Indocyanine green fluorescence–guided laparoscopic anatomical hepatectomy for radical resection of hepatocellular carcinoma: a clinical study

Abstract

INTRODUCTION:

Anatomical hepatectomy improves outcomes in hepatocellular carcinoma (HCC), but remains technically demanding laparoscopically. Indocyanine green (ICG) fluorescence may aid navigation, yet its clinical value remains unproven.

AIM:

We aimed to evaluate whether ICG-guided laparoscopic anatomical hepatectomy improved resection quality and short-term outcomes, as compared with conventional high-definition laparoscopy.

MATERIALS AND METHODS:

In this single-center, prospective, open-label, parallel-group randomized controlled trial (October 2023–December 2025), 100 patients with resectable primary HCC were randomized at a 1:1 ratio to the ICG-guided laparoscopic anatomical hepatectomy (n = 50) or conventional laparoscopy (n = 50) cohorts. The ICG protocol included preoperative intravenous ICG plus intraoperative ultrasound-guided positive / negative staining. Primary outcomes were anatomical resection and R0 rates. Secondary outcomes included blood loss, complications, 1-year disease-free survival (DFS), and overall survival (OS).

RESULTS:

Baseline characteristics of the study population were balanced. Tumor fluorescence and segmental staining were successful in 94% and 98% of the ICG patients, respectively. The ICG group achieved higher anatomical resection (94% vs 58%; P <0.001) and R0 rates (100% vs 86%; P = 0.01), wider margins (mean [SD], 10.27 [3.39] vs 6.97 [4.11] mm; P <0.001), less mean blood loss (217.2 [88.76] vs 298.44 [156.19] ml; P = 0.002), fewer complications (18% vs 50%; P <0.001), and shorter hospital stay (mean, 7.65 [1.68] vs 8.93 [2.84] d; P = 0.008) than the controls. ICG detected additional lesions in 20% of the patients. One-year DFS (88% vs 75.04%) and OS (93.82% vs 91.49%; P = 0.8) were similar in both groups. Adjusted recurrence hazard ratio was 2.146 (95% CI, 0.932–4.938; P = 0.07).

CONCLUSIONS:

ICG fluorescence-guided laparoscopic anatomical hepatectomy is safe and feasible, significantly improving radical resection quality with more anatomical dissections, higher R0 rates, wider margins, and fewer complications, as compared with the conventional approach. Longer follow-up is needed to confirm survival benefits.

INTRODUCTION

Hepatocellular carcinoma (HCC) is one of the dominant oncological challenges worldwide, accounting for approximately 80%–90% of primary liver cancers and ranking among the leading causes of cancer-related deaths.^1,2 Although contemporary surveillance programs, antiviral therapy, and expanding systemic options have altered its trajectory, HCC incidence continues to rise in many regions, driven in particular by chronic hepatitis B virus (HBV) infections in East Asia and metabolic dysfunction–associated steatotic liver disease in Western populations.^3,4 In China, the disease remains a public-health priority, contributing disproportionately to the national cancer burden and claiming a large share of hepatobiliary surgical workload.

Surgical resection continues to offer the greatest probability of long-term survival for patients with preserved hepatic function and localized disease. Anatomical hepatectomy—en-bloc removal of the portal territory that drains the tumor—is founded on the pathobiological concept that HCC disseminates along tumor-bearing portal branches. A growing body of propensity-matched cohort evidence, corroborated by systematic reviews and meta-analyses, supports the view that anatomical resection offers superior overall survival (OS) and recurrence-free survival, as compared with nonanatomical resection, especially for tumors of moderate size and those localized in noncirrhotic or compensated cirrhotic livers.^5,6 Performance of a true anatomical plane, however, requires confident identification of segmental boundaries. Even though it is fairly straightforward when performed through an open approach using inflow occlusion and ischemia lines, it is considerably more demanding when performed laparoscopically.

Laparoscopic liver resection has shifted from a pioneering technique to a standard-of-care option in high-volume centers since the publication of the Louisville and Morioka international consensus statements,⁷ and currently extends across all hepatic segments. Its advantages, such as reduced surgical access trauma, diminished postoperative pain, shortened recovery, and oncological efficacy comparable to open surgery, are well documented. Yet the restricted range of instrument motion, lack of tactile feedback, rigid orientation imposed by fixed trocar sites, and frequent coexistence of macronodular cirrhosis contribute to the limitation of intraoperative tumor localization and blurring of anatomical resection planes. Consequently, nonanatomical wedge procedures have often been performed laparoscopically when surgeons could not confidently demarcate the intended hepatic territory.

Indocyanine green (ICG) is a near-infrared fluorophore that binds rapidly to plasma proteins after intravenous injection, is selectively taken up by hepatocytes, and excreted into bile.⁸ When it is injected days before operation, delayed biliary excretion within the tumor and in the noncancerous parenchyma surrounding metastatic foci produces intratumoral or peritumoral fluorescence, allowing real-time identification of subcapsular HCC nodules with high sensitivity. When injected intraoperatively either into a target portal branch (positive staining) or intravenously after selective Glissonean pedicle occlusion (negative staining), ICG instantly demarcates the corresponding portal territory on the hepatic surface and, crucially, on the parenchymal transection plane. The European Association for Endoscopic Surgery and other professional bodies have formally endorsed fluorescence-guided image navigation as a valuable adjunct to laparoscopic liver surgery.⁹

The question of whether ICG guidance improves not just the mechanics of laparoscopic dissection but the actual short-term and oncological outcomes of HCC patients has been pursued through retrospective cohorts and propensity-matched analyses.^10,11 Meta-analyses have reported less blood loss, fewer transfusions, reduced length of hospital stay (LOS), and higher R0 resection rates with ICG-guided laparoscopic hepatectomy, without extending operative time.^12,13 Nonetheless, much of the evidence base is retrospective, limited, or restricted to nonanatomical resection; prospective randomized data focusing on anatomical hepatectomy are still scarce, and the relative contribution of combined tumor-and-territory fluorescence to radical HCC resection remains undefined.^14,15

AIM

We designed a prospective randomized trial to evaluate whether ICG fluorescence–guided laparoscopic anatomical hepatectomy, incorporating both preoperative intravenous tumor staining and intraoperative positive- or negative-staining territory mapping, improved the rate and quality of radical resection of HCC relative to conventional high-definition laparoscopic hepatectomy. We also aimed to describe the safety profile and short-term oncological outcomes of the technique in a Chinese population with predominantly HBV-related HCC and cirrhosis.

MATERIALS AND METHODS

Study design

This was a prospective, single-center, open-label, parallel-group randomized controlled trial conducted at the Department of Hepatopancreatobiliary Surgery of Ganzhou People’s Hospital, China. The protocol complied with the Declaration of Helsinki¹⁶ and the International Council for Harmonisation Guideline for Good Clinical Practice E6 (R2),¹⁷ and followed the CONSORT reporting guidelines (FIGURE 1).

FIGURE 1.

CONSORT flow diagram of the study

Participants

Adults aged 18 to 75 years with a clinical or pathological diagnosis of primary HCC, considered eligible for laparoscopic liver resection by a multidisciplinary hepatobiliary team were screened between October 2023 and December 2025. The inclusion criteria were: 1) preoperative diagnosis of primary HCC based on typical imaging features or biopsy results according to contemporary guidelines; 2) tumor technically amenable to laparoscopic anatomical resection; 3) Child–Pugh class A or B; and 4) American Society of Anesthesiologists score of III or lower. The key exclusion criteria comprised: 1) contraindication to laparoscopic surgery; 2) known hypersensitivity to ICG or iodine; 3) any previous local therapy for the index tumor (transarterial chemoembolization, percutaneous ablation, or local radiotherapy); 4) severe psychiatric illness or other active malignancy within the preceding 5 years; and 5) pregnancy or lactation. Patients with portal vein tumor thrombus in the main trunk or first-order branches were excluded.

Randomization and blinding

Eligible patients were randomly allocated at a 1:1 ratio to the ICG group or the control group using computer-generated block randomization (block size of 4), prepared by an independent biostatistician and concealed within sequentially numbered opaque envelopes that were opened in the operating room only after induction of anesthesia. Because the nature of the intervention precluded blinding of the operating team, outcome assessors and the biostatistician performing the final analysis were blinded to group allocation; the pathologists reviewing the specimens had no access to the fluorescence data.

Indocyanine green fluorescence protocol (indocyanine green group)

Preoperative tumor staining

All patients in the ICG group underwent the routine preoperative ICG clearance test (ICG retention rate at 15 min [ICG-R15]). For tumor fluorescence imaging, a further fixed intravenous dose of 0.5 mg/kg was administered 48 to 96 hours before surgery in the patients with ICG-R15 below 10%, and 72 to 120 hours before surgery in the patients with ICG-R15 between 10% and 20%, so as to minimize nonspecific parenchymal background fluorescence in the individuals with reduced liver clearance.¹⁸ The timing was individualized by the treating surgeon in consultation with hepatologists, reflecting ongoing uncertainty in the literature regarding the optimal dosing interval in the cirrhotic liver.^19,20

Intraoperative territorial staining

After pneumoperitoneum (12–14 mm Hg) and laparoscopic intraoperative ultrasound to confirm tumor location, exclude satellite lesions, and plan the resection, 1 of 2 staining strategies was selected according to the target hepatic territory. For segmentectomy or subsegmentectomy, positive staining was performed. Under laparoscopic ultrasound guidance, a percutaneous transhepatic cholangiography needle was advanced into the tumor-feeding portal branch. Correct placement was confirmed by back-bleed, and 0.5–1 ml of ICG solution at 0.025 mg/ml was slowly injected, immediately delineating the corresponding territory on the hepatic surface. For sectionectomy, hemihepatectomy, or combined segmental resection, negative staining was used. The Glissonean pedicle of the target territory was isolated extrahepatically through the Takasaki approach and temporarily clamped, and 1 ml of ICG at 2.5 mg/ml was administered through a peripheral vein, immediately rendering the surrounding nontarget liver fluorescent, while leaving the occluded target territory dark.²¹

Image acquisition

Fluorescence imaging was obtained with a 4K near-infrared laparoscopic system (OPTO-CAM 214K, Opto Medical), which overlays pseudocolor fluorescence images onto the standard white-light video in real time in a 4-panel display. Parenchymal transection proceeded along the fluorescence-defined boundary using an ultrasonic dissector and a harmonic device; hepatic-venous branches were divided with clips or a vascular stapler, and the target Glissonean pedicle was transected as the final step.

Surgical technique (control group)

The patients allocated to the control group underwent standard high-definition laparoscopic hepatectomy without ICG guidance. Liver segmentation was attempted using conventional methods. Extrahepatic Glissonean pedicle dissection was performed, with transient arterial inflow occlusion to produce a demarcation ischemia line on the liver surface, and the intended resection plane was intraoperatively mapped under ultrasound guidance. A 2-cm macroscopic margin from the tumor was targeted and electrocoagulated when anatomical landmarks were indistinct. All operations in both arms were performed by a unified team of 2 board-certified hepatobiliary surgeons (YH and LZ) with individual experience of more than 150 laparoscopic hepatectomies.

Outcomes and definitions

The primary outcomes were: 1) anatomical resection rate (defined as complete removal of the portal territory draining the tumor, verified on the resection specimen and operative video by 2 independent reviewers [XD and YL]) and 2) the R0 resection rate (defined as a histologically negative surgical margin ≥1 mm). Secondary outcomes comprised: 1) intraoperative blood loss; 2) operative time; 3) Pringle maneuver use and duration; 4) conversion to open surgery; 5) surgical margin distance (mm); 6) postoperative complications, graded according to the Clavien–Dindo classification²²; 7) bile leakage, defined according to the International Study Group of Liver Surgery (ISGLS) criteria²³; 8) posthepatectomy liver failure, defined according to the ISGLS criteria²⁴; 9) 90-day mortality; 10) time to first oral intake; 11) LOS; and 12) short-term oncological outcomes (1-y and 18-mo disease-free survival [DFS] and OS). The following ICG-specific performance metrics were recorded in the ICG group: 1) the rate of tumor fluorescence visualization; 2) staining pattern (total, partial, rim, or no fluorescence); 3) the rate of successful target-segment territorial staining; and 4) the number of additional hepatic lesions detected intraoperatively by fluorescence that had not been visualized on preoperative cross-sectional imaging or intraoperative ultrasound.

Follow-up

The patients were followed in an outpatient clinic every 3 months during the first 2 years and every 6 months thereafter. Contrast-enhanced abdominal computed tomography or magnetic resonance imaging with serum α-fetoprotein (AFP) was performed at each visit. Recurrence was defined as a new hepatic or extrahepatic lesion with typical HCC imaging features or biopsy confirmation. The administrative data cutoff for this analysis was May 31, 2026.

Sample size and statistical analysis

The sample size was calculated for the comparison of 2 independent proportions using the χ² test. The expected rate of achieved anatomical resection in the control group (60%) was based on the anatomical resection rates reported for conventional laparoscopic hepatectomy in previous comparative studies and meta-analyses,^12,13 and the expected rate in the ICG group (85%) was based on the improvement observed with ICG fluorescence guidance in prior propensity-matched cohorts and on institutional pilot experience.^10,25 To detect an absolute 25% improvement (from 60% to 85%) with a 2-sided α of 0.05 and 80% power, the standard formula for 2 independent proportions with pooled variance yielded 49 patients per group; 50 patients per group were enrolled to accommodate incomplete data, yielding a total sample size of 100 participants, which was consistent with the prespecified protocol. Numerical variables were tested for normality with the Shapiro–Wilk test. Normally distributed variables were presented as mean (SD) and compared using the independent-samples t test, whereas non-normally distributed variables were expressed as median (interquartile range [IQR]) and compared using the Mann–Whitney test. Categorical variables were presented as counts (percentages) and compared using the χ² test or the Fisher exact test when expected cell counts were below 5. Survival was estimated by the Kaplan–Meier method and compared using the log-rank test. Cox proportional hazards regression was used to estimate adjusted hazard ratios (HRs) for recurrence, with prespecified covariates of group allocation, cirrhosis, and tumor size. Analyses followed the intention-to-treat principle. A 2-sided P value below 0.05 was considered significant. All analyses were performed with SPSS software, version 26.0 (IBM Corp.) and Python 3.11 (scipy 1.11, NumPy 1.26).

Ethics

The study protocol was approved by the Institutional Review Board of Ganzhou People’s Hospital (20230712128801). Written informed consent was obtained from every participant. All procedures were performed in accordance with the ethical standards of the institutional research committee and the Declaration of Helsinki.¹⁶

RESULTS

Group allocation and baseline characteristics

Between October 2023 and December 2025, 166 consecutive patients were screened and 100 were randomized: 50 to the ICG group and 50 to the control group (FIGURE 1). All allocated patients received the assigned intervention and were included in the intention-to-treat analysis; none were lost to follow-up by the data cutoff (May 31, 2026). Median (IQR) follow-up was 19.3 (14.43–25) months in the ICG group and 18.25 (14.7–22.15) months in the control group.

Baseline demographic, clinical, and tumor characteristics are summarized in TABLE 1. The groups were well matched for age (mean [SD], 56.48 [9.47] vs 56.06 [9.63] y), sex distribution (74% in both arms were men), body mass index (mean, 23.44 [2.77] vs 23.52 [2.85] kg/m²), etiology (HBV predominant in both), Child–Pugh classification (class A predominant in both), ICG-R15 (mean, 9.79% [3.95%] vs 9.76% [3.86%]), serum AFP level (median, 29.75 [16.09–146.75] vs 42.67 [10.48–99.13] ng/ml), tumor size (mean, 3.72 [1.2] vs 3.78 [1.31] cm) and number, location (posterosuperior segments in 42% vs 40%), Barcelona Clinic Liver Cancer stage, and IWATE difficulty index. The only baseline covariate to differ between the cohorts was the prevalence of histologically confirmed cirrhosis, which was higher in the ICG group (88% vs 70%; P = 0.03), a finding that was carried forward into the adjusted survival analyses.

TABLE 1.

Baseline demographic and clinical characteristics of the study population

Characteristic		ICG group (n = 50)	Control group (n = 50)	P value
Age, y		56.48 (9.47)	56.06 (9.63)	0.83
Men		37 (74)	37 (74)	>0.99
BMI, kg/m2		23.44 (2.77)	23.52 (2.85)	0.88
Etiology	HBV	40 (80)	37 (74)	0.41
	HCV	4 (8)	2 (4)
	Alcohol abuse	5 (10)	7 (14)
	Other	1 (2)	4 (8)
Child–Pugh class A		46 (92)	45 (90)	>0.99
Liver cirrhosis		44 (88)	35 (70)	0.03
ICG-R15		9.79 (3.95)	9.76 (3.86)	0.97
AFP, ng/ml		29.75 (16.09–146.75)	42.67 (10.48–99.13)	0.44
Tumor size, cm		3.72 (1.2)	3.78 (1.31)	0.8
Solitary tumor		40 (80)	45 (90)	0.16
Posterosuperior location		21 (42)	20 (40)	0.84
BCLC stage	0	7 (14)	6 (12)	0.91
	A	35 (70)	37 (74)
	B	8 (16)	7 (14)
IWATE difficulty score, points	Low	10 (20)	8 (16)	0.23
	Intermediate	19 (38)	28 (56)
	Advanced	19 (38)	11 (22)
	Expert	2 (4)	3 (6)

Data are presented as number (percentage), mean (SD), or median (interquartile range).

Abbreviations: AFP, α-fetoprotein; BCLC, Barcelona Clinic Liver Cancer; BMI, body mass index; HBV, hepatitis B virus; HCV, hepatitis C virus; ICG, indocyanine green; ICG-R15, indocyanine green retention rate at 15 minutes; IWATE, IWATE difficulty score for laparoscopic hepatectomy

Intraoperative outcomes

Intraoperative outcomes are reported in TABLE 2 and visually summarized in FIGURE 2. Operative time was comparable between the groups (median, 246.45 [218.8–284.35] vs 236.35 [199.15–279.07] min; P = 0.55), indicating that the added maneuvers required for ICG staining did not prolong surgery. Conversely, intraoperative blood loss was lower in the ICG group (mean, 217.2 [88.76] vs 298.44 [156.19] ml; P = 0.002), as was the duration of the Pringle maneuver when used (mean, 31.06 [11.4] vs 36.33 [10.48] min; P = 0.04). Intraoperative transfusion (10% vs 16%; P = 0.37) and conversion rates to open surgery (4% vs 6%) were lower in the ICG group but the latter did not reach significance. Most importantly, the rate of achieved anatomical resection was markedly higher in the ICG group (94% vs 58%; P <0.001), corresponding to an absolute increase of 36%.

TABLE 2.

Intraoperative outcomes

Outcome	ICG group (n = 50)	Control group (n = 50)	P value
Operative time, min	246.45 (218.8–284.35)	236.35 (199.15–279.07)	0.55
Estimated blood loss, ml	217.2 (88.76)	298.44 (156.19)	0.002
Pringle maneuver used	34 (68)	41 (82)	0.11
Pringle maneuver duration, mina	31.06 (11.4)	36.33 (10.48)	0.04
Intraoperative transfusion	5 (10)	8 (16)	0.37
Conversion to open surgery	2 (4)	3 (6)	>0.99
Anatomical resection achieved	47 (94)	29 (58)	<0.001

Data are presented as number (percentage), mean (SD), or median (interquartile range).

a The Pringle maneuver duration is reported only for the patients in whom it was used.

Abbreviations: see TABLE 1

FIGURE 2.

Perioperative and pathological outcomes of the study groups; A – numerical perioperative outcomes plotted as standardized mean differences (Cohen d), with the direction of effect aligned so that values to the right of the dashed reference line favor the ICG group and those to the left favor the controls; B – categorical outcomes plotted as risk ratios (ICG vs control group) on a logarithmic scale with 95% CIs. The dashed reference line at 1 corresponds to no effect, with values to the left favoring the ICG group and those to the right the controls. The size of each point is proportional to the total event count, with raw event numerators (ICG vs control) presented to the right of each row.

Abbreviations: see TABLE 1

Short-term oncological outcomes

Over median follow-up of approximately 18 months, recurrence was documented in 9/50 patients in the ICG group and 16/50 in the control group, and death occurred in 5/50 and 4/50 patients, respectively (FIGURE 3). Kaplan–Meier 1-year DFS was 88% in the ICG group, as compared with 75.04% in the controls (log-rank P = 0.09), and 1-year OS was 93.82% vs 91.49%, respectively (log-rank P = 0.8). At 18 months, the respective DFS estimates were 80.18% vs 67.23%, and OS estimates were 88.02% vs 91.49%, respectively. The DFS curve consistently favored the ICG arm throughout the follow-up window. The OS estimates were closely overlapping and nonsignificantly different between the groups, reflecting the limited number of events. Neither comparison achieved conventional significance within the observed period.

FIGURE 3.

Kaplan–Meier curves of DFS (A) and OS (B), stratified by treatment allocation.

Abbreavitations: DFS, disease-free survival; OS, overall survival; others, see TABLE 1

To mitigate the baseline cirrhosis imbalance, multivariable Cox regression was performed for DFS, adjusting for treatment allocation, cirrhosis status, and tumor size. The adjusted HR for recurrence in the control arm relative to the ICG arm was 2.146 (95% CI, 0.932–4.938; P = 0.07); neither cirrhosis (HR, 1.636; 95% CI, 0.549–4.871; P = 0.38) nor tumor size (HR, 0.977 per cm; 95% CI, 0.704–1.357; P = 0.89) was independently prognostic within this short observation window.

Performance of indocyanine green fluorescence imaging

The performance of ICG fluorescence imaging in the ICG group is detailed in TABLE 3 and FIGURE 4A and 4B. Tumor fluorescence was successfully visualized in 47/50 patients (94%). The dominant staining pattern was total-type fluorescence (n = 33; 66%), followed by partial-type (n = 7; 14%) and rim-type (n = 7; 14%); 3 tumors (6%) did not fluoresce. Two of the 3 nonvisualized tumors arose in poorly-differentiated lesions with preoperative ICG-R15 above 10%, consistent with the recognized phenomenon of background parenchymal signal-swamping tumor contrast in advanced cirrhosis (FIGURE 4B). Successful demarcation of the target hepatic territory (segment or section) via intraoperative ICG staining was obtained in 49/50 patients (98%); positive staining via portal-branch puncture was used in 24 patients (48%), and negative staining via Glissonean pedicle occlusion in 26 individuals (52%). Intraoperative fluorescence disclosed 10 additional lesions (20%) that had not been identified on preoperative imaging or intraoperative ultrasound alone; 8 were subcapsular micronodules (≤10 mm), and 5 proved to be well-differentiated HCC or dysplastic foci on final pathological examination.

TABLE 3.

Performance of indocyanine green fluorescence imaging in the indocyanine green group (n = 50)^a

ICG performance metric		Value
Tumor fluorescence visualization	Overall	47 (94)
	Total-type	33 (66)
	Partial-type	7 (14)
	Rim-type	7 (14)
	No fluorescence	3 (6)
Segment territory staining success	Overall	49 (98)
	Positive-staining method	24 (48)
	Negative-staining method	26 (52)
Additional lesions detected intraoperativelyb		10 (20)

Data are presented as number (percentage).

a Staining patterns are mutually exclusive.

b Additional lesions denote hepatic foci identified intraoperatively via ICG fluorescence that had not been detected on preoperative cross-sectional imaging or intraoperative laparoscopic ultrasound.

Abbreviations: see TABLE 1

FIGURE 4.

Performance of ICG fluorescence imaging in the ICG group; A – 4 key performance metrics of intraoperative ICG fluorescence imaging displayed as horizontal stacked bars and expressed as percentages; B – a dot plot of preoperative ICG-R15 stratified by whether intraoperative tumor fluorescence was successfully visualized. All 3 visualization failures occurred in patients with ICG-R15 ≥5%, with 2 of 3 concentrated above 10%, illustrating the impact of impaired hepatic clearance on tumor-to-background fluorescence contrast in cirrhotic livers.

Abbreavitations: see TABLE 1

Pathological and postoperative outcomes

Pathological and postoperative outcomes are shown in TABLE 4. The R0 resection rate was higher in the ICG group (100% vs 86%; P = 0.01), and the width of the parenchymal margin was also greater in this cohort (mean, 10.27 [3.39] vs 6.97 [4.11] mm; P <0.001), as compared with the controls. These histological gains translated into a favorable short-term morbidity profile. The overall incidence of any postoperative complication (Clavien–Dindo grade ≥I) was less than half that observed in the control arm (18% vs 50%; P <0.001), driven chiefly by reductions in bile leakage (8% vs 18%; P = 0.14), pleural effusion (6% vs 10%; P = 0.72) and refractory ascites (6% vs 10%; P = 0.72). Major complications (Clavien–Dindo grade ≥IIIa) occurred in 5 (10%) vs 8 patients (16%; P = 0.37). The incidence of ISGLS-defined posthepatectomy liver failure was similar between the groups (12% vs 10%; P = 0.14), reflecting the comparable baseline hepatic reserve. No deaths occurred within 90 days in the ICG group, whereas 3 of the control patients died within this window (0% vs 6%; P = 0.24). LOS was shorter in the ICG group (mean, 7.65 [1.68] vs 8.93 [2.84] d; P = 0.008), and time to first oral intake was similar in both cohorts (mean, 2.56 [0.83] vs 2.53 [0.9] d).

TABLE 4.

Pathological and postoperative outcomes

Characteristic		ICG group (n = 50)	Control group (n = 50)	P value
R0 resection		50 (100)	43 (86)	0.01
Microvascular invasion		12 (24)	14 (28)	0.65
Surgical margin distance, mm		10.27 (3.39)	6.97 (4.11)	<0.001
Tumor differentiation	Good	9 (18)	9 (18)	0.35
	Moderate	26 (52)	32 (64)
	Poor	15 (30)	9 (18)
Bile leakage (ISGLS)		4 (8)	9 (18)	0.14
Posthepatectomy liver failure		6 (12)	5 (10)	0.75
Pleural effusion		3 (6)	5 (10)	0.72
Refractory ascites		3 (6)	5 (10)	0.72
Any postoperative complication		9 (18)	25 (50)	<0.001
Major complications (Clavien–Dindo ≥IIIa)		5 (10)	8 (16)	0.37
90-day mortality		0	3 (6)	0.24
Time to first oral intake, d		2.56 (0.83)	2.53 (0.9)	0.88
Length of hospital stay, d		7.65 (1.68)	8.93 (2.84)	0.008

Data are presented as number (percentage) or mean (SD).

Abbreviations: ISGLS, International Study Group of Liver Surgery; others, see TABLE 1

DISCUSSION

In this prospective randomized controlled trial of 100 patients with resectable primary HCC, the addition of ICG near-infrared fluorescence into laparoscopic anatomical hepatectomy yielded 3 consistent, interlocking benefits: surgical precision, perioperative safety, and favorable early oncological trends. First, ICG nearly eliminated the operational gap between the intention to perform anatomical resection and the ability to achieve one, raising the rate of completed anatomical resection from 58% with conventional laparoscopy to 94%. Second, it converted an 86% R0 resection rate into a 100% R0 resection rate, while increasing median parenchymal margin by more than 3 mm. Third, it halved the overall postoperative complication rate, reduced blood loss by roughly 80 ml, and shortened LOS by more than a day, without extending operative time. These results are internally coherent: a better-defined transection plane generates less bleeding, less bile leakage, and less parenchymal trauma, which translates to a smoother recovery—the very chain of events that clinical fluorescence navigation was designed to unlock.

Our findings extend the existing evidence base in several important ways. Prior retrospective and propensity-matched cohort studies have documented that ICG guidance improved short-term outcomes in laparoscopic hepatectomy, with multiple meta-analyses reporting reductions in blood loss, transfusions, and LOS, as well as higher R0 rates with ICG.^12,13,25 A 2025 systematic review and meta-analysis of 12 comparative studies¹² similarly found that ICG fluorescence navigation significantly increased the R0 resection rate and improved perioperative outcomes relative to conventional laparoscopic hepatectomy for HCC. The magnitude of the R0 improvement in our trial (100% vs 86%) is larger than the pooled effect reported in that meta-analysis, most plausibly because we combined tumor-level and territory-level fluorescence strategies, and focused exclusively on anatomical resection, in which the margin-defining plane is a structural feature of the dissection rather than an ad hoc 2-cm cuff. Consistent with the recent retrospective cohort study by Mao et al,²⁶ who identified enhanced recurrence-free survival in ICG-guided difficult laparoscopic HCC resections, our Kaplan–Meier curves showed a similar absolute separation of 13% in 1-year DFS, favoring ICG, even if significance was not reached within the analyzed follow-up.

The mechanistic logic supporting these results is well grounded. HCC exhibits impaired biliary excretion of ICG due to loss of canalicular organic anion-transporting polypeptide expression, so intravenously administered ICG accumulates preferentially within tumor cytoplasm and pseudoglands, yielding real-time transcapsular fluorescence of most superficial lesions.^8,27 Partial and rim staining patterns—observed in 28% of the tumors—reflect histological heterogeneity and a mixed contribution from peritumoral compressed hepatocytes; the fluorescence pattern has also been reported to correlate with the degree of tumor differentiation.²⁸ At the territorial level, either portal-vein injection (positive staining) or Glissonean pedicle occlusion with peripheral ICG (negative staining) produces an in-situ 3-dimensional segmental map that persists for hours, greatly facilitating parenchymal transection along a true intersegmental plane rather than along a surface ischemia line that tends to fade and provides no information below the capsule.^29,30 Our 98% territorial staining success rate is in line with the 94% success rate reported for open anatomical hepatectomy by Berardi et al³¹ in the original description of the technique, and with more recent laparoscopic reference.

Two specific features of the present study deserve emphasis. First, incorporating preoperative tumor ICG and intraoperative territorial ICG in the same protocol creates complementary information: the preoperative injection localizes and characterizes the lesion (including microfoci), while the intraoperative injection defines the area to be removed. The added value of this combined strategy is evident in the 20% rate of additional lesions detected intraoperatively, missed by preoperative imaging—a number consistent with independent reports of 10% to 15% and sufficient, when present, to alter the resection plan on the operating table. Second, the trial was conducted in a real-world Chinese HBV-dominant cohort, in which cirrhosis was present in roughly 80% of the patients. The well-documented observation that cirrhotic parenchymal fluorescence can obscure tumor contrast was evident in our 3 failures of tumor visualization, all occurring in the patients with ICG-R15 of 5% or higher, reinforcing the practical recommendation to delay ICG administration in individuals with impaired hepatic clearance.¹⁹

From a safety perspective, major adverse events were infrequent and evenly distributed, consistent with the sparing physiological profile of laparoscopic hepatectomy and the absence of known systemic toxicity of ICG at the doses used. The significant 32% reduction in overall complications reflects chiefly the avoidance of smaller, clinically meaningful adverse events—bile leakage, postoperative ascites, and pleural effusions—whose incidence is tied directly to the precision of parenchymal transection and the handling of pedicle structures. The shorter LOS (by 1.3 d) in the ICG group is correspondingly intuitive and translates into measurable economic value in high-volume units.³²

Limitations

Several limitations must be acknowledged. First, this was a single-center trial with a modest sample size and an inevitable learning-curve bias, since both operators were already experienced with ICG-guided surgery before enrollment. Second, although follow-up extended for up to 38 months and median follow-up exceeded 18 months, the absolute number of recurrences and deaths remained small, so 1-year DFS and OS differences that visually favored ICG did not reach significance; whether the surgical and pathological advantages we observed will translate into durable oncological gains awaits maturation of the cohort. Third, our pragmatic randomization produced a small but significant imbalance in cirrhosis prevalence, which we addressed through adjusted Cox modelling, but which cannot be completely erased. Fourth, the lack of blinding for the operating surgeons, unavoidable for a fluorescence trial, introduced potential performance bias that could only be mitigated by blinded outcome assessment, as we did for pathology and survival. Fifth, although both positive and negative staining were applied pragmatically, we did not power the trial to compare these 2 strategies directly—a question that larger multicenter studies, such as the recently registered ICG–human serum albumin trial, will need to address.¹⁵

Despite these caveats, the clinical implications are clear. At a minimum, ICG fluorescence should be considered a standard adjunct to laparoscopic anatomical hepatectomy for HCC in centers with appropriate equipment and expertise, as reflected in the recent international and Asian professional guidelines.⁹ Future work should validate our findings in multicenter randomized trials, explore second-generation fluorophores or ICG-albumin conjugates that enhance contrast in cirrhotic liver,¹⁵ and incorporate ICG-based margin assessment into real-time intraoperative quality metrics for hepatobiliary surgery. Recent work published in this journal further illustrates the expanding role of ICG fluorescence across minimally-invasive surgery, from artificial intelligence–based recognition of the hepatocystic triangle during cholecystectomy³³ to augmented-reality lymphatic mapping in colorectal cancer,³⁴ and reinforces the need for continued standardization of fluorescence protocols.

CONCLUSIONS

In the patients with resectable primary HCC, ICG fluorescence–guided laparoscopic anatomical hepatectomy produced large and clinically meaningful improvements over conventional high-definition laparoscopic hepatectomy: a higher rate of anatomical resection, a 100% R0 resection rate, wider surgical margins, less intraoperative blood loss, halved overall complications, and shorter LOS, without prolonging operative time. Early DFS trended in favor of the ICG strategy, with adjustment for cirrhosis and tumor size, preserving the direction of the DFS effect. OS curves were closely overlapping and nonsignificantly different between the groups. These findings support the incorporation of ICG fluorescence imaging into the standard operating protocol for laparoscopic anatomical hepatectomy in HCC and provide a practical basis for a prospective standardized operating procedure that other hepatobiliary units can adopt.

Funding Statement

This work was supported by the Jiangxi Provincial Natural Science Foundation Project (20202BABL206076; to LZ), Leading Talent Project of Ganzhou Science and Technology Bureau (2024030080; to LZ), and the Science and Technology Plan of Jiangxi Provincial Administration of Traditional Chinese Medicine (2024B0295; to LZ).

CONTRIBUTION STATEMENT

YH: conceptualization, methodology, formal analysis, and writing of the original draft. YXiao and YXie: methodology, investigation, data curation, and writing of the original draft. XD, YL, and WC: investigation, data curation, validation, and writing of the original draft. LZ: conceptualization, formal analysis, project administration, review, and editing. All authors read and approved the final version of the manuscript.

AI STATEMENT

Artificial intelligence was not used in the preparation of this manuscript.

HOW TO CITE

He Y, Xiao Y, Xie Y, et al. Indocyanine green fluorescence–guided laparoscopic anatomical hepatectomy for radical resection of hepatocellular carcinoma: a clinical study. Wideochir Inne Tech Maloinwazyjne. 2026; 21: 233-243. doi:10.20452/wiitm.2026.18047

JOURNAL INFORMATION

Videosurgery and Other Miniinvasive Techniques is an official journal of the Videosurgery Foundation.