Liver resection is superior to tumor ablation in patients with multifocal hepatocellular carcinoma

The management of T2 multifocal hepatocellular carcinoma (MHCC) is controversial, and the comparative impact of liver resection (LR) versus tumor ablation (TA) on survival continues to be debated. The aim of our study was to examine short‐ and long‐term survival for LR and TA in a nationally representative cohort. We hypothesized that patients who underwent LR would have improved survival.


| INTRODUCTION
Hepatocellular carcinoma (HCC) is the most common primary liver tumor globally, and it is the second leading cause of cancer-related mortality. 1,2The incidence of HCC has been increasing in the United States. 3It is estimated that there were 25 606 new cases of HCC diagnosed in 2013, but in 2030, over 55 000 new cases are expected. 46][7] The Barcelona Clinic Liver Classification is the most commonly used staging system for determining the optimal treatment strategy. 8Of note, the management of T2 MHCC (multiple tumors with none >5 cm) is particularly controversial as liver resection (LR), tumor ablation (TA), and liver transplant are considered acceptable treatment choices, although the evidence continues to be debated. 9,10Liver transplant is often considered the optimal treatment modality for MHCC as it, in principle, cures cancer and eliminates the underlying liver disease; however, organ shortages and cumbersome allocation processes constrain its utilization in clinical practice. 11,12Given these challenges, many patients undergo LR and TA.
4][15] However, there is a lack of robust data evaluating their comparative impact on survival for T2 MHCC.In this study, we The NCDB collects data from >1500 institutions accredited by the Commission on Cancer, capturing >70% of cancer diagnoses in the United States annually. 16It is publicly available deidentified data; therefore, IRB review and approval were not required for this study.

STROBE (Strengthening the Reporting of Observational Studies in
Epidemiology) guidelines were reviewed and complied with for the purpose of the study. 17the following groups: 18-49, 50-59, 60-69, and ≥70 years.We also categorized race (non-Hispanic White, Black, Asian, Hispanic White, and other/unknown), Charlson-Deyo score (0, 1, and ≥2), and insurance status (none, private, Medicaid, Medicare, and other/ unknown).County designation of patients' residences was categorized into metropolitan (high-population counties with >250 000 individuals) and nonmetropolitan (both medium-population counties with 2500-250 000 residents and small-population counties with <2500 residents). 18,19Facility types included community cancer programs, comprehensive community cancer programs, academic cancer programs, and integrated network cancer programs. 19Finally, we categorized hospital volume based on the total HCC case volume per facility from 2004 to 2020, in which <25th percentile was designated as low volume, 25th-75th percentile was designated as medium volume, and >75th percentile was designated as high volume.

| Statistical analysis
Categorical variables were compared using Pearson's chi-square test; continuous variables were described using the median and interquartile range (IQR).Kaplan-Meier survival curves were generated to compare 10-year OS between patients who underwent LR versus TA.Stratified analyses were also performed based on LVI, resection margin status, and Charlson-Deyo score.Cox proportional hazard models were used in multivariable analysis.
The covariates in the regression model included age, sex, race, Charlson-Deyo score, LVI, and type of treatment. 20We did not include facility features, insurance, county designation of patient residence, and AFP level due to multicollinearity (high variance inflation factor).All statistical analyses were performed using Stata 17. 21 T A B L E 1 Rates of tumor ablation or liver resection for patients in the National Cancer Database diagnosed with T2, multifocal hepatocellular carcinoma between 2004 and 2015, according to patient, tumor, and hospital characteristics.
The median age of patients in our cohort was 62 years (56-70 years).
The majority of patients were male (75.4%), non-Hispanic White (58.8%), and had a Charlson-Deyo score ≤1 (75%).In addition, most (66.8%) were treated at academic centers.There were no differences in baseline demographics, tumor, and facility characteristics between patients who underwent LR versus TA (Table 1).About 84% of the LR cohort had negative margins, while 17% had the presence of lymphovascular invasion reported in their surgical specimen.
Similarly, a lower number of comorbidities, female sex, and Asian race were associated with improved OS.

| DISCUSSION
In this study, we investigated short-and long-term survival outcomes among T2 MHCC who underwent tumor resection versus ablation.We found that the 30-day mortality rate was significantly higher among patients who underwent LR.However, LR patients had significantly better 10-year OS, and their mortality risk was reduced by about 20% after adjusting for confounders.Importantly, we found that about 85% of the resected cohort had margin-negative resection.For the purpose of this study, patients who underwent liver transplantation were excluded.Therefore, the paper's take-home message would be that patients with Multifocal HCC-LR are associated with better outcomes than TA when the transplant option is excluded.In fact, it has been shown that for those who undergo transplants for HCC, their 10-year survival is significantly higher than that of our study patients.Although our study patients are a highly selected group of multifocal T2 patients, it gives us a good insight into the fact that LR provides better outcomes if a transplant option is unavailable.
To date, most studies comparing LR and TA have focused on patients within the Milan criteria, combining a heterogeneous group of patients with solitary and multifocal tumors.Shin et al. recently performed a systematic review and meta-analysis of published literature on LR versus TA for HCC. 22Six randomized studies and 12 non-randomized matched studies were included in their analysis.
In their pooled analysis of non-randomized matched studies, they found that 3-year and 5-year OS was significantly better for LR patients.In contrast, their pooled analysis of the randomized studies revealed no significant differences in 3-year and 5-year OS between LR and TA.Nevertheless, there was significant heterogeneity among the included studies, and all were reported to have a high risk of bias.
Notably, they found strong evidence that LR was associated with significantly lower odds of local recurrence in an analysis of both the non-randomized and randomized studies.
The data for LR versus TA for MHCC specifically is even more survival among patients with MHCC using the Surveillance, Epidemiology, and End Results database. 23They performed a propensity score matching of the cohorts and found disease-specific survival and OS to be significantly better for LR patients compared to TA patients after adjusting for confounders.In a smaller study by Liu et al., they found that LR patients had significantly better recurrence-free survival than TA patients; however, 5-year OS did not significantly differ between both groups. 24They reported complication rates of 3.2% and 14% in the TA and LR groups, respectively.
In clinical practice, the decision to pursue LR and TA is complex and influenced by several factors, including surgeon experience, geographic location, institution experience, patient comorbidities, tumor location, tumor size, and expected future liver remnant (FLR). 9,25,26Historically, proponents of ablation have often cited the minimal invasiveness of ablation and obviation of extensive LR that increases morbidity and mortality risk for a disease with an already dire prognosis.Conversely, proponents of resection now often emphasize longer recurrence-free survival and the potential of laparoscopic hepatic resections to mitigate some risks associated with open surgery. 23,26Additionally, liver ablation is not without risks, with reported complications in literature including hemorrhage, bile leak, and hepatic abscesses. 279][30] Zhou et al. recently reported 1-year, 2year, and 3-year OS of 82%, 68%, and 64% receiving both modalities. 29This approach has primarily been used for patients with insufficient FLR.
From a clinicopathological perspective, we found that patients who had positive postoperative surgical margins had equal long-term outcomes.2][33] Clinically, there are situations where positive margins are more likely, such as based on tumor location, liver cirrhosis, or comorbidities that preclude LR, making ablation.
Additionally, we found that LR and LA had equal long-term outcomes in patients with a high Charles Deyo index score.This is an important finding.As we continue to see patients presenting with more comorbidities, patient selection will be a key factor in determining management, as these patients may benefit from ablation.This goes along with two studies that showed excellent outcomes in ablation in patients with comorbidities. 34,35rthermore, our results reveal that female sex and Asian race were independently associated with improved overall survival.For gender, these findings are consistent with prior studies that looked at overall survival between men and women who underwent LR for HCC.7][38][39] For race, our results agree with previous studies highlighting racial disparities in terms of access to resection, and this still held true after fixing for other variables. 40,41r study is not without limitations.Briefly, there is an inherent risk of selection bias in the treatment cohort and data missingness, given its retrospective design.The NCDB also captures data only from CoC-accredited centers, which may affect the generalizability of our results.Additionally, most patients were elderly white males.We did not utilize propensity score matching (PSM) in this study.if a transplant option is unavailable.Prospective studies are warranted to confirm these findings.These findings are crucially informative to discussions between patients and providers when assessing the best management approach.Well-powered prospective studies are warranted to confirm these findings.
queried the National Cancer Database (NCDB) to determine shortand long-term outcomes in patients with MHCC who underwent surgical treatment: LR versus TA.Second, we sought to determine clinicodemographic factors associated with receiving LR and TA. 2 | MATERIALS AND METHODS 2.1 | Data source and patient selection Using the International Classification of Disease (ICD) codes (ICD-10, C22.0) and histology codes (8170 and 8172-8175), we identified patients diagnosed with T2 MHCC between 2004 and 2015 in the NCDB.The T2 designation was determined using the NCDB "TNM_PATH_T" variable, while MHCC was determined using the "CS_EXTENSION" (390, 400, 420, and 440) variable.The NCDB collected the "CS_EXTENSION" variable up to 2015.Patients with T1, T2 single lesions, T3, T4, or metastatic cancer (M+) were excluded.Additionally, patients who underwent liver transplants were excluded.

2. 2 |
Variables We collected baseline patient demographic (age at diagnosis, sex, race, Charlson-Deyo score, insurance status, country designation of patients' residence, and insurance status), disease (blood levels for creatinine and bilirubin, international normalized ratio [INR], margin status, lymphovascular invasion [LVI] for resected patients, Alfa fetoprotein [AFP] ≥ 500 ng/mL, and 30-day mortality), and facility (facility type and volume) data.Patients ages were categorized into further stratified the cohort based on the Charlson-Deyo Comorbidity index.The 30-day mortality in the Charlson-Deyo score ≤1 LR cohort was 4.8%, while in those with a Charlson-Deyo score ≥2, it was 7.3%.

3. 3 |
Long-term overall survival among LR and TA patients Kaplan-Meier curves for patients who underwent LR versus TA for the entire cohort are shown in Figure 2. The 10-year OS was 27.5% Kaplan-Meier survival curves among patients in the NCDB with T2, multifocal hepatocellular carcinoma for type of surgery performed (A: overall; B: Vascular involvement for resection).
sparse.Yue et al. assessed the impact of LR versus TA on long-term F I G U R E 3 Kaplan-Meier survival curves among patients in the NCDB with T2, multifocal hepatocellular carcinoma for type of surgery performed (A: Negative margins; B: Positive margins; C: Charlson-Deyo score ≤1; D: Charlson-Deyo score ≥2).
Multivariable Cox proportional hazard model for overall survival among patients in the NCDB diagnosed with nonmetastatic T2, multifocal hepatocellular 2004-2015.
The two groups were similar in age, sex, race, Charlson-Deyo score, Alfa fetoprotein, county, insurance, facility volume, and facility type.5 | CONCLUSIONLR poses a higher long-term survival benefit than TA.Although our study patients are a highly selected group of multifocal T2 patients, it gives us a good insight into the fact that LR provides better outcomesT A B L E 2