According to data from the World Health Organization's Global Cancer Statistics 2020 (Globocan), liver cancer presents an incidence rate of 4.7% and a mortality rate of 8.3%, ranking it as the fifth most common cancer and the third leading cause of cancer-related mortality worldwide [1]. Hepatocellular carcinoma (HCC) is the predominant type of malignant liver tumor [2]. A study from the National Referral Center, Dr. Cipto Mangunkusumo Hospital, showed that the one-year survival rate for HCC patients increased from 24.1% in 1998 to 1999 to 29.4% in 2013 to 2014, while the median survival time decreased from 146 days to 138 days over the same period [3]. In Indonesia, the age-standardized mortality rate for HCC is 7.7 per 100,000 population [1]. This figure aligns with the 3-year HCC mortality rate of 94.4% reported in two cancer centers in Indonesia [4]. HCC patients in Southeast Asia (Thailand and Malaysia) exhibit low survival rates and high mortality due to late diagnosis. HCC remains a significant health concern in Southeast Asia, especially in Indonesia, given its large population [3].

In Indonesia and other Southeast Asian countries, Hepatitis B virus (HBV) and Hepatitis C virus (HCV) are the primary causes of HCC [3,5-9]. Together, HBV and HCV infections contribute to approximately 80% of HCC cases, especially in patients with advanced cirrhosis or fibrosis. The incidence of these infections is expected to rise over the next 10 to 20 years, peaking around 2030 [10]. In resectable HCC with Child-Pugh A, liver resection is regarded as the first-line curative treatment, although the optimal surgical strategy remains under debate as indications for liver resection vary. Current guidelines suggest that surgery is limited to patients with early-stage liver cancer (the Barcelona Clinic Liver Cancer [BCLC] score 0-A) [11,12]. Despite these guidelines, adhering to the BCLC strategy for prognosis prediction and treatment recommendations proves challenging, as 50% of patients undergoing surgery at tertiary hospitals globally are in advanced stages of HCC [13]. According to guidelines from the American Association for the Study of Liver Diseases (AASLD) and the European Association for the Study of the Liver (EASL), criteria for liver resection are increasingly restricted to patients with single nodules and Child-Pugh A who show no signs of portal hypertension [12,14].

Several Asian countries’ guidelines suggest that liver resection is an appropriate therapy for HCC patients with (1) more advanced tumor stages and (2) poorer liver function [15]. In Japan, the treatment algorithm advocates liver resection for patients either with a single nodule of any size, disregarding macrovascular invasion, or for those with up to 3 nodules, regardless of the size [16-18]. Conversely, the algorithm by Hong Kong Liver Cancer endorses liver resection for early-stage tumors (≤ 3 nodules, ≤ 5 cm, without intrahepatic venous invasion) and intermediate tumors (> 3 nodules, ≤ 5 cm, with intrahepatic venous invasion; or 3 nodules, > 5 cm, without intrahepatic venous invasion) [19]. Meanwhile, the Korean algorithm imposes fewer limitations, recommending liver resection whenever the procedure is feasible for the patient [20].

Until then, the data on risk factors contributing to the high mortality rate among HCC patients in Indonesia have been limited, particularly in those who have undergone liver resection. The reluctance to perform liver resection on large HCC tumors stems from a scarcity of survival analysis. Surgeons at our institution are experienced and adept, frequently encountering large tumors and proficient at performing resections. Thus, it is pertinent to investigate whether this option is the optimal recommendation for patients. This study aimed to investigate mortality outcomes and survival while analyzing the risk factors affecting these outcomes in HCC patients who underwent liver resection at National Referral Center, Dr. Cipto Mangunkusumo Hospital.

Study design and patients

We conducted a retrospective cohort study at Dr. Cipto Mangunkusumo Hospital from 2010 to 2020. A total of 85 subjects with a confirmed diagnosis of HCC who underwent liver resection met the inclusion criteria. Patients with incomplete data on subject characteristics, clinical features, or mortality outcomes post-liver resection were excluded. The indication for liver resection was resectable HCC of any size, with or without prior neoadjuvant therapy. All results were reviewed and reported according to the STROBE reporting guidelines for retrospective studies (Supplementary Table 1) [21]. Ethical approval was granted by the Institutional Review Board (IRB) of the Ethics Committee of the Dr. Cipto Mangunkusumo Hospital, Faculty of Medicine, Universitas Indonesia (protocol number: 24-01-0149).

Data collection

All data were obtained from the hospital’s medical records. The collected characteristics of subjects included sex, age at surgery, hepatitis status, tumor size, Edmondson-Steiner (ES) grade, tumor differentiation, Child-Pugh classification, and BCLC stage. Tumor sizes were categorized into extra-large (> 10 cm), large (> 5–10 cm), medium (3–5 cm), and small (< 3 cm) [22-27]. The diagnosis of HCC was confirmed through radiological and histopathological examinations. Data on mortality outcome following liver resection and survival time (in months) were recorded up to January 1, 2022. Survival time was investigated through medical records, routine outpatient visits, and follow-up phone calls.

Statistical analysis

Statistical analysis was conducted using IBM SPSS version 26.0 (IBM Corp.). Continuous data were presented as median (min–max), and categorical data were shown as frequency (percentage). Survival time and mortality rate were analyzed using the Kaplan-Meier method, along with the log-rank test and Tarone-Ware test. Significant risk factors related to HCC mortality were identified through bivariate analysis. Risk factors with p < 0.25 or that were clinically significant were included in the multivariate analysis. Both bivariate and multivariate analyses utilized the Cox proportional hazards model. All variables with p < 0.05 were considered statistically significant. The unadjusted and adjusted hazard ratios (HRs) with 95% confidence intervals (95% CIs) were provided for each risk factor.

Seventy participants from among 85 HCC patients who underwent liver resection surgery between 2010 and 2020 were included in this study. Fifteen patients were excluded due to incomplete medical records or being unreachable (loss to follow-up). The characteristics of liver resection for HCC patients are summarized in Table 1.

Table 1 . Characteristics of liver resection patients with hepatocellular carcinoma

	All subjects (n = 70)	Survivor (n = 20)	Non-survivor (n = 50)
Sex
Female	17 (24.3)	5 (29.4)	12 (70.6)
Male	53 (75.7)	15 (28.3)	38 (71.7)
Age at surgery (yr)	53.5 (33–76)	53.5 (33–76)	53.5 (34–74)
Age at surgery
< 40 years	7 (10.0)	2 (28.6)	5 (71.4)
40–60 years	38 (54.3)	11 (28.9)	27 (71.1)
> 60 years	25 (35.7)	7 (28.0)	18 (72.0)
Hepatitis status
Non-reactive	21 (30.0)	6 (28.6)	15 (71.4)
Hepatitis B	42 (60.0)	11 (26.2)	31 (73.8)
Hepatitis C	7 (10.0)	3 (42.9)	4 (57.1)
Tumor size
Small (< 3 cm)	12 (17.1)	4 (33.3)	8 (66.7)
Medium (3–5 cm)	13 (18.6)	5 (38.5)	8 (61.5)
Large (> 5–10 cm)	14 (20.0)	3 (21.4)	11 (78.6)
Extra-large (> 10 cm)	31 (44.3)	8 (25.8)	23 (74.2)
Edmondson-Steiner grade
1	17 (24.3)	4 (23.5)	13 (76.5)
2	13 (18.6)	5 (38.5)	8 (61.5)
3	31 (44.3)	11 (35.5)	20 (64.5)
4	9 (12.9)	0 (0.0)	9 (100)
Tumor differentiation
Well	13 (18.6)	3 (23.1)	10 (76.9)
Moderate	33 (47.1)	11 (33.3)	22 (66.7)
Moderate to poor	7 (10.0)	3 (42.9)	4 (57.1)
Poor	17 (24.3)	3 (17.6)	14 (82.4)
Child-Pugh classification
A	54 (77.1)	15 (27.8)	39 (72.2)
B	13 (18.6)	2 (15.4)	11 (84.6)
C	3 (4.3)	3 (100)	0 (0.0)
BCLC stage
A	46 (65.7)	13 (28.3)	33 (71.7)
B	23 (32.9)	6 (26.1)	17 (73.9)
C	1 (1.4)	1 (100)	0 (0.0)
Mortality
No	20 (28.6)
Yes	50 (71.4)

Values are presented as number (%) or median (range).

BCLC, Barcelona Clinic Liver Cancer.

The median age of all participants was 53.5 years (range, 33–76 years), with 38 patients (54.3%) falling within the 40–60 age group. Based on tumor size, 31 patients (44.3%) had an extra-large tumor size (> 10 cm). Fig. 1 displays a case of extra-large HCC. Hepatitis B was reactive in 42 patients (60.0%). A majority of participants were classified as ES grade 3 (n = 31, 44.3%), moderately differentiated tumors (n = 33, 47.1%), Child-Pugh A (n = 54, 77.1%), and BCLC stage A (n = 46, 65.7%).

Fig 1. Extra-large tumor size (19.5 cm × 25.0 cm) resected from a patient with hepatocellular carcinoma in segments 4, 5, 6, 7, and 8.

A higher proportion of mortality was observed among males (71.7%), patients older than 60 years (72.0%), those with hepatitis B (73.8%), those with extra-large tumors (74.2%), ES grade 4 (100%), moderate tumor differentiation (66.7%), Child-Pugh B (84.6%), and BCLC stage B (73.9%).

Mortality rate, survival rate, and median survival time

The mortality rate of HCC following liver resection in this study was 71.4% (n = 50). The overall survival analysis (Fig. 2 and Table 2) indicated a 1-year survival rate of 55.7%, and the 5-year survival rate decreased to 28.6%. For participants with extra-large HCC, the 1-year survival rate was lower, at 38.7%. Nonetheless, eight participants with extra-large tumor sizes survived over three years post-resection, categorized as BCLC stage A and Child-Pugh A. Table 2 details annual survival rates post-liver resection.

Fig 2. Survival of hepatocellular carcinoma patients following liver resection, categorized by (A) overall survival; (B) sex; (C) age at surgery; (D) hepatitis status; (E) tumor size; (F) Edmonson-Steiner grade; (G) tumor differentiation; (H) Child-Pugh classification; and (I) Barcelona Clinic Liver Cancer (BCLC) stage.

Table 2 . Survival rate of liver resection patients with HCC

	Time of follow-up after liver resection
	Initial	1 year	2 years	3 years	5 years	> 5 years
Overall
Survive (n)	70	39	33	26	21	20
Survival rate (%)	100	55.7	47.1	37.1	30.0	28.6
Extra-large HCC
Survive (n)	31	12	11	9	8	8
Survival rate (%)	100	38.7	35.5	29.0	25.8	25.8

HCC, hepatocellular carcinoma.

Median survival time, during which half of the subjects died, is a standard metric for comparing survival rates across groups with variability. According to Table 3, the median overall survival for HCC patients post-liver resection was 19.0 months. Higher median survival rates were noted in males (21.0 months), patients older than 60 (21.0 months), those with hepatitis C (79.0 months), patients with medium tumor size (33.0 months), ES grade 3 (32.0 months), moderate-to-poor tumor differentiation (33.0 months), Child-Pugh C (30.0 months), and BCLC stage A (19.5 months). The Kaplan-Meier curve (Fig. 2) also mirrors these trends. Survival rate declined more rapidly among females, those under 40, patients with hepatitis B, extra-large tumor, ES grade 4, poor tumor differentiation, Child-Pugh B, and BCLC stage B. Only Child-Pugh was found to be statistically significant in association with survival outcomes (log-rank p = 0.011 and Tarone-Ware p = 0.006).

Table 3 . Median survival time of patients with hepatocellular carcinoma after liver resection

	Median survival (mon)	95% CI	p-value
			Log-ranka)	Tarone-Ware
Overall survival	19.0	6.8–31.2
Sex			0.791	0.656
Female	19.0	0.3–37.7
Male	21.0	6.3–35.7
Age			0.966	0.853
< 40 years	5.0	0.0–12.7
40–60 years	15.0	1.9–28.0
> 60 years	21.0	2.8–39.2
Hepatitis status			0.829	0.900
Non-reactive	21.0	6.5–35.5
Hepatitis B	16.0	1.2–30.8
Hepatitis C	79.0	-
Tumor size			0.327	0.168
Small (< 3 cm)	5.0	0.0–15.2
Medium (3–5 cm)	33.0	16.5–49.5
Large (> 5–10 cm)	21.0	8.8–33.2
Extra-large (> 10 cm)	8.0	1.5–14.5
Edmondson-Steiner grade			0.318	0.229
1	13.0	2.2–23.8
2	27.0	18.2–35.8
3	32.0	15.9–48.0
4	1.0	0.3–1.7
Tumor differentiation			0.361	0.349
Well	15.0	0.0–40.8
Moderate	21.0	12.3–29.7
Moderate to poor	33.0	0.0–84.3
Poor	2.0	0.0–8.1
Child-Pugh classification			0.011*	0.006*
A	20.0	18.2–33.8
B	5.0	0.6–16.8
C	30.0	0.4–62.9
BCLC stage			0.316	0.314
A	19.5	16.6–33.4
B	12.0	8.2–27.9
C	-	-

BCLC, Barcelona Clinic Liver Cancer; CI, confidence interval.

^a)Mantel-Cox test.

*Statistically significance (p < 0.05).

Risk factor of mortality

In the bivariate analysis, ES grade 4 had an unadjusted HR of 3.2 (95% CI: 1.3–7.7; p = 0.011), while Child-Pugh B had an unadjusted HR of 2.2 (95% CI: 1.1–4.3; p = 0.026), both statistically significant in increasing mortality risk. Despite these findings, all variables were included in the multivariate analysis due to their clinical relevance (Table 4). Consequently, only the Child-Pugh B classification significantly increased the mortality risk, with an HR of 2.3 (95% CI: 1.0–5.2; p = 0.046).

Table 4 . Bivariate and multivariate analysis of mortality risk factors

	Unadjusted hazard ratio (95% CI)	p	Adjusted hazard ratio (95% CI)	p
Sex
Female	1	-	1	-
Male	1.090 (0.576–2.095)	0.795	0.837 (0.392–1.786)	0.645
Age
< 40 years	1	-	1	-
40–60 years	0.887 (0.340–2.314)	0.807	0.837 (0.289–2.422)	0.742
> 60 years	0.882 (0.327–2.381)	0.804	0.791 (0.258–2.421)	0.681
Hepatitis status
Non-reactive	1	-	1	-
Hepatitis B	1.018 (0.548–1.892)	0.954	0.995 (0.468–2.114)	0.990
Hepatitis C	0.741 (0.245–2.238)	0.595	1.354 (0.388–4.725)	0.634
Tumor size
Small (< 3 cm)	1	-	1	-
Medium (3–5 cm)	0.610 (0.228–1.633)	0.610	0.717 (0.234–2.190)	0.559
Large (> 5–10 cm)	0.997 (0.400–2.483)	0.997	0.965 (0.328–2.840)	0.948
Extra-large (> 10 cm)	1.279 (0.571–2.866)	1.279	1.692 (0.622–4.603)	0.303
Edmondson-Steiner grade
1	1	-	1	-
2	0.643 (0.263–1.570)	0.332	0.430 (0.088–2.103)	0.297
3	0.703 (0.348–1.422)	0.327	0.332 (0.073–1.508)	0.153
4	3.158 (1.298–7.683)	0.011*	1.367 (0.276–6.780)	0.702
Tumor differentiation
Well	1	-	1	-
Moderate	0.822 (0.388–1.742)	0.610	1.474 (0.306–7.102)	0.629
Moderate to poor	0.579 (0.178–1.879)	0.363	1.018 (0.137–7.571)	0.986
Poor	1.328 (0.586–3.009)	0.497	2.429 (0.443–13.308)	0.307
Child-Pugh classification
A	1	-	1	-
B	2.178 (1.095–4.331)	0.026*	2.301 (1.015–5.215)	0.046*
C	0.000 (0.000)	0.974	0.000 (0.000)	0.983
BCLC stage
A	1	-	1	-
B	1.297 (0.720–2.338)	0.386	1.595 (0.806–3.153)	0.173
C	0.000 (0.000)	0.972	0.000 (0.000)	> 0.999

BCLC, Barcelona Clinic Liver Cancer; CI, confidence interval.

*Statistically significant (p < 0.05).

The median age of liver resection patients with HCC in this study was 53.5 years (range: 33–76 years), closely aligning with other Indonesian studies that reported ages around 54 to 55 years. In this study, patients aged < 40 years had the lowest median survival, whereas those aged > 60 years had the highest median survival, despite a 72.2% mortality rate in this group. The high prevalence of hepatitis B in Indonesia contributes to the high number of liver resection patients under 60 years old. Hepatitis B infection in Indonesia is significantly high compared to other Asian countries [4,28]. A previous study by Sanyal et al. [29] did not demonstrate differences in mortality or overall survival rates, although there was a trend toward lower mortality and better survival in patients older than 80 years. This suggests that age should not be a limiting indication for liver resection in HCC patients. Yasuda et al. [30] also found that age did not impact liver regeneration rate. The increasing percentage volume of resected liver was directly associated with an increased liver regeneration rate. Most subjects older than 60 years had small tumor sizes, which may explain their higher survival times.

Patients undergoing liver resection due to HCC were predominantly male. There is no statistical difference between male and female to survival or mortality. Other studies have also reported a large number of male patients undergoing liver resection for HCC [3,4]. In Southeast Asia, the incidence of liver resection among males exceeded 20 cases per 100,000 population. Hormonal differences between the sexes are significant risk factors for HCC. Testosterone acts as a positive regulator of the hepatocyte cell cycle, thereby accelerating hepatocarcinogenesis. Moreover, androgens inhibit p53 function, impair DNA repair mechanisms, and increase oxidative stress. Males are also more likely to smoke and consume alcohol, further contributing to the higher prevalence of HCC in this group. In contrast, estradiol suppresses the cell cycle and slows HCC development [4,28].

Patients with extra-large tumor sizes experienced a lower survival rate, reduced median survival, and a higher mortality risk. These results are clinically relevant, despite not being statistically significant. Generally, a larger tumor size is a factor that decreases the survival rate following liver resection [31,32]. Tumors larger than 5 cm predict 2-year recurrence and significantly raise the mortality risk compared to tumors smaller than 5 cm [33,34]. The number of patients with extra-large tumor sizes was the highest in this study. Many patients referred to our hospital arrived too late, by which time the tumors had grown significantly in size. This indicates that although liver resection is possible for tumors larger than 10 cm, it doesn’t guarantee improved long-term survival outcomes. Retnowulan et al.’s study [35] indicated that larger tumor sizes were significantly associated with poor differentiation of HCC and microvascular invasion. HCC tumors larger than 3 cm had diploid DNA, which is of a more benign and non-aggressive nature. Endothelial cell proliferation and migration are promoted by overexpression of vascular endothelial growth factors (VEGFs), which also increases vascular permeability.

The high number of extra-large HCC tumors was associated with elevated mortality rates. Patients with more advanced tumor grades exhibited lower median survival, lower survival rates, and higher mortality risks. The mortality rate of patients post-liver resection in this study was notably high. Survival rates varied among patients who underwent liver resection. Almost 50% of patients died within the first year following liver resection. The mortality rate was even higher among patients with large and extra-large tumors. Kawasaki et al. [22] found that tumors larger than 3 cm were significant predictors of poor outcomes in recurrence-free survival. Their results align with those of Lu et al. [26], who identified larger tumor size as an independent factor for overall survival. They also discovered that increasing tumor sizes significantly correlate with the pathologic state, growth patterns, clinical stages, and biological stages (capsular invasion, tumor thrombi, and satellite nodules) [23-26]. Liver resection is typically performed as a first-line treatment for cirrhotic patients with early or very-early HCC stages and preserved liver function regardless of size [25,36]. In Indonesia, the national guideline for HCC management recommends resection in non-cirrhotic patients with a solitary tumor, good liver function (Child-Pugh A), no portal hypertension, tumor size under 5 cm, and anatomical resection. Resection of multifocal tumors is considered, with a maximum of three nodules, involving anatomical 2–3 segments, or in patients with mild portal hypertension who are not candidates for transplantation. This guideline incorporates the management algorithm from the Asian Pacific Association for the Study of the Liver (APASL) with some modifications (Supplementary Fig. 1) [37]. The Multidisciplinary Team approach is implemented at our institution as the new standard for managing HCC. Various specialties within the team discuss treatment plans during weekly meetings, which span from 3 to 9 months from the initial assignment to surgery (Supplementary Fig. 2) [27].

Although hepatitis B is frequently cited as the most common cause of HCC in Indonesia, supported by the fact that 60.0% (n = 42) of patients had reactive hepatitis B status. A correlation was observed between reactive hepatitis B status and lower median survival, although this result was not statistically significant [4,28].

Long-term survival rates for HCC patients across all stages of the BCLC improved following liver resection. In 2012, a single large HCC tumor (> 5 cm) was classified as BCLC stage A. The AASLD and the EASL suggested revising the BCLC classification. Clinically, tumors larger than 5 cm are considered a criterion for liver resection in early-stage HCC (BCLC-A). However, the prognosis for patients with BCLC-A was poorer than predicted by the Milan criteria. Intriguingly, research by Liu et al. [38] indicated no significant differences in outcomes.

This study indicated that both overall and extra-large HCC survival rates were lower than those reported in 15 international studies of large HCC, as reviewed by Cauchy et al. [39]. In that review, mortality rates ranged from 0.0% to 10.6%, with 5-year overall survival rates varying between 30% and 35%. Common risk factors for poor long-term survival included macroscopic vascular invasion, multiple lesions, underlying cirrhosis, and significant blood loss during surgery. In practice, liver resection for large HCC carries no inherent risk. Tumor size should not be a deterrent for liver resection [23]. However, as observed in our study, significant operative mortality continues to be a major challenge in performing liver resections for large HCC.

The limitations of this study were shaped by the characteristics of patients who were diagnosed late with HCC upon arrival at our hospital, a national referral tertiary center. Furthermore, limited data on patient characteristics and comorbidities were available due to incomplete medical records, and selection biases were inevitable in this retrospective, single-center study. Most clinical features from Indonesian patients did not significantly correlate with survival outcomes. Further research into tumor biology, type of surgery, and molecular and genomic profiles could reveal significant predictive factors for survival outcomes [26]. To validate the findings of this study, additional multicenter prospective studies and detailed examinations of tumor biology are necessary. Many HCC patients had extra-large tumors, providing a unique perspective on the outcomes of liver resection management for extra-large HCC.

We continue to perform liver resections for large HCC at our center because transcatheter arterial chemoembolization (TACE) is not routinely available, and transcatheter arterial radioembolization is not offered. Our registry indicates that two subjects underwent TACE prior to liver resection, and one subject received radio frequency ablation. We have also recognized that liver resection can be effectively used as follow-up management after TACE [40]. Another management strategy for HCC is liver transplantation. Since 2010, we have been transitioning towards transplantation for HCC management. Our center’s successfully perform 11 adult living donor liver transplants under supervision from Japan. However, the number of adult liver transplants in Indonesia remains low due to a shortage of living donors [41].

In conclusion, although it is technically feasible to resect extra-large HCC, this does not guarantee a cure or improved survival outcomes. Despite successful surgical interventions, neither larger tumor sizes nor higher tumor grades can increase mortality risk. The only factor that significantly impacts mortality risk and survival outcomes is the Child-Pugh class. These data suggest a trend of late-stage diagnosis, posing significant challenges. To enhance survival rates, it is vital to prioritize early diagnosis and timely management of HCC. Early intervention, particularly the surgical resection of smaller, lower-grade tumors, has demonstrated a notable improvement in survival rates. Thus, advancing early detection techniques and implementing swift treatments are imperative for reducing mortality and enhancing patient outcomes.

Supplementary data related to this article can be found at https://doi.org/10.14701/ahbps.24-179.

Authors thank to all staff of Digestive Division, Department of Surgery, Dr. Cipto Mangunkusumo Hospital who performed liver resection during 2010–2020. Also, Muhammad Luthfi Prasetyo, MD and Natan Kevin Partogu Siagian, MD for their assistance in HCC registry management during this research.

None.

No potential conflict of interest relevant to this article was reported.

Conceptualization: LS, ANLL. Methodology: LS, ANLL, ABP. Validation: LS, ANLL, RAS. Formal analysis: ABP. Investigation: LS, ANLL, RAS, YM, ASP, TJML. Data curation: ABP. Visualization: ABP. Supervision: ASP, TJML. Project administration: ABP. Writing – original draft: LS, ANLL, ABP. Writing – review & editing: RAS, YM, ASP, TJML.