Real-world treatment patterns and outcomes in patients with early-stage HCC in the US treated with resection or ablation

Abstract

Aim

Real-world outcomes in early-stage hepatocellular carcinoma (eHCC) are not well characterized. We aimed to evaluate treatment patterns and long-term outcomes in patients with eHCC treated with resection or ablation in the United States.

Materials and methods

We conducted a retrospective study with Optum’s de-identified Market Clarity Data. Patient characteristics, treatment patterns, and overall survival (OS) were assessed in adults with eHCC treated with resection or ablation between July 2016 and March 2021.

Results

Of 649 patients who met inclusion criteria, 59.3%, 37.3%, and 3.4% underwent ablation only, resection only, or both, as their initial treatment, respectively. Median age was 64.0 years; most patients were male (72.9%) and White (65.5%). Subsequent treatment was received in 47.1% of patients. The median (quartile 1–3) time to first subsequent treatment was 216 (89.3–414.3) days. The most common subsequent treatments included embolization (22.7%) and ablation (15.6%). In total, 35.7% of patients died post-index. Median OS was 67.7 (95% CI: 56.4–not estimable) months. Estimated 24-month OS was 79.0% (95% CI: 75.0–82.0).

Conclusions

Our results highlight the need for post-treatment surveillance and the potential role for neoadjuvant and/or adjuvant treatments to improve outcomes in patients with eHCC treated with resection or ablation.

PLAIN-LANGUAGE SUMMARY

Hepatocellular carcinoma (HCC) is the most common type of primary liver cancer, which is the third leading cause of cancer-related death worldwide. Early detection of HCC is associated with improved outcomes because of the availability of therapies that aim to cure the cancer. Curative-intent treatments for early-stage HCC (eHCC) include liver transplantation, surgery to remove part of the liver (resection), or treatment which directs energy (heat or cold) into liver tumors to destroy them without removing them (ablation). Although liver transplantation is associated with the best survival, livers available for transplantation are limited, and not all patients are eligible. Despite being treated early, many people with eHCC have their cancer come back after initial treatment (recurrence). We performed this research to understand more about people with eHCC in the United States treated with resection or ablation.

We used a healthcare database to find records of people in the United States diagnosed with eHCC who had been treated with resection or ablation. We assessed the characteristics of these people, and the most common subsequent treatments they received and the length of time they were alive after their initial resection or ablation.

The study included 649 patients with eHCC. Ablation was the most common initial treatment for people with eHCC. In people who received treatment after initial resection or ablation, embolization and ablation (both treatments that directly target the tumor, also known as locoregional therapies) were used most often, followed by liver transplantation, and treatments that target the entire body (systemic therapies). The average length of time that people were alive after receiving ablation or resection as their initial treatment was 67.7 months; similar to other studies in people with eHCC.

The results of this real-world study suggest more research is needed to improve outcomes for people with eHCC in the United States treated with resection or ablation as treatments of curative intent.

HIGHLIGHTS

• Data on real-world treatment patterns and outcomes among patients with early-stage hepatocellular carcinoma (eHCC) treated with surgical resection or ablation in clinical practice are lacking and may be useful for contextualizing data on neoadjuvant and/or adjuvant immunotherapy and targeted therapies.

• This study assessed patient characteristics, treatment patterns, and overall survival (OS) in a retrospective cohort of 649 patients with eHCC in the United States who were treated with initial resection, ablation, or both.

• 47.1% of patients received subsequent treatment, most commonly embolization (22.7%) and ablation (15.6%).

• In total, 35.7% of patients died post-index. Median OS was 67.7 months and the estimated 24-month OS rate was 79.0%.

• These findings highlight the need for close surveillance of patients with eHCC after treatment and the potential need for neoadjuvant and/or adjuvant treatments to improve outcomes in the early disease setting.

Graphical Abstract

1. Introduction

Hepatocellular carcinoma (HCC) is the most common form of liver cancer, which is the third most common cause of cancer-related death worldwide [1,2]. In 2025, there are an estimated 42,240 new cases and 30,090 deaths related to liver cancer in the United States [3]. The prognosis for patients with HCC depends on tumor stage at diagnosis [3]. Five-year survival is approximately 37% in patients with early-stage disease and 3% in patients with advanced-stage disease [3]. Ineffective early detection strategies leading to late diagnosis of HCC contribute to the poor prognosis and high mortality of patients [4].

Treatment guidelines for early-stage HCC (eHCC) in the United States recommend therapies with curative intent, traditionally comprising three primary strategies: surgical resection, liver transplantation, or ablation [5]. Locoregional therapy may be considered in patients who are not candidates for surgical curative treatments or as part of a strategy to bridge or downstage patients to other curative therapies [5]. For patients with unresectable eHCC, liver transplantation is associated with the best outcomes; however, graft availability and recipient eligibility are barriers to transplantation [5]. Five-year overall survival (OS) is approximately 80%, 60%, and 51% in patients with eHCC who have undergone liver transplantation, resection, and ablation, respectively [6,7]. While treatment modalities for eHCC have demonstrated local tumor control, a significant proportion of patients experience disease recurrence following initial treatment, with estimated 5-year recurrence rates of approximately 13%, 78%, and 66% in patients with eHCC who have undergone liver transplantation, resection, and ablation, respectively [6,7]. There is a potential unmet need for effective neoadjuvant and/or adjuvant treatments that can reduce disease recurrence in eHCC, particularly for patients at high risk of recurrence following surgical resection or ablation.

The treatment landscape for HCC has expanded in recent years due to the advent of targeted therapies and immunotherapy [8]. Immune checkpoint inhibitor (ICI) combinations have demonstrated efficacy comparable to standard treatments across various stages of HCC [9–12]. Several Phase III clinical trials are evaluating ICIs as adjuvant therapy in patients with HCC [9,13–15]. The IMbrave050 study (NCT04102098) is comparing adjuvant atezolizumab plus bevacizumab versus active surveillance in patients with eHCC who are at high risk of recurrence following curative resection or ablation [9]. Although initial results from IMbrave050 showed a significant improvement in recurrence-free survival with adjuvant atezolizumab plus bevacizumab versus active surveillance, this benefit was not sustained with longer follow-up [9,16]. Other ongoing Phase III clinical trials in patients with HCC who are at high risk of recurrence following curative resection or ablation include the EMERALD-2 study (NCT03847428), which is assessing adjuvant durvalumab with or without bevacizumab versus placebo [13], and the CheckMate 9DX study (NCT03383458), which is assessing adjuvant nivolumab versus placebo [14]. The KEYNOTE-937 study (NCT03867084) is assessing adjuvant pembrolizumab versus placebo in patients with HCC and complete radiological response after surgical resection or local ablation [15].

Data on treatment patterns and outcomes among patients treated with surgical resection or ablation in clinical practice are lacking and can be useful for contextualizing data on neoadjuvant and/or adjuvant therapies in patients with HCC. Here, we performed a retrospective cohort study to descriptively summarize patient characteristics, treatment patterns, and outcomes in patients with HCC treated with resection or ablation in the United States.

2. Methods

2.1. Data source

This retrospective study used Optum’s de-identified Market Clarity Data (Optum Market Clarity), “a database that deterministically links medical and pharmacy claims with electronic health record (EHR) data from providers across the continuum of care in the United States.” Access to Optum’s de-identified Market Clarity Data (Optum Market Clarity) was purchased with permission to use. Data were certified as de-identified by an independent statistical expert following the statistical de-identification rules of the Health Insurance Portability and Accountability Act, and managed according to customer data use agreements. Ethics Committee approval was not required.

2.2. Study population

HCC treated with resection or ablation was used as a proxy to define an eHCC population in this study. Treatment was used as a proxy to define patients with eHCC because cancer staging information was poorly recorded in Optum Market Clarity and resection and ablation are the standard of care for patients with early-stage disease. Adult patients (≥18 years of age) diagnosed with HCC with evidence of first resection or ablation between July 01, 2016, and March 31, 2021, were selected. HCC diagnosis was captured using International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) criteria: C22.0, C22.7, C22.8, C22.9; and resection or ablation were captured using Current Procedural Terminology or ICD-10-Procedure Coding System codes.

Patients were required to have at least two diagnosis codes for HCC, 1–90 days apart, in their EHR or medical claims before the index date, which was the date of first observed resection or ablation procedure following first HCC diagnosis. Additional eligibility criteria included continuous enrollment in a medical and pharmacy plan for at least 6 months before the index date (baseline period) and for at least 1 month post index. The short minimum duration of follow-up is intended to reduce survival bias in the sample by allowing the inclusion of patients who die following index. Patients with evidence of other primary cancers or liver-related procedures and those treated with chemotherapy, immunotherapy, or systemic therapy for HCC during the baseline period were excluded. An overview of patient selection criteria is shown in Figure 1.

Figure 1.

Flow chart depicting the patient selection criteria of adults diagnosed with eHCC who underwent ablation, resection, or both between July 01, 2016 and March 31, 2021. Overall, 649 patients met the study selection criteria with 385 patients treated with ablation, 242 patients treated with resection, and 22 patients treated with ablation and resection.

^aLiver-related procedures included ablation, resection, transarterial embolization, transarterial chemoembolization, or transarterial radioembolization.

BTC: Biliary tract cancer; CTLA-4: Cytotoxic T-lymphocyte-associated antigen 4; eHCC: Early-stage hepatocellular carcinoma; EHR: Electronic health record; HCC: Hepatocellular carcinoma; ICD-10-CM: International Classification of Diseases, Tenth Revision, Clinical Modification; PD-1: Programmed cell death-1; PD-L1: Programmed cell death ligand-1; TKI: Tyrosine kinase inhibitor.

Patients were followed from the index date until the end of continuous enrollment, death, or end of study period, whichever occurred first (follow-up period). The study design is shown in Supplementary Figure S1.

2.3. Study variables

Patients were stratified into two cohorts, based on whether they received ablation only or resection only as their first treatment at index between July 01, 2016, and March 31, 2021. Patients who received both ablation and resection at index were excluded from the ablation and resection cohorts. Patient demographics, including age, sex, race, region, insurance type, and index year, were assessed on the index date. Clinical characteristics, including the Charlson Comorbidity Index, etiology of HCC (hepatitis C virus, hepatitis B virus, alcohol-related liver disease, and metabolic dysfunction-associated steatotic liver disease [MASLD]), covariates of interest (hypertension, portal hypertension, diabetes, obesity, encephalopathy, and bleeding), albumin-bilirubin (ALBI) grade, and alpha-fetoprotein levels were evaluated during the baseline period. Etiology of HCC and factors of interest were identified using ICD-10-CM codes. ALBI scores were calculated using baseline albumin (g/L) and bilirubin (µmol/L) levels using the following formula: (log₁₀ bilirubin × 0.66) + (albumin × −0.085).

Treatment patterns were assessed overall and by treatment sequence during the follow-up period. Subsequent procedures included embolization (transarterial embolization [TAE], transarterial chemoembolization [TACE], or transarterial radioembolization [TARE]), liver transplant, hepatic resection, ablation, and radiation (external beam radiation therapy and stereotactic body radiation therapy). Subsequent systemic therapies included tyrosine kinase inhibitors (TKIs; cabozantinib, lenvatinib, regorafenib, and sorafenib), ICIs (atezolizumab, ipilimumab, nivolumab, pembrolizumab, durvalumab, and tremelimumab), or angiogenesis inhibitors (bevacizumab and ramucirumab).

The first procedure or systemic therapy administered post-index was considered the start of the first treatment sequence. All repeat procedures or systemic therapies performed or administered within 35 days of starting the treatment sequence were considered part of the same treatment sequence. A treatment sequence continued until discontinuation, switch, or end of follow-up. Discontinuation was defined as a gap longer than 60 days after the allowable refill gap period, after which the medication was meant to be depleted for all therapies within the treatment sequence. Switching was defined as initiating a new therapy more than 35 days after starting the treatment sequence.

Time to first subsequent treatment was defined as the number of months from the index date to the first treatment sequence post-index or death while continuously enrolled in a healthcare plan. Time to first subsequent treatment was calculated using the Kaplan-Meier (KM) method, where an event was defined as the initiation of a subsequent treatment or death during the follow-up period. Patients who did not receive subsequent treatment and those who were alive at the end of the follow-up period were censored. The proportion of patients without subsequent treatment was reported at 3, 6, 12, 18, and 24 months.

OS was calculated as the time from first resection/ablation to death. Patients with no evidence of death were censored at their last known alive date (defined as the latest date of service from the medical/pharmacy claims and the last encounter date from the EHR). Estimated OS rates were reported at 6, 12, 24, 36, and 48 months.

2.4. Statistical analyses

All study variables were exploratory and assessed using descriptive statistics. Continuous variables were summarized by the number of observations, median, and first and third quartiles (Q1, Q3). Categorical variables were summarized by frequency counts and percentages for each category. All time-to-event analyses were estimated using KM methods, with the median time-to-event, cumulative probability of survival, and 95% confidence intervals (CIs) presented. No data imputation was performed. Analyses were performed using Instant Health Data software (Pangalo, Boston, MA, USA) and Statistical Analysis System version 9.4 or higher.

3. Results

3.1. Patient demographics and clinical characteristics

In total, 649 patients with eHCC met inclusion criteria. At the index date, 59.3% (n = 385) of patients had undergone ablation only, 37.3% (n = 242) of patients had undergone resection only, and 3.4% (n = 22) of patients had undergone both ablation and resection as their initial treatment (Figure 1). Median (Q1, Q3) age at index across study cohorts was 64.0 (59.0, 71.0) years; most patients were male (72.9%, n = 473), White (65.5%, n = 425), and from the Midwest (34.5%, n = 224) and Northeast (29.6%, n = 192) regions of the United States (Table 1). Commercial insurance was more common in the resection cohort compared with the ablation cohort (47.5%, n = 115 vs 35.8%, n = 138, respectively) (Table 1). Treating physician specialties at index are summarized in Supplementary Table S1. During the baseline period, numerically, more patients in the ablation cohort had a Charlson Comorbidity Index category ≥3 (75.6%, n = 291 vs 52.5%, n = 127), MASLD (83.1%, n = 320 vs 45.5%, n = 110), alcohol-related liver disease (32.2%, n = 124 vs 6.6%, n = 16), and portal hypertension (46.5%, n = 179 vs 14.9%, n = 36) when compared with the resection cohort (Table 2). ALBI grade was available for 62.4% (n = 405) of patients across study cohorts (Table 2). Among the evaluable patients, the majority (98.7%, n = 400) underwent baseline assessments for albumin and bilirubin within +/− 2 weeks of each other. Most (369/405) patients were ALBI grade 1 (40.2%, n = 163) or grade 2 (50.9%, n = 206) (Table 2). The resection cohort had a higher percentage of ALBI grade 1 patients (48.5%, n = 83) than the ablation cohort (33.6%, n = 72) (Table 2). At baseline, the median alpha-fetoprotein concentration, available in ≤50% of patients, was 8.7 ng/mL in the ablation cohort, 8.0 ng/mL in the resection cohort and 8.1 ng/mL in the total population (Table 2).

Table 1.

Baseline demographics.

Demographic	Patients who underwent ablation only at index (n = 385)	Patients who underwent resection only at index (n = 242)	Total patients (N = 649)a
Time from index until end of continuous enrollment, death, or end of study period: months, median (Q1, Q3)	21.9 (9.5, 36.1)	24.4 (10.4, 42.6)	23.0 (10.0, 38.6)
Index year, n (%)
Q3–Q4 2016	46 (12.0)	24 (9.9)	72 (11.1)
Q1–Q4 2017	85 (22.1)	54 (22.3)	144 (22.2)
Q1–Q4 2018	93 (24.2)	59 (24.4)	158 (24.4)
Q1–Q4 2019	87 (22.6)	53 (21.9)	145 (22.3)
Q1–Q4 2020	59 (15.3)	46 (19.0)	108 (16.6)
Q1 2021	15 (3.9)	6 (2.5)	22 (3.4)
Age at index: years, median (Q1, Q3)	65.0 (59.0, 72.0)	63.0 (56.3, 69.0)	64.0 (59.0, 71.0)
Age categories at index: years, n (%)
18–64	192 (49.9)	138 (57.0)	344 (53.0)
≥65	193 (50.1)	104 (43.0)	305 (47.0)
Sex, n (%)
Male	280 (72.7)	178 (73.6)	473 (72.9)
Female	105 (27.3)	64 (26.5)	176 (27.1)
Race, n (%)
White	257 (66.8)	153 (63.2)	425 (65.5)
Black	45 (11.7)	45 (18.6)	95 (14.6)
Asian	23 (6.0)	21 (8.7)	45 (6.9)
Missing	60 (15.6)	23 (9.5)	84 (12.9)
Census region at index, n (%)
Midwest	131 (34.0)	84 (34.7)	224 (34.5)
Northeast	111 (28.8)	77 (31.8)	192 (29.6)
South	74 (19.2)	58 (24.0)	140 (21.6)
West	59 (15.3)	13 (5.4)	72 (11.1)
Missing	10 (2.6)	10 (4.1)	21 (3.2)
Insurance type close to index, n (%)
Commercial	138 (35.8)	115 (47.5)	262 (40.4)
Medicare	133 (34.6)	67 (27.7)	204 (31.4)
Medicaid	77 (20.0)	39 (16.1)	123 (19.0)
Unknown payor/other	37 (9.6)	21 (8.7)	60 (9.2)

^a A total of 22 patients underwent both ablation and resection at index and are included in the “Total patients” cohort only.

Q: Quartile.

Table 2.

Baseline clinical characteristics.

Characteristic	Patients who underwent ablation only at index (n = 385)	Patients who underwent resection only at index (n = 242)	Total patients (N = 649)a
Etiology of HCC during the baseline period, n (%)b
Hepatitis C virus	198 (51.4)	110 (45.5)	320 (49.3)
Hepatitis B virus	21 (5.5)	27 (11.2)	52 (8.0)
Alcohol-related liver disease	124 (32.2)	16 (6.6)	140 (21.6)
MASLD	320 (83.1)	110 (45.5)	445 (68.6)
ALBI grade, n (%)c	n = 214	n = 171	n = 405
1	72 (33.6)	83 (48.5)	163 (40.3)
2	114 (53.3)	80 (46.8)	206 (50.9)
3	28 (13.1)	8 (4.7)	36 (8.9)
Alpha-fetoprotein, n	164	121	299
Median, ng/mL (Q1, Q3)	8.7 (3.5, 24.9)	8.0 (3.5, 63.8)	8.1 (3.4, 29.8)
Covariates of interest during the baseline period, n (%)b
Hypertension	305 (79.2)	199 (82.2)	522 (80.4)
Portal hypertension	179 (46.5)	36 (14.9)	217 (33.4)
Diabetes	167 (43.4)	87 (36.0)	260 (40.1)
Obesity	146 (37.9)	74 (30.6)	225 (34.7)
Encephalopathy	81 (21.0)	35 (14.5)	119 (18.3)
Bleeding	70 (18.2)	35 (14.5)	107 (16.5)
Charlson Comorbidity Index category during the baseline period, n (%)
0–1	10 (2.6)	11 (4.5)	22 (3.4)
2	84 (21.8)	104 (43.0)	194 (29.9)
≥3	291 (75.6)	127 (52.5)	433 (66.7)

^a A total of 22 patients underwent both ablation and resection at index and are included in the “Total patients” cohort only.

^b Not mutually exclusive.

^c Using albumin and bilirubin scores assessed within +/− 8 weeks of each other.

ALBI: Albumin-bilirubin; HCC: Hepatocellular carcinoma; MASLD: Metabolic dysfunction-associated steatotic liver disease; Q: Quartile.

3.2. Treatment patterns

The median (Q1, Q3) duration of follow-up was 23.0 (10.0, 38.6) months in all patients, 21.9 (9.5, 36.1) months in the ablation cohort, and 24.4 (10.4, 42.6) months in the resection cohort. During the follow-up period, 306 (47.1%) patients received at least one post-index subsequent treatment, including 212 (55.1%) patients in the ablation cohort and 80 (33.1%) patients in the resection cohort. We limited the follow-up time and the number of subsequent treatment sequences to four. The median (Q1, Q3) number of post-index subsequent treatments received during follow-up was 1.0 (1.0, 2.0) in all patients, 1.0 (1.0, 2.0) in the ablation cohort, and 1.0 (1.0, 2.3) in the resection cohort.

The median (95% CI) time to first subsequent treatment or death from the index date was 17.2 (14.7–21.5) months in all patients, 12.7 (10.7–15.4) months in the ablation cohort, and 43.5 (32.4–55.7) months in the resection cohort (Supplementary Figure S2A–C). Overall, 42.0% of all patients received subsequent treatment or died by 12 months post-index (Supplementary Figure S2A).

Figure 2.

OS in all patients, patients treated with ablation only, and patients treated with resection only, depicted in subfigures labeled A to C, illustrating the exploratory assessment of median OS and estimated OS rates at 6, 12, 24, 36, and 48 months.

CI: Confidence interval; NE: Not estimable; OS: Overall survival.

In patients who received at least one post-index subsequent treatment, the most common subsequent procedures were embolization (22.7%, n = 147) and ablation (15.6%, n = 101) (Table 3). In follow-up, 10.3% (n = 67) of patients underwent liver transplantation (Table 3). A minority of patients received post-index subsequent treatment with systemic therapy, including TKIs (8.6%, n = 56), ICIs (7.4%, n = 48), or angiogenesis inhibitors (4.6%, n = 30) (Table 3).

Table 3.

Subsequent procedures and systemic therapy received in any sequence post-index during the follow-up period.

Procedures and systemic therapy	Patients who underwent ablation only at index (n = 385)	Patients who underwent resection only at index (n = 242)	Total patients (N = 649)a
Received subsequent procedure, n (%)b
Any embolization	114 (29.6)	29 (12.0)	147 (22.7)
TAE	43 (11.2)	14 (5.8)	59 (9.1)
TACE	42 (10.9)	8 (3.3)	53 (8.2)
TARE	47 (12.2)	17 (7.0)	67 (10.3)
Any ablation	88 (22.9)	8 (3.3)	101 (15.6)
Any liver transplant	47 (12.2)	18 (7.4)	67 (10.3)
Any radiation (SBRT or EBRT)	17 (4.4)	21 (8.7)	41 (6.3)
Any hepatic resection	5 (1.3)	NA	NA
Received subsequent systemic therapy, n (%)b
Any TKI	30 (7.8)	21 (8.7)	56 (8.6)
Sorafenib	21 (5.5)	10 (4.1)	34 (5.2)
Lenvatinib	11 (2.9)	8 (3.3)	22 (3.4)
Regorafenib	NA	NA	NA
Cabozantinib	NA	NA	NA
Any ICI	25 (6.5)	21 (8.7)	48 (7.4)
Nivolumab	14 (3.6)	12 (5.0)	28 (4.3)
Atezolizumab	10 (2.6)	11 (4.6)	21 (3.2)
Pembrolizumab	NA	NA	NA
Ipilimumab	NA	NA	NA
Any angiogenesis inhibitor	15 (3.9)	14 (5.8)	30 (4.6)
Bevacizumab	13 (3.4)	14 (5.8)	28 (4.3)
Ramucirumab	NA	NA	NA

Subsequent procedures included embolization (TAE, TACE, or TARE), liver transplant, hepatic resection, ablation, and radiation (SBRT and EBRT). Subsequent systemic therapies included TKIs (cabozantinib, lenvatinib, regorafenib, and sorafenib), ICIs (atezolizumab, ipilimumab, nivolumab, and pembrolizumab), or angiogenesis inhibitors (bevacizumab and ramucirumab). NA, data are not presented in order to preserve patient identity due to low patient numbers (≤5).

^a A total of 22 patients underwent both ablation and resection at index and are included in the “Total patients” cohort only.

^b Patients may have received more than one procedure or therapy.

EBRT: External beam radiation therapy; ICI: Immune checkpoint inhibitor; NA: Not available; SBRT: Stereotactic body radiation therapy; TAE: Transarterial embolization; TACE: Transarterial chemoembolization; TARE: Transarterial radioembolization; TKI: Tyrosine kinase inhibitor.

In patients who had undergone ablation only at index date, the most common post-index subsequent treatment administered in the first sequence was ablation (27.4%, n = 58), while TARE was the most common subsequent treatment in the resection cohort (16.3%, n = 13) (Table 4). In the ablation cohort, 25.5% (n = 98) of patients received at least two post-index treatment sequences, and 10.4% (n = 40) of patients received at least three post-index treatment sequences (Table 4). Ablation was the most common post-index second and third treatment in the ablation cohort (32.7%, n = 32 and 25.0%, n = 10, respectively) (Table 4). In the resection cohort, 14.5% (n = 35) of patients received at least two post-index treatment sequences, and 8.3% (n = 20) of patients received at least three post-index treatment sequences (Table 4). Data for the most common post-index second and third treatments were unavailable for the resection cohort due to small patient numbers across treatment groups (Table 4). Treatment discontinuation was the most common reason for the end of post-index subsequent treatment across study cohorts (Table 4).

Table 4.

Treatment sequences during the follow-up period.

Treatment sequences	Patients who underwent ablation only at index (n = 385)	Patients who underwent resection only at index (n = 242)	Total patients (N = 649)a
Received ≥1 treatment sequence during the follow-up period, n (%)	212 (55.1)	80 (33.1)	306 (47.1)
Treatments administered in first sequence, n (%)
Ablation only	58 (27.4)	NA	NA
TARE only	27 (12.7)	13 (16.3)	42 (13.7)
Liver transplant only	26 (12.3)	12 (15.0)	39 (12.7)
TAE only	26 (12.3)	6 (7.5)	32 (10.5)
TACE only	25 (11.8)	NA	NA
SBRT only	9 (4.2)	7 (8.8)	17 (5.6)
Sorafenib monotherapy	8 (3.8)	7 (8.8)	17 (5.6)
Otherb	33 (15.6)	28 (35.0)	66 (21.6)
Reason for end of first treatment sequence, n (%)
Treatment discontinuation	194 (91.5)	62 (77.5)	266 (86.9)
Treatment switch or remained on treatment at end of follow-up	18 (4.7)	18 (7.4)	40 (6.2)
Received ≥2 treatment sequences during the follow-up period, n (%)	98 (25.5)	35 (14.5)	143 (22.0)
Treatments administered in second sequence, n (%)
Ablation only	32 (32.7)	NA	NA
TACE only	12 (12.2)	NA	NA
TARE only	10 (10.2)	NA	NA
Liver transplant only	9 (9.2)	NA	NA
TAE only	8 (8.2)	NA	NA
Sorafenib monotherapy	7 (7.1)	NA	NA
Otherc	20 (20.4)	18 (51.4)	42 (29.4)
Reason for end of second treatment sequence, n (%)
Treatment discontinuation	90 (91.8)	27 (77.1)	125 (87.4)
Treatment switch	NA	6 (17.1)	NA
Remained on treatment at end of follow-up	NA	NA	NA
Received ≥3 treatment sequences during the follow-up period, n (%)	40 (10.4)	20 (8.3)	65 (10.0)
Treatments administered in third sequence, n (%)
Ablation only	10 (25.0)	NA	NA
TARE only	6 (15.0)	NA	NA
Otherd	24 (60.0)	15 (75.0)	43 (66.2)
Reason for end of third treatment sequence, n (%)
Treatment discontinuation	33 (82.5)	14 (70.0)	51 (78.5)
Treatment switch	6 (15.0)	NA	NA
Remained on treatment at end of follow-up	NA	NA	NA

NA, data are not presented in order to preserve patient identity due to low patient numbers (≤5).

^a A total of 22 patients underwent both ablation and resection at index and are included in the “Total patients” cohort only.

^b Treatments administered to ≤5 patients, including resection, resection plus liver transplant, ablation plus liver transplant, ablation plus resection, ablation plus SBRT, ablation plus TACE, ablation plus TAE, liver transplant plus TARE, liver transplant plus TAE, TARE plus TAE, TARE plus TACE, TAE plus nivolumab, TACE plus sorafenib, EBRT, SBRT plus lenvatinib, SBRT plus atezolizumab plus bevacizumab, SBRT plus regorafenib, lenvatinib, lenvatinib plus pembrolizumab, sorafenib plus nivolumab, atezolizumab plus bevacizumab, bevacizumab, nivolumab, and cabozantinib.

^c Treatments administered to ≤5 patients, including resection plus liver transplant, ablation plus TAE, ablation plus TACE, ablation plus TARE, ablation plus sorafenib, EBRT, SBRT, SBRT plus bevacizumab, SBRT plus cabozantinib, TARE plus TACE, lenvatinib, atezolizumab plus bevacizumab, bevacizumab, sorafenib plus regorafenib, sorafenib plus nivolumab, regorafenib, ramucirumab, nivolumab, and ipilimumab plus nivolumab.

^d Treatments administered to ≤5 patients, including liver transplant, resection, ablation plus TAE, TAE, TARE plus TAE, TACE, TACE plus nivolumab, EBRT, SBRT, sorafenib, lenvatinib, bevacizumab, atezolizumab plus bevacizumab, cabozantinib, regorafenib, pembrolizumab, nivolumab, and ipilimumab plus nivolumab.

EBRT: External beam radiation therapy; NA: Not available; TAE: Transarterial embolization; TACE: Transarterial chemoembolization; TARE: Transarterial radioembolization; SBRT: Stereotactic body radiation therapy.

3.3. Overall survival

Across study cohorts, 35.7% (n = 232) of patients died at any time after the index date, including 41.8% (n = 161) of patients in the ablation cohort and 26.0% (n = 63) of patients in the resection cohort (Figure 2A–C). Median OS (95% CI) was 67.7 (56.4–not estimable) months in all patients, 52.1 (45.2–68.6) months in the ablation cohort, and not estimable (as the median was not reached) in the resection cohort (Figure 2A–C). Estimated 12-month OS (95% CI) rates were 88.0% (86.0–91.0) in all patients, 87.0% (84.0–91.0) in the ablation cohort, and 90.0% (87.0–94.0) in the resection cohort (Figure 2A–C). Estimated 24-month OS (95% CI) rates were 79.0% (75.0–82.0) in all patients, 75.0% (71.0–80.0) in the ablation cohort, and 84.0% (79.0–88.0) in the resection cohort (Figure 2A–C). Estimated 48-month OS (95% CI) rates were 61.0% (57.0–65.0) in all patients, 54.0% (48.0–60.0) in the ablation cohort, and 71.0% (64.0–78.0) in the resection cohort (Figure 2A–C).

4. Discussion

In this study, we retrospectively evaluated patient demographics, clinical characteristics, treatment patterns, and OS in patients with eHCC treated with resection or ablation in the United States.

Findings from our study showed that ablation was more common than resection as the initial curative intent therapy for patients with eHCC. The choice between resection and ablation for eHCC may depend on several factors, including tumor burden, liver function, and comorbidities [17]. Indeed, more patients treated with ablation had a Charlson Comorbidity Index category ≥3 (75.6% vs 52.5%), MASLD (83.1% vs 45.5%), portal hypertension (46.5% vs 14.9%), and alcohol-related liver disease (32.2% vs 6.6%) when compared with patients treated with resection. The presence of comorbidities and impaired liver function in our study population may have influenced the physician’s preference for ablation over resection, as described in other real-world studies of eHCC [18]. Of note, our finding that ablation was more common than resection as curative intent therapy for patients with eHCC differs when compared with the patient population enrolled in the IMbrave050 study, which reported that resection (n = 585/668 [87.6%]) was more common than ablation (n = 83/668 [12.4%]) [16]. IMBrave050 enrolled a large Asian population, and resection is the local standard practice in Asia [16,19]. While cross-study comparisons are limited due to different study designs, geographical locations, and patient populations, enrollment criteria for clinical trials of patients with eHCC may not be reflective of real-world clinical practice [16]. Although resection and ablation have both demonstrated efficacy in patients with eHCC, there is a lack of consensus on which has better outcomes as first-line treatment in eHCC [20]. Retrospective analyses have shown that resection can improve outcomes in eHCC, compared with ablation, and randomized controlled trials either support these findings [21,22] or suggest that outcomes are similar in the resection and ablation groups [23,24]. Factors associated with resection versus ablation include patient characteristics (including age >65 years, race, Charlson/Deyo Comorbidity Score, insurance status, and facility type and region), tumor size, and fibrosis score; ultimately, most patients are not surgical candidates at the time of diagnosis [20]. While the assessment of OS in our study was exploratory, with no formal comparison of ablation versus resection performed, median OS was 67.7 months in all patients, 52.1 months in the ablation cohort, and not estimable in the resection cohort (due to the low number of events recorded post-index), with 2-year OS rates of 75.0% in patients treated with ablation and 84.0% in patients treated with resection, and 4-year OS rates of 54.0% in patients treated with ablation and 71.0% in patients treated with resection. These findings are generally consistent with other studies reporting survival in patients with eHCC following resection or ablation, which have reported 5-year median OS rates of 57.9% and 67.0%, respectively [25,26].

Approximately half (47.1%) of the patients included in our study received post-index subsequent treatment. Embolization procedures, including TAE, TACE, and TARE, were the most common post-index subsequent treatments in our study population. Post-index subsequent treatment with systemic therapy, including TKIs, ICIs, and angiogenesis inhibitors, was not commonly observed. However, this may be due to the study period (2016–2021), when ICI use in patients with HCC was limited.

More patients received post-index subsequent treatment in the ablation (55.1%) versus the resection (33.1%) cohort, and the median (95% CI) time to first subsequent treatment or death from the index date was shorter in the ablation (12.7 [10.7–15.4] months) versus the resection (43.5 [32.4–55.7] months) cohort. The results from our study differ compared with a recent retrospective chart review study of patients from Europe, Asia, and North America with localized HCC at high risk of recurrence (based on the high risk criteria from the CheckMate 9DX clinical trial [NCT03383458]) who were treated with resection or ablation [27]. In the study, subsequent treatment use was reported in 90.9% of patients, including systemic therapy (38.2% of patients), locoregional therapy (26.8% of patients), and a second resection or ablation (26.8% of patients) [27]. Of note, the incidence of subsequent therapy use appears approximately two-fold higher in that study compared with our study (90.9% vs 47.1%, respectively), with differences in the type of the most common subsequent treatment administered (systemic therapy [38.2%] vs embolization procedures [22.7%]) [27]. Several reasons may explain the different results reported between studies, including differences in the proportion of patients that were retreated with resection, ablation, or locoregional therapy, differences in study type (chart review vs observational study) and patient characteristics (eHCC at high risk of recurrence vs heterogenous population of eHCC) [27]. Regardless, findings indicate that subsequent therapy use is common in patients with eHCC who are treated with initial resection or ablation.

In total, 42.0% of patients in our study either received subsequent treatment (including procedures) or died at 12 months post-index. Notably, 10.3% of patients underwent subsequent liver transplantation. It is important to note that some patients may have initially received resection or ablation as a bridging strategy due to long wait times for liver transplantation. These factors make it challenging to use subsequent treatment as an indicator of disease progression or recurrence. Previous studies have reported 1-year disease-free and recurrence-free survival rates of 79.0–86.6% in patients with eHCC treated with resection or ablation [7,28]. Based on these findings, HCC progression or recurrence may have accounted for a large proportion of subsequent treatment use in our study. Together, these findings highlight the need for close surveillance of patients with eHCC after treatment and the potential role for neoadjuvant and/or adjuvant treatments to improve outcomes in the early disease setting.

Limitations of our study include the non-interventional retrospective design, which meant that findings were dependent on the accuracy and completeness of the medical records and data reported by the physicians. As with any observational study, caution must be taken when interpreting results due to potential selection bias. The results reported in this study are from patients in the United States, and the analysis was restricted to patients who had linked EHR and claims data (commercial, Medicaid, and Medicare) and enrollment data, which is a small proportion of patients in the Optum Market Clarity database. Therefore, the results of this analysis may not be generalizable to patients from other countries outside of the United States or to patients in the United States who are either uninsured or have Medicare fee-for-service. Additional limitations include that data were not well captured for certain variables of interest, such as Child-Pugh scores, alpha-fetoprotein levels, and tumor burden or tumor imaging data, and reliance on ICD-10-CM codes for HCC diagnosis. Another limitation is that the initiation of resection and ablation was treated as a proxy for identifying eHCC, which may have led to patients being included in our study who did not have eHCC. In addition, resection and ablation may have been intended for bridging to liver transplantation rather than curative-intent therapy.

5. Conclusions

The findings from our study show the high rates of subsequent therapies in patients receiving initial resection or ablation for eHCC. This highlights the potential unmet need for neoadjuvant and/or adjuvant treatments that can reduce disease recurrence and improve overall outcomes. As the treatment landscape evolves, future studies are needed to evaluate the benefits versus risks of neoadjuvant and/or adjuvant therapy for patients with eHCC at high risk of recurrence, including identifying which patients may benefit from these treatments and whether they can enhance current resection or ablation options to match the clinical benefit that may be provided by liver transplantation.

Funding Statement

This work was funded by AstraZeneca.

Author contributions

Conception and design of the study: Neehar D Parikh, Jenny Hu, Heide A Stirnadel-Farrant, Nehemiah Kebede, Kirema Garcia-Reyes.

Data acquisition: Nehemiah Kebede.

Analysis and interpretation of data: Neehar D Parikh, Ravi Patel, Heide A Stirnadel-Farrant, Nehemiah Kebede, Kirema Garcia-Reyes.

Administrative, technical, or material support: Ravi Patel.

Study supervision: Ravi Patel, Kirema Garcia-Reyes.

Writing (critical reviews of all manuscript drafts, confirmation on target journal, accountable for contents of and responsible for addressing queries on the manuscript): Neehar D Parikh, Ravi Patel, Jenny Hu, Heide A Stirnadel-Farrant, Nehemiah Kebede, Cindy Wang, Kirema Garcia-Reyes.

Disclosure statement

Neehar D Parikh reports consultant fees from Exact Sciences and Exelixis; advisory fees from Bayer, Genentech, Eisai, Exelixis, Sirtex, and Wako/Fujifilm; and research funding from Bayer, Exact Sciences, Glycotest, and Target Pharmasolutions. Ravi Patel, Jenny Hu, Heide A Stirnadel-Farrant, Nehemiah Kebede, and Cindy Wang are employees of and hold stock in AstraZeneca. Kirema Garcia-Reyes reports consultant or advisory fees from AstraZeneca, Boston Scientific, Johnson & Johnson, and Varian.

Medical writing support, under the direction of the authors, was provided by Eilidh McLachlan, PhD, on behalf of CMC Connect, a division of IPG Health Medical Communications, and was funded by AstraZeneca, in accordance with Good Publication Practice (GPP2 2022) guidelines (Ann Intern Med. 175[9],1298–1304 [2022]).

Ethical declaration

Access to Optum’s de-identified Market Clarity Data (Optum Market Clarity) was purchased with permission to use. This study was conducted using de-identified participant data.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.