Hepatocellular carcinoma (HCC) is the fourth most common cause of cancer-related death and currently ranks sixth globally in terms of tumor incidence.1 Hepatic resection is the established treatment of choice for HCC as a potentially curative therapy among several treatment options, such as resection, liver transplantation, local ablation, transarterial chemoembolization (TACE), and systemic therapy.2 However, the surgical indications for HCC differ between Western and East-Asian countries. The European Association for the Study of the Liver (EASL) and the American Association for the Study of Liver Diseases (AASLD) recommend the Barcelona Clinic Liver Cancer (BCLC) staging system for the management of HCC.2,3 According to this staging system, liver resection is only indicated for a single HCC of BCLC stages 0 or A. On the other hand, the Asian Pacific Association for the Study of the Liver (APASL) and the Japan Society of Hepatology (JSH) do not necessarily limit the indications for hepatic resection to solitary HCC cases.4,5

Many prior reports have also suggested that hepatic resection should be indicated for a single HCC even if large in size.6,7 These studies have indicated that liver resection is an effective treatment for a single HCC of any size. In the case of multiple HCCs, it has been reported that hepatic resections yield satisfactory results if they fall within the Milan criteria.8 In contrast, the effectiveness of these surgeries for multiple HCCs beyond the Milan criteria, i.e. intermediate-stage cases (BCLC stage B), is controversial and remains unclear.

In our present study, we examined the long-term outcomes among Japanese liver resection cases in accordance with the BCLC stage. We retrospectively reviewed a cohort of BCLC stage 0, A, and B HCC patients who underwent liver resection at our institution over a 30 year period. We conducted further subclassifications, analyzed the surgical outcomes, and investigated the validity of hepatic resection for multiple HCCs in comparison with single HCC cases.

Patients and Methods

Between 1991 and 2020, 1088 patients comprising very early (BCLC stage 0), early (BCLC stage A), or intermediate (BCLC stage B) stage HCCs underwent a liver resection at the Department of Gastroenterological Surgery I, Hokkaido University Hospital. A modified BCLC staging system was used for these patients as follows: BCLC 0 was defined as a single tumor ≤2 cm; BCLC A was defined as a single tumor >2 cm, or two to three nodules all <3 cm; and BCLC B was defined as two to three nodules ≥3 cm, or four or more nodules.2,3 In our present analyses, we further subclassified patients in the study cohort into five groups as follows: stage 0 (single nodule ≤2 cm), stage A1 (single 2–5 cm. or three or fewer nodules ≤3 cm), stage A2 (single nodule 5–10 cm), stage A3 (single nodule ≥10 cm), stage B1 (two to three nodules >3 cm), and stage B2 (four or more nodules). All cases had received a pathological diagnosis of HCC and any cases of pathological necrosis were excluded.

This study was approved by the Institutional Review Board of Hokkaido University Hospital (approval number: 021-0075). All analyses were performed in accordance with the ethical guidelines of Hokkaido University Hospital.

Preoperative Management

The surgical indications in our present study series were determined using an algorithm we developed and that we have described previously.9 The absence of uncontrolled ascites and a total bilirubin level of <2 mg/dL were required criteria for a subsequent hepatectomy. The specific liver resection procedure was then determined by measuring the indocyanine green retention rate at 15 min (ICGR15) and remnant liver volume by volumetric computed tomography (CT) prior to surgery.

Surgical Methods

We have previously described the surgical methods used for liver resection.9 An intraoperative ultrasound was performed in all patients. Transection of the liver parenchyma was conducted using the hook spatula of an ultrasonic harmonic scalpel (Ethicon EndoSurgery, San Angelo, TX, USA) and either a DS3.0 Dissecting Sealer (Medtronic, Minneapolis, MN, USA) or a bipolar cautery with a saline irrigation system. Inflow occlusion was applied in an intermittent manner with 15 min of occlusion alternating with 5 min of reperfusion. We defined anatomical resection in our current study as the complete removal of the lesion based on Couinaud’s classification.

Postoperative Management

Follow-up studies using CT or magnetic resonance imaging and measurement of α-fetoprotein (AFP) were conducted 1 month after the operation and at 3-month intervals for the first 3 years. After 3 years, routine follow-ups were performed once every 4 months and then every 6 months after 5 years, using CT scans and AFP assays.

Statistical Analysis

Categorical variables were compared between the study groups using Fisher’s exact test, and continuous variables were expressed as medians with ranges and compared using the Mann–Whitney U test. The overall survival (OS) and recurrence-free survival (RFS) rates were calculated using the Kaplan–Meier method and compared between groups using the log-rank test. Potential prognostic factors were identified by univariate analysis using the log-rank test. Independent prognostic factors were evaluated using a Cox proportional hazards regression model. A p-value <0.05 was considered statistically significant and all statistical analyses were performed using JMP version 14 for Windows (SAS Institute, Cary, NC, USA).

Results

Characteristics of the Entire Study Cohort

The clinicopathological features of patients in the entire cohort analyzed in this study are presented in Table 1. The study population included 88 stage 0, 750 stage A, and 250 stage B cases, comprising 892 men (82.0%) and 196 women (18.0%) aged from 18 to 92 years, with a median age of 65 years. A total of 371 patients (34.1%) were positive for hepatitis B surface antigen and 366 cases (33.6%) were positive for hepatitis C virus (HCV) antibody. We categorized 799 patients (73.4%) as no cirrhosis, 272 patients (25.0%) as Child–Pugh A cirrhosis, and 17 (1.6%) as Child–Pugh B cirrhosis. The median ICGR15 was 13.6% (range 0.8–94.4%) and the median AFP was 14.5 ng/mL (range 0–5,986,980 ng/mL). The median tumor size in the entire cohort was 4.0 cm (range 0.5–35.0 cm), with 751 patients (69.0%) having a single tumor and 337 (31.0%) showing multiple tumors. A non-anatomical liver resection was conducted in 315 patients (29.0%), whereas 773 patients (71.0%) underwent an anatomical liver resection. There were 88 (8.1%) stage 0, 485 (44.6%) stage A1, 190 (17.5%) stage A2, 75 (6.9%) stage A3, 166 (15.2%) stage B1, and 84 (7.7%) stage B2 patients in our current series. There were 896 patients (82.4%) showing a well to moderately differentiated HCC and 192 (17.6%) with poor to undifferentiated differentiation; 303 cases (27.8%) were positive for pathological microvascular invasion. The median operation time and blood loss were 313 min (range 88–1335 min) and 420 mL (range 0–61,350 mL), respectively. When we divided the patients into two groups by era of surgery, 480 patients (44.1%) were included in the first period (1991–2005) and 608 patients (55.9%) in the second period (2006–2020).

Table 1 Clinicopathological characteristics of the entire cohort
Full size table

Prognostic Factors Associated with Overall and Recurrence-Free Survival in the Entire Cohort

In our entire cohort of BCLC stage 0, A, and B HCC patients, univariate analysis revealed that HCV antibody, Child–Pugh B cirrhosis, ICGR15, AFP, BCLC stage, type of liver resection, histological differentiation, microvascular invasion, and era of surgery were significant prognostic indicators of OS. Multivariate analysis indicated that HCV antibody, Child–Pugh B cirrhosis, ICGR15, AFP, BCLC stage, type of liver resection, microvascular invasion, and era of surgery were independent prognostic indicators of OS (Table 2). Univariate analysis of the whole study population revealed that Child–Pugh B cirrhosis, ICGR15, AFP, BCLC stage, type of liver resection, histological differentiation, and microvascular invasion were significant prognostic indicators of RFS. By multivariate analysis, Child–Pugh B cirrhosis, ICGR15, BCLC stage, type of liver resection, and microvascular invasion were independent prognostic indicators of RFS (Table 2). The BCLC stage thus affected both the OS and RFS outcomes.

Table 2 Univariate and multivariate analyses of the prognostic factors in the entire cohort
Full size table

Overall and Recurrence-Free Survival Analysis

The 5-year OS rates for the stage 0, A1, A2, A3, B1, and B2 cases were 70.4%, 74.2%, 63.8%, 47.7%, 47.5%, and 31.9%, respectively (p < 0.0001) [Fig. 1a], and the 5 year RFS rates were 41.9%, 36.5%, 34.8%, 18.3%, 15.3%, and 0.0%, respectively (p < 0.0001) [Fig. 1b]. There were significant differences in the OS outcomes between stages A1 and A2 (p = 0.0118), A2 and A3 (p = 0.0013), and B1 and B2 (p = 0.0050), but not between stages A3 and B1 (p = 0.4742) (Fig. 1a). There were also significant differences in the RFS rates between stages A2 and A3 (p < 0.0001) and B1 and B2 (p = 0.0047), but not between stages A1 and A2 (p = 0.5940) or A3 and B1 (p = 0.5126) (Fig. 1b). No differences were found in either OS or RFS between stages A3 and B1.

Fig. 1
figure 1

a Significant differences in overall survival between stage A1 and A2 (p = 0.0118), A2 and A3 (p = 0.0013), and B1 and B2 (p = 0.0050) HCC patients. There were no significant differences between stage 0 and A1 (p = 0.8679) or stage A3 and B1 (p = 0.4742) cases. b Significant differences in recurrence-free survival between stages A2 and A3 (p < 0.0001) and B1 and B2 (p = 0.0047). There were no significant differences between stages 0 and A1 (p = 0.2150), A1 and A2 (p = 0.5940), or A3 and B1 (p = 0.5129).

Full size image

Risk Factors in Patients with a Stage B1 Hepatocellular Carcinoma (HCC)

Because no significant differences were found in either the OS or RFS outcomes between stage A3, which is indicated for a liver resection by the BCLC classification system, and stage B1, we performed subgroup analysis of the prognoses in our stage B1 study patients. Univariate analysis of these B1 cases revealed that Child–Pugh B cirrhosis was a significant prognostic factor for the OS rate. Multivariate analysis of this subgroup further revealed that Child–Pugh B cirrhosis was an independent prognostic factor for OS (Table 3). In the same manner, univariate analysis revealed that Child–Pugh B cirrhosis and microvascular invasion were significant prognostic factors for the RFS rate in stage B1 HCC patients, with multivariate analysis indicating that both of these variables were independent prognostic factors for the RFS outcome (Table 3).

Table 3 Univariate and multivariate analyses of prognostic factors in the BCLC-B1 cohort
Full size table

Recurrence Sites in Barcelona Clinic Liver Cancer (BCLC)-A3 and -B1 HCC Patients

Because there were no differences between stages A3 and B1 with respect to either OS or RFS, we further analyzed the sites of HCC recurrence for both of these HCC classifications. As indicated in Table 4, more patients in the B1 group experienced intrahepatic recurrence, whereas more patients in the A3 group experienced lung recurrence. On the other hand, there were no significant differences between the A3 and B1 groups in terms of other extrahepatic recurrences at sites such as the bone, lymph node, brain, adrenal gland, or peritoneum.

Table 4 Recurrence sites of BCLC-A3 and BCLC-B1 HCCs
Full size table

Discussion

The surgical indications for a hepatectomy to treat a single HCC are widely accepted, whereas those for multiple HCCs remain unclear and controversial. The BCLC staging system recommends liver resection for BCLC-0 and BCLC-A patients only, and not for BCLC-B cases. In contrast, Asian guidelines, including those from the APASL and the JSH, suggest liver resection as a treatment option for BCLC-B patients.4,5 In this present study, we retrospectively assessed the therapeutic value of liver resection for BCLC-B HCC patients, and found that it yields an acceptable surgical outcome for select BCLC-B cases.

In accordance with this staging system, the treatment options in the past for BCLC-B patients have been TACE only, and either TACE or systemic therapy in more recent years.2,3 However, the BCLC-B stage comprises a highly heterogeneous population of HCC cases,10 for example containing both Child–Pugh class A and B patients, resulting in an extremely large patient population even from the perspective of a hepatic functional reserve alone. Multiple HCCs beyond the Milan criteria are classified as BCLC stage B and thus involve various sizes and numbers of tumors. Several previous reports have shown that some populations benefit from hepatic resection, even in BCLC stage B patients. In their previous study, Zhong et al. insisted that a BCLC-B classification is not a contraindication for hepatic resection from an assessment of the therapeutic value of this surgical approach and comparing it with TACE among BCLC-B and -C patients.11 Wang et al. recommended a resection for BCLC-B patients when there is no microvascular invasion,12 while Wada et al. also contended that a hepatic resection should be considered as a radical treatment for certain patients with multiple BCLC-B HCCs.13 Our current study findings also suggest that the long-term results of hepatic resection for BCLC-B HCC with three or fewer nodules are equivalent to those for a single large HCC. JSH guidelines also recommend hepatectomy as a treatment option for HCC cases with three or fewer nodules, but recommend other interventions in cases of four or more nodules.5 Tsilimigras et al. reported that the prognosis for a single large HCC was poorer than in other BCLC stage A cases, but was similar to patients presenting with BCLC stage B tumors following a liver resection,14 which is consistent with our current findings. Liver resection for a single large HCC has been previously associated with acceptable long-term outcomes.6,7,15 Therefore, taken together, the cumulative evidence to date suggests that a BCLC-B HCC should not be comprehensively regarded as a contraindication for surgery if there are three or fewer tumors.

Our current results from multivariate analysis further indicated that only a Child–Pugh B cirrhosis is an independent prognostic factor for OS in patients with a Stage B1 HCC. Liver resection for HCC in a Child–Pugh B cirrhosis background is generally controversial but can be acceptable in select cases, although the prognosis is generally poor. Taura et al. reported that the OS rate following a hepatic resection in Child–Pugh class B cirrhotic patients was poorer than in both non-cirrhotic and Child–Pugh class A cirrhotic patients,16 while Berardi et al. reported that liver resection should be considered for HCC in cases with a Child–Pugh B cirrhosis after careful selection in accordance with the patient characteristics, tumor pattern, and liver function.17 Harimoto et al. stated that a hepatic resection for recurrent HCC and excessive blood loss should be avoided in patients with Child–Pugh class B cirrhosis18; the prognosis of a hepatectomy for Child–Pugh B HCC is not always acceptable. The same theory applies to liver resection for patients with cirrhosis. Taura et al. reported that co-existing cirrhosis is associated with a higher mortality and recurrence rate and that this limits the efficacy of hepatic resection.16 These authors insisted that hepatic resection should be the treatment of choice for HCC patients without cirrhosis. Hence, based on the results of our current study and other reports, a liver resection for a BCLC-B HCC should be limited to no cirrhosis or Child–Pugh A cases. Fukami et al. recently reported that a liver resection could offer a good long-term survival outcome for patients with multiple HCCs with up to three tumors with a Child–Pugh A grading,19 which is consistent with our present findings. Also based on our present observations, even in BCLC stage B1 cases, the 5-year OS rate for the second surgical period was 60.7%, compared with 38.5% of cases in the first period (Table 3). Furthermore, our current analyses indicated a 5 year OS rate of 62.3% from a second period hepatectomy in BCLC stage B1/no cirrhosis or Child–Pugh A patients. In contrast, Fukami et al. reported a 5 year OS rate of 41.6% with TACE for HCC patients with up to three tumors and a Child–Pugh A grade.19 Taken together, and since the results from these surgeries have been improving in recent years, a liver resection may even be considered for a BCLC-B HCC if there are three or fewer tumors, especially in patients with a no cirrhosis or Child–Pugh A classification.

Our present observations have also indicated that patients with a single large HCC had similar long-term results as those with multiple HCCs, i.e. three or fewer. In 2011, the original BCLC staging system was updated to define a single large HCC (≥5 cm) as BCLC stage A rather than stage B.20 The current BCLC staging system also follows that definition.2,3 Among several factors, tumor size has been reported to correlate with a poor prognosis in HCC patients.21,22 Jung et al. suggested that a single large HCC should be classified as BCLC stage B, rather than stage A.23 In our present analyses, the long-term results after hepatic resection were found to be similar for stage A3 and B1 patients; however, in terms of recurrence, pulmonary recurrence was more common in stage A3 cases, while hepatic recurrence was more common in our stage B1 patients. Hence, stage A3 and B1 HCCs cannot be regarded as the same group. However, many previous reports have suggested that liver resection should be indicated for a single HCC even if it is large in size.6,7 In addition, our present study found that the results of a hepatectomy for multiple HCCs in select patients are comparable with those for a single large HCC treated in this way. Hence, a designation of BCLC stage B should not be considered an a priori contraindication for liver resection.

The treatment of HCC has evolved dramatically and has diversified in recent years. In particular, the development of systemic therapy has changed the treatment systems available for HCC. In the treatment of BCLC-B HCCs, only TACE has been applied previously3; however, both TACE and systemic therapy can now be indicated for this grade of HCC in accordance with the revised BCLC staging system.2 In the field of HCC, the possibility of conversion surgery has recently been explored with the development of systemic therapy, although the preoperative treatments for HCC have not yet been standardized.24,25,26,27 Currently, the mainstays of the systemic therapies for HCC are atezolizumab plus bevacizumab, as well as sorafenib, and lenvatinib. Atezolizumab plus bevacizumab has now become the first‐line systemic treatment for unresectable HCC but its impact on conversion surgery is still unknown. However, knowledge of the impacts of lenvatinib has been accumulating with regard to conversion surgery for HCC. Shindoh et al. recently reported clinical data from conversion surgeries after lenvatinib treatment for HCCs, including BCLC stages A–C.27 These authors concluded that conversion surgery after lenvatinib treatment may offer a significant survival benefit in select patients as long as an R0 resection is achieved. Therefore, in the future, conversion surgery may offer a better prognosis for patients with multiple HCCs, and liver resection may even be indicated for cases of four or more nodules with preoperative treatment.

Our present study had some notable limitations, including its retrospective nature and examination of patients from a single center. Hence, a potential bias may have existed in relation to the enrolled cohort. In addition, this study lacked a control group that received TACE or systemic therapy, and we could not make definitive conclusions regarding the superiority of different treatment approaches for patients with multiple HCCs. Furthermore, this study only included patients who were eligible for hepatic resection. Since this study did not examine the total patient population, including those treated with therapies other than hepatic resection, especially for stage B1 and B2 cases, patients who underwent liver resection for multiple HCCs are a highly select population. Hence, selection bias also existed when liver resection was chosen instead of TACE or systemic therapy. However, while the indications for HCC surgery differ between Europe, the US, and Asia, our present investigation was a valuable examination of Japanese liver resection cases from the perspective of a Western staging system.

Conclusion

The long-term results of a hepatectomy for multiple HCCs are equivalent to those for a single HCC if there are three or fewer tumors, and a good prognosis can be expected for patients with a no cirrhosis or Child–Pugh A grading. Hence, hepatic resection should be considered for patients with multiple HCCs if they have no cirrhosis or a good functional liver reserve and there are three or fewer tumors.