Prognostic Impact of Surgical Margin in Hepatectomy on Patients With Hepatocellular Carcinoma: A Meta-Analysis of Observational Studies

Objective This study aims to comprehensively evaluate the prognostic impact of the surgical margin in hepatectomy on patients diagnosed with hepatocellular carcinoma (HCC). Methods A comprehensive and systematic search for eligible articles published in English before July 2021 was conducted across PubMed, Cochrane Library, Web of Science, and Embase electronic databases. The overall survival (OS) and disease-free survival (DFS) were the primary endpoints. Results In total, 37 observational studies with 12,295 cases were included in this meta-analysis. The results revealed that a wide surgical margin (≥1 cm) was associated with better OS (hazard ration (HR), 0.70; 95% confidence interval (CI), 0.63–0.77) and DFS (HR, 0.66; 95% CI, 0.61–0.71) compared to a narrow surgical margin (<1 cm). Subgroup analyses were conducted based on median follow-up time, gender, country, hepatitis B surface antigen (HBsAg) status, tumor number, and liver cirrhosis. The prognostic benefit of a wide surgical margin was consistent in most subgroups, however, analysis of studies from Western countries showed that margin width was not associated with prognosis. Conclusion In summary, a surgical margin wider than 1 cm prolongs the long-term prognosis of HCC patients compared to a surgical margin narrower than 1 cm.


INTRODUCTION
Although hepatocellular carcinoma (HCC) has the 5 th highest incidence across the globe, it is currently the 3 rd leading cause of cancer-related deaths (1,2). So far, liver transplantation, hepatic resection and radiofrequency ablation are the few treatment strategies for HCC. Although hepatectomy is the first-line therapeutic intervention, the prognosis of patients is unsatisfactory due to the high risk of recurrence (70% in the 5 th year after surgery) and metastasis (3).
The long-term prognosis of patients with HCC is influenced by several factors, among them, liver cirrhosis is a main factor, and the surgical margin is considered a potential prognostic factor (4,5). Curative hepatectomy is complete resection of all visible tumors without residual tumor cells at the resection margin (6). As such, an adequate resection margin is vital in preventing tumor recurrence (7). Nonetheless, minimizing the removal of the non-malignant parenchyma tissue and protecting the residual liver of liver resection is necessary for many HCC patients with liver cirrhosis or other liver diseases. This is because the capacity for liver regeneration is impaired among these patients and excessive liver tissue removal leads to severe consequences including liver failure (8,9). Thus, controversies on the width of the surgical margin have been reported under the premise of R0 resection. Many studies reveal that the width of the resection margin narrower than 1 cm is a risk factor for the long-term prognosis of HCC patients after surgery (4,10). Nevertheless, a number of articles found that a wide surgical margin did not improve the prognosis of HCC patients after hepatectomy (11,12).
Therefore, this meta-analysis seeks to assess the correlation between surgical margins (a surgical margin wider than 1 cm; a surgical margin narrower than 1 cm) and long-term prognosis of HCC patients after hepatectomy.

Literature Search Strategy
This meta-analysis adhered to the guidelines from the Preferred Reporting Items for Systematic Review and Meta-Analysis (13). A comprehensive and systematic literature search for articles published in English before July 2021 was conducted in four online electronic databases including PubMed, Cochrane Library, Web of Science, and Embase. The search terminologies included: "Hepatocellular Carcinoma" OR "Liver Cell Carcinomas" OR "Hepatoma" OR "HCC" AND "Resection Margin" OR "Surgical Margin" OR "Margin Width." Besides, reference lists of all retrieved papers were inspected to identify potentially eligible but uncaptured literature in the primary search.

Inclusion Criteria
Studies were included if they met the following criteria: (1) The cancer type was primary HCC and hepatectomy was performed on patients; (2) Patients received different surgical margins in the study (a wide surgical margin, ≥1 cm) and control (a narrow surgical margin, <1 cm) groups; (3) The study was original, including retrospective and prospective observational studies (OBS); (4) Extractable outcomes were in the studies.

Exclusion Criteria
The exclusion criteria for this meta-analysis included: (1) HCC was recurrent; (2) The patients received palliative hepatectomy or had extrahepatic metastases; (3) The study did not divide the study group and the control group into larger than 1 cm and smaller than 1 cm; (4) Duplicate article or repeat analyses using similar data.

Data Extraction and Quality Evaluation
Data extracted from eligible studies included study characteristics (author, country, publication year, study design, median followup time, and mentioned outcome measures), demographic data of patients (age, gender, and the number of patients), and clinicopathological features (liver cirrhosis, virus status, tumor number and size, and serum alpha-fetoprotein (AFP), and survival outcomes.
The quality of incorporated OBSs was evaluated using the Newcastle-Ottawa Scale (NOS) based on three aspects i.e., patient selection, comparability of groups, and outcome evaluation. The scores of papers >6 were considered high-quality.

Statistical Analysis
To evaluate the relationship between surgical margins and HCC prognosis, the overall survival (OS) and disease-free survival (DFS) in the wide surgical margin group vs. the narrow surgical margin group was compared using a pooled hazard ratio (HR) with its corresponding 95% confidence interval (CI). The degree of heterogeneity across included literature was assessed using the I 2 statistic. Considering the potential heterogeneity, randomeffect model was applied to all analyses. To assess the robustness of conclusions, a sensitivity analysis was conducted. P < 0.05 was considered statistically significant.

Data Collection and Characteristics
A total of 6,864 records were initially identified by the literature search. Out of these, 4,743 records were excluded because of duplication, and 2,050 records were eliminated after evaluating their titles or abstracts. The remaining 71 records were carefully inspected by full-text reading. Finally, 37 articles (4, 5, 7, 10-12, 14-44) were included. The comprehensive search and selection process is shown in (Figure 1).
The comprehensive characteristics of the included studies are summarized in (Table 1-1). The included articles were published between 1993 and 2021. A total of 12,295 patients from Western and Asian countries were enrolled in 37 OBSs; two studies of these were prospective, while the rest were retrospective. The majority of articles were from Asia, with China representing the most (24 articles). The demographic and clinicopathological characteristics of patients are presented in ( Table 1-2). Based on a qualitative assessment by NOS criteria, the results revealed that all included OBSs were of higher quality (Supplementary Table 1).

Correlation Between Surgical Margin and OS
A total of 28 studies reported on OS outcomes and pooling analysis of these data revealed that a wide surgical margin is associated with better OS (HR, 0.70; 95% CI, 0.63-0.77) compared to a narrow surgical margin (Figure 2). Subgroups analyses were conducted to explore the potential factors that might affect the impact of the surgical margin on the prognosis ( Table 2). This was based on the reported median follow-up time. The studies were divided into 3-year OS and 5-year OS subgroups. The result showed that patients who received a wide resection margin had better mid-and long-term prognosis than those who received a narrow resection margin. Moreover, the gender factor in the subgroups was analyzed and the findings revealed that a narrow surgical margin was a risk factor for OS of patients regardless of men and women. For patients from China or Non-Chinese Asian countries, a wide resection margin was associated with better OS than a narrow resection margin. However, a pooled analysis of three studies from western countries showed that margin width was not associated with prognosis. Additionally, the wide surgical margin group obtained greater OS than that of the narrow surgical margin group in subgroups of hepatitis B surface antigen status (HBsAg) positive/negative and single/multiple tumors.

Correlation Between Surgical Margin and DFS
A pooled analysis of DFS data from 27 studies including 9,443 patients revealed that a wide surgical margin was related to better DFS (HR, 0.66; 95% CI, 0.61-0.71) (Figure 3). Further, subgroup analyses were performed based on reported median followup time (3-year DFS/5-year DFS), gender (male/female), country (China/Non-Chinese Asian countries/Western countries), HBsAg status (positive/negative), tumor number (single/multiple), liver cirrhosis (patients with/without). As a consequence, a wide surgical margin provided patients with better DFS compared to a narrow surgical margin ( Table 2).

Sensitivity Analysis
After careful selection of studies in sequence, sensitivity analysis outcomes confirmed the excellent robustness of the conclusion that a wide surgical margin could benefit the OS and DFS of patients (Supplementary Figures 1, 2).

DISCUSSION
The findings of this meta-analysis revealed that surgical margins correlate with the prognosis of HCC patients; besides, a wide surgical margin (≥1 cm) could improve long-term prognosis compared to a narrow surgical margin (<1 cm). This is in line with the results reported in previous articles (39,40). Through subgroups analyses, we found that the above outcome showed a similar phenomenon in different subgroups except for studies from Western countries. In this analysis, a wide surgical margin    Frontiers in Surgery | www.frontiersin.org did not prolong the OS of patients compared to a narrow surgical margin. This is potentially attributed to the inclusion of a small number of studies (five articles). No consensus has been reached in academia on whether gender is an independent risk factor for the prognosis of HCC patients after hepatectomy (45). Although there is no direct evidence that gender is a risk factor for HCC prognosis, men have higher smoking rates, alcohol consumption rates, and tumor burden than women (46). A different study found that women have a better long-term prognosis than men, but without statistical difference among patients with HCC lesions maximum size<3 cm or with solitary HCC (47). Notably, regional factors were also considered in subgroup analysis. The etiology of HCC in different regions is remarkably different. Asian countries, specifically East Asia are dominated by viral hepatitis, whereas HCC etiology in Western countries is mostly related to alcohol (48). Subgroup analyses revealed that despite HCC patients with/without hepatitis B virus (HBV) and liver cirrhosis, a wide surgical margin prolonged the prognosis of patients than a narrow surgical margin. HBV-liver cirrhosis-HCC progression is a vital approach for HCC occurrence. High HBsAg level, lack of antiviral treatment, severe liver cirrhosis are risk factors affecting this process (49)(50)(51). However, in single or multiple HCC populations, the wide surgical margin group still yields a better prognosis than the narrow surgical margin group. Nevertheless, a study on a patients with solitary HCC lesions revealed that a wide surgical margin was not a prognostic factor. However, after propensity score matching (PSM), a wide surgical margin still prolongs the prognosis of patients (44). This is possibly because PSM could reduce the confounding bias of OBS and improve the research efficacy by omitting the unmatched study subjects.
Microvascular invasion (MVI) is the presence of tumor emboli in vascular spaces rowed by endothelial cells from the tumor capsule into the liver parenchyma (either hepatic vein or portal vein branches) (52). Based on the distribution and number of MVI, MVI is classified into the following grades, M0: no MVI; M1 (low risk): MVI <5 and the distance from adjacent liver tissues ≤1 cm; and M2 (high risk): MVI >5 or the distance from adjacent liver tissues >1 cm (53). Researchers attempted to develop a preoperative model integrating laboratory examinations and imaging examinations to predict MVI. However, its accuracy requires additional validation by large-scale prospective multi-center studies (54). At present, MVI can only be diagnosed by postoperative histopathological examination; this significantly limits the application of MVI in guiding diagnosis and treatment. From MVI to macrovascular invasion, the malignant degree of HCC cells gradually increases and destroys the surrounding tissues; the chance of radical surgery is lost if a macrovascular invasion is formed (55). Therefore, effective surgical plans and postoperative adjuvant treatment can be adopted if timely interventions are implemented at the MVI stage of HCC. This thus minimizes metastasis and HCC recurrence as well as significantly improves the prognosis of patients.
To survive and metastasize, cancer cells must evade the immune system. After cancer cells invade the bloodstream, the classic hematological mechanism believes that platelets, leukocytes, and endothelial cells mediate the related process of metastasis and recurrence (56). New research indicates that MVI provides another path for HCC recurrence and metastasis; besides, HCC cell clusters obtain endothelial coating by protruding into the vessels. This enables evasion of the immune surveillance mechanism and thereby preventing the activation of the coagulation cascade (57)(58)(59)(60). Thus, if a liver resection with a narrower surgical margin is performed on patients, theoretically, the residual micrometastasis increases the risk of recurrence (37). Besides, 90% of MVI occurs in the range narrower than 1 cm from the edge of the tumor. If a wider margin is achieved, the incidence of MVI can be reduced, hence significantly preventing tumor recurrence and metastasis (61). However, due to data unavailability, we were unable to analyze the influence of MVI on the results in subgroup analysis. On the other hand, the liver status may be another mechanism of the prognostic influence of the resection margin. Patients who received a wide resection margin tend to have better liver reserves than patients who received a narrow resection margin. Therefore, compared with the narrow surgical margin group, the wide surgical margin group could achieve better OS and DFS.
The surgical margin should however not be blindly enlarged for preventing the recurrence and metastasis of HCC after surgery. Because of the excessively wide surgical margin, more normal liver parenchyma will be removed, causing serious postoperative complications including liver failure, and eventually death (8,9,11,12). Poon et al. (12) revealed that the relatively healthy liver parenchyma should not be sacrificed for obtaining the wider margin, particularly in cirrhotic patients with limited hepatic functional reserves. Another study (25) showed that a wide surgical margin did not improve the OS of patients compared to a narrow surgical margin. This was because of different baselines of the study group and the control group. This was largely reflected in liver cirrhosis, large and multiple tumors.
Previous research evaluated the relationship between surgical margins and prognosis by systematic review and meta-analysis (62,63). The findings (62) are inconsistent with this metaanalysis and suggests that prognostic benefits are not achieved in patients receiving a resection margin≥1 cm. A small number of articles (5 articles) included a potential reason. The study by Zhong et al. (63) lacked sensitivity analysis, therefore, the reliability and stability of its findings are uncertain. Yet, its results were consistent with this paper's findings. However, it had its limitations. Primarily, although the number of included studies is more than that of previous studies, it is still a relatively small amount when compared to the number of studies in our article (37 articles vs. 7 articles). Besides, subgroup analysis was not performed by Zhong et al. (63). It, therefore, remains unknown whether the conclusion (the prognostic benefit of a wide surgical margin) is affected by other factors.
Our study has worth-mentioning limitations. Firstly, because of the limited number of related studies, comprehensive analysis of different resection margin width could not be performed. Secondly, the study population is from Asia, therefore the results cannot be directly applied to the population in Western countries. Thirdly, most of the included literature is retrospective, thereby hinting a possibility of the potential risk of information bias. Fourthly, because of the non-availability of relevant data, we were unable to perform additional subgroup analyses including MVI and kind of resection (anatomical vs. non-anatomical).

CONCLUSION
In conclusion, our meta-analysis revealed that a wide surgical margin (≥1 cm) potentially prolongs the long-term prognosis of HCC patients than a narrow surgical margin (<1 cm). This metaanalysis conducted various subgroup analyses, and the results remained consistent across most factors of median follow-up time, gender, country, hepatitis B surface antigen status, tumor number, and liver cirrhosis.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.