This retrospective population-based cohort study used data extracted from the HIRA claims database. HIRA maintains a National Health Insurance (NHI) database that covers all medical claims in Korea and as they are generated in the process of reimbursing claims for healthcare services under the NHI system in Korea.13 It contains healthcare information for almost the Korean population, including demographics, in-hospital treatment, disease diagnosis, prescription, and provider information.14 The NHI system has covered almost all of Korea’s total population, which numbered approximately 50 million as of 2014.14 This study was conducted per the declaration of Helsinki and was approved by the Institutional Review Board (IRB) of the Seoul National University Bundang Hospital with a waiver of informed consent (IRB number, X-2312-868-903).

This study aimed to investigate trends in treatment patterns for HCC in South Korea. Disease diagnosis was based on the International Statistical Classification of Diseases and Related Health Problems, 10th revision (ICD-10) diagnosis codes, with data on procedures and prescriptions in the HIRA database. A total of 242,796 patients with HCC were enrolled from January 1, 2008 to December 31, 2022. In Korea, HCC was defined as ICD-10 code C22.0 with a cancer-specific deductible insurance code (V193). We excluded 71,794 patients for the following reasons: 1) history of prior treatment for HCC within 1 year before January 1, 2008 (washout period) (n=25,433), 2) diagnosis of other malignancies before HCC diagnosis (n=42,707), and 3) primary treatment for malignancies other than HCC in patients with multiple primary cancers (n=3,654). In the final analysis, 171,002 patients newly diagnosed with HCC were enrolled from January 1, 2008 to December 31, 2022 (Fig. 1).

The etiologies of HCC are categorized as hepatitis B virus (HBV), hepatitis C virus (HCV), alcoholic liver disease (ALD), nonalcoholic fatty liver disease (NAFLD), and others (autoimmune hepatitis, primary biliary cholangitis, Wilson’s disease, Budd-Chiari syndrome, and cryptogenic). Cases in which HBV-or HCV-related diagnostic codes or the corresponding antiviral agents for HBV or HCV were used once during the study period were regarded as HBV- or HCV-related, respectively (Supplementary Table 1). Cases in which ALD diagnostic codes were present at least once were defined as ALD-related. NAFLD was determined based on diagnostic codes. We also defined NAFLD as cases with diagnostic codes for diabetes mellitus if the patients did not have any cause for the above-mentioned chronic liver diseases (Supplementary Table 1). When patients had more than two of the etiology of HCC, we classified them in the following order: HBV, HCV, others (Wilson’s disease, autoimmune hepatitis, primary biliary cholangitis, and Budd-Chiari syndrome), ALD, NAFLD, and others (cryptogenic), which prioritizes causes based on the relative strength of their association with HCC.15 For instance, we regarded HBV as the etiology of HCC if the patient has chronic hepatitis B and NAFLD in the present study. However, patients with autoimmune hepatitis, primary biliary cholangitis, Wilson’s disease, or Budd-Chiari syndrome were excluded from the NAFLD group and were considered as part of the other group.

Treatment modalities for HCC were classified into surgical resection (including laparoscopic surgical resection), liver transplantation, local ablation therapy (radiofrequency ablation [RFA], microwave ablation [MWA], and cryoablation), transarterial therapy (conventional transarterial chemoembolization [cTACE], drug-eluting beads transarterial chemoembolization [DEB-TACE], transarterial radioembolization [TARE], or other transarterial therapy), radiation therapy, and systemic treatment (sorafenib, lenvatinib, and AB), and best supportive care. cTACE was defined as the use of lipiodol and the chemotherapeutic agents doxorubicin or cisplatin. DEB-TACE was defined as the use of DC beads or HepaSpheres (Supplementary Table 2). Cases in which either lipiodol or chemotherapeutic agents such as doxorubicin, cisplatin, or 5-fluorouracil alone were considered as other transarterial therapies. Best supportive care (BSC) was defined as the absence of treatment for HCC after its diagnosis. We excluded patients who received 1st line systemic treatment other than HCC (sorafenib for thyroid cancer and renal cell carcinoma, lenvatinib for thyroid cancer, atezolizumab for lung cancer, and bevacizumab for lung cancer, colorectal cancer, breast cancer, ovarian cancer, cervical cancer, and renal cell carcinoma).

The clinical outcomes in the present study were analyzed as follows. We investigated the annual and crude incidence rates of HCC. The annual incidence of HCC was determined based on the number of patients with HCC each year. The crude incidence rate was defined as cases per 100,000 persons, which was calculated using the mid-year resident registration population each year.16 The annual incidence rate and the crude incidence rate of HCC were analyzed by etiology of HCC (HBV, HCV, alcohol, NAFLD, and others), and age group (<29, 30-39, 40-49, 50-59, 60-69, 70-79, and ≥80 years). Second, the trend in initial treatment for HCC was investigated between 2008 and 2022. Initial treatments refers to the first therapeutic intervention after diagnosis, including surgery, transarterial therapy, or systemic therapy. First-line systemic therapy, on the other hand, refers to the first systemic therapy administered as an initial treatment or following other treatments. The treatment modalities were classified as surgical resection, local ablation therapy, liver transplantation, transarterial therapy, systemic treatment, radiation therapy, or BSC. We also analyzed the trend of initial treatment for HCC according to age group, HCC etiology (viral vs. non-viral), and place of residence (metropolitan [Seoul, Incheon, and Gyeonggi-do] vs. non-metropolitan areas). The index year of the initial treatment for HCC was determined based on the year of HCC diagnosis, excluding the analysis of first-line systemic therapy. First-line systemic therapy was calculated based on the date of the initial prescription of each systemic agent. Finally, detailed subgroup analyses were conducted on the initial HCC treatment according to each modality: surgical resection (open and laparoscopic approaches), local ablation therapy (RFA, MWA, and cryoablation), transarterial therapy (cTACE, DEB-TACE, TARE, and other transarterial therapies), and first-line systemic treatment (sorafenib, lenvatinib, and AB). Additionally, the annual number of treatments performed and the number of patients treated with each modality were investigated by year. R software (version 3.5.1; R Foundation for Statistical Computing, Vienna, Austria; http://cran.r-project.org/) was used for the statistical analyses. GraphPad Prism (version 6.0; GraphPad Software, La Jolla, CA, USA) was used for the statistical analyses and graph generation.

We first investigated whether the epidemiology of HCC in Korea changed during the study period, as it can affect treatment patterns. Regarding the annual incidence, there was a gradual decrease in the total number of newly diagnosed HCC cases from 2008 to 2022. In 2008, there were 11,814 cases, which slightly decreased to 10,443 cases by 2022 (Supplementary Fig. 1A). The male-to-female ratio is 3.56:1, showing a slight decrease from 3.62:1 in 2008 to 3.39:1 in 2022. In addition, the crude incidence rate (CR) per 100,000 people, calculated using the midyear resident registration population for each year, also showed a declining trend. In 2008, the CR was 23.9 cases per 100,000 people, which is reduced to 20.4 cases per 100,000 by 2022 (Supplementary Fig. 1A), indicating a slight overall decrease in the incidence of HCC over the observed period.

Next, we investigated the age-specific trends in the number of cases and CRs. The distribution of HCC cases across different age groups showed a significant concentration in older age groups. Specifically, the 70s and 80s or older age groups showed a marked increase in cases compared to previous years (Supplementary Fig. 1B). Furthermore, the CR increased from 61.1 to 67.4 per 100,000 people from 2008 to 2022 for the 80s or older age group. In contrast, other age groups showed a gradual decrease in CR (Supplementary Fig. 1B). Importantly, in 2008, the 80s or older group accounted for 3.8% of HCC cases, which grew to 13.1% by 2022. Similarly, the proportion of 60s or more patients from 48.0% in 2008 to 67.0% in 2022 (Fig. 2A). Collectively, these data suggest a notable increase in the proportion of the elderly population with HCC.

In terms of the etiology of HCC, the distribution of HCC cases by etiology also revealed a shift over the years. HBV infection remains the predominant cause, accounting for the majority of cases. However, the number of cases attributed to ALD and NAFLD is expected to increase significantly by 2022 (Supplementary Fig. 1C). The percentage distribution reflected these changes. In 2008, HBV accounted for 68.9% of all HCC cases; however, this percentage will decrease to 59.7% by 2022. Meanwhile, the proportion of cases attributed to NAFLD increased from 7.5% in 2008 to 11.8% in 2022, and that of ALD increased from 8.9% to 15.8% during the same period (Fig. 2B, Supplementary Fig. 1C). These trends indicate a shifting landscape in the etiology of HCC, with a growing emphasis on nonviral etiologies, including ALD and NAFLD.

Importantly, we analyzed the distribution of initial treatment modalities for HCC, which have evolved over the years (Fig. 3). In patients diagnosed with HCC in 2008, transarterial therapy was the predominant treatment, accounting for 5,890 cases (49.9%). However, by 2022, the proportion of transarterial therapy had decreased to 3,820 cases (36.6%). Surgical resection increased from 1,435 cases (12.2%) in 2008 to 2,221 cases (21.3%) in 2022. Only 21 cases (0.2%) of patients diagnosed with HCC in 2008 received systemic therapy, but this proportion increased to 1,002 cases (9.6%) by 2022. These trends highlight the increasing use of surgical resection and systemic therapy for HCC treatment, reflecting advancements in these treatments and their broader acceptance. The BSC decreased from 3,742 cases (31.7%) in 2008 to 2,226 cases (21.3%) in 2022, indicating a shift towards more active treatments.

We then analyzed each treatment method by subdividing it into more specific categories. Surgical resection methods have seen significant changes between 2008 and 2022. The proportion of laparoscopic resections increased substantially from 10.6% in 2008 to 60.6% in 2022, indicating a shift towards minimally invasive surgical techniques (Fig. 4A). Local ablation techniques, including RFA, MWA, and cryoablation, have been introduced, and evolved over time. Although RFA remained dominant, accounting for 100% of the local ablations in 2008, its proportion decreased to 75.3% by 2022. MWA and cryoablation have been increasingly adopted, reaching 24.3% and 0.3% by 2022 (Fig. 4B). Transarterial therapy methods, such as cTACE, DEBTACE, and TARE, have also evolved. cTACE, which accounted for 88.1% of transarterial therapies in 2008, decreased to 79.0% by 2022. DEB-TACE and TARE were introduced later, accounting for 7.7% and 8.4%, respectively, by 2022 (Fig. 4C).

The pattern of systemic therapy as the initial treatment for HCC has also changed markedly. Notably, the introduction of new systemic therapies, such as AB and lenvatinib, has significantly altered the treatment landscape. Lenvatinib accounted for 35.7% and 36.5% in 2020 and 2021, respectively; however, this proportion decreased to 15.9% by 2022. AB accounted for 18.2% of systemic therapy cases in 2021 following approval but rose to 63.1% by 2022 following the introduction of national health insurance reimbursement (Fig. 4D).

We also performed subgroup analyses of the initial treatment according to age, etiology, and place of residence. In the 30s, 40s, and 50s age groups, treatment patterns remained relatively stable; however, in the 70s and 80s age groups, surgical resection increased significantly (from 6.3% to 19.2% in the 70s and from 0.7% to 6.8% in the 80s or older). BSC was also significantly decreased in the 70s and 80s age groups (Supplementary Fig. 2A). For viral etiologies, the treatment pattern was similar to the overall trend, with minor variations. In non-viral etiologies, the proportion of BSC was higher. The proportions of other treatments, including surgical resection and systemic therapy, were generally lower than those in viral etiologies (Supplementary Fig. 2B). In addition, the differences in treatment patterns between the metropolitan and non-metropolitan areas were diminished (Supplementary Fig. 2C). Metropolitan areas showed an increase in surgical resection and a decrease in transarterial therapy. Nonmetropolitan areas also showed a slight decline in transarterial therapy and increased surgical resection and systemic therapy.

Next, we investigated the changes in the patterns of first-line systemic therapies, including sorafenib, lenvatinib, and AB, following their approval for national health insurance reimbursement, which is not restricted to the initial treatment for HCC at diagnosis. In 2008, nine patients with HCC were treated with first-line systemic therapy, which increased markedly to 3,249 by 2022 (Fig. 5A). Sorafenib was the sole therapy used until 2018, followed by lenvatinib in 2018 and AB in 2020. Following its approval and insurance coverage in 2020, AB soared to 1,954 cases by 2022, becoming the most frequently used first-line systemic therapy. Furthermore, a more granular analysis from 2019 to 2023 by quarter showed the significant impact of these new therapies on usage patterns (Fig. 5B). Sorafenib usage decreased consistently from 568 cases in Q1 2019 to 86 cases in Q1 2023. In contrast, lenvatinib usage peaked at 350 cases in Q4 2019 but decreased consistently to Q1 2023. AB showed a dramatic increase in post-approval and insurance coverage, becoming the dominant therapy, with 618 cases in Q1 2023. The proportion of each systemic therapy used as first-line treatment has shifted significantly. Sorafenib accounted for 100% of the cases in Q1 2019; sorafenib dominated with 66.4% in Q2 2020, with lenvatinib accounting for 33.6%; and AB increased to 78.2% in Q1 2023, with lenvatinib and sorafenib accounting for 10.9% each (Fig. 5C). These trends underscore the rapid evolution of first-line systemic therapies for HCC and highlight the increasing preference for newer therapies through the government’s approval of national insurance coverage.

To investigate the total annual cases of each treatment, regardless of the initial or treatment line, we included patients diagnosed before 2008. A patient flowchart is presented in Supplementary Fig. 3. The number of procedures performed and the number of patients receiving therapy showed distinct patterns over time. Transarterial therapy remains the most commonly performed procedure, slightly decreasing from 26,895 in 2008 to 23,335 in 2022. Surgical resections have markedly increased from 1,724 in 2008 to 3,587 in 2022. In contrast, local ablation decreased from a peak of 4,032 in 2015 to 3,201 in 2022. Both systemic and radiation therapies resulted in significant increases. The number of systemic therapy cases increased sharply from 23 in 2008 to 4,501 in 2022, whereas the number of radiation therapy cases increased from 917 in 2008 to 4,126 in 2022 (Supplementary Fig. 4).

In the subgroup analyses (Supplementary Fig. 5), the number of laparoscopic resections increased significantly, peaking at 1,909 cases (53.2%) in 2022. Open approaches decreased in comparison with the initial treatment. RFA remained the dominant local ablation method, although its use has slightly decreased. MWA and cryoablation have been increasingly adopted, with MWA reaching 735 cases (23.0%) by 2022. cTACE remained the predominant transarterial therapy, although its use has decreased slightly. TARE showed a gradual increase in adoption, reaching 474 cases (2.0%) by 2022. Systemic therapy showed a pattern similar to that observed with first-line systemic therapy.