Resection vs. Sorafenib for Hepatocellular Carcinoma With Macroscopic Vascular Invasion: A Real World, Propensity Score Matched Analytic Study

Background: Macroscopic vascular invasion (MVI) commonly occurs in patients with advanced hepatocellular carcinoma (HCC) for which resection and sorafenib are the common therapies prescribed. Here, we aimed to compare the survival outcomes of these two therapies in HCC patients with MVI. Methods: In total, 496 patients diagnosed with HCC and MVI without extrahepatic metastasis, treated with resection (resection-based group, n = 388) and sorafenib (sorafenib-based group, n = 108) were included in this study. A one-to-one propensity score-matching analysis (PSM) was performed to minimize the effect of potential confounders. Results: The median OS in the resection- and sorafenib-based group was 20.7 months (95% CI: 16.9–24.5) and 11.6 months (95% CI: 8.4–14.9) (p < 0.001), respectively. The median PFS was 4.7 months (95% CI: 3.8–5.5) in the resection-based group and 4.4 months (95% CI: 3.6–5.2) in the sorafenib-based group (p < 0.001). After PSM, 72 patients from each group were matched. The median OS was 27.2 months (95% CI: 16.4–38.0) in the resection-based group and 13.0 months (95% CI: 9.6–16.3) in the sorafenib-based group (p < 0.001). The median PFS was 5.3 months (95% CI: 3.2–7.4) in the resection-based group and 4.8 months (95% CI: 3.6–6.0) in the sorafenib-based group (p = 0.061). Conclusion: Findings from this study showed that, compared with sorafenib-based treatment, surgical resection might be associated with better survival benefits to HCC patients with MVI.


INTRODUCTION
Hepatocellular carcinoma (HCC) is one of the most common malignancies, and the fourth leading cause of cancer-related deaths worldwide (1). Because of its concealed onset, HCC often progresses to macroscopic vascular invasion (MVI) at the time of diagnosis (2). Before new targeted drugs such as lenvatinib and immune checkpoint inhibitors were available, the first-line of treatment for HCC patients with MVI recommended by the Barcelona guideline was systemic therapy with sorafenib (3) and had a median overall survival (mOS) ranging from 5.6 to 8.1 months (4,5). However, in the Asia-Pacific region, some patients with MVI, especially those without extrahepatic metastases, could still benefit from survival through resection, with mOS ranging from 8.9 to 33 months (6)(7)(8)(9)(10)(11). Therefore, the optimal choice between the two therapies for HCC patients with MVI was controversal.
To date, only a few studies have investigated the prognosis of surgical resection in comparison to sorafenib in HCC patients with MVI and inconsistent results have been reported, possibly, due to imbalanced patients characteristics between the investigated cohorts and a limited number of enrolled patients (8,12). Nowadays, the vast majority of patients with MVI were not in the initial treatment state when receiving surgery or sorafenib. Late-stage HCC patients are referred to combined therapies, instead of surgery or sorafenib alone. Therefore, relevant clinical researches were valuable.
The purpose of this study was to compare the prognoses of HCC patients with MVI undergoing surgical resection and sorafenib, aiming to provide a reference for the treatment of advanced HCC patients.

MATERIALS AND METHODS
This study was conducted according to the ethical guidelines of the 1975 Declaration of Helsinki. The analysis of the patient data was reviewed and approved by the Institutional Review Board and Human Ethics Committee at the Sun Yat-sen University Cancer Center (SYSUCC; Guangzhou, China).

Patients
The medical records of patients diagnosed with HCC and MVI without extrahepatic metastasis who underwent surgical resection or were prescribed sorafenib as part of standard therapy at the Department of Liver Surgery (SYSUCC), between 2005 and 2017, were reviewed for eligibility. Some patients were excluded based on the following criteria: (a) diagnosed with malignant diseases other than HCC; (b) aged >80 or <18; (c) had a performance status score >1; (d) had incomplete followup or medical information; (e) had sorafenib treatment for <2 months. Those who first received surgical resection treatment were classified into a resection-based group, while those who first received sorafenib-based treatment were classified into a sorafenib-based group. The patient enrolment and categorization flow chart is shown in Figure 1. All laboratory serum test data was collected within 3 days before treatment (resection or sorafenib). Preoperative imaging examinations included contrast-enhanced computed tomography (CT) or magnetic resonance imaging (MRI) within a week before treatment.

Treatment Procedure
Hepatic resection was performed as previously described (13). Intraoperative ultrasound was routinely performed to evaluate the tumor burden, remnant liver, and possibility of a negative resection margin. Anatomic hepatectomy with mass tumor thrombectomy was the preferred method of liver resection. Depending on its location and extent, the tumor thrombus was removed by en-bloc resected with the tumor tissue or extracted from the lumen of the blood vessel. Tumor thrombus was confirmed by rinsing with normal saline, and the absence of tumor thrombus formation was confirmed by intraoperative ultrasound.
Sorafenib (Bayer, Leverkusen, Germany) was initially orally administered 200 or 400 mg twice daily and continued for at least 2 months. Withdrawal and reduction of the drug depended on unacceptable toxicities or untreatable disease progression.
The final follow-up ended on July 31, 2019. Enhanced CT or MRI was performed every 2 or 3 months after surgery or sorafenib according to subsequent therapies. Follow-ups were performed as previously described (14), unless judged otherwise by the treating physicians.

Diagnosis and Definitions
The diagnosis of MVI was based on standard radiological imaging prior to treatment (resection or sorafenib prescription) (15,16). Portal vein tumor thrombus (PVTT) was graded according to the classification suggested by the Liver Cancer Study Group of Japan (17). Based on liver vessel structure and prognosis for different location of vascular tumor thrombus (18), we combined hepatic vein tumor thrombus (HVTT) into PVTT classification, based on which, Vp1 represented the invasion of a third-order branch or distal to the second branch of the portal vein; Vp2, invasion of a second-order branch of the portal vein, or branch of the hepatic vein; Vp3, invasion of in the first branch of the portal vein, or hepatic vein trunk or the short hepatic vein; Vp4, invasion of the main trunk/controlateral branch of the portal vein, or inferior vena cava. Overall survival (OS) was defined as the time interval from treatment initiation to cancerrelated death. Progression-free survival (PFS) was defined as the time interval from treatment initiation to tumor progression. For the resection-based group, tumor progression was defined as intrahepatic recurrence or new intrahepatic or extrahepatic lesions developed. For the sorafenib-based group, progression was defined as progressive disease (PD) according to the modified response evaluation criteria in solid tumors (mRECIST) (19), progressive intrahepatic tumor thrombus or extrahepatic metastasis.

Identification of Study Patients
From 2005 to 2017, 488 patients with HCC who underwent surgical resection (n = 388) or sorafenib (n = 108) treatment after a diagnosis of MVI without extrahepatic metastasis were identified. Of note, all patients in the sorafenib-based group were treated since January 2009 because sorafenib was available only from that year. In the resection-based group, 88 (22.7%) patients underwent surgical resection before January 2009 and the rest after January 2009.

Characteristics of the Study Patients
Between 2005 and 2017, 691 patients were reviewed for eligibility and 496 patients were ultimately included in this study (388 in resection-based group, 108 in sorafenib-based group).
The clinical pre-treatment characteristics of the patients in the resection-based and sorafenib-based groups are summarized in Table 1. In general, patients who underwent surgical resection had smaller tumor burden and better liver function. In the resection-based group, a smaller proportion of patients had severer liver cirrhosis (51.0 vs. 78.7%, p < 0.001), higher childpugh score (11.3 vs. 24.1%, p < 0.001), higher AST (61.1 vs. 75.0%, p = 0.008), and higher TBIL (13.7 vs. 28.7, p < 0.001), as compared to the sorafenib-based group. Meanwhile, larger proportion of patients in the sorafenib-based group were with multiple (67.6 vs. 32.2%, p < 0.001) or bilateral tumors (50.9 vs. 9.5%, p < 0.001), and had higher tumor thrombus grade (Vp3 and Vp4, 90.8 vs. 65.7%, p < 0.001). 313(80.7%) patients received surgical resection as their first treatment in the resection-based group, while only 40 (37.0%) patients were first treated with sorafenib in sorafenib-based treatment. In this study, 25 (6.6%) patients received subsequent sorafenib treatment in  the resection-based group, while 3 (2.8%) patients chose surgical resection afterward in the sorafenib-based group. After a 1:1 PSM, 72 pairs of patients were selected. The basic clinical characteristics between the two groups were almost consistent ( Table 2). Initially treated patients still differed, for 50 (69.4%) in the resection-based group and 20 (27.8%) in the sorafenib-based group. As for additional treatments, 4 patients received sorafenib after surgery and 3 patients received surgical resection after sorafenib treatment.
Forest plot analyses of factors associated with OS showed that resection provided a superior clinical benefit in most pre-planned subgroups except in female patients and those with tumor size <5 cm and Vp4 thrombus (Figure 4), as compared to sorafenib. ForPFS, resection only benefited patients with a single tumor ( Figure S1).

Prognostic Factors Analysis of Matched Patients
The risk factors for OS and PFS were analyzed in the matched cohorts (Tables S1, S2)   = 0.063) was a unique factor for RFS in both the univariate and multivariate analysis.

Progression Analysis of Matched Patients
The position of assessable tumor progression was analyzed in the matched cohorts (Table S3). Patients treated with sorafenib tended to have more intrahepatic progression (91.2 vs. 68.5%, p = 0.003). However, there was no statistically significant difference for extrahepatic progression between patients who underwent resection and sorafenib treatment (40.7 vs. 26.3%, p = 0.107).

Survival Analysis of Confounding Factors
Due to the actual treatment, the proportion of initial treated patients in the two groups was inconsistent. Thus, we further explore the prognosis of primary and non-primary patients in the two groups. Patients who underwent non-primary resection had better OS (mOS, 34.0 vs. 18.2 months, p = 0.005) but similar PFS (mPFS, 4.1 vs. 4.8 months, p = 0.885) as compared to those who received primary resection ( Figure S2). No significant survival difference was found between patients received primary and nonprimary sorafenib treatment (mOS, 13.6 vs. 10.1 months, p = 0.565; mPFS, 5.0 vs. 3.9 months, p = 0.407; Figure S3). In the resection-based group, 25 patients were treated with sorafenib after surgical resection and our findings showed that they had no superior OS or PFS (mOS, 24.1 vs. 20.7 months, p = 0.900; mPFS, 3.7 vs. 4.7 months, p = 0.077) than those who did not take sorafenib during the follow-up treatment ( Figure S4).

DISCUSSION
At present, the optimal therapy for advanced HCC remained uncertain. Although most European guidelines recommend targeted therapy as the first-line therapy, there were still a large number of studies confirming that surgery could bring survival benefit (6-8, 17, 22). Our study proved that in the real world, for some selected patients with good liver function and low tumor burden, surgical resection could have significant benefits of survival and disease control. After PSM, in the two groups of patients with similar baseline levels, surgery was still associated with significant increase in OS.
To our knowledge, there are currently a few studies comparing the efficacy of surgery with sorafenib. Costentin CE reported that OS of patients with HCC and MVI undergoing surgical resection was similar to that treated with sorafenib in a multicenter retrospective study, but the tumor states of patients after matching were not consistent and the result was based on a small sample of patients (46 patients vs. 39 patients) (12). Lee et al. and Wang et al. suggested that surgery offered more survival benefits to advanced HCC patients than other treatments including sorafenib and TACE (8,23). Kokudo et al. reported their results of surgery in a large cohort of more than 2,000 HCC patients with PVTT (17). The mOS in the resection group was 0.88 years longer than that in the non-resection group (2.45 vs. 1.57 years) in a propensity score-matched cohort. However, sorafenib was not included in the non-surgical treatments.
This study included initially treated and non-initially treated HCC patients with MVI. In our study, there were more initially treated patients who received resection than those who received sorafenib, which was in line with the actual treatment process for HCC patients in the Asia-Pacific region. In the sorafenib-based group, initially or non-initially treatment had no effect on OS and PFS. Several researchers indicated that proper therapies prior to sorafenib led to better survival outcomes for HCC patients (24,25). It implied that in the treatment strategy of advanced HCC, multiple combinations of locoreginal and systemic modalities treatment could be applied, which may have a beneficial therapeutic impact. For the resection-based group, patients who received non-first-line resection had better prognosis. This might suggest that those patients who could undergo surgery could benefit from preoperative neoadjuvant therapies.
Multivariate analysis showed that indicators of liver function seemed to have a greater impact on prognosis than tumor burden, distribution, and tumor thrombus levels. It was confirmed by other researches that attention should be not only paid to tumorrelated conditions, but also to the liver function of patients (26,27).
The results of subgroup analyses showed that almost all subgroups of patients could have greater overall survival benefits from surgery, except for female patients. This was possibly due to the limited number of identifiable and enrollable cases. Patients with all tumor thrombus levels except Vp4 could benefit from surgery and therefore, systemic therapy would be recommendable for advanced patients with PVTT reaching the main portal vein or HVTT reaching the inferior vena cava. It was confusing that patients with tumor size within 5 centimeters had not a significant OS benefit from surgical resection. Given the relatively small number of patients with small tumors in the analyzed two groups (resection vs. sorafenib, 17 vs. 16), it might accidentally cause statistical bias. The wide range of CI (0.25-1.61) indicated the poor sample representation, which probably could not reveal a true clinical phenomenon. Thus, more cases needed to be accumulated to confirm this part of the issue.
In this study, 25 patients received sorafenib after surgery but their prognosis (mOS, 24.1 months, mPFS, 3.7 months) was not significantly different from those who did not take post-operative sorafenib. Bruix Jordi's phase 3, double-blind, placebo-controlled study showed sorafenib is not an effective intervention in the adjuvant setting for HCC following resection or ablation (28). So, the application of post-operative sorafenib was controversial.
To note, the characteristics showed that patients with a higher tumor burden and more severe liver pathology were given sorafenib-based treatment. This was also reflected by the result that sorafenib-treatment as a negative prognostic factor for this group of patients. So resection shall be given careful evaluation for advanced HCC patients.
This study had some limitations. First, this was a retrospective study and although PSM was applied, there may still be some inevitable selection biases. Second, due to multidisciplinary comprehensive treatment strategy for HCC with MVI, patients included in this study had strong heterogeneity. The final result could not get rid of the influence of confounding factors including previous, concomitant and subsequent treatment after surgery and sorafenib. In addition, after the PSM, the number of cases was relatively small. Findings from this study should be further expanded to multicenter to obtain higher-level medical evidence. Besides, it was worth noting that the analyzed patients were coming from an Asia-Pacific region which was known to have a high incidence of hepatitis B virus (HBV) associated HCCs. Given the high prevalence of HBV associated HCC in Caucasians, Hong Kong Liver Cancer (HKLC) score could be operationprone due to a better stratification for these patients, where BCLC score might be conservative and unbefitting (29, 30). To better manage the treatment for patients with advanced HCC, especially in Caucasian, HKLC score was expected to applied.
In conclusion, our study indicated that, compared with sorafenib, surgical resection might be associated with better survival benefits in resectable HCC patients with MVI and adequate liver function, and should be considered as an important reliable therapy.

DATA AVAILABILITY STATEMENT
The datasets generated for this study are available on request to the corresponding author.

ETHICS STATEMENT
The analysis of the patient data was reviewed and approved by the Institutional Review Board and Human Ethics Committee at the Sun Yat-sen University Cancer Center (SYSUCC; Guangzhou, China).

AUTHOR CONTRIBUTIONS
JM, WW, and R-PG conceived and designed the experiments. JM, S-HL, Q-XW, J-WZ, L-HL, and Y-HL collected the data.