Robotic liver resection for hepatic focal nodular hyperplasia in children: comparison with open surgery

Abstract

Background

Robot-assisted surgery is becoming increasingly used in pediatric oncology. The present study aimed to evaluate the feasibility and safety of robotic liver resection (RLR) for hepatic focal nodular hyperplasia (FNH) in children and compare surgical outcomes between RLR and open liver resection (OLR).

Methods

Pediatric patients with liver FNH undergoing lesion resection between January 2020 and June 2024 were included in the study. Patient demographics, operative details, postoperative outcomes, and follow-up were recorded and analyzed.

Results

A total of 20 patients were included in this study. Twelve patients underwent RLR and eight underwent OLR. In the RLR group, the median age was 93.1 months (range, 28–134 months) with a median weight of 32.4 kg (range, 9.7–80 kg). The median maximum tumor diameter at operation was 62.6 mm (range, 49–80 mm) and the median tumor volume was 94.3 mL (range, 35–254.1 mL). Operative time was 168.5 min (range, 116–245 min), intraoperative blood loss was 23.3 mL (range, 5–50 mL) and the length of postoperative hospital stay was 5.7 days (range, 4–11 days). There was a significant difference (p<0.05) between the RLR and OLR groups for: age (93.1 months vs. 137.6 months), maximum tumor diameter (62.6 mm vs. 98 mm), tumor volume (94.3 mL vs. 496.2 mL), operative time (168.5 min vs. 281.4 min), blood loss (23.3 mL vs. 288.7 mL), and length of postoperative hospital stay (5.7 days vs. 9.5 days). There was a borderline significant association between surgical approach (RLR/OLR) and fluctuation in the magnitude of Alanine aminotransferase (ALT) (odds ratio=0.004, 95% confidence interval: 0.000 to 1.096, p=0.05).

Conclusions

Our initial experience suggested that RLR for hepatic FNH in children was both feasible and safe. Tumors in the RLR group were significantly smaller than the OLR group: it proved possible to excise tumors larger than 250 mL in volume.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Focal nodular hyperplasia (FNH) is a common benign liver tumor. Traditionally, open surgery has been the standard treatment. In recent years, robotic surgery has demonstrated advantages such as enhanced surgical precision and reduced invasiveness and has been increasingly applied in the management of pediatric tumors. However, evidence supporting its efficacy in the treatment of FNH in children remains limited.

WHAT THIS STUDY ADDS

• This study provides a comprehensive comparison of clinical outcomes between robotic liver resection (RLR) and open liver resection in pediatric patients with FNH. The findings suggest that RLR may provide significant advantages in terms of reduced surgical invasiveness and faster recovery for pediatric FNH patients. Moreover, this study indicates that RLR may have a positive impact on the recovery of postoperative liver function.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• This study reinforces the feasibility and safety of RLR for pediatric FNH. For practice, this study supports the adoption of RLR in pediatric surgery, provided that the necessary skills and technology are available. It emphasizes the need for further research to evaluate the long-term benefits of RLR, particularly in terms of the recovery of liver function and overall patient outcomes.

Introduction

Focal nodular hyperplasia (FNH) is the second most frequent benign liver lesion after hepatic hemangioma, which is often found incidentally with imaging.¹ FNH accounts for 8% of all primary liver lesions and approximately 25% of benign lesions.² Contrast-enhancement CT can assist diagnosis and aid surgical resection decision making for patients with symptoms or gradual enlargement of the lesion.¹

Laparoscopic liver surgery has been increasingly adopted in recent years. With the development of minimally invasive surgical (MIS) technology the invention and application of robotic surgical systems have been gradually utilized for liver tumors with good outcomes.^{3 4}

However, there is a paucity of published data regarding robotic resection of pediatric liver tumors, especially FNH. The aim of the present study was to verify the feasibility and safety of robotic liver resection (RLR) for FNH, present our initial experiences and discuss technical issues. We will also compare surgical outcomes between RLR and open liver resection (OLR).

Patients and Methods

Patients

We performed a retrospective study of patients who underwent surgical treatment for FNH between January 2020 and June 2024. All patients underwent preoperative ultrasound and enhanced CT scanning. Tumor volumes were calculated using the 3-dimensional (3D) Slicer toolkit and units expressed in cubic millimeters (mL).

Patients were divided into two groups according to the operative types: RLR and OLR groups. Group assignment was based on: (1) Parental preference, including economic considerations (as robotic surgery required partial self-payment) and acceptance of innovative techniques. (2) Assessment of surgical feasibility assessment by the operating surgeon, considering tumor characteristics. Specifically this was evaluated based on tumor location, exophytic growth pattern and size. Data collection included demographic characteristics, tumor size and volume, operative time, intraoperative blood loss, postoperative hospital length of stay and perioperative complications.

Surgical procedures

Robotic liver resection

We used the da Vinci Xi Surgical System with four ports to perform the surgery: three 8 mm robotic arm ports and one 5 mm auxiliary port (figure 1A).

Figure 1. Port placement for da Vinci robotic-assisted surgery. T: Surface projection of the target lesion. Ports 1–3: 8 mm robotic trocars. Port 1: connected to robotic arm 1 (bipolar forceps). Port 2: main camera port. Port 3: connected to robotic arm 3 (monopolar hook, electric scissors, or harmonic scalpel). Port 4: 5 mm assistant port (flexible instrumentation: graspers, Hem-o-lok clip applier, suction, or scissors).

Following endotracheal intubation under general anesthesia and establishment of a pneumoperitoneum via the umbilicus the patient was placed in the supine position. An 8 mm da Vinci trocar was inserted into the umbilical incision to serve as the port for the 3D camera. Two additional 8 mm trocars were placed in the left upper abdomen and right lower abdomen, and a 5 mm laparoscopic trocar was placed in the left abdomen as an auxiliary port for patients with right hepatic FNH (figure 1B,C).

The falciform ligament and triangular ligament were dissected to improve exposure (figure 2A). Parenchymal transection was carried out using a clamp crushing technique with bipolar precise forceps and an ultrasonic scalpel (figure 2B). Intermittent Pringle maneuver was performed in cycles of 10/5/15/5/15 min (clamping/reperfusion) (figure 2C). The neoplasm was gradually separated from the edge of the tumor, the blood vessels were ligated using Hem-o-lock clips and then the tumor was removed, as seen in figure 2D. Hemostasis on the transection surface was completed using electrocoagulation. The specimen was placed into an endobag and extracted through the enlarged umbilical port or a Pfannenstiel incision. A closed suction drain was placed adjacent to the resection area.

Figure 2. Intraoperative images in da Vinci robotic-assisted enucleation of focal nodular hyperplasia. (A) Dissection of the falciform ligament to expose the hepatic dome. (B) Circumferential dissection along the tumor pseudocapsule, preserving parenchymal integrity. (C) Pringle maneuver application. (D) Selective ligation of the central feeding artery.

Open liver resection

The patient was placed in the supine position with the upper abdomen elevated. An L-shaped incision was made as shown in figure 3. First, complete release of hepatic ligaments (falciform, triangular, and coronary) was performed and subsequent manual delivery of the liver to achieve optimal tumor exposure. The transection plane beyond the tumor pseudocapsule was maked. Subsequently, intermittent Pringle maneuver harmonic scalpel and electric knife were used to completely resect the tumor. Hemostasis was applied to the wound and visible pipeline structures were clamped with titanium clips. Finally, abdominal layers were closed.

Figure 3. Postoperative L-shaped incision following open liver resection.

Statistical analysis

For statistical analysis, we utilized both IBM SPSS software (V.26.0.0) and RStudio (V.4.1.3). Two independent grouping strategies were employed: (1) by surgical approach (RLR vs. OLR) and (2) by the magnitude ALT fluctuation (high-difference vs. low-difference), where ΔALT was calculated as (postoperative peak ALT−preoperative ALT) and standardized using composite-score normalization prior to grouping. Continuous variables were reported as median and range, and the Mann-Whitney U test was used to compare data from two groups. Categorical variables were expressed by counts and percentages and comparison between the two groups used Pearson’s χ² and Fisher’s exact tests. We performed univariate followed by multivariate logistic regression, with surgical method (RLR/OLR) included as an independent variable alongside other clinically relevant covariates. P value less than 0.05 was considered statistically significant.

Results

Robotic liver resection

There were 12 patients (4 boys and 8 girls) underwent RLR. Median age was 93.1 months (range, 28–134 months) with a median weight of 32.4 kg (range, 9.7–80 kg). Median maximum tumor diameter at operation was 62.6 mm (range, 49–80 mm) and the median tumor volume was 94.3 mL (range, 35–254.1 mL). Operative time was 168.5 min (range, 116–245 min), intraoperative blood loss was 23.3 mL (range, 5–50 mL) and the length of postoperative hospital stay was 5.7 days (range 4–11 days).

Open liver resection

Eight patients including 5 boys and 3 girls received OLR. Median age was 137.6 months (range, 82–180 months) with a median weight of 45.3 kg (range, 19.8–78.5 kg). Median maximum tumor diameter at operation was 98 mm (range, 63–139 mm) and the median tumor volume was 496.2 mL (range, 95.7–1130.7 mL). Operation time was 281.4 min (range, 95–410 min) the intraoperative blood loss was 288.7 mL (range, 30–1000 mL) and the length of postoperative hospital stay was 9.5 days (range 5–15 days).

Comparison of RLR and OLR groups

Despite significant baseline differences in tumor size (p=0.003) and volume (p=0.002), the comparative analysis provides insights into evolving technical capabilities. All patients successfully completed the procedure successfully and there was no conversion in the RLR group. Surgical margins were negative in all cases. Median age, maximum diameter, and tumor volume were smaller in the RLR group compared with the OLR group (p<0.05). The RLR group had a significantly shorter operative time, less intraoperative blood loss, and shorter length of postoperative stay (p<0.05) as seen in table 1.

Table 1. Clinical parameters of the patients grouped according to surgical approach.

Variables	RLR group (n=12)	OLR group (n=8)	P value
Patient characteristics
Gender			0.170
Male	4 (33.3%)	6 (75.0%)
Female	8 (66.7%)	2 (25.0%)
Age (month)	93.1±33.0	138.0±31.7	0.008
Weight (kg)	32.4±18.3	45.3±17.2	0.129
Tumor and operative characteristics
Maximum diameter (cm)	6.0 (5.3, 7.2)	9.8 (7.8, 11.8)	0.003
Volume (mL)	80.6 (50.1, 138.0)	315.0 (208.0, 767.0)	0.002
Location by Couinaud segments			0.212
I	0 (0.0%)	1 (12.5%)
II	0 (0.0%)	1 (12.5%)
III	2 (16.7%)	0 (0.0%)
IV	4 (33.3%)	2 (25.0%)
V	3 (25.0%)	0 (0.0%)
Multiple	3 (25.0%)	4 (50.0%)
Operative time (min)	160.0 (140.0, 186.0)	309.0 (172.0, 394.0)	0.049
Intraoperative blood loss (mL)	20.0 (16.2, 27.5)	90.0 (50.0, 550.0)	0.001
Postoperative hospital stay (days)	4.5 (4.0, 7.0)	9.0 (6.8, 11.8)	0.017
Complications			0.325
Yes	5 (41.7%)	1 (12.5%)
No	7 (58.3%)	7 (87.5%)
ALT difference group			0.650
High	5 (41.7%)	5 (62.5%)
Low	7 (58.3%)	3 (37.5%)

Data was presented as n (%) or mean±standard deviation (sd) or median (range).

ALT, alanine aminotransferase; OLR, open liver resection; RLR, robotic liver resection.

In the OLR group, one patient had massive intraoperative hemorrhage (over 1000 mL).In this patient the tumor was completely resected after massive blood transfusion and no life-threatening conditions occurred post-operatively. The remaining patients in both groups recovered well without any other complications. Median follow-up time did not differ significantly between groups (95 (7–161) months in RLR group vs. 46 (8–183) months in OLR group; p=0.400). None of the patients had local recurrence during follow-up.

Comparison of ALT high-difference low-difference groups

Given the statistically significant group difference in composite score (p<0.001) presented in table 2, we considered this stratification clinically meaningful and subsequently performed both univariate and multivariate regression analyses using this grouping method.

Table 2. Clinical parameters of patients grouped according to magnitude of ALT fluctuation.

Variables	High group (n=10)	Low group (n=10)	P value
Patient characteristics
Gender			0.655
Male	6 (60.0%)	4 (40.0%)
Female	4 (40.0%)	6 (60.0%)
Age (month)	106±50.8	116 (22.9)	0.558
Weight (kg)	38.8±25.9)	36.2 (7.29)	0.770
Tumor and operative characteristics
Maximum diameter (cm)	6.65 (5.18, 9.12)	7.30 (6.35, 8.00)	0.405
Volume (mL)	131 (47.5, 276)	141 (88.3, 241)	0.545
Locations by Couinaud segments			0.413
I	1 (10.0%)	0 (0.0%)
II	0 (0.0%)	1 (10.0%)
III	0 (0.0%)	2 (20.0%)
IV	3 (30.0%)	3 (30.0%)
V	1 (10.0%)	2 (20.0%)
Multiple	5 (50.0%)	2 (20.0%)
Operative time (min)	166.0 (142.0, 340.0)	180.0 (149.0, 221.0)	0.910
Intraoperative blood loss (mL)	50.0 (20.0, 320)	25.0 (8.75, 50.0)	0.122
Postoperative hospital stay (day)	7.5 (4.3, 9.8)	6.5 (4.3, 7.0)	0.702
Complications			0.628
Yes	2 (20.0%)	4 (40.0%)
No	8 (80.0%)	6 (60.0%)
Method			0.650
RLR	5 (50.0%)	7 (70.0%)
OLR	5 (50.0%)	3 (30.0%)
Composite Score	−0.16 (−0.06, −0.17)	−0.41 (−0.54, −0.39)	<0.001

Data was presented as n (%), mean±sd, or median (range).

ALT, alanine aminotransferase; OLR, open liver resection; RLR, robotic liver resection.

Univariate logistic regression revealed that none of the variables reached statistical significance. On multivariate analysis adjusting for tumor maximum diameter, volume, complications, Couinaud segment locations, and surgical approach (RLR/OLR), only surgical approach showed a borderline significant association with ALT the magnitude of fluctuation (odds ratio (OR)=0.004, 95% confidence interval (CI): 0.000 to 1.096, p=0.050). This suggests a potentially protective effect of RLR against hepatic injury. There was a non-significant trend for tumor maximum diameter (OR=4.491, 95% CI: 0.759 to 26.568, p=0.090), which may warrant further investigation in larger cohorts tables3 4.

Table 3. Univariable analysis of hepatic injury predictors.

Variables	OR	95%CI	P value
Gender	2.25	0.38 to 14.66	0.374
Age	1.01	0.98 to 1.03	0.534
Weight	0.99	0.94 to 1.04	0.753
Maximum diameter	1.13	0.79 to 1.69	0.528
Volume	1.00	1.00 to 1.00	0.746
Locations by Couinaud segments	0.74	0.36 to 1.38	0.359
Operative time	1.00	0.99 to 1.00	0.483
Intraoperative blood loss	0.99	0.96 to 1.00	0.390
Postoperative hospital stay	0.89	0.66 to 1.16	0.401
Complications	0.38	0.04 to 2.61	0.337
Method	0.43	0.06 to 2.60	0.365

CI, confidence interval; OR, odds ratio.

Table 4. Multivariable analysis of hepatic injury predictors.

Variables	OR	95%CI	P value
Maximum diameter	4.491	0.759 to 26.568	0.090
Volume	0.997	0.987 to 1.007	0.550
Complications	0.143	0.006 to 3.170	0.220
Locations by Couinaud segments	0.452	0.169 to 1.206	0.110
Method	0.004	0.000 to 1.096	0.050

CI, confidence interval; OR, odds ratio.

Discussion

To the best of our knowledge this cohort study is the first series of consecutive pediatric and adolescent patients undergoing full RLR for FNH. Our results demonstrate that RLR for FNH is feasible and safe, with significant advantages over OLR including: less intraoperative blood loss (20 vs. 90 mL, p=0.001) and shorter hospitalization (4.5 vs. 9 days, p=0.017) . Whilst the smaller tumor size in the RLR group may explain these results, the observed advantages are primarily attributable to the application of MIS techniques. This aligns with reported outcomes in other surgical contexts.5,7 To mitigate potential confounding by tumor size, we developed a novel biomarker—the ALT differential (ΔALT, calculated as (peak postoperative ALT)−(preoperative ALT))—and restratified patients using surgical approach as the dependent variable in our regression models. Multivariate analysis revealed surgical approach (RLR vs. OLR) as an independent predictor (OR=0.004, 95% CI:0.000 to 1.096, p=0.05), whereas tumor size demonstrated only a non-significant trend (OR=4.491, 95% CI:0.759 to 26.568, p=0.090).

Minimally invasive liver resection is a challenging surgical procedure due to the complex architecture and rich vascularity of the liver, increasing the risk of life-threatening complications, both during and after surgery.⁸ A clear surgical field is the key to the success of MIS. Characteristic features of FNH namely solitary subcapsular lesions with well-defined borders, facilitates minimally invasive resection. In addition, the clinical advantages of robotic surgical systems, including image stability, 3D view, flexible instruments, the abolition of physiological tremor, better comfort for the surgeon and shorter learning curve,^{9 10} increase the success rate of MIS.

Appropriate treatment of FNH is controversial and the surgical indication is uncertain. Surgical resection is considered for progressive growth or large lesions>10 cm in size with symptoms of compression, increased risk of hemorrhage due to trauma, or occurrence of a complications.¹ However, surgery for FNH lesions was more frequently performed in children because of symptoms, increasing size or inability to confidently rule out malignancy.¹¹ The relatively large median tumor size observed in children (6 cm) may increase the likelihood of symptoms.¹¹

Valuable experience has been gained from handling dozens of cases at our center, enabling the successful procedure of RLR. To ensure optimal surgical outcomes, the following technical points require attention: (1) The FNH boundary often exhibits a false capsule. Irregular resection along this pseudo-capsule significantly reduces intraoperative bleeding. (2) FNH typically contains a dominant feeding vessel at its core, which should be localized preoperatively via imaging and meticulously ligated prior to parenchymal transection.(3) A standardized Pringle maneuver using 6–8 Fr catheters with Hem-o-lok clips should be used: 10 min occlusion, 5 min reperfusion, 15 min occlusion, 5 min reperfusion, 15 min occlusion. During the reperfusion phases we inspect for minor vessel/bile duct leakage and apply targeted electrocoagulation to active bleeders. (4) Compared with open surgery requiring complete ligament release, the robotic approach allows targeted mobilization based on tumor location (table 5). (5) vascular embolization can be performed before surgery if necessary.

Table 5. Ligament release and traction methods required for liver tumor resection by segment.

Tumor segment	Required ligament release	Traction method
III (exophytic)	None	None
II/VII	Triangular ligament	Lateral retraction
IVa/VIII	Falciform ligament	Inferior traction
I/IVb/V/VI	Gallbladder fossa suspension	Anterior wall traction

The methods recommended in this table are not absolute and should be adapted according to intraoperative findings.

This study has several limitations. First, the RLR group inherently contained smaller tumors located in more accessible segments. Selection is significantly influenced by a combination of parental socioeconomic factors and technical confidence of surgeon. Second, the non-randomized grouping reflects real-world clinical decision-making but resulted in significantly larger tumors in the OLR group (median 9.75 vs. 6.0 cm), potentially magnifying outcome differences. Third, due to the retrospective nature of our study, artificial group adjustment was not feasible. Whilst the smaller surgical space in younger patients brings challenges to MIS, it underscores the advantages of robotic surgery. Besides, surgeon’s focusing attention to prevent conversion to open surgery, along with refined surgical techniques, plays a critical role in ensuring successful outcomes. Tumor-to-liver volume ratio may reduce the impact of patient weight and age and can substitute for tumor size as a reference index to some extent.

In future studies we plan to use a multicenter approach to examine differences in robot models and variations in surgical data across different centers. Baseline comparability will be enhanced with a larger overall cohort and achieve balanced cohorts through propensity score matching. Furthermore, we will adopt tumor-to-liver volume ratio (TLVR) as a standardized measure (TLVR = (tumor volume)/(total liver volume−tumor volume)) to refine outcome metrics.

In conclusion, RLR for hepatic FNH in children was feasible and safe. Tumors in the RLR group were significantly smaller, but it was possible to excise tumors that were larger than 250 mL in volume.

Data availability statement

Data are available on reasonable request.