss Statistics for Macintosh Version 20.0 (IBM Corp., Armonk, NY, USA) and sas Version 9.2 (SAS Institute, Inc., Cary, NC, USA) were used for statistical analyses.

Results

Of the 1534 patients in whom liver resections were performed between 1989 and 2012, a total of 62 underwent surgical resection for HCA. Only five patients presented with polyadenomatosis; these patients were excluded from this analysis.

Patient demographics by operative approach are shown in Table 1. Overall, 96.6% of female patients had taken oestrogen-based oral contraceptive pills at some time during their lives. At the time of HCA diagnosis, 23.7% of the women were in menopause. None had familial histories of benign or malign liver tumours. Five (8.1%) patients had diabetes, five (8.1%) patients had arterial hypertension, one (1.6%) patient presented with coronary disease, one (1.6%) patient had chronic kidney failure, and three (4.8%) patients were diagnosed with dyslipidaemia. Overall, the demographic and clinical variables did not differ between the group operated on by laparoscopy and the group operated on by laparotomy. Conversely, the two groups showed significant differences in lesion size and type of liver resection; these variables were taken into account in the regression analysis (Table 1).

Table 1.

Demographics, presentation and treatment in 62 patients submitted to liver resection (LR) for hepatocellular adenoma

	Laparoscopic LR (n = 36)	Open LR (n = 26)	P-value
Age, years, mean ± SD	35.4 ± 9.3	37.7 ± 8.3	0.382
Female, n	35	24	0.374
BMI, kg/m2, mean ± SD	19.8 ± 4.7	22. ± 5.6	0.101
More than one comorbidity, n	10	5	0.375
Use of oral contraceptives, n	35	22	0.072
Increased AFP, >10 ng/ml, n	0	0	0.996
Child–Pugh class A status, n	36	26	0.994
Fatty liver disease, n	3	2	0.882
Bleeding adenoma, n	3	4	0.383
Symptomatic patients, n	20	14	0.894
Number of lesions, median (range)	2 (1–6)	1.5 (1–4)	0.369
Diameter of lesion, mm, median (range)	50.5 (25–130)	95.0 (20–180)	0.030
Emergency hepatectomy, n	0	1	0.424
Type of resection, n			0.044
Major	9	14
Bisegmentectomy	12	3
Segmentectomy	3	1
Wedge	12	8

P-values in bold indicate statistical significance (P ≤ 0.05).

SD, standard deviation; BMI, body mass index; AFP, α-fetoprotein.

Of the seven patients presenting with tumour rupture and haemorrhage, six were first treated with selective arterial embolization and then submitted to surgery in elective settings (three by laparoscopic and three by open approach), whereas one patient, in whom haemodynamic stability was not achieved, underwent surgery in an emergency setting (open liver resection).

Operative factors by type of intervention are shown in Table 2. In the laparoscopic group, three interventions were converted to laparotomy as a result of haemorrhage (caused by endovascular stapler malfunction, difficult exposition, and extreme parenchyma fragility during transection). All three patients were undergoing right hepatectomy.

Table 2.

Operative variables in 62 patients submitted to liver resection (LR) for hepatocellular adenoma

	Laparoscopic LR (n = 36)	Open LR (n = 26)	P-value	Adjusted P-valuea
Duration of hepatectomy, min, median (range)	165 (90–390)	180 (90–320)	0.829	0.214
Pringle manoeuvre, n	10	15	0.021	0.009
Conversion, n	3			NA
Blood loss, ml, median (range)	200 (50–1600)	250 (50–2500)	0.428	0.883
RBC units transfused, n	0	2	0.175	NA

P-values in bold indicate statistical significance (P ≤ 0.05).

^aRegression analysis for type of liver resection (major versus minor) and tumour size.

RBC, red blood cells; NA, not applicable.

Overall, 22 specimens were not available for re-evaluation or re-classification. Of the remaining 40 that were retrieved, HCAs were classified as follows: I-HCA, n = 19; H-HCA, n = 9; U-HCA, n = 9; β-HCA, n = 1, and HCA/borderline HCC, n = 2. The only two patients with HCA/borderline HCC lesions had α-fetoprotein levels within the normal limits and were asymptomatic (no bleeding). They were both managed in elective surgical settings using a laparoscopic approach in one and laparotomy in the other.

Postoperative variables did not differ between the two groups (Table 3). Of the three patients in the laparoscopic liver resection group in whom postoperative complications occurred (within 1 month), two were converted to laparotomy during the surgical intervention (both major liver resections). The overall mortality was nil.

Table 3.

Postoperative variables in 62 patients submitted to liver resection (LR) for hepatocellular adenoma

	Laparoscopic LR (n = 36)	Open LR (n = 26)	P-value	Adjusted P-valuea
Duration of hospitalization, days, median (range)	7.5 (3–26)	6.5 (2–28)	0.528	0.955
Postoperative complications, n	3	5	0.206	0.152
Haemorrhage	1	2
Liver failure	0	2
Biliary leak	1b	1
Sepsis	1b	0
Complications Clavien–Dindo Grade III, IV, n	3	4	0.439	1

^aRegression analysis for type of liver resection (major versus minor) and tumour size.

^bConverted to laparotomy.

At the 1-year follow-up (data available for all 62 patients), all patients were alive without recurrence of HCA. Three incisional hernias were observed in the open liver resection group (11.5%) and one (2.8%) in the laparoscopic group (P = 0.346). At the 3-year follow-up (data available for 53 patients only), all patients were alive with no recurrence of HCA.

Discussion

The present study describes the clinical outcomes of open and laparoscopic liver resection performed in 62 patients with HCA. Over the last two decades, the incidence of HCA appears to have increased, but, by contrast with findings reported in a recent review,³³ surgical need does not seem to have been influenced drastically by improvements in the diagnosis of HCA or by the introduction of laparoscopy.²² Indeed, the availability of a different surgical approach cannot change the indications for surgery. Today, in the management of benign liver pathologies, it is widely accepted that surgery should be reserved for selected patients with symptomatic tumours or tumours larger than 5 cm in size, or in whom there is diagnostic doubt regarding the presence of HCC. In fact, it has been shown that the great majority (96%) of HCAs complicated by malignant transformation measure > 5 cm.¹⁸ For asymptomatic HCAs or lesions of <5 cm, a ‘wait-and-see’ approach can be applied and the discontinuation of oral contraceptives is recommended.^34,35 Moreover, less invasive approaches, such as radiofrequency ablation and arterial embolization, may be favourable although their efficacy is still under evaluation.^36,37

In liver resections for benign liver tumours, morbidity rates have ranged from 10% to 27% and mortality rates from 0% to 3%.^{22,23,29,38,39} From this perspective, the correct selection of candidate patients for surgery is essential,^22,40 but the choice between the laparoscopic and open surgical approaches is still mainly empirical. According to some authors, laparoscopic liver resection is more favourably indicated for isolated lesions, lesions of ≤5 cm, and lesions located in peripheral liver segments II–VI.²⁴ However, successful major laparoscopic hepatectomies (e.g. right hepatectomy and extended liver resection) have been reported in specialized high-volume liver and laparoscopic surgical centres.^19,23,26,28

According to the present results, despite the different tumour sizes and consequent types of liver resection, no differences in clinical outcomes or complication rates were observed between open and laparoscopic liver resections for HCA. The only difference was observed operatively in terms of the frequency of the Pringle manoeuvre, which was more frequently performed during open surgery.

It appears that open and laparoscopic liver resections are equivalently successful approaches for the surgical management of HCA. However, whereas in minor hepatectomy tumour size may not be a limiting factor for laparoscopy,^23,41 tumour size and proximity to hepatic vessels remain the principal criteria guiding surgical decisions in major hepatectomy.²⁹ In this specialist hepatobiliary unit, major hepatectomies were performed using laparoscopy for lesions of up to 13 cm with successful clinical outcomes. The only three laparoscopic conversions (4.8%) in the present series occurred in the first major hepatectomies to be performed via laparoscopy and thus may reflect the effect of a surgical learning curve. However, laparoscopic major hepatectomy remains a complex and challenging procedure.

In the present study, seven patients (11.3%) presented with bleeding HCA, representing a lower percentage than previously reported for haemorrhagic HCA in the literature (assessed at approximately 20–30%).^23,29 All bleeding adenomas were managed by selective arterial embolization to achieve haemodynamic stability before the patient was submitted to elective laparoscopic or open surgery. Only one patient underwent emergency surgery; arterial embolization was not possible as a result of this patient's extremely unstable condition. This emergency surgery was performed using an open surgery approach, but other previous patient series have reported the successful management of haemorrhagic HCA by laparoscopy.¹⁹

Of the 62 patients with HCA submitted to surgery, only two (3.2%) presented with a borderline malignant transformation. This finding supports the unknown but certainly low rate of malignancy in this liver tumour,^18,29,34 as well as the clear advantage of surgical resection in preventing disease complications and evolution. To date, surgical indications have been based mainly on the morphological and clinical characteristics of HCAs, but it is possible that in the near future the HCA genetic profile may be considered within the assessment of need for surgery and patient risk because the H-HCA subtype is known to pose a greater risk for bleeding and the β-catenin mutation HCA is more likely to develop malignant transformation.^4,42 However, the innate limitations of hepatic biopsy are associated with a low sensitivity for the detection of HCA malignancy and are aggravated by an increased risk, albeit small, for major haemorrhage.^15,43 At present, surgical resection is the only way to ensure the definitive diagnosis, characterization and treatment of HCAs.

In conclusion, in the absence of an international consensus and guidelines, each centre may define its own protocol for the indications for and type of resection, based on its own surgical experience and on the morphological characteristics of the HCA. It appears that for minor liver resections, the laparoscopic approach is safe and feasible, and should be routinely performed. Major hepatectomy, however, requires advanced surgical skills if it is to be performed through laparoscopy and the feasibility of this approach appears to be limited by lesion size and location.

Conflicts of interest

None declared.