As the sixth most common cancer and the third leading cause of cancer-related deaths worldwide, hepatocellular carcinoma (HCC) imposes a significant burden on health care [1]. The geographical distribution of HCC is extremely uneven. Despite the increased frequency of surveillance programs, the majority of patients have an intermediate or advanced stage HCC at diagnosis due to no significant symptoms in early stage. Of these, quite a number of them are with multinodular HCC (≥2). It is therefore vital to choose the most effective and appropriate therapy for these patients.

Hepatic resection (HR) is the standard therapy for early stage HCC in non-cirrhotics with a well-preserved liver function [2]. However, the role of HR in the treatment of multinodular HCC remains controversial. According to the widely used Barcelona Clinic liver cancer (BCLC) staging system [3] and its updated review [4], HR should not be recommended to the patients with multinodular HCC [3]. Although the recommendations of BCLC staging system are endorsed by the American Association for the Study of Liver Diseases (AASLD) [5] and the European Association for the Study of the Liver (EASL) [6], many large and qualified liver centers in treating HCC especially those in Asia do not subscribe to these guidelines. Some retrospective studies have demonstrated that HR is superior to palliative treatments for multinodular HCC [7–10]. Although these retrospective studies may have some unintentional selection bias, the first randomized controlled trial (RCT) by Yin and coworkers definitely support this conclusion [11]. These studies highlight one of the ongoing controversies that surround the BCLC system.

The fact that AASLD/EASL guidelines do not recommend HR for patients with intermediate or advanced stage HCC reflects primarily concerns over high recurrence and perioperative mortality, rather than direct prospective clinical evidence. However, recent advances in perioperative management and surgical techniques, as well as more restrictive selection of patients for HR, have rapidly reduced perioperative mortality, which was only 1.1 % in the RCT [11]. Moreover, HR can achieve satisfactory overall survival [12] in patients with recurrent HCC. Aggressive treatment of recurrence by repeat HR, radiofrequency ablation (RFA), and adjuvant therapies, such as transarterial chemoembolization (TACE), can offer satisfactory overall survival (OS) [13–15]. Therefore, this trial and other retrospective studies concluding HR superior to TACE makes sense [7–9, 11].

RCTs directly comparing TACE and HR in such patients have been lacking because clinicians are loathe to treat them exclusively with TACE, given the demonstrated survival benefit of HR in many retrospective studies [7–9]. For this reason, this RCT [11] stands out as providing clear, direct evidence against current guidelines. Therefore, it is a very valuable study.

This RCT was conducted on 173 patients with multinodular HCC outside of Milan criteria who were treated in a big liver center in eastern China. The patients were included in a short time period from 2008 to 2010 and underwent intention-to-treat analysis. Most of the included patients (91.3 %) were positive for hepatitis B surface antigen and with cirrhosis (82.7 %). There were no significant differences in the 30-day and in the 90-day mortality between the HR and TACE groups. However, the HR group had significantly better OS than the TACE group (P < 0.001). The 1-, 2-, and 3-year OS and median survival were 76, 64, and 52 % and 41 months, respectively, in the HR group. The trial did not report the 5-year OS.

Though this is a well-conducted RCT that challenges current thinking about the appropriateness of HR for patients with multinodular HCC, the evidence of this RCT suffers from important weaknesses. First, median 3-year survival in their TACE cohort was 14 months (range, 5–47 months), and OS at 1, 2, and 3 years was 52, 35, and 18 %, respectively. These survival rates are significantly lower than those reported in other large retrospective studies of Asian cohorts [7, 16–19] (Table 1). The authors [11] attribute this discrepancy to the fact that their TACE patients had larger tumors than those of similar cohorts in other studies [7, 16–19], yet other studies involving similarly large tumors in fact showed better OS (Table 1). It seems quite possible that the results of this trial reflect the presence of confounding factors in their TACE cohort. Although the TACE and HR patient groups did not show any statistically significant differences in baseline characteristics, risk factors for poor prognosis were more prevalent in the TACE group: TACE patients tended to be older than HR patients (54.0 vs. 51.6 years), to show higher incidence of cirrhosis (87 vs. 78 %) and Child-Pugh B liver function (5.9 vs. 1.1 %), and to have larger tumors (10.4 vs. 9.5 cm). These characteristics together may have contributed to the relatively low OS in the TACE cohort, and they call into question the generalizability of the study’s conclusions for patients with multinodular HCC.

Table 1 Efficacy of transarterial chemoembolization for Asian patients with BCLC-B/C hepatocellular carcinoma
Full size table

Further complicating interpretation of the survival outcomes in the RCT is the fact that although TACE is one of the primary methods for treating recurrent HCC, none of the 52 patients in the HR group who suffered intrahepatic recurrence (including 12 who suffered both intra- and extrahepatic recurrence) was treated with TACE. Moreover, although the authors themselves indicated in the methods section their intention to administer sorafenib to patients with extrahepatic recurrence, only five of 22 such patients (23 %) actually received this therapy.

Therefore, unintentional selection bias may also exist. The efficacy of HR may be amplified. In addition, because the indications for HR in Asia are broader than specified by the BCLC, it is easier to compare outcomes in patients with multinodular HCC who were treated with HR or TACE. A comprehensive review in the same topic may help to solve these problems.

After systematically searched PubMed database, 21 studies involving a total of 4,945 patients with multinodular HCC who underwent initial HR were included into analysis [7–9, 11, 12, 20–35]. All the 21 studies were published after January 2000 and in English (Table 2). Almost all the included patients were with preserved liver function. The median mortality and survival were 2.7 % (0–15.0 %) and 41 months, respectively. The median 1-, 3-, and 5-year OS and disease-free survival following HR were 75, 48, and 30 % and 60, 32, and 25 % (Fig. 1). The median survivals of our comprehensive review and the RCT are similar [36].

Table 2 In-hospital mortality and survival of patients with multinodular hepatocellular carcinoma after hepatic resection
Full size table
Fig. 1
figure 1

The median 1-, 3-, and 5-year OS and disease-free survival following HR

Full size image

The Asia-Pacific Association for the Study of the Liver [37], the Clinical Practice Guidelines for Hepatocellular Carcinoma (HCC) in Japan [38], the American Hepato-Pancreato-Biliary Association [39], and the Asian Oncology Summit 2009 [40] state that multifocality of tumor is not absolute contraindications to HR. The NCCN version 2.2014 guideline for hepatobiliary cancers recommends that limited and resectable multifocal HCC with no portal hypertension can be considered for HR [41]. Moreover, the combination of HR and intraoperative RFA widens the applicability of surgical intervention for patients with multinodular HCC [20, 21, 29]. Therefore, as long as HR is technically feasible and safe, HR was associated with good survival for selected patients with multinodular HCC.