Introduction

Benign liver tumors (BLTs) comprise a group of lesions with different characteristics in imaging tests, clinical course and treatment options. They are classified in solid and cystic tumors. The most common solid BLT include hemangioma, focal nodular hyperplasia (FNH) and adenoma, while simple hepatic cyst is the most common cystic lesion. Most tumors are asymptomatic, and their diagnosis is incidental^1,2.

While conservative treatment is indicated in most cases, the main indications for surgical treatment are the presence of symptoms, risk of intratumoral bleeding, rupture, malignant transformation or the impossibility of ruling out a malignant tumor (diagnostic uncertainty)^3-5.

The advantages of laparospcopic liver resection (LLR) as less intraoperative bleeding, fewer postoperative complications, less need for pain-killers, shorter time to initiate oral feeding and shorter length of hospital stay, in addition to better cosmetic results, have led these techniques to gain more ground for the management of BLTs^6,7. The evidence published supports the use and safety of LLR in the surgical treatment of BLT which may currently be considered the gold standard procedure^8,9. However, the benefit of LLR remains under permanent debate since there are no prospective and randomized studies evaluating its results.

The aim of the present study was to compare the perioperative outcomes of patients undergoing LLR with those of open liver resection (OLR) for benign liver disease using propensity score matching (PSM).

Material and methods

We conducted a descriptive and retrospective, study comparing OLR and LLR for benign liver lesions performed between August 2010 and June 2021. The data analyzed were obtained from a prospective database.

Patients who had undergone a procedure associated with liver resection, as hepaticojejunostomy or biliary tract exploration were excluded. All the procedures were performed by the same surgical team. All the patients underwent clinical examination, laboratory tests and imaging tests (ultrasound, computed tomography scan and/or magnetic resonance imaging) to define the indication for surgery. The open or laparoscopic approach was decided in each case according to a comprehensive evaluation of the patient.

The lesions were divided considering their etiology into solid or cystic tumors and non-tumoral lesions. Major liver resection was considered when > 3 segments or greater were resected¹⁰. The group of patients who underwent laparoscopic surgery included those who were operated on with pure laparoscopy or hand-assisted laparoscopy¹¹. Intraoperative ultrasound was used in all the resections for a correct characterization of the lesion.

We arbitrarily divided the study period into two 6-year periods for PSM to standardize the samples before comparing them: period I from 2010 to 2015, and period II from 2016 to 2021.

The variables analyzed were age, sex, body mass index (BMI), type of liver resection, tumor size, type of tumor, period in which the procedure was performed, medical history, operative time, transfusion of red blood cells, use of Pringle maneuver, complications, postoperative abnormalities in the liver panel and coagulogram, reoperation, hospitalization on general ward or intensive care unit (ICU) and total length of hospital stay. Complications were categorized using the Clavien-Dindo (CD) classification. Grade I to V complications were considered general complications (GC) and major complications (MC) included grade III-b to grade V¹².

We used PSM with 1:1 matching to avoid biases of the different covariates between the groups. The criteria used for the PSM were age, sex, BMI, type of liver resection (major or minor), tumor size, type of lesion (solid, cystic or non-tumoral) and period during which the procedure was performed. Homogeneity of samples was assessed using the Shapiro-Wilk test. Categorical variables were expressed as percentages and compared using the chi-square test or Fisher’s test, as applicable. Numerical variables were expressed as mean ± standard deviation (SD), or median and interquartile range (IQR), and were compared with the t test of Mann-Whitney U test, as applicable. A p value < 0.05 was considered statistically significant. All the statistical calculations were performed using SPSS 25^® software package.

Results

Of 403 liver resections performed in patients >18 years, 89 were due to benign lesions (Fig. 1).

After applying the exclusion criteria, 82 liver resections were considered: 36 (44%) OLRs and 46 (56%) LLRs. Of all the LLRs, 35 (76%) were pure laparoscopies and 11 (24%) were hand-assisted laparoscopies. Mean age was 45 ±14 years and 65% were women. Forty-three (52%) were solid tumors, 23 (28%) were cystic tumors and 16 (20%) corresponded to non-tumoral lesions.

Among BLTs, adenoma was the most common indication for liver resection (18 cases, 22%), followed by FNH (15 cases, 18%). The remaining indications were polycystic liver disease (9 cases), hemangioma (8 cases), liver abscess (8 cases), complex cyst (6 cases), intrahepatic duct lithiasis (6 cases), hydatid cyst (5 cases), cystadenoma (3 cases), granuloma (two cases), hamartoma (1 case) and solitary fibrous tumor (1 case).

Table 1 shows the variables analyzed before PSM. Evidently, the groups are different. Diabetes and major resections were statistically higher in the OLR group.

After PSM, two groups of 28 patients each were constituted. The density curves after PSM can be observed in Figure 2. The comparison of the variables used for PSM and how the samples were matched are shown in Table 2. Furthermore, the differences in the history of diabetes were matched.

When the perioperative results were analyzed, the Pringle maneuver was more commonly used in the LLR group, and the operative time was longer; nevertheless, these differences were non-significant. Likewise, there were non-significant differences in the results of postoperative liver panel and coagulogram. There were significant differences in the need for transfusions, general complications, major complications, reoperations, length of hospital stay in the ICU and total length of hospital stay (Table 3).

As for the surgical indications in both groups after PSM, of 11 patients with FNH, resection was performed in 6 patients due to diagnostic uncertainty (suspected adenoma in 4, suspected hepatocellular carcinoma [HCC] and suspected neuroendocrine tumor metastasis in 1), in 3 because of tumor growth during follow-up, and in 2 due to symptoms. Resection was indicated in the 10 adenomas due risk or presence of complications. Of the 6 hemangiomas, the indication was the presence of symptoms in 3, tumor growth in 2 and suspected HCC in 1. The fibrous solitary tumor was resected due to symptoms. The 5 resections due to polycystic liver disease were performed in symptomatic patients. Of the patients with complex cysts, 2 presented symptoms and one underwent surgery with a diagnosis of hydatid cyst. Three resections were performed for hydatid cyst and 3 for cystadenomas. The six abscess resections were due to poor response to medical treatment and lack of response to percutaneous drainage. Six liver resections were performed for intrahepatic duct lithiasis: 4 of them were patients with Caroli’s disease with repeated cholangitis and the other 2 had associated bile duct strictures. In two patients with a history of cancer, 2 granulomas were resected for suspected liver metastases; both had PET-CT scans with increased FDG uptake.

Th OLR group included 9 hemihepatectomies, 9 bisegmentectomies, 5 atypical resections, 4 segmentectomies and 1 trisegmentectomy, while the LLR group included 9 hemihepatectomies, 7 bisegmentectomies, 7 atypical resections, 4 segmentectomies and 1 trisegmentectomy.

The GCs of LLRs were 3 CD grade II and 4 CD grade III-a complications, which corresponded to 1 biliary fistula that only solved with surgical drainage, 1 atelectasis, and 1 collection at the cut surface that solved with antibiotic therapy. The 4 CD grade III-a complications were collections that required percutaneous drainage. The GC of OLR were 5 CD grade II, 1 CD grade III-a and 4 CD grade III-b. The 5 CD grade II complications included 1 hemoperitoneum requiring red blood cell (RBC) transfusion, 1 biloma requiring antibiotic therapy, 1 pneumonia requiring antibiotic therapy, 1 arrhythmia (paroxysmal supraventricular tachycardia) that prolonged hospitalization, and 1 biliary fistula that solved spontaneously. The grade III-a complication was one collection that required percutaneous drainage and the 4 grade III-b complications were 1 intra-abdominal collection, 1 hemoperitoneum, 1 biloma and 1 surgical site infection that required laparotomy. There were no deaths within 90 days in any of the groups.

Discussion

Over the years, LLR has proved to be a safe procedure for the patient and has overcome the barriers and adapted the maneuvers from open surgery, such as hepatic mobilization, parenchymal transection and bleeding control^5,13. Undoubtedly, the incorporation of new technologies has favored the acceptance of LLR¹⁴. In this study, we use the term benign lesions because we include patients with BLT and other etiologies as abscesses, granulomas, intrahepatic duct lithiasis and intrahepatic bile duct strictures.

Benign lesions represent between 7% and 40% of the indications for liver resections in modern liver surgery^15,16. In our group, benign liver lesions accounted for 22% of liver resections. In 2016, Ciria et al. published a review of the literature on LLR; 35% of the cases (3337 of 9527) were due to benign lesions, mainly intrahepatic lithiasis, complex cysts, FNH and hemangiomas¹⁷. Our initial experience with 119 LLRs recently published in Revista Argentina de Cirugía, included 44 (36.5%) procedures performed due to benign diseases¹⁸. In a survey conducted in different hepatobiliary centers in South America, published in 2020 by J. Pekolj et al., 43% of LLRs were performed for benign diseases¹⁹. We should point out that the fact of counting with LLR capabilities, which has advantages over OLR, is not in itself an indication for resection of BLTs. We must be very selective and convinced of the true surgical indication, providing patients with the necessary explanations so that they understand the risks and benefits of surgery. In the presence of a patient with symptomatic BLT, medical treatment should be the first approach.

The indications for resection in patients with benign lesions vary according to the type of tumor²⁰. Women with adenomas who are taking oral contraceptives (OCPs) should stop taking them, since there are reports describing tumor size decrease. If tumor size decreases after 6 months, they can be managed with conservative treatment. Surgical treatment is recommended if there is no history of OCP intake and the size is larger than 5 cm^21,22. Adenomas in male patients should be resected regardless of their size, as they have higher risk of malignant transformation²³. Symptomatic FNH unresponsive to medical treatment is an indication for surgery²⁴. In the case of hemangiomas, surgical resection is reserved in case of symptoms or complications, such as bleeding, rupture or infarction. The association with Kasabach- Merritt syndrome is another indication for surgery²⁵.

A few years ago, the treatment strategies for patients with BLTs were published in the United States. The number of hepatic resections performed increased by > 50% from 2000 to 2011 which was attributed to more diagnosis of lesions due to better quality of imaging tests and to increase in the use of LLR²⁶. Several studies have demonstrated improvement in quality of life after liver resection due to BLTs. A systematic review recently published concludes that more than 82% of patients improve their symptoms after surgery. Two of the studies analyzed in this review have demonstrated the superiority of LLRs for improving symptoms after liver resection^27-29.

A study published by Assis et al. in 2020 compared OLR versus LLR after PSM and demonstrated favorable results in favor of laparoscopic resection, with shorter operative time, fewer complications and shorter length of hospital stay¹³. Elfrink et al. have recently published a multicenter study analyzing the postoperative results of 415 open and laparoscopic liver resections for BLTs, demonstrating shorter length of hospital stay, lower morbidity and no differences in mortality in the laparoscopic group^30,31. Our study shows similar results in terms of morbidity and length of hospital stay. These advantages of laparoscopic surgery not only provide benefits for the patient, in terms of less morbidity and faster return to work, but also reduces health care costs. In our study we did not analyze the presence of incisional hernias at long-term follow-up. Laparoscopic surgery significantly reduces their incidence, and, therefore, this could be considered as another benefit of this type of approach³².

The weaknesses of are study are:

▪ Its retrospective nature. We understand that prospective randomized studies in symptomatic patients are difficult to perform, deciding on conservative treatment for one group or surgical treatment for the other, and on which approach to use. Propensity score matching is a valid tool to reduce this type of bias in prospective studies.

▪ The low number of patients analyzed. This is a consequence of PSM. In this type of study, the samples are standardized, but the number of cases decreases.

▪ The lack of measurement of quality of life after liver resection as a parameter for the evaluation of longterm results which are mentioned in the discussion, is a matter to address to in future studies.

Based on the results presented, we may conclude that LLR for BLTs is a safe and effective technique, with shorter length of hospital stay, lower requirement for transfusions, fewer reoperations and fewer major complications than the open approach. Therefore, LLR could be considered the surgical technique of choice for benign liver disease.