Despite advancements in screening and treatment for hepatocellular carcinoma (HCC), overall 5-year survival remains poor at approximately 20%.¹ According to the National Comprehensive Cancer Network guidelines for HCC treatment, patients with potentially curative disease should be offered surgical resection, locoregional therapies or liver transplant as curative-intent therapy.² Despite that early stage HCC represents a population of patients who have the highest potential for cure, nationwide population-based research demonstrates broad underutilization and well documented disparities in treatments offered for these patients.^3–6 The disparities in receiving curative-intent treatment have been demonstrated with regards to race/ethnicity⁷, insurance coverage⁸, resident location⁹, and specialization of treating facility.¹⁰ Furthermore, type of hospital and hospital volume have been shown to affect the receipt of treatment, amplifying the end effects on overall early-stage HCC survival. Specifically, some studies have shown that hospital volume appears to outweigh the inconvenience of longer travel distances.¹¹

In California, significant disparities in early-stage HCC incidence and treatment have been well documented in racial/ethnic minority populations^12,13 and those with lower socioeconomic status or who lack health insurance.^14,15 However, the impact of travel distance to transplant center and type of center (NCI-designation versus not) on utilization in curative-intent treatment and survival of early-stage HCC patients in California has not been evaluated. We sought to demonstrate whether treatment facility type and distance from treatment center with transplant capabilities play a role in disparities in curative-intent treatment and survival for early-stage HCC in California. We hypothesized that patients with access to an NCI-designated center and shorter travel distance to a transplant center would have higher odds of receiving curative-intent treatment with associated improved overall survival, potentially exposing a possible implementation strategy for future early intervention programs.

This retrospective cohort study used data from the California Cancer Registry (CCR) for patients diagnosed with stage I or II primary invasive HCC between January 1, 2005, and December 31, 2017. Our analysis was restricted to stage I or II disease to focus on HCC patients that would be amenable to curative-intent treatments. We defined curative-intent treatments as surgical resection, ablation, or transplantation; hereon, curative-intent treatments will be denotated as treatment. Locoregional therapies including transarterial chemoembolization (TACE), Y90, and radiation therapy (SBRT) are not captured as individualized codes as part of the CCR and therefore cannot be analyzed independently. Systemic therapy was not included as it is not the standard of care for early stage hepatocellular carcinoma. The primary outcome was receipt of treatment, and the secondary outcome was overall survival. All analyses were overseen by the institutional review board of University of California, Davis.

Hepatocellular carcinoma diagnoses were based on International Classification of Diseases for Oncology codes for site (C220) and histology (8170, 8172, 8173, 8174, and 8175). Clinical stage I or II patients were identified according to the American Joint Committee on Cancer (AJCC) (seventh edition) TNM staging criteria.¹⁶ Patients with unknown or unspecified surgical coding (n=121), with missing race/ethnicity data (n=49), or with missing zip code data (n=17) were excluded. A total of 19,059 patients were included in the analysis.

Patient demographic, clinical and treatment characteristics were obtained from the CCR, including race/ethnicity, age, sex, marital status, date of diagnosis, tumor size, comorbidities, rural or urban residence, minimum distance to the nearest transplant center, type of health insurance, receipt of treatment at an NCI-designated cancer center; and neighborhood socioeconomic status (SES) quintile. Minimum distance to the nearest transplant center was calculated based on the geodetic distance in miles between two zip code locations. Private insurance included Health Maintenance Organization (HMO), Preferred Provider Organization (PPO), military funded, and Medicare with supplement. Public insurance included Medicare without supplemental insurance, Medicaid, Medicare/Medicaid dual eligible, county funded, and Indian/public health service.

Of 19,059 patients with early-stage HCC, only 36% (6,778) received treatment. The majority of patients presented with stage I disease (65%), were non-Hispanic White (39%), male (73%), and had more than one comorbidity (56%). Patients more commonly lived in urban areas (88%), with a mean distance of 37.2 miles (±42.9) from the nearest transplant center (Table 1). Nearly all patients had either private (51%) or public insurance (46%), with a very small proportion of uninsured patients (1.5%). Most patients did not receive care at an NCI-designated cancer center (64%). Of patients who received treatment, ablation (46%) and resection (36%) were more common than liver transplantation (18%). Univariate analyses by treatment status are shown in Table 1. Stage at diagnosis, tumor size, T category, race/ethnicity, mean age at diagnosis, marital status, comorbidities, residence type, distance from the nearest transplant center, insurance status, neighborhood SES quintile, and receipt of care at an NCI-designated center differed by treatment status (p<0.05).

Demographic characteristics associated with a decreased odds of receiving treatment were Hispanic ethnicity (OR 0.78, CI0.71–0.84; vs. non-Hispanic White), increasing age (OR 0.98, CI 0.98–0.98), male sex (OR 0.91, CI 0.85–0.98), being unmarried (OR 0.74, 0.69–0.79), rural residence (OR 0.88, CI 0.78–0.98), any type of non-private insurance, and lower neighborhood SES quintile (Table 2). Clinical characteristics associated with a decreased odds of receiving treatment included stage II disease (OR 0.73, CI 0.68–0.78), larger tumor size (OR 0.86, CI O.79–0.94 for 2–5 cm; OR 0.54, CI 0.49–0.60 for > 5cm), at least one comorbidity (OR 0.79, CI 0.74–0.85), or care at a non-NCI designated cancer center (OR 0.48, CI 0.45–0.51). In contrast, several characteristics were associated with greater odds of receiving treatment, specifically, Asian/Pacific Islander descent (OR 1.40, CI 1.28–1.53) and a residence located 20 to 50 miles from the nearest transplant center (OR 1.2, CI 1.10–1.31).

Lack of treatment was strongly associated with worse overall survival (multivariable-adjusted HR 2.85, CI 2.67–3.04). Treatment type was associated with survival (Figure 1). In multivariable survival models (Table 3), compared with resection, ablative treatments were associated with worse survival (HR 1.63, CI 1.51–1.76), while transplantation was associated with improved survival (HR 0.57, CI 0.51–0.65). Demographic characteristics associated with decreased survival were increasing age (HR 1.02, CI 1.02–1.02), male sex (HR 1.07, CI 1.03–1.11), being unmarried (HR 1.18, CI 1.14–1.23), a far travel distance to the nearest transplant center (>50 miles HR 1.11, CI 1.04–1.19), any non-private insurance type, and residing in the three lowest neighborhood SES quintiles. Clinical characteristics associated with decreased survival included stage II (vs. stage I) disease (HR 1.25, CI 1.20–1.30), larger tumor size (HR 1.33, CI 1.26–1.41 for 2–5 cm; HR 2.23, CI 2.10–2.38 for > 5cm), at least one comorbidity (HR 1.48, CI 1.42–1.53), and care at a non-NCI designated cancer center (1.33, CI 1.28–1.39). Hispanic ethnicity (vs. non-Hispanic White) was associated with improved overall survival (HR 0.94, CI 0.89–0.98).

In this large, population-based study of early-stage HCC patients in California, we observed that 64% of patients did not receive treatment, with disparities in treatment observed by sociodemographic factors and type of treatment center. Hispanic patients were less likely and Asian/Pacific Islander patients were more likely to receive treatment than Non-Hispanic White patients. In addition, patients who did not receive treatment at an NCI-designated cancer center, lived in a lower SES neighborhood, and had any type of non-private insurance were less likely to receive treatment. Further, patients who lived farther from a transplant center, had any non-private insurance type, did not received treatment at an NCI-designated cancer center, and lived in a lower SES neighborhood experienced worse overall survival. Additionally, treatment was strongly associated with improved survival, with transplantation associated with superior overall survival compared with resection or ablation, as found previously.¹⁷ Our findings suggest that early intervention programs should target travel barriers and referral and access to specialist care to help improve oncologic outcomes.

Our findings confirm the main hypothesis that travel distance to transplant center was associated with poorer survival in patients with early HCC in the state of California. This finding is consistent with prior studies in other phases of HCC care. In a recent population-based study, Goldberg et. Al¹⁸ found that “remoteness” of specialty care was a predictor of decreased survival. Similarly, Moon et al¹⁹ demonstrated that patients residing 50 miles or more from a transplant center had substantially decreased odds of visiting a transplant center for evaluation. Interestingly, other studies have shown some incongruent findings showing that hospital volume, rather than longer travel distance, appears to impact oncologic outcomes.^9,11 Interestingly, we found that receipt of treatment was associated with being farther away from a transplant center. This factor alone can therefore not explain the full picture. This is likely due to the complexity of care involved in HCC, such as diagnosis, surveillance, and other treatments not included in our analysis. For instance, several reports have shown rural residents were found to have lower rates of anti-viral treatment for hepatitis, decreased rates of surveillance, and higher stages of disease at diagnosis.^19–21 Therefore, the disparities noted in the current study of patients with early-stage HCC are likely multi-factorial and partially contributed to increased distance from specialty care. Furthermore, additional complexity is added when one considers that while distance from treatment centers is often measured in miles, this often understates travel difficulty.²² Travel to urban centers within California from rural mountainous regions can be time consuming or challenging in certain months of the year. When designating critical access hospitals, Medicare recognizes facilities 35 miles or greater from the nearest other hospital as critical access; within mountainous regions, this distance is decreased to only 15 miles.²³

Considering our second main finding that patients who received care at NCI-designated cancer centers were more likely to receive curative-intent treatment and experience improved survival, we must examine the unique characteristics of NCI-designated cancer centers in tandem with access to care issues. California is home to eight NCI-designated cancer centers, all of which are academic-affiliated, and either close to or are themselves, transplant facilities.²⁴ NCI-designated cancer centers bring together physicians skilled in complex clinical care and scientists working on novel treatments, such as advanced procedural techniques and access to clinical trials opportunities. Furthermore, NCI-designated cancer centers often use a multidisciplinary team approach with discussion of patient cases at tumor boards with real‐time peer review.²⁵ NCI-designated cancer centers have been shown to treat younger and healthier patients, but also perform more complex procedures with similar complication rates (except for surgical site infections).²⁶ These practices have been shown to contribute to improve survival.²⁷ Some studies suggest increased survival at NCI-designated cancer centers may result secondary to delayed dissemination of treatment advances in a rapidly advancing field.^28,29 Others have shown also that structural, organizational, or provider characteristics at centers with and without NCI-designation could explain observed associations in survival difference.³⁰ Nevertheless, efforts should be made to standardize care between all levels of oncologic institutes through the dissemination of best practices using common benchmarks and guidelines.

Lower neighborhood SES and lack of private insurance were independently associated with not receiving curative-intent treatment and worse overall survival. These results are consistent with findings in HCC and other advanced cancers where sociodemographic factors, such as lack of private insurance and lower SES, have been shown to influence survival.^3,8,18,19,31 Additionally, we demonstrated that unmarried patients were less likely to receive treatment and had worse overall survival, consistent with prior reports.^32,33 Treatment-based racial disparities reported in our findings were in congruence with national data in which Hispanic patients had significantly lower rates of curative resection or ablation. Similar patterns were also observed in association of SES, insurance, marital status, and survival in HCC patients.^3,34 Finally, we show that compared to Non-Hispanic Whites, Asian patients had the best survival. This could be explained due to the fact that they were likely to receive hepatectomy as described by others⁶, underlying HCC etiology, access to screening programs³⁵, and have favorable SES factors.³⁶ Taken together, race/ethnicity, SES, and marital status are important demographic factors that contribute to one’s social vulnerability, which has been consistently associated with poorer surgical treatment outcomes and worse survival.^37,38 In the context of HCC treatment, social vulnerability can impact the ease at which patients obtain prompt evaluation and care within certain hospital systems.

Our findings suggest that the majority of early-stage HCC patients in California do not receive therapy, which is associated with lower overall survival. We found that patients who received care at an NCI-designated cancer center and were in closer proximity to a transplant center had better overall survival even after controlling for known confounders. Considering that patients who received liver transplantation had substantially improved survival in our study, it is paramount that all eligible patients have equitable access to transplant specialist referral and resources to mitigate potential travel barriers to a transplant center. Improved partnership between community and NCI-designated cancer centers and regionalization of care has improved cancer care in other malignancies. Implementation of a rapid quality reporting system through the National Cancer Institute Community Care Centers Program showed significant improvements in several aspects of breast and colon cancer care.^40,41 Such quality improvement partnerships specific to prioritizing curative-intent treatments in early-stage HCC should be strongly considered to improve oncologic outcomes.