Laparoscopic microwave ablation for hepatic hemangiomas with different locations: A comparative study

Abstract

BACKGROUND

Hepatic hemangioma (HH) is the most common benign liver tumor. However, the relative impact of lesion location and tumor size on postoperative outcomes following trans-arterial embolization (TAE) combined with laparoscopic microwave ablation (LMWA) remains incompletely defined. In addition, heat-sink-related effects during ablation may contribute to perioperative complications, particularly in large or deeply located lesions.

AIM

To evaluate the association of lesion location and tumor size with short-term postoperative outcomes following TAE + LMWA under a standardized procedural protocol.

METHODS

This retrospective study included 152 patients with large or giant HH (GHH) who underwent TAE + LMWA. Lesions were categorized by location as subcapsular or parenchymal, and by size as large or GHHs. Perioperative parameters, postoperative complications, inflammatory markers, and hepatic and renal function indices were compared between groups. All patients were treated using a standardized ablation protocol incorporating a gas-blood outflow channel.

RESULTS

Lesion location was not significantly associated with most postoperative laboratory indices, although postoperative fever occurred more frequently in liver parenchymal than in liver subcapsular lesions. Tumor size demonstrated a more pronounced association with postoperative outcomes. Compared with large HH, GHH were associated with longer ablation times, greater intraoperative blood loss, delayed gastrointestinal recovery, and higher incidences of postoperative fever and hypoproteinemia.

CONCLUSION

Tumor size appears to be an important factor influencing short-term postoperative recovery following TAE + LMWA, whereas lesion location shows a limited and selective association. Under a standardized protocol incorporating a gas-blood decompression channel, a low incidence of major heat-sink-related complications was observed. However, given the retrospective design and absence of a control group, the potential role of this technical modification should be regarded as exploratory and requires prospective validation.

Key Words: Hepatic hemangioma; Microwave ablation; Lesion location; Heat-sink effect; Trans-arterial embolization; Postoperative outcomes

Core Tip: Laparoscopic microwave ablation (LMWA) patients have less surgical bleeding and recover faster. Symptom improvement rates are similar between the LMWA and large hepatic hemangiomas resection groups, with no significant difference. The LMWA group has a lower rate of postoperative complications. Transient liver function changes after LMWA usually resolve within a week. Transient liver function changes may result from heat damage and tissue thermal coagulation.

INTRODUCTION

Hepatic hemangioma (HH) is the most common benign hepatic tumor and is increasingly detected incidentally with the widespread use of abdominal imaging[1]. Although most HH are asymptomatic and require no intervention, large lesions may cause compressive symptoms, rapid growth, or complications such as Kasabach-Merritt syndrome, necessitating active treatment[2,3].

Minimally invasive thermal ablation, particularly microwave ablation (MWA), has emerged as an alternative to surgical resection for selected patients with HH[4]. Compared with hepatectomy, MWA offers the advantages of reduced invasiveness and faster recovery[5,6]. However, the heat-sink effect remains a major concern, especially in large or deeply located lesions, where thermal energy dissipation and intravascular hemolysis may lead to complications such as hemoglobinuria and acute kidney injury[7].

To address these limitations, trans-arterial embolization (TAE) has been increasingly combined with laparoscopic MWA (LMWA)[8,9]. By reducing tumor perfusion, TAE may improve ablation efficiency and decrease intraoperative bleeding[10]. Based on this strategy, our center further standardized a technical modification involving the construction of a gas-blood outflow channel during ablation, intended to facilitate intratumoral decompression. Nevertheless, the clinical impact of this approach has not been systematically evaluated.

Despite the growing use of TAE + LMWA, the influence of tumor anatomical location on postoperative recovery remains insufficiently defined. Liver subcapsular lesions may allow more effective heat dissipation, whereas liver parenchymal lesions may retain thermal energy and provoke greater systemic responses. At the same time, tumor size - commonly classified as large HHs (LHHs; 5-10 cm) or giant HHs (GHHs; > 10 cm) - is widely recognized as a key determinant of procedural complexity and postoperative outcomes[11]. However, existing studies rarely distinguish or simultaneously evaluate the relative contributions of lesion size and location.

Therefore, this retrospective study aimed to assess postoperative outcomes following a standardized TAE + LMWA protocol in patients with large and GHH, with a specific focus on disentangling the effects of tumor size and anatomical location. By comparing perioperative parameters, laboratory changes, and postoperative complications across liver subcapsular vs liver parenchymal lesions and LHH vs GHH groups, we sought to provide clinically relevant insights to inform the optimization of minimally invasive treatment strategies for HH (Figure 1).

Figure 1 Schematic diagram illustrating microwave ablation of a hepatic hemangioma. The image shows a microwave electrode inserted into the hemangioma and the generated heat distribution. It depicts heat transfer along the portal vein and heat accumulation within surrounding blood. During the ablation process, the heat from the vessel’s spreads to other liver segments, causing thermal damage to the liver and the accumulation of heat leading to intravascular hemolysis. A channel pre-drilled near the ablation point, known as the gas-blood outflow channel, is conducive to the outflow of hot oxygen bubbles and blood. The image created by FigDraw (Supplementary material).

MATERIALS AND METHODS

Study design and patient selection

This retrospective study was approved by the Institutional Ethics Committee of the Chinese PLA General Hospital (Approval No. S2025-024-01). Clinical data from consecutive patients diagnosed with HH who underwent TAE combined with LMWA (TAE + LMWA) between November 2021 and August 2024 were reviewed.

Patients were eligible for inclusion if they met the following criteria: (1) Radiologically confirmed HH on contrast-enhanced computed tomography or magnetic resonance imaging; (2) A maximal tumor diameter greater than 5 cm; and (3) Complete perioperative clinical and imaging records. Patients were excluded if they had concomitant benign or malignant hepatic diseases, significant systemic comorbidities precluding surgery, prior hepatic surgery, pregnancy, incomplete baseline or perioperative data, or had received management strategies other than TAE + LMWA. After application of these criteria, a total of 152 patients were included in the final analysis (Figure 2).

Figure 2 A total of 180 patients who underwent trans-arterial embolization + laparoscopic microwave ablation surgery for hepatic hemangioma were included in this study. After screening, 28 patients who did not meet the criteria were excluded. Among them, 12 had other benign liver diseases, and the other 16 had partial data missing. Ultimately, 152 patients were randomly assigned to the giant hepatic hemangioma (HH) group (72 cases) and the large HH group (80 cases). Specifically, the giant HH group included 55 cases of liver subcapsular of the liver and 17 instances of lesions within the liver parenchymal. In the large HH group, there were 63 cases of liver subcapsular of the liver and 17 instances of lesions within the liver parenchymal. This study aims to conduct a comparative analysis of the liver and kidney functions, coagulation functions, inflammatory indicators, perioperative indicators, and complications of these two groups of patients after surgery. TAE: Trans-arterial embolization; LMWA: Laparoscopic microwave ablation; GHH: Giant hepatic hemangioma; LHH: Large hepatic hemangioma.

Lesions were categorized according to both anatomical location and maximal diameter. Subcapsular HH was defined as a lesion adjacent to or involving the liver capsule, whereas parenchymal HH was defined as a lesion located entirely within the hepatic parenchyma without capsular contact. Tumors measuring 5-10 cm were classified as LHH, and those larger than 10 cm were classified as GHH. All classifications were determined independently by two senior radiologists based on preoperative imaging, with discrepancies resolved by consensus.

TAE + LMWA procedure

All patients first underwent selective TAE approximately 48-72 hours before ablation. Using the Seldinger technique, a 5F catheter was advanced via the right femoral artery, and angiography was performed to identify feeding arteries originating from the hepatic artery, superior mesenteric artery, or accessory sources. Super selective catheterization of the feeding vessels was then achieved using a microcatheter, followed by slow infusion of an iodized oil-gelatin sponge emulsion until satisfactory tumor devascularization was confirmed. This step was performed to reduce intraoperative bleeding and attenuate intratumoral perfusion before MWA.

LMWA was carried out under general anesthesia using a standardized three-port approach. After the creation of pneumoperitoneum, the lesion was exposed and evaluated for optimal puncture trajectory. Ablation was performed using a 2450 MHz dual-source microwave system equipped with internally cooled applicators. Ablation parameters, including power and duration, were adjusted based on tumor size and intraoperative response.

To improve the efficiency and safety of the procedure, a gas-blood decompression channel was established at the beginning of the ablation. A MWA needle was first inserted shallowly into the lesion to provide a channel for the egress of thermally induced gas and intratumoral blood, thereby facilitating pressure release. A second ablation needle was then inserted at a planned ablation point, allowing thermal gas and blood to escape through the decompression channel during heating. The channel was sequentially repositioned as ablation progressed across the lesion to maintain adequate decompression and reduce thermal accumulation. Upon completion, the ablation field and puncture sites were irrigated and inspected to ensure hemostasis and absence of bile leakage (Figure 3).

Figure 3 Establishment of the gas-blood outflow channel. A: Prior to the initiation of each microwave ablation cycle, a gas-blood outflow channel is created using a standard microwave ablation needle. The specific depth of this channel is determined based on the maximum diameter of the hemangioma as measured from preoperative imaging, rather than a fixed length, to ensure tailored channel placement according to individual tumor characteristics (the depth is approximately half of the maximum diameter of the tumor); B: The blood flowing out of the channel, and the thermal oxygen bubbles caused by the heat conduction effect are observed.

Postoperative evaluation

All patients underwent standardized postoperative monitoring, including daily laboratory assessments of inflammatory markers, hematologic indices, hepatic and renal function, and coagulation parameters. Particular attention was paid to the detection of hemoglobinuria and urine output within the first 24 hours after ablation, given the known association between thermal injury and intravascular hemolysis. Acute kidney injury was diagnosed according to the Kidney Disease: Improving Global Outcomes criteria[12,13].

Postoperative fever was defined as a body temperature ≥ 38.0 °C on at least two occasions separated by at least 24 hours. Postoperative anemia was defined as a hemoglobin level < 90 g/L, and hypoproteinemia was defined as a serum albumin level < 30 g/L within 7 days after the procedure. Inflammatory markers evaluated included white blood cell count, neutrophil percentage, and D-dimer concentration[14].

Postoperative gastrointestinal recovery was assessed by the time to first flatus, and postoperative length of hospital stay was recorded. Before discharge, all patients underwent contrast-enhanced computed tomography or magnetic resonance imaging to evaluate the completeness of ablation.

The primary outcomes of this study were postoperative complications, particularly fever and major heat-sink-related events, including hemoglobinuria and acute kidney injury. Secondary outcomes included perioperative parameters (ablation time and intraoperative blood loss), postoperative laboratory changes, and recovery indicators.

Statistical analysis

Statistical analyses were performed using SPSS software (version 23.0; IBM Corp, Armonk, NY, United States). Continuous variables were expressed as mean ± SD or median (interquartile range,) as appropriate. Data distribution was assessed using the Shapiro-Wilk test in combination with visual inspection of histograms. Comparisons between two independent groups were conducted using the independent-sample t test for normally distributed variables and the Mann-Whitney U test for non-normally distributed variables. Categorical variables were compared using the χ² test or Fisher’s exact test, as appropriate.

Given the retrospective design and the limited sample size in certain subgroups, particularly for lesion location-based analyses, no multivariable regression analysis was performed. All statistical tests were two-sided, and a P-value < 0.05 was considered statistically significant. Patients with incomplete baseline or perioperative data were excluded from the final analysis.

RESULTS

Baseline characteristics

A total of 152 patients met the inclusion criteria, including 80 with LHH and 72 with GHH. Among them, 118 lesions were located liver subcapsular and 34 within the liver parenchymal. Baseline demographic characteristics, lesion distributions, and preoperative laboratory parameters were comparable between liver subcapsular and liver parenchymal groups, as well as between LHH and GHH groups (all P > 0.05). These findings indicated that the groups were comparable for subsequent analyses (Tables 1 and 2).

Table 1 Baseline comparison table in giant hepatic hemangioma group, mean ± SD/n (%)/median (interquartile rage).

	Liver subcapsular (n = 55)	Liver parenchymal (n = 17)	t/Z/χ2	P value
Age (years)	48.00 ± 9.12	47.47 ± 7.54	-0.217	0.829
Sex			0.872	0.350
Female	39 (72.72)	10 (58.82)
Male	16 (27.27)	7 (41.18)
BMI	24.26 ± 3.10	25.56 ± 3.48	1.469	0.146
Size (cm)	10.83 (10.33-13.80)	10.50 (10.11-10.90)	-1.711	0.087
Location of main lesion			3.535	0.060
Left lobe	18 (30.91)	1 (5.88)
Right lobe	37 (69.09)	16 (94.12)
Hepatitis B virus	3 (5.45)	0 (0)
Reasons for operation			0.062	0.804
Symptomatic	10 (18.2)	2 (11.8)
Enlargement	45 (81.8)	15 (88.2)
Number of lesions			2.103	0.349
1	19 (34.5)	8 (47.1)
2	9 (16.4)	4 (23.5)
Multiple	27 (49.1)	5 (29.4)
Inflammatory
WBC (109/L)	5.34 ± 1.82	5.06 ± 1.30	-0.581	0.563
Neutrophil (%)	57.60 (49.43-62.25)	55.40 (49.28-63.80)	-1.048	0.295
Hb (g/L)	127.65 ± 21.51	135.29 ± 17.14	1.337	0.186
Liver function
TB (μmol/L)	12.54 ± 4.56	12.41 ± 5.72	-0.101	0.920
ALT (U/L)	15.30 (10.25-28.00)	24.5 (11.75-42.75)	-1.533	0.125
AST (U/L)	15.50 (14.00-19.75)	20.50 (16.25-26.50)	-1.920	0.055
GGT (U/L)	18.00 (11.25-33.75)	37.00 (17.75-54.50)	-1.360	0.174
Coagulation
PT (seconds)	11.62 ± 0.72	11.69 ± 1.30	0.299	0.766
D-D (μg/L)	292.00 (145.75-828.25)	242.50 (141.00-526.50)	-0.945	0.345
Renal function
Cr (μmol/L)	65.30 ± 13.40	67.41 ± 10.76	0.591	0.556
BUN (mg/dL)	4.86 ± 1.21	4.79 ± 1.12	-0.207	0.837

BMI: Body mass index; WBC: White blood cell; Hb: Hemoglobin; TB: Total bilirubin; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; GGT: Gamma-glutamyl transferase; PT: Prothrombin time; D-D: D-dimer; Cr: Creatinine; BUN: Blood urea nitrogen.

Table 2 Baseline comparison table in large hepatic hemangioma group, mean ± SD/n (%)/median (interquartile rage).

	Liver subcapsular (n = 63)	Liver parenchymal (n = 17)	t/Z/χ2	P value
Age (years)	46.92 ± 10.762	46.29 ± 11.22	-0.211	0.833
Sex			0.000	1.000
Male	14 (22.2)	4 (23.5)
Female	49 (77.8)	13 (76.5)
BMI	24.44 ± 3.07	24.68 ± 4.58	-0.731	0.803
Size (cm)	7.20 ± 1.24	6.94 ± 1.18	0.782	0.436
Location of main lesion			1.396	0.237
Left lobe	14 (22.2)	1 (5.88)
Right lobe	49 (77.8)	16 (94.12)
Number of lesions			0.720	0.698
1	25 (39.7)	7 (41.2)
2	10 (15.9)	4 (23.5)
Multiple	28 (44.4)	6 (35.3)
Hepatitis B virus	1 (1.59)	0 (0)
Reasons for operation			0.990
Symptomatic	5 (7.9)	2 (11.8)
Enlargement	58 (92.1)	15 (88.2)
Inflammatory
WBC (109/L)	5.26 ± 1.60	5.75 ± 1.61	-1.126	0.297
Neutrophil (%)	128.22 ± 15.58	134.12 ± 12.65	-1.436	0.155
Hb (g/L)	51.40 (47.18-59.85)	50.95 (46.08-61.18)	-0.112	0.911
Liver function
TB (μmol/L)	14.08 ± 13.41	13.21 ± 6.10	0.258	0.797
ALT (U/L)	12.00 (10.00-15.75)	13.30 (11.00-16.75)	-0.807	0.420
AST (U/L)	15.00 (13.00-16.75)	15.00 (12.88-19.00)	-0.605	0.545
GGT (U/L)	15.00 (13.00-19.50)	25.50 (14.00-31.50)	-1.434	0.152
Coagulation
PT (seconds)	11.54 ± 0.59	11.88 ± 1.16	-1.656	0.102
D-D (μg/L)	155.50 (133.25-214.50)	183.00 (88.75-545.00)	-0.957	0.338
Renal function
Cr (μmol/L)	65.30 ± 11.22	62.35 ± 14.42	0.902	0.370
BUN (mg/dL)	4.81 ± 1.42	4.86 ± 1.49	-0.130	0.897

BMI: Body mass index; WBC: White blood cell; Hb: Hemoglobin; TB: Total bilirubin; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; GGT: Gamma-glutamyl transferase; PT: Prothrombin time; D-D: D-dimer; Cr: Creatinine; BUN: Blood urea nitrogen.

Postoperative outcomes by lesion location

Among patients with LHH, postoperative inflammatory markers, hepatic and renal function indices, coagulation profiles, and perioperative recovery variables did not significantly differ between liver subcapsular and liver parenchymal lesions (P > 0.05). The only exception was a slightly higher intraoperative blood loss in the subcapsular lesion group, although the absolute difference remained clinically modest.

Similarly, in patients with GHH, no relevant differences were observed between the liver subcapsular and liver parenchymal lesions subgroups across these laboratory indices or perioperative parameters. Overall, lesion location - liver subcapsular vs liver parenchymal - was not associated with significant differences in postoperative laboratory recovery (Tables 3 and 4).

Table 3 Compare the postoperative recovery conditions of the liver subcapsular group and the liver parenchymal group in the large hepatic hemangioma group, mean ± SD/median (interquartile rage).

	Liver subcapsular (n = 63)	Liver parenchymal (n = 17)	t/Z/χ2	P value
Inflammatory
WBC (109/L)	9.61 ± 2.83	9.14 ± 2.84	0.607	0.545
Neutrophil (%)	83.13 ± 6.67	80.85 ± 8.96	1.162	0.249
Liver function
TB (μmol/L)	31.17 ± 14.06	31.41 ± 10.21	-0.066	0.948
ALT (U/L)	281.70 (180.95-452.83)	328.95 (191.15-596.55)	-0.747	0.455
AST (U/L)	413.30 (251.15-617.20)	318.50 (221.43-705.93)	-0.006	0.955
GGT (U/L)	24.50 (17.25-39.75)	43.50 (22.75-68.50)	-1.808	0.071
Renal function
Cr (μmol/L)	53.95 (47.65-64.58)	58.50 (50.43-65.65)	-0.968	0.333
BUN (mg/dL)	3.30 (2.60-3.98)	3.65 (2.82-4.25)	-0.341	0.733
Coagulation
PT (seconds)	12.18 ± 0.64	12.70 ± 0.56	-3.026	0.003
D-D (μg/L)	2026 (1158-2534)	2403 (2001-2645)	-1.607	0.108
Perioperative
Ablation time (minutes)	32.82 ± 9.50	32.36 ± 9.20	0.167	0.868
Clinical stay (day)	3.00 (3.00-3.00)	3.00 (3.00-4.00)	-0.721	0.471
Exhaust time (day)	1.00 (1.00-2.00)	1.50 (1.00-2.00)	-0.431	0.667
Surgical blood loss (mL)	10.00 (9.00-20.00)	5.00 (5.00-10.00)	-2.053	0.040
Spent	61352 (47405-65220)	56366 (47355-60906)	-0.489	0.625

WBC: White blood cell; TB: Total bilirubin; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; GGT: Gamma-glutamyl transferase; Cr: Creatinine; BUN: Blood urea nitrogen; PT: Prothrombin time; D-D: D-dimer.

Table 4 Compare the postoperative recovery conditions of the liver subcapsular group and the liver parenchymal group in the giant hepatic hemangioma group, mean ± SD/median (interquartile rage).

	Liver subcapsular (n = 55)	Liver parenchymal (n = 17)	t/Z/χ2	P value
Inflammatory
WBC (109/L)	9.98 ± 3.58	9.58 ± 3.03	0.408	0.685
Neutrophil (%)	86.40 (81.50-89.95)	85.00 (80.25-90.85)	-0.097	0.923
Liver function	2.00 (2.00-2.00)	2.00 (1.50-2.50)	-0.885	0.376
TB (μmol/L)	41.60 (25.05-54.05)	38.70 (24.70-52.25)	-0.225	0.822
ALT (U/L)	237.60 (157.30-473.90)	338.60 (261.75-418.75)	-1.492	0.136
AST (U/L)	498.80 (284.75-685.65)	482.20 (381.70-640.35)	-0.577	0.564
GGT (U/L)	40.00 (24.00-46.50)	47.00 (23.50-60.50)	-0.736	0.462
Renal function
Cr (μmol/L)	56.20 (47.60-65.82)	58.1 (48.90-68.25)	-0.245	0.806
BUN (mg/dL)	4.40 (3.60-5.40)	4.00 (3.30-4.95)	-0.763	0.445
Coagulation
PT (seconds)	12.72 (12.40-12.72)	12.58 (12.00-13.10)	-1.152	0.249
D-D (μg/L)	2822 (1597-2822)	1936 (1442-3131)	-1.506	0.132
Perioperative
Ablation time (minutes)	50.09 ± 14.48	43.71 ± 10.89	1.671	0.099
Clinical stay (day)	3.00 (3.00-4.00)	3.00 (2.00-4.00)	-1.010	0.312
Spent	60455 (57106-64849)	60716 (56262-62966)	-0.166	0.868
Exhaust time (day)	2 (2-3)	1.5 (1.25-2.5)	-0.885	0.376
Surgical blood loss (mL)	20.00 (10.00-50.00)	20.00 (10.00-35.00)	-0.097	0.923

WBC: White blood cell; TB: Total bilirubin; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; GGT: Gamma-glutamyl transferase; Cr: Creatinine; BUN: Blood urea nitrogen; PT: Prothrombin time; D-D: D-dimer.

Postoperative outcomes by tumor size

Tumor size demonstrated a more pronounced association with postoperative outcomes than lesion location. Compared with patients in the LHH group, those with GHH experienced significantly longer ablation times and greater intraoperative blood loss, reflecting the increased procedural complexity associated with larger lesions.

Several postoperative laboratory parameters differed significantly between the two groups. Patients with GHH showed higher postoperative neutrophil percentages, elevated total bilirubin and γ-glutamyltransferase levels, and increased D-dimer concentrations (all P < 0.05). Gastrointestinal functional recovery was also slower in the GHH group, indicating a broader impact of tumor size on early postoperative recovery. In addition, postoperative blood urea nitrogen and prothrombin time differed statistically between groups (P < 0.05), although these values remained within clinically acceptable ranges.

Given the retrospective nature of this study, these findings suggest that a greater tumor burden is associated with a trend toward more pronounced surgical stress, heightened systemic response, and delayed short-term recovery following TAE + LMWA, rather than establishing a causal relationship (Tables 5 and 6).

Table 5 Compare the postoperative recovery conditions of the giant hepatic hemangioma and large hepatic hemangioma groups in the liver parenchymal group, mean ± SD/median (interquartile rage).

	GHH (n = 17)	LHH (n = 17)	t/Z/χ2	P value
Inflammatory
WBC (109/L)	9.58 ± 3.03	9.14 ± 2.84	0.441	0.662
Neutrophil (%)	85.00 (80.25-89.93)	82.40 (73.75-89.70)	-1.051	0.293
Liver function
TB (μmol/L)	42.61 ± 23.37	31.41 ± 10.21	1.809	0.080
ALT (U/L)	375.63 ± 157.13	378.32 ± 234.54	-0.039	0.969
AST (U/L)	559.81 ± 257.56	479.50 ± 337.71	0.780	0.441
GGT (U/L)	41.00 (19.50-79.75)	42.50 (25.75-63.00)	-0.689	0.491
Renal function
Cr (μmol/L)	59.50 (52.60-65.50)	58.1 (48.90-68.25)	-0.362	0.718
BUN (mg/dL)	5.60 (3.33-12.38)	3.35 (2.95-4.07)	-1.758	0.079
Coagulation
D-D (μg/L)	1935.05 ± 1315.62	2403.25 ± 2441.19	-1.003	0.596
PT (seconds)	12.58 ± 1.12	12.70 ± 0.85	-0.449	0.793
Perioperative
Spent	60716.60 (56261.64-62966.02)	56527.12 (47523.46-60886.15)	-1.705	0.088
Ablation time (minutes)	43.71 ± 10.89	32.29 ± 8.29	3.438	0.002
Exhaust time (day)	2.00 (1.25-2.75)	1.50 (1.00-2.00)	-2.925	0.003
Surgical blood loss (mL)	30.00 (6.25-50.00)	15.00 (6.25-42.50)	-2.553	0.011
Clinical stay (day)	3.00 (2.00-4.00)	3.00 (3.00-4.00)	-0.829	0.407

WBC: White blood cell; TB: Total bilirubin; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; GGT: Gamma-glutamyl transferase; Cr: Creatinine; BUN: Blood urea nitrogen; D-D: D-dimer; PT: Prothrombin time; GHH: Giant hepatic hemangioma; LHH: Large hepatic hemangioma.

Table 6 Compare the postoperative recovery conditions of the giant hepatic hemangioma and large hepatic hemangioma groups within the liver subcapsular group, mean ± SD/median (interquartile rage).

	GHH (n = 55)	LHH (n = 63)	t/Z/χ2	P value
Inflammatory
WBC (109/L)	85.58 ± 6.13	83.11 ± 6.67	2.071	0.041
Neutrophil (%)	9.98 ± 3.58	9.61 ± 2.83	0.618	0.538
Liver function
TB (μmol/L)	39.35 (26.30-55.60)	32.10 (22.23-37.23)	-3.067	0.002
ALT (U/L)	201.40 (131.40-372.95)	280.55 (181.88-460.90)	-0.310	0.756
AST (U/L)	462.40 (240.05-626.98)	396.35 (271.80-647.13)	-1.387	0.165
GGT (U/L)	38.00 (22.25-61.75)	22.00 (15.00-27.25)	-2.928	0.003
Renal function
Cr (μmol/L)	61.15 (44.65-88.10)	55.75 (47.80-64.58)	-0.758	0.448
BUN (mg/dL)	4.89 (3.85-6.37)	3.60 (2.80-3.98)	-4.466	< 0.001
Coagulation
D-D (μg/L)	1909.50 (996.25-5124.50)	1339.50 (719.50-3002.00)	-2.485	0.013
PT (seconds)	12.72 (12.40-12.72)	12.18 (11.90-15.50)	-3.985	< 0.001
Perioperative
Spent	60846.12 (53898.31-64930.00)	55564.91 (36598.66-62983.01)	-1.513	0.130
Exhaust time (day)	2.00 (2.00-2.00)	1.00 (1.00-2.00)	-4.606	< 0.001
Ablation time (minutes)	50.00 (40.00-60.00)	30.50 (28.00-37.75)	-6.346	< 0.001
Clinical stay (day)	3.00 (3.00-4.00)	3.00 (2.00-3.00)	-1.361	0.173
Surgical blood loss (mL)	20.00 (10.00-45.00)	10.00 (5.25-20.00)	-2.179	0.029

WBC: White blood cell; TB: Total bilirubin; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; GGT: Gamma-glutamyl transferase; Cr: Creatinine; BUN: Blood urea nitrogen; D-D: D-dimer; PT: Prothrombin time; GHH: Giant hepatic hemangioma; LHH: Large hepatic hemangioma.

Postoperative complications

When classified by lesion location, postoperative fever occurred more frequently in the liver parenchymal group than in the liver subcapsular group (P < 0.01). The rates of anemia, hemoglobinuria, hypoproteinemia, and acute kidney injury were similar between groups.

When classified by tumor size, complications were more common in patients with GHH. The incidence of fever (P < 0.001) and hypoproteinemia (P < 0.05) was significantly higher in GHH than in LHH. Other complications - hemoglobinuria, anemia, and acute kidney injury - did not differ between size groups (Tables 7 and 8).

Table 7 Comparison of postoperative complications between giant hepatic hemangioma and large hepatic hemangioma groups, n (%).

	GHH (n = 72)	LHH (n = 80)	χ2	P value
Anemia	19 (26.39)	12 (15)	3.027	0.082
Fever	40 (55.56)	8 (10)	36.398	< 0.001
Hemoglobinuria	8 (11.11)	9 (11.25)	0.001	0.987
Acute kidney injury	1 (1.39)	1 (1.25)		1.000
Hypoproteinemia	3 (4.17)	0 (0)	4.550	0.033

GHH: Giant hepatic hemangioma; LHH: Large hepatic hemangioma.

Table 8 Comparison of postoperative complications between liver subcapsular and liver parenchymal groups, n (%).

	Liver subcapsular (n = 118)	Liver parenchymal (n = 34)	χ2	P value
Anemia	23 (19.49)	8 (23.53)	0.265	0.607
Fever	31 (26.27)	17 (50)	6.878	0.009
Hemoglobinuria	15 (12.71)	1 (2.94)	1.739	0.187
Acute kidney injury	1 (0.8)	1 (2.94)		0.398
Hypoproteinemia	2 (1.69)	1 (2.94)		0.535

Summary of findings

Our retrospective analysis showed that lesion location (subcapsular vs parenchymal) was not associated with significant differences in most postoperative outcomes. However, this finding should be interpreted with caution due to the relatively small sample size of the liver parenchymal subgroup, which may limit statistical power. In contrast, lesion size demonstrated a more consistent association with perioperative and postoperative parameters. Compared with LHHs, giant lesions were associated with longer ablation duration, greater intraoperative blood loss, more pronounced postoperative changes in inflammatory and hepatic function markers, and delayed overall recovery. Differences in complication profiles also tended to be more evident between the LHH and GHH groups, whereas lesion location showed no such clear pattern, with the exception of postoperative fever.

Notably, under the standardized ablation protocol incorporating a gas-blood outflow channel, a relatively low incidence of major heat-sink-related complications, including hemoglobinuria and acute kidney injury, was observed in this cohort. Although no control group without decompression was available for direct comparison, this retrospective observation suggests a potential association between the use of the gas-blood outflow channel and attenuation of heat-sink-related effects during TAE + LMWA. This finding should be regarded as hypothesis-generating rather than evidence of causality.

Given the retrospective nature of the study and the imbalance between comparison groups, these findings should be interpreted cautiously. Overall, the results suggest that tumor burden, primarily reflected by lesion size, may play a more prominent role than anatomical location in influencing short-term outcomes following TAE + LMWA. Further prospective, controlled studies are required to validate these observed associations and to clarify the clinical value of technical refinements such as the gas-blood outflow channel.

DISCUSSION

Primary factor influencing postoperative recovery

In this retrospective analysis of 152 patients, our findings consistently suggest that tumor size was the primary factor influencing postoperative recovery following TAE + LMWA, with lesion location having a more limited impact.

GHH, with their extensive vascularity, likely lead to greater heat absorption and tissue destruction during ablation, contributing to a heavier thermal burden. This explains the longer ablation duration and increased intraoperative blood loss observed in the GHH group. Postoperative elevations in neutrophil percentage, bilirubin, γ-glutamyltransferase, and D-dimer levels in GHH patients indicated a stronger systemic inflammatory and hepatic response. While some of these parameters did not always reach statistical significance in all stratified analyses, their consistently higher absolute values and more dynamic changes in the GHH cohort suggest a greater clinical impact and physiological burden. Slower gastrointestinal recovery further supports the heightened physiological stress associated with treating larger lesions.

Lesion location: Limited but real effect

The observed limited effect of lesion location on postoperative recovery can be primarily associated with the technical advantages of the laparoscopic approach. This method provides comprehensive liver surface access and allows flexible puncture trajectory adjustment, thereby minimizing differences caused by anatomical depth. Notably, for LHH in the LP group, their deeper location during ablation might have contributed to reduced intraoperative bleeding. Because intraparenchymal hemangiomas are surrounded by intact hepatic tissue, intratumoral blood drainage may be less efficient during ablation. Additionally, preoperative TAE uniformly reduces intratumoral perfusion, further mitigating positional impact. These combined procedural factors likely explain the largely similar recovery profiles between liver subcapsular and liver parenchymal lesions. However, the substantial imbalance between liver subcapsular and liver parenchymal, particularly within size-stratified subgroups, significantly limited our statistical power. Consequently, the absence of statistically significant differences should be interpreted cautiously and not as definitive evidence of equivalence, as a type II error cannot be excluded. Future studies with larger, balanced cohorts are warranted to precisely assess the impact of lesion location.

The decompression channel: A hypothesis-generating observation

A notable observation of this study is the relatively low incidence of major heat-sink-related complications, particularly hemoglobinuria and acute kidney injury, following TAE + LMWA. All patients were treated using a standardized protocol incorporating a gas-blood outflow channel, which was intended to facilitate intratumoral decompression during ablation. From a mechanistic standpoint, this modification may allow pressure release and potentially attenuate hemolysis; however, this interpretation remains speculative[15].

Importantly, given the retrospective design and the absence of a control group without a decompression channel, no causal inference can be made regarding its protective effect. The observed complication profile may also be influenced by other factors, including preoperative embolization, standardized energy delivery, perioperative management, and patient selection. Therefore, the decompression channel should be regarded as a hypothesis-generating technical refinement. Prospective, controlled studies are required to clarify its independent role and clinical value.

Integration with recent evidence

Historically, the management of LHH and GHH predominantly involved surgical resection[16]. Over time, scholars introduced various ablative techniques, such as radiofrequency ablation, and these methods showed promising initial efficacy[17,18]. Specifically, bipolar radiofrequency ablation, TAE, and MWA were adopted for LHH treatment[19]. However, complications related to thermal damage from ablation have consistently been a concern. In light of this, combined approaches like TAE + LMWA for LHH and GHH have been explored, which appear to be associated with a reduced incidence of postoperative complications[20,21].

The present findings contribute to a growing body of literature supporting the potential utility of TAE combined with LMWA as a minimally invasive option for LHH and GHH. Compared with surgical resection, this combined approach may offer reduced trauma, sparing of functional parenchyma, and rapid postoperative recovery, while potentially avoiding some risks associated with major hepatic surgery[22,23]. Given the inherent variability in liver hemangioma location and the recognized importance of size and position on prognosis - as suggested by previous comparative studies on surgically treated hemangiomas - we undertook this retrospective study to further investigate these relationships within the context of combined TAE + LMWA. Our results further refine the understanding of factors that may influence postoperative outcomes and underscore the importance of considering tumor size when counseling patients and planning perioperative management.

Limitations and future directions

This study is subject to several important limitations. Firstly, its retrospective and single-center design introduced inherent selection bias and limits generalizability. Secondly, a major methodological constraint was the marked imbalance in lesion location subgroups (118 subcapsular vs 34 parenchymal lesions). This disparity severely limited statistical power for reliable comparisons, increasing the risk of type II errors. It also precluded robust statistical methods like propensity score matching and made comprehensive multivariable analyses for interaction effects between tumor size and location challenging due to the skewed distribution. Thirdly, the absence of dedicated control channel groups restricts our ability to definitively isolate specific treatment contributions. Consequently, observations regarding the gas-blood decompression channel remain descriptive and hypothesis-generating, not causally conclusive. Finally, our focus was on short-term perioperative outcomes. Crucially, data on long-term durability, recurrence, re-intervention rates, and patient-reported quality of life were not collected, preventing a comprehensive assessment of long-term efficacy.

Future prospective multicenter studies with larger, balanced cohorts and dedicated control groups are warranted. These should incorporate comprehensive long-term follow-up for recurrence, re-intervention, alongside advanced statistical modeling to explore interaction effects between tumor size and location.

CONCLUSION

Tumor size appears to be an important factor influencing short-term postoperative recovery following TAE + LMWA, whereas lesion location shows a limited and selective association. Under a standardized protocol incorporating a gas-blood decompression channel, a low incidence of major heat-sink-related complications was observed. However, given the retrospective design and absence of a control group, the potential role of this technical modification should be regarded as exploratory and requires prospective validation.

ACKNOWLEDGEMENTS

We sincerely thank all the staff who contributed to this study, whose dedication and teamwork made its successful completion possible.