Early adoption of laparoscopic liver surgery in Poland: a national retrospective cohort study

Abstract

Background:

The need for safe and efficient dissemination of minimally invasive approach in liver surgery is among the current challenges for hepatobiliary surgeons. After the stage of innovators and pioneers, the following countries should adopt a laparoscopic approach. The aim of this study was to assess the national experience and trend in implementing laparoscopic liver resection (LLR) in Poland.

Materials and methods:

A national registry of LLR performed in Poland was established in June 2020. All LLR cases performed before were included retrospectively, followed by prospectively collected new cases. Baseline characteristics, preoperative and intraoperative data, short-term results and long-term follow-up were recorded.

Results:

Since 2010 up to the end of 2022 there were 718 LLRs performed in Poland. The national rate of laparoscopic approach has gradually increased since 2017 (P<0.001), reaching the rate of 11.7% in 2022. There were 443 (61.7%), 107 (14.9%), and 168 (23.4%) LLRs performed in accordance to increasing grades of difficulty. The move towards more demanding cases had an increasing trend over the years (P<0.001). Total intraoperative adverse event and postoperative severe complications rates were estimated for 13.5% (n=97) and 6.7% (n=48), respectively. 30-day reoperation, readmission and postoperative mortality rates were 3.6% (n=26), 2.8% (n=20), and 0.8% (n=6), respectively. While the R0 resection margin was assessed in 643 (89.6%) cases, the total textbook outcomes (TO) were achieved in 525 (74.5%) cases. Overcoming the learning curve of 60 LLRs, resulted in an increasing TO rate from 72.3 to 80.6% (P=0.024).

Conclusions:

It is the first national analysis of a laparoscopic approach in liver surgery in Poland. An increasing trend of minimizing invasiveness in liver resection has been observed. Responsible selection of cases in accordance with difficulty may provide results within global benchmark values and textbook outcomes already during the learning curve.

Introduction

Highlights

• The presented study is the first national analysis regarding the state of laparoscopic approach in liver surgery in Poland.

• The rate of a minimally invasive approach in liver resection has been an increasing trend in the last years, reaching 11.7% in 2022.

• Proper case selection in accordance with difficulty grade may provide results within global benchmark values and textbook outcomes already during the learning curve.

• Overcoming the learning curve for laparoscopic liver resection may additionally increase the rate of textbook outcome achievement.

The first laparoscopic liver resections (LLR) were reported in the early 1990s^1–4. Experiencing its feasibility with simultaneous improvement of outcomes in terms of complications, postoperative pain, length of hospital stay, or cosmetic effect have driven hepatobiliary surgeons to go further with applying this approach for even more complex cases. Establishing the safety of LLR was finally expressed by the implementation of laparoscopic living donor hepatectomy for pediatric and later also adult recipients^5–7. In the last few years, there were also arguments for technical feasibility of laparoscopic approach in liver transplantation; however, more studies need to be performed to prove its safety, efficacy and benefits^8–10.

The world pioneers of LLR, straight after initial procedures, intensively acted for the world dissemination of minimizing invasiveness in liver surgery. The dissemination was supported by organized consensus meetings^11–13, designing randomized control trials^14,15 and associating all enthusiasts through the International Laparoscopic Liver Society¹⁶. However, rapid implementation of LLR in the following departments may be challenging and limited due to its high technical complexity and long learning curve^17,18. According to the available data from France, Italy, and the Netherlands, the national overall use of minimally invasive approach for liver resection in European countries varies between 10.3 and 23.0%^19–21. Taking advantage of the initial experience, several difficulty classifications were proposed to predict the surgical complexity of the planned procedures^22–24. They aimed for facilitating proper selection of LLR cases in arising new surgical teams devoted for minimally invasive liver surgery.

The period of the most intensive development of hepatobiliary surgery in Poland has been seen in the last of the 20th and first decade of the 21st century. It was expressed by establishing active liver transplant programs and reaching a significant number of open liver resections^25,26. Inspired by world trends, LLRs slowly but successively gained popularity in our country in the second decade of 21st century. While an increasing number of Polish surgeons were involved in performing such cases, a need for the assessment of the national dissemination and results of laparoscopic liver surgery performed in Poland has arisen.

The aim of this study was to assess the national experience in performing LLRs in Poland. Secondly, it aims to assess the short-term results and to give an overview of the national dissemination of LLRs.

Methods

A national registry for minimally invasive liver surgery in Poland was established in June 2020. An open calling for participation was sent through the Association of Polish Surgeons, followed with personal invitations sent directly to every department, where liver surgery is regularly performed. The main aims of the registry were to assess the evolution and expansion of minimally invasive liver surgery in Poland, to evaluate the short-term and long-term results of LLR performed in Poland and to prepare the background for prospective studies on LLR in our country.

At the beginning of establishing the registry seven departments joined the study. All LLR cases performed before June 2020 were included retrospectively, followed by prospectively collected new cases. In 2021, the eighth department started performing LLRs and joined the registry. To the best of our knowledge all departments, where regularly LLR is performed, have joined the registry. No minimal cutoff in the number of cases was established for registry inclusion, resulting in an even smaller series being considered. Indication for liver resection were solid malignant tumors or benign which were symptomatic or of an uncertain diagnosis. A minimally invasive approach was contraindicated in patients with respiratory or circulatory insufficiency, as in other procedures with a laparoscopic approach. Cases scheduled for LLR were selected in accordance to its difficulty and the main surgeon’s experience. The surgical technique for parenchyma transection differed between surgeons. General assumptions for maintaining low central venous pressure and low positive end-expiratory pressure were common in all centers. Intermittent inflow occlusion (Pringle maneuver) was applied based on surgeons’ preferences. Only cases where liver parenchyma transection was initiated from a laparoscopic approach were included. From 2010 up to the end of 2022 there were 718 LLRs performed in Poland.

Study data was collected and managed using REDCap electronic data capture tools hosted at the university server at https://watroba-lap.wum.edu.pl ^27,28. Baseline characteristics, preoperative (comorbidities, tumor characteristics, and laboratory tests), intraoperative (type of resection, transection technique, and intraoperative adverse events), postoperative (morbidity and mortality rates, reoperation and 30-day readmission rates, and histopathological examination), and long-term follow-up were recorded. Intraoperative adverse events were defined according to the Oslo classification²⁹. Postoperative complications were classified in accordance to Clavien–Dindo classification³⁰. Postoperative bile leak was assessed based on the International Study Group of Liver Surgery grading system³¹. For the complex measurement of surgical outcome Textbook Outcome (TO) was evaluated, which was defined as the absence of intraoperative adverse events of grade 2 or higher, postoperative bile leak of grade B or C, severe complications (Clavien–Dindo ≥3), postoperative reintervention, readmission within 30 days after discharge, in-hospital mortality and the presence of R0 resection margin³². Data about the type of surgical experience of the main surgeons was collected based on personal declaration.

This study included all LLR performed up to the end of 2022; however, the registry is still ongoing. Descriptive statistics were used to summarize the data [median, proportion, interquartile range (IQR), 95% CI]. The total number of liver resections performed in Poland, including open technique, was acquired from the National Health Fund (pol. Narodowy Fundusz Zdrowia), based on reported procedures 50.221, 50.31, 50.32 in accordance with the ICD-9 coding system. For a description of performed procedures the difficulty classification proposed by Kawaguchi et al.²⁴ was applied. In accordance to it, group 1 includes wedge, nonanatomical resections and left lateral sectionectomy, group 2 includes anterolateral segmentectomy and left hepatectomy, while group 3 includes posterosuperior segmentectomy, right hepatectomy, extended right hepatectomy, right posterior sectionectomy, central hepatectomy, and extended left hepatectomy. In case of simultaneous two different types, the procedure was classified based on more difficult type. The final size of compared group was 443, 107, and 168 for groups 1, 2, and 3, respectively. Selected short-term results were compared in accordance with the type of surgical experience and stage of their learning curve. Based on the literature, the cutoff for fulfilling the individual learning curve was set for 60 cases³³.

Data analysis was performed using STATISTICA 13 and Microsoft Excel 365. Continuous data were compared using the Mann–Whitney U test or Kruskal–Wallis test as appropriate. Categorical data were compared using Pearson χ² statistic or Fisher exact test, as appropriate. The yearly trend in the percentage of LLR was tested with the Cochrane–Armitage test. The work has been reported in line with the strengthening the reporting of cohort, cross-sectional, and case–control studies in surgery (STROCSS) criteria³⁴. The study is registered at https://clinicaltrials.gov/. Ethical approval for this study was provided by the local committee.

Results

National dissemination

In 2010, LLRs started to be performed in Poland. Nonanatomical resections were first done, followed by the left lateral sectionectomy the following year and the first major LLR- right hepatectomy, in 2013. The number of LLRs performed nationally per year gradually increased, reaching a total number of 718 cases by the end of 2022 (Fig. 1). A significant number of procedures were performed in years 2017–2022, with the rate of laparoscopic approaches growing from 1.7 to 11.7% in the total number of liver resections performed in Poland (Fig. 2), showing a yearly increasing trend (P<0.001). The rates of more complex cases successively increased over the years, reaching 21.1 and 30.2% for difficulty groups 2 and 3, respectively, in 2022 (Fig. 3). The yearly, increasing trend of the most demanding cases was significant (P=0.001).

Figure 1.

Number of laparoscopic liver resections performed in Poland.

Figure 2.

Rates of laparoscopic liver resection in relation to the total number of liver resections performed in Poland in years 2017–2022.

Figure 3.

Distribution of LLR cases over the years in accordance with IMM difficulty score.

Up to the end of 2022, there were eight departments where LLRs were performed. The median number of cases performed per department was 58, while three departments had an experience of >100 cases. Among 718 LLRs, 85.8% of the cases were performed by 10 surgeons, who registered >10 performed cases, creating their own personal learning curves (Fig. 4). The type of their surgical experience was different. Years of surgical practice at the moment of performing the first LLR ranged from 4 to 26 years with a median of 12.5 years. Experience of liver transplantation, advanced laparoscopic oncological cases and advanced laparoscopic benign cases was declared in 40, 50, and 70% of them, respectively. Four of them presented with a dominance of liver surgery experience at the start of the LLR learning curve, while the rest with proficiency of advanced laparoscopic surgery, referred to other organs than the liver.

Figure 4.

Learning curves for surgeons performing laparoscopic liver resections.

Preoperative data

The median age of patients scheduled for LLR was 62 years, 362 (50.4%) of them were female (Table 1). The median BMI of all patients was 26.6 kg/m² and was comparable in each group in accordance with the difficulty classification. The most common comorbidity was cardiovascular disease (325, 45.2%), while liver cirrhosis was diagnosed in 31 (4.3%) of patients. Almost half of the patients had a history of abdominal surgery, with 175 (24.3%) and 132 (18.4%) of the cases with previous laparotomy and laparoscopy, respectively. The liver segments most frequently affected by the tumor in preoperative imaging were segments V, VI, and VII (Fig. 5).

Table 1.

Preoperative assessment of patients scheduled for laparoscopic liver resection.

	Median (IQR)/ n (%) [95% CI]
	Difficulty group
Preoperative data	1 n=443	2 n=107	3 n=168	Total n=718	P
Age [years]	61 (48–68) [59–63]	67 (57–74) [61–71]	63 (53–70) [60–65]	62 (50–70) [60–63]	0.002
Sex (female)	239 (54.0%) [49.3–58.6%]	52 (48.6%) [39.1–58.1%]	71 (42.3%) [34.8–49.7%]	362 (50.4%) [46.7–54.0%]	0.031
BMI [kg/m2]	26.8 (23.4–29.8) [26.6–27.6]	26.4 (23.9–30.1) [26.5–28.5]	26.4 (23.7–29.0) [26.1–27.5]	26.6 (23.7–29.7) [26.7–27.5]	0.578
Cardiovascular disease	194 (43.8%) [39.2–48.4%]	57 (53.3%) [43.8–62.7%]	74 (44.1%) [36.5–51.6%]	325 (45.2%) [41.6–48.8%]	0.196
Pulmonary disease	8 (1.8%) [0.6–3.1%]	1 (0.9%) [0.0–2.8%]	8 (4.8%) [1.5–8.0%]	17 (2.4%) [1.3–3.5%]	0.057
Liver cirrhosis	22 (5.0%) [2.9–7.0%]	5 (4.7%) [0.7–8.7%]	4 (2.4%) [0.1–4.7%]	31 (4.3%) [2.8–5.8%]	0.366
Preoperative chemotherapy	105 (23.7%) [19.7–27.7%]	32 (29.9%) [21.2–38.6%]	68 (40.5%) [33.1–47.9%]	205 (28.5%) [25.2–31.8%]	0.001
Previous laparotomy	107 (24.2%) [20.2–28.1%]	25 (23.4%) [15.4–31.4%]	43 (25.6%) [19.0–32.2%]	175 (24.3%) [21.2–27.5%]	0.902
Previous laparoscopy	76 (17.2%) [13.7–20.7%]	19 (17.8%) [10.5–25.0%]	37 (22.0%) [15.8–28.3%]	132 (18.4%) [15.5–21.2%]	0.376

(P<0.05) values are in bold.

Figure 5.

Liver segments affected by tumors based on preoperative imaging.

Intraoperative data

According to the Institute Mutualiste Montsouris (IMM) difficulty classification of LLR procedures there were 443 (61.7%) and 168 (23.4%) of group 1 and 3, respectively (Table 2). Nonanatomical resection in anterolateral segments was the most common in group 1 (286, 39.8%). Among more challenging procedures, the most frequently performed were anterolateral segmentectomies (71, 9.9%) and right hepatectomies (88, 12.3%) in groups 2 and 3, respectively. In 476 (66.2%) cases, LLRs were performed with the assistance of intraoperative ultrasonography, other methods of intraoperative imaging such as cholangiography or indocyanine green staining were used occasionally, in 7 (1.0%) cases each. In selected cases, additional procedures were performed, among which the most common were cholecystectomy (254, 35.3%), lymphadenectomy (33, 4.6%), and colon resection (19, 2.6%) (Table 3).

Table 2.

Type of laparoscopic liver resections performed in Poland.

Type of procedure	n (%) [95% CI]
Nonanatomical resection in anterolateral segments (II, III, IVb, V, or VI)	286 (39.8%) [36.3–43.4%]
Nonanatomical resection in posterosuperior segments (I, IVa, VII, VIII)	84 (11.7%) [9.4–14.1%]
Left lateral sectionectomy	73 (10.2%) [8.0–12.4%]
Anterolateral segmentectomy	71 (9.9%) [7.7–12.1%]
Left hepatectomy (II, III, and IV±I)	36 (5.0%) [3.4–6.6%]
Posterosuperior segmentectomy	41 (5.7%) [4.0–7.4%]
Right hepatectomy (V, VI, VII, and VIII)	88 (12.3%) [9.9–14.7%]
Extended right hepatectomy (IV, V, VI, VII, and VIII±I)	8 (1.1%) [0.4–1.9%]
Right posterior sectionectomy (VI, VII)	21 (2.9%) [1.7–4.2%]
Central hepatectomy (V, VIII or IV, V, VIII)	4 (0.6%) [0.0–1.1%]
Extended left hepatectomy (I, II, III, IV, V, and VIII)	6 (0.8%) [0.2–1.5%]
Total	718 (100%)

Table 3.

Additional procedures during laparoscopic liver resection.

Additional procedure	n (%) [95% CI]
Intraoperative ultrasonography	476 (66.2%) [62.8–69.7%]
Cholecystectomy	254 (35.3%) [31.8–38.8%]
Lymphadenectomy	33 (4.6%) [3.1–6.1%]
Colon resection	19 (2.6%) [1.5–3.8%]
Intraoperative ablation	18 (2.5%) [1.4–3.7%]
Diaphragm resection	10 (1.4%) [0.5–2.3%]
Closure of colostomy	7 (1.0%) [0.3–1.7%]
Indocyanine green staining	7 (1.0%) [0.3–1.7%]
Cholangiography	7 (1.0%) [0.3–1.7%]
Hernia repair	5 (0.7%) [0.1–1.3%]

Liver parenchyma transection was the most commonly performed with usage of clips (89.0%), the ultrasonic surgical aspirator (66.4%), and the ultrasonic coagulating shear device (46.7%) (Table 4). In 345 (48.1%) cases the Pringle maneuver was applied during transection, while only in group 3 in 115 (68.5%) cases this maneuver was performed (Table 5). The median total time of the maneuver was 40 min (IQR=30–50). Median blood loss was 200 ml (IQR=150–400) with a need for blood transfusion in 64 (8.9%) patients. The total conversion rate was 6.3% (n=45), while the most common reason for laparotomies was intraoperative hemorrhage and concerns for the resection margin in 19 (42.2%) and 11 (24.4%) cases, respectively. The median operative time ranged from 185 min (IQR=130–270) in group 1 to 340 min (IQR=260–420) in group 3. There were four types of specimen extraction sites: widening trocar incision (364, 50.7%), Pfannenstiel incision (225, 31.3%), previous surgery incision (79, 11.0%), and midline incision (50, 7.0%). The total postoperative drainage rate was 80.4%, while for group 3 it was 94.0%.

Table 4.

Surgical instruments used for liver parenchyma transection.

	n (%) [95% CI]
	Difficulty group
Transection technique	1 n=443	2 n=107	3 n=168	Total n=718	P
Clips	381 (86.0%) [82.8–89.2%]	98 (91.6%) 86.3–96.9%]	160 (95.2%) [92.0–98.5%]	639 (89.0%) [86.7–91.3%]	0.303
Ultrasonic surgical aspirator	267 (60.3%) [55.7–64.8%]	75 (70.1%) (61.4–78.8%)	135 (80.4%) (74.4–86.4%]	477 (66.4%) [63.0–69.9%]	0.001
Ultrasonic coagulating shear device	234 (52.8%) [48.2–57.5%]	44 (41.1%) [31.8–50.4%]	57 (33.9%) [26.8–41.1%]	335 (46.7%) [43.0–50.3%]	0.003
Bipolar electrocautery	140 (31.6%) [27.3–35.9%]	51 (47.7%) [38.2–57.1%]	72 (42.9%) [35.4–50.3%]	263 (36.6%) [33.1–40.2%]	0.262
Hybrid energy devicea	135 (30.5%) [26.2–34.8%]	40 (37.4%) [28.2–46.6%]	76 (45.2%) [37.7–52.8%]	252 (35.1%) [31.6–38.6%]	0.001
Staplers	82 (18.5%) [14.9–22.1%]	35 (32.7%) [23.8–41.6%]	101(60.1%) [52.7–67.5%]	218 (30.4%) [27.0–33.7%]	0.001
Monopolar electrocautery	35 (7.9%) [5.4–10.4%]	18 (16.8%) [9.7–23.9%)	35 (20.8%) [14.7–27.0%]	88 (12.3%) [9.9–14.7%]	0.601
Electrothermal bipolar vessel sealer	37 (8.4%) [5.8–10.9%]	11 (10.3%) [4.5–16.0%]	19 (11.3%) [6.5–16.1%]	67 (9.3%) [7.2–11.5%]	0.019
Argon plasma coagulation	52 (11.7%) [8.7–14.7%]	9 (8.4%) [3.2–13.7%]	5 (3.0%) [0.4–5.6%]	66 (9.2%) [7.1–11.3%]	0.001
Sutures	26 (5.9%) [3.7–8.1%]	6 (5.6%) [1.3–10.0%]	21 (12.5%) [7.5–17.5%]	53 (7.4%) [5.5–9.3%]	0.149
Water-jet device	11 (2.5%) [1.0–3.9%]	0	0	11 (1.5%) [0.6–2.4%)	n/a

(P<0.05) values are in bold.

^aCombining ultrasonic and electrothermal energy.

Table 5.

Intraoperative results.

	Median (IQR)/ n (%) [95% CI]
	Difficulty group
Intraoperative data	1 n=443	2 n=107	3 n=168	Total n=718	P
Operative time [min]	185 (130–270) [198–218]	270 (213–360) [265–307]	340 (260–420) [335–375]	235 (160–330) [244–264]	0.001
Pringle’s maneuver	167 (37.7%) [33.2–42.2%]	63 (58.9%) [49.6–68.2%]	115 (68.5%) [61.4–75.5%]	345 (48.1%) [44.4–51.7%]	0.001
Blood loss [ml]	200 (100–300) [216–264]	350 (250–500) [384–532]	400 (300–600) [448–572]	200 (150–400) [308–358]	0.001
Blood transfusion	25 (5.6%) [3.5–7.8%]	12 (11.2%) [5.2–17.2%]	27 (16.2%) [10.5–21.6%]	64 (8.9%) [6.8–11.0%]	0.001
Conversion	18 (4.1%) [2.2–5.9%]	13 (12.2%) [6.0–18.3%]	14 (8.3%) [4.2–12.5%]	45 (6.3%) [4.5–8.0%]	0.003
Intraoperative adverse eventsa	43 (9.7%) [7.0–12.5%]	19 (17.8%) [10.5–25.0%]	35 (20.8%) [14.7–27.0%]	97 (13.5%) [11.0–16.0%]	0.001

(P<0.05) values are in bold.

^aOslo classification grade ≥2.

Postoperative data

The total morbidity rate was evaluated as 13.5% (Table 6). In accordance with the Clavien–Dindo classification, 50.0% (n=48) of registered postoperative complications were classified as grade ≥3. The total mortality rate was assessed as 0.8% (n=6). Postoperative deaths were related to liver remnant insufficiency (n=2), respiratory failure (n=1), sepsis (n=2), and excessive blood loss (n=1). The median hospital stay was 5 (IQR=3–6) days. Total 30-day reoperation and 30-day readmission rates were 3.6% (n=26) and 2.8% (n=20), respectively. The most common indications for reoperation were biliary fistula (n=9), postoperative hemorrhage (n=6), and bowel perforation (n=3). Most common indications for readmission were biloma (n=5), intra-abdominal abscess (n=3), intra-abdominal hematoma (n=3), wound infection (n=3), and R1 margin (n=2).

Table 6.

Postoperative results.

	Median (IQR)/ n (%) [95% CI]
	Difficulty group
Postoperative data	1 n=443	2 n=107	3 n=168	Total n=718	P
Hospital stay [days]	4 (3–5) [4.6–5.4]	5 (4–6) [5.5–8.5]	7 (5–9) [7.3–10.7]	5 (3–6) [5.5–6.5]	0.001
General postoperative complication rate	45 (10.2%) [7.3–13.0%]	17 (15.9%) [9.0–22.8%]	35 (20.8%) [14.7–27.0%]	97 (13.5%) [11.0–16.0%]	0.004
Severe postoperative complication ratea	19 (4.3%) [2.4–6.2%]	8 (7.5%) [2.5–12.5%]	21 (12.5%) [7.5–17.5%]	48 (6.7%) [4.9–8.5%]	0.001
Bile leakb	8 (1.8%) [0.6–3.1%]	4 (3.7%) [0.1–7.3%]	9 (5.4%) [2.0–8.8%]	21 (2.9%) [1.7–4.2%]	0.056
Mortality rate	0	2 (1.9%) [0.0–4.4%]	4 (2.4%) [0.1–4.7%]	6 (0.8%) [0.2–1.5%]	0.001
30-day reoperation rate	12 (2.7%) [1.2–4.2%]	5 (4.7%) [0.7–8.7%]	9 (5.4%) [2.0–8.8%]	26 (3.6%) [2.3–5.0%]	0.259
30-day readmission rate	10 (2.3%) [0.9–3.6%]	5 (4.7%) [0.7–8.7%]	5 (3.0%) [0.4–5.6%]	20 (2.8%) [1.6–4.0%]	0.375
Textbook outcome	343 (77.4%) [73.5–81.3%]	82 (76.6%) [68.6–84.7%]	110 (65.5%) [58.3–72.7%]	535 (74.5%) [71.3–77.7%]	0.009

(P<0.05) values are in bold.

^aClavien–Dindo classification ≥3.

^bISGLS grade B or C.

According to the histopathological results R0 margin was achieved in 89.6% (n=643) of cases (Table 7). The median number of tumors was 1 (IQR=1–1) with a median maximum size of 30 mm (IQR=20–45). Malignant tumors were diagnosed in 79.4% (n=570) of cases. The most frequent type of tumor was colorectal liver metastasis (n=324, 45.1%).

Table 7.

Histopathological results.

	Median (IQR)/ n (%) [95% CI]
	Difficulty group
Histopathological data	1 n=443	2 n=107	3 n=168	Total n=718	P
Number of tumors	1 (1–1) [0.9–1.1]	1 (1–1) [0.8–1.2]	1 (1–2) [1.7–2.3]	1 (1–1) [0.9–1.1]	0.001
Maximum size of the tumor [mm]	28 (17–40) [29.0–33.1]	39 (25–54) [36.4–45.6]	39 (23–60) [40.6–49.4]	30 (20–45) [34.2–37.8]	0.001
R0 resection margin	392 (88.5%) [85.5–91.5%]	100 (93.5%) [88.8–98.1%]	151 (89.9%) [85.3–94.4%]	643 (89.6%) [87.3–91.8%]	0.317
Colorectal liver metastasis	175 (39.5%) [34.9–44.1%]	52 (48.6%) [39.1–58.1%]	100 (59.5%) [52.1–67.0%]	324 (45.1%) [41.5–48.8%]	0.001
Hepatocellular carcinoma	51 (11.5%) [8.5–14.5%]	14 (13.1%) [6.7–19.5%]	16 (9.5%) [5.1–14.0%]	81 (11.3%) [9.0–13.6%]	0.001
Noncolorectal liver metastasis	41 (9.3%) [6.6–12.0%]	10 (9.4%) [3.8–14.9%]	17 (10.1%) [5.6–14.7%]	65 (9.1%) [7.0–11.2%]	0.947
Intrahepatic cholangiocarcinoma	9 (2.0%) [0.7–3.3%]	8 (7.5%) [2.5–12.5%]	15 (9.0%) [4.6–13.2%]	31 (4.3%) [2.8%-5.8%]	0.001
Benign tumor	99 (22.4%) [18.5–26.2%]	23 (21.5%) [13.7–29.3%]	20 (11.9%) [7.0–16.8%]	148 (20.6%) [17.7–23.6%]	0.014

(P<0.05) values are in bold.

Due to the heterogeneity of indications and a relatively short observation period of most recorded cases, overall survival was not analyzed in this study.

Surgical experience influence

Among registered LLRs, 527 (73.4%) cases were performed before reaching the personal learning curve number of 60 cases for each surgeon. Most of the cases (319, 60.5%) were performed by surgeons who declared dominance of advanced laparoscopic surgery in their experience in the beginning of their learning curve (Table 8). There was no significant difference in the selection of the most demanding cases depending on the type of surgeon’s experience (P=0.224), similarly in the rates of conversion (P=0.694), general complication (P=0.503), and severe complication (P=0.479) in the postoperative period. Liver surgeons applied Pringle’s maneuver less frequently (P=0.001), nonetheless median blood loss was comparable in both groups (P=0.084). Procedures performed by surgeons with a dominance of advanced laparoscopic surgery experience lasted shorter (P=0.001) and more frequently were assessed as a R0 resection margin (P=0.008).

Table 8.

Comparison of selected rates during the first 60 cases in accordance with surgeons’ experience.

	Median (IQR)/ n (%) [95% CI]
	Type of surgeons’ experience
Selected rates	Dominance of liver surgery experience (n=208)	Dominance of advanced laparoscopic surgery experience (n=319)	P
Difficult cases ratea	33 (15.9%) [10.9–20.8%]	64 (20.1%) [15.7–24.5%]	0.224
Operative time [min]	270 (180–390) [279.9–324.1]	210 (136–290) [211.3–236.7]	0.001
Blood loss [ml]	150 (100–300) [215.0–271.0]	200 (100–400) [282.7–365.3]	0.084
Pringle’s maneuver	66 (31.7%) [25.4–38.1%]	161 (50.5%) [45.0–56.0%]	0.001
Conversion rate	15 (7.2%) [3.7–10.7%]	26 (8.2%) [5.2–11.2%]	0.694
Intraoperative adverse eventsb	23 (11.1%) [6.8–15.3%]	51 (16.0%) [12.0–20.0%]	0.111
General postoperative complication rate	31 (14.9%) [10.1–19.7%]	41 (12.9%) [9.2–16.5%]	0.503
Severe postoperative complication ratec	10 (4.8%) [1.9–7.7%]	20 (6.3%) [3.6–8.9%]	0.479
Bile leakd	8 (3.9%) [1.2–6.5%]	6 (1.9%) [0.4–3.4%]	0.170
30-day readmission rate	3 (1.4%) [0.0–3.1%]	12 (3.8%) [1.7–5.9%]	0.118
Mortality rate	0	3 (0.9%) [0.0–2.0%]	n/a
R0 resection margin	171 (82.2%) [77.0–87.4%]	285 (89.3%) [86.0–92.7%]	0.008
Textbook outcome	148 (71.2%) [65.0–77.3%]	233 (73.0%) [68.2–77.9%]	0.636

(P<0.05) values are in bold.

^aInstitute Mutualiste Montsouris difficulty group 3.

^bOslo classification grade ≥2.

^cClavien–Dindo classification ≥3.

^dISGLS grade B or C.

There were 191 (26.6%) LLRs performed after completing 60 cases on the personal learning curve (Table 9). Although these procedures were significantly more demanding in accordance with the IMM difficulty score (0.001), general postoperative complication rates and severe complication rates were similar (P=0.843 and 0.077, respectively). Even though surgeons applied Pringle’s maneuver more frequently (P=0.001), median blood loss was increased compared to cases performed before reaching the number of 60. After the 60th case on the personal learning curve, the conversion rate was significantly decreased (P=0.005) and the R0 resection margin was significantly increased (P=0.001).

Table 9.

Comparison of selected rates during learning curve.

	Median (IQR)/n (%) [95% CI]
	Stage of learning curve
Selected rates	≤60 cases n=527	>60 cases n=191	P
Difficult cases ratea	97 (18.4%) [15.1–21.7%]	71 (37.2%) [30.3–44.0%]	0.001
Operative time [min]	230 (150–330) [239.2–262.8]	245 (180-330) [244.9–275.1]	0.084
Blood loss [ml]	200 (100–350) [303.7–354.3]	300 (200-500) [288.6–409.4]	0.001
Pringle’s maneuver	227 (43.1%) [38.9–47.3%]	118 (61.8%) [54.9–68.7%]	0.001
Conversion rate	41 (7.8%) [5.5–10.1%]	4 (2.1%) [0.1–4.2%]	0.005
Intraoperative adverse eventsb	74 (14.0%) [11.1–17.0%]	23 (12.0%) [7.4–16.7%]	0.488
General postoperative complication rate	72 (13.7%) [10.7–16.6%]	25 (13.1%) [8.3–17.9%]	0.843
Severe postoperative complication rateb	30 (5.7%) [3.7–7.7%]	18 (9.4%) [5.3–13.6%]	0.077
Bile leakd	14 (2.7%) [1.3–4.0%]	7 (3.7%) [1.0–6.3%]	0.479
30-day readmission rate	15 (2.9%) [1.4–4.3%]	5 (2.6%) [0.4–4.9%]	0.869
Mortality rate	3 (0.6%) [0.0–1.2%]	3 (1.6%) [0.0–3.3%]	0.193
R0 resection margin	456 (86.5%) [83.6–89.4%]	187 (97.9%) [95.9–99.9%]	0.001
Textbook outcome	381 (72.3%) [68.5–76.1%]	154 (80.6%) [75.0%-86.2%]	0.024

(P<0.05) values are in bold.

^aInstitute Mutualiste Montsouris difficulty group 3.

^bOslo classification grade ≥2.

^cClavien–Dindo classification ≥3.

^dISGLS grade B or C.

Discussion

Laparoscopic liver surgery is among the most challenging group of procedures performed with a minimally invasive approach. After the period of LLR pioneers, the safety and efficacy of this procedure were established in an increasing number of large cohort studies and randomized controlled trials. Consensus meetings on LLR were aimed not only on summarizing actual results but also at scratching the directions for following surgeons to promote this approach around the world. The need for safe and efficient dissemination throughout the world of laparoscopic liver surgery is among the current, most important challenge for LLR leaders. Even though the new centers starting laparoscopic liver surgery have the opportunity to skip some of the pioneering efforts and may potentially shorten the learning curve, in reality the pace of LLR dissemination still remains rather slow.

This study is a representative picture of the actual practice in laparoscopic liver surgery in Poland. It also gives a perspective on the dissemination of this technique since the first case was performed in our country. Simultaneously, it is a real-life example of the implementation pace, even though already existing proofs of LLR safety and efficacy. The first cases of LLR performed in Poland were registered almost 20 years after pioneer cases performed worldwide³⁵. T. Wróblewski with the first cases in 2010 and A. Budzyński with the first major, laparoscopic right hepatectomy in 2013 were the first early adopters of LLRs in Poland. After a few years they were followed by several surgeons who together achieved a 11.7% rate of minimally invasive approach in the total number of liver resections performed nationally in the year 2022.

There are few studies presenting national data about LLR dissemination in different countries^{19–21,36–38}. They were performed in countries where world pioneers implemented LLR as a new approach. Early analysis includes roughly period of the first decade in the 21st century. Rates of LLR in relation to the total number of liver resections ranged from 3.9 to 14.8%^20,21,36. With the next decade, national LLR rates almost doubled, reaching even 24.8% in the biggest national cohort from Japan^19,37,38. While analyzing more selected cohorts and differentiating centers in accordance to the number of cases performed per year, there are studies which presents up to 42.9% rate in high volume centers³⁹. This study presents a regular increase of LLR dissemination in our country over the last 6 years, with persistent potential for higher LLR national rates. The most active department in the registry had a minimally invasive approach rate of 26.7% in 2022.

Huge work done by world pioneers proved the feasibility and safety of LLRs and facilitated further dissemination of this approach in the following countries. An indispensable factor which enhanced this process was a continuous development of surgical devices and new technologies. The pace of the popularization process results not only from the difficulty of the learned procedures per se, but also from the available methods for gaining knowledge and experience that allow for the dissemination of the laparoscopic approach. Despite the increasing universality of the exchange of knowledge and video materials, practical training is irreplaceable and crucial. In 2020, International Laparoscopic Liver Society has described the fellow skills curriculum, which should be acquired during training in laparoscopic liver surgery¹⁷. It was pointed, that some advanced substeps are too rare and too difficult to might be trained during 2-year fellowship and should be continued afterwards. Limited access to the 2-year international fellowships for surgeons in Poland resulted in adopting minimally invasive approach mainly based on short-term clinical visits and specialized courses. It has additionally prolonged the learning curves of Polish early adopters to even greater than 5 years (Fig. 4), which potentially could negatively affect the postoperative results. For that very reason, awareness of LLR difficulty grading systems⁴⁰ was crucial in maintaining safety during the early adoption phase in Poland. Based on known preoperative factors, a surgeon could select cases appropriate for her or his skillset and experience. It helped in gradual experience gaining, but additionally lengthen personal learning curves. Responsible selection was expressed by the trend of the increasing rate of the most demanding LLR cases in the last years of the registered period (Fig. 3), while tumors were located mainly in anterolateral segments (Fig. 5).

In 2023, based on an international multicenter study, global benchmarks for LLR were established⁴¹. Outcome measures were evaluated in accordance to the IMM difficulty classification; however, for unifying purposes, cases with one range of liver resection for each difficulty group was included (left lateral sectionectomy, left hepatectomy and right hepatectomy). In accordance with these benchmarks median operative time, hospital stay, and the rate of general postoperative complications were below the cutoff for all three difficulty groups of LLRs performed in Poland. Median blood loss in group 1 (200 ml) and 2 (350 ml) exceeded benchmark values which were 100 ml and 300 ml, respectively, while in group 3 it was equalized. It has resulted also in increased rates of blood transfusion; however, policy for intraoperative blood transfusion may differ between centers. The conversion rate was increased in accordance with the benchmark value only in group 1, what could be affected by selecting these cases as the very early phase of adopting laparoscopic approach for liver resection. In the same difficulty group, the severe postoperative complications rate (4.3%) also exceeded the benchmark value, which was established for 0%. However, it is important to notice, that most of the significant additional procedures like colon resection (n=14, 77.8%) or hepatoduodenal lymphadenectomy (n=18, 54.5%) were performed in the least demanding LLRs (Table 3). Most concerning was the increased mortality observed in group 2 (1.9%) and 3 (2.4%) in accordance with the global benchmark cutoff (0% for both). With great regret, zero mortality has not been achieved in the Polish cohort; however, despite such a benchmark, postoperative deaths were also observed in the benchmarking cohort. In other, nationwide study, which also aim for establishing benchmark outcomes for left lateral sectionectomy and right hepatectomy, assessed postoperative mortality for 1.2 and 2.8% respectively⁴². Assuming that resection margin status is the most significant, surgical factor affecting long-term results after liver resection, oncological results may be compromised only in group 1. Benchmark R1 resection in the least demanding cases was set on 7.1%, while in the Polish national cohort a rate of 11.5% was observed. As in the conversion rate, such observation could be affected by performing cases in the very beginning of the learning curve. Fortunately, patients in group 1 have the most probable chance for following reresections in case of recurrence. Intermediate and most difficult groups were far below benchmarking cutoffs regarding resection margin. Combining several outcome parameters, TO were achieved in 77.4, 76.6, and 65.5% for increasing difficulty groups, respectively. Görgec et al.³² have not distinguished the TO rate in accordance to LLR difficulty; however, in general population of patients who underwent LLR TO was achieved in 70.7%.

According to the IDEAL paradigm for surgical innovation⁴³, most of the registered LLRs in Poland were performed in period, when laparoscopic liver surgery was in the stage of, early adopters of its worldwide evolution⁴⁴. Benefiting from the innovators’ and pioneers’ experience, early adopters may overcome the learning curve faster, by focusing on the acquisition and development of their skills, rather than development novel surgical techniques or evolving proper indications. Preferably, for starting a new LLR program, surgeons should have comprehensive experience in open liver surgery and advanced laparoscopic procedures, simultaneously^13,44. Practically, based on personal declaration, early adopters in Poland presented with a dominance of experience in open liver surgery or advanced laparoscopic procedures other than liver resection at the moment of starting their individual learning curve. Assessing the results during the learning curve in accordance to type of surgical experience, a significant difference was observed in operative time (P<0.001) and R0 resection margin (P=0.008); however, TO were achieved with a comparable frequency of 71.2% and 73.0% (P=0.636) (Table 8). Overcoming personal learning curves in LLR is mostly observed with the decrease in blood loss, operative time, and conversion rate^45–47. Even though vast majority of cases registered in Poland were performed before overcoming personal learning curves (n=527, 73.4%), some significant differences were observed after completing 60 LLRs (Table 9). The conversion rate has significantly dropped (P=0.005), while median operative time was similar (P=0.084) and median blood loss was even increased (P<0.001). However, the last two variables need to be interpreted with the background of a significantly increased percentage of the most demanding LLRs after overcoming the learning curve (P<0.001). Despite this fact, the TO rate has significantly increased up to 80.6% (P=0.024) with increasing experience of performing LLRs.

This study has several limitations. Firstly, most of the registered cases (up to June 2020) were included retrospectively, which may result in an increased risk of selection bias. However, further realization of a nationwide registry was made prospectively, so that future updates of status of LLR practice in Poland would be possible. Second, the overall national number of liver resections, including open surgery, was acquired from the National Health Fund based on data possessed in the refunding hospital treatment process. Unfortunately, there is no other registry which may more precisely estimate the actual number and type of liver resections performed in Poland. Third, surgeons were individually selecting patients for LLRs in accordance with their individual sense of increasing experience and based on the available grading systems, while the IMM difficulty system was implemented retrospectively to assess recorded LLRs. Forth, the learning curve was arbitrarily set for 60 cases based on available publications, however there was no distinction between minor or major LLRs in overcoming this number, which obviously may affect the results of the analysis. Nevertheless, it presents the real-life structure of case selection, where postponing more demanding cases only after completely overcoming a minor LLR learning curve would additionally prolong the phase of adapting LLR in Poland. Noteworthy, the postoperative outcomes were within generally accepted standards.

Conclusions

The presented study is the first attempt to analyze the dissemination of the laparoscopic approach of liver surgery in Poland. Polish surgeons benefit from world inventors and pioneers to implement LLR safely and efficiently. Simultaneously, achieved outcomes proved, that the move to more complex cases was timed correctly. For international observers it is a representative picture of the national practice in adopting LLR and dissemination of minimally invasive approach for the following country. Responsible selection of cases in accordance with difficulty may provide results within global benchmark values and textbook outcomes, which are achievable already during the learning curve.

Ethical approval

Ethical approval for this study (AKBE/71/2020) was provided by the Ethical Committee of Medical University of Warsaw, Warsaw, Poland on 9 March 2020.

Consent

There are no identifying details in the manuscript, which may impar patients anonymity.

Sources of funding

This is a noncommercial study, without need for external funding.

Author contribution

W.H., M.P., M.G.: study concept and design; W.S., M.B., M.W., M.D., J.M., S.P., P.H., Ł.M., P.K., M.W., M.F., and A.T.: data collection; W.H., M.B., M.W., M.P., M.G.: data analysis; Wacław Hołówko, M.W., M.F., T.W., A.B., M.P., M.G.: data interpretation; W.H., A.T., M.P., M.G.: writing the paper.

Conflicts of interest disclosure

The authors declare that they have no financial conflict of interest with regard to the content of this report.

Presentation

International Laparoscopic Liver Society 2021 3rd World Congress, New York, USA.

Research registration unique identifying number (UIN)

The study is registered at https://clinicaltrials.gov/study/NCT05516394?id=NCT05516394&rank=1.

Guarantor

Wacław Hołówko, Klinika Chirurgii Ogólnej, Transplantacyjnej i Wątroby, Uniwersyteckie Centrum Kliniczne Warszawskiego Uniwersytetu Medycznego, Ul. Banacha 1a, 02-097 Warszawa, Poland. E-mail: waclaw.holowko@wum.edu.pl.

Data availability statement

Datasets generated during the current study are available upon reasonable request.

Provenance and peer review

Not commissioned, externally peer-reviewed.