Perioperative and prognostic implication of albumin‐bilirubin‐TNM score in Child‐Pugh class A hepatocellular carcinoma

Abstract Background and Aim A reliable classification for predicting postoperative prognosis and perioperative risk of hepatocellular carcinoma (HCC) patients is required to make a precise decision for HCC treatment. In the present study, we assessed the perioperative and prognostic importance of indocyanine green (ICG) testing, tumor‐node‐metastasis (TNM) stage, albumin‐bilirubin (ALBI) grade, and ALBI‐TNM (ALBI‐T) score using consecutive resected HCC cases. Methods Between 1998 and 2011, 273 consecutive patients who underwent primary and curative hepatectomy for HCC were identified. Among these 273 cases, 235 Child‐Pugh class A patients were enrolled in the present study. Results Correlation analysis showed that the value of linear predictor for ALBI grade was significantly correlated with ICG 15‐minute retention rates (r = 0.51, P < 0.0001). Survival analysis for both recurrence‐free survival (RFS) and overall survival (OS) showed there were significant differences between the two groups stratified by stage or ALBI‐T score (stage, RFS: P = 0.01, OS: P = 0.003; ALBI‐T, RFS: P < 0.0001, OS: P < 0.0001). In addition, Cox proportional hazard model identified ALBI‐T score was a significant predictor for both RFS and OS (RFS, P = 0.001; OS, P = 0.004). Furthermore, ALBI‐T score could predict perioperative risk in hepatectomy such as longer operation time and excessive intraoperative blood loss. Conclusions This study showed a robust association of ALBI‐T score with postoperative HCC patient survival and perioperative risk in hepatectomy. ALBI‐T score can be used as a simple and powerful tool for assessing HCC patients with further study.


| INTRODUCTION
Hepatocellular carcinoma (HCC) is a lethal disease and the second leading cause of cancer death worldwide. 1 For curative treatment of HCC, hepatectomy (hepatic resection) is a major and desirable strategy. [2][3][4] However, even after curative hepatectomy, 80% of patients develop HCC recurrence in the remnant liver and 50% die within 5 years. 5 Hetero chronological multiple HCC occurrences are generally associated with background hepatitis caused by virus, alcohol, and non-alcoholic fatty liver disease 6 whereas intrahepatic tumor metastasis is mainly attributed to invasiveness of primary loci and relatively early recurrence. 7 According to these two types of hepatic recurrence, both background liver status and tumor factor of HCC should be considered to make a precise decision for HCC treatment.
Currently, the American Joint Committee on Cancer (AJCC)/ Union for International Cancer Control (UICC) tumor-node-metastasis (TNM) staging system is commonly used for evaluating pretreatment tumor status, and Child-Pugh (CP) classification is applied for evaluating background liver status. 8 However, as previously mentioned, CP classification and the integrated scoring system using CP classification have several limitations as a result of including subjective values such as grading of ascites and hepatic encephalopathy. 9 Indocyanine green (ICG) testing is usually carried out prior to HCC treatment for assessing background liver status especially in Asian countries and in several institutes in Europe. 9,10 Although ICG testing well reflects the liver function, this examination requires well-organized preparation to obtain an accurate result such as preexamination bed-rest whether or not using an ICG clearance meter that does not require multiple blood sampling. Hence, a simple and accurate way to evaluate both tumor and background liver status of HCC prior to treatment has been strongly desired.
Lately, albumin-bilirubin (ALBI) grade 11 and ALBI-TNM (ALBI-T) score 12 have attracted clinicians' attention as more convenient and precise methods to evaluate HCC and background liver. Although there are some important studies showing the prognostic impact of ALBI grade and ALBI-T score on HCC treatment, 11,12 their actual significance to HCC surgery is still being considered. In the present study, we retrospectively assessed the usefulness of ICG testing, TNM stage of the Liver Cancer Study Group of Japan (LCSGJ), 13 ALBI grade, and ALBI-T score to predict HCC prognosis after hepatectomy and to evaluate the risk at hepatectomy by conducting a search of consecutive resected HCC cases from our institute.

| Patients enrolled in the present study
Between January 1998 and December 2012, 273 consecutive patients who underwent curative hepatectomy for HCC at the Department of Gastroenterological Surgery, Nagoya University Hospital (Nagoya, Japan) were identified. The 235 cases identified as CP class A were enrolled in this study. Written informed consent, as required by the Institutional Review Board, was obtained from all patients for use of the anonymized information.  16 were carried out both respectively and jointly. Resection was defined as curative when all gross tumors were completely removed; cases of incidentally found small lesions suspected to be HCC that were treated by radiofrequency therapy or microwave coagulation therapy during the surgery were regarded as curative cases.

| Clinical examination and hepatectomy
Surgery-related variables included operation time, intraoperative blood loss (IOBL), and requirement for intraoperative blood transfusion. Tumor-related variables included tumor number and size, and postoperative pathological variables (tumor differentiation, serosal invasion, capsule formation, capsule infiltration, septal formation, vascular invasion, bile duct invasion, and surgical margin). Tumors were categorized as well/ moderately or poorly differentiated, whereas the other pathological variables were categorized as positive or negative, as described by the guidelines of the LCSGJ. 13 We used the definition of the Clavien-Dindo classification to assess postoperative ascites, pleural effusion, bile leakage, and surgical site infection and grade IIIa or greater was considered positive. 17 As for postoperative liver failure, we referred to the grading by the International Study Group of Liver Surgery and, in the present study, grade B and C were considered positive. 18

| Follow up after surgery
After discharge, patients were followed up once per month for 3 months and every 3 months thereafter. Blood examination, including those for serum AFP and des-gamma-carboxy prothrombin, was carried out at every outpatient care visit, and dynamic contrastenhanced CT was done every 6 months. Patients with abnormal data or suspected lesions underwent further examinations, including contrast ultrasonography, magnetic resonance imaging with gadoliniumethoxybenzyl-diethylenetriamine pentaacetic acid, CT with hepatic arterioportography, and/or positron emission tomography for the diagnosis of HCC recurrence.

| Classifications
Albumin-bilirubin score was calculated and patients were classified with the cut-off points as previously reported. 11 Linear predictor for ALBI grade was calculated with the following equation: Guidelines by LCSGJ. 13 T factor for TNM stage of LCSGJ is as shown in Table S1. ALBI-T score was calculated using the following equation: ALBI-T score = ALBI grade + TNM stage of LCSGJ − 2. 12

| Statistical analysis
All statistical analyses were carried out using R version 3.4.3 (https://www.r-project.org/). Continuous variables were expressed as medians (ranges) and compared using the Wilcoxon rank-sum test, and categorical variables were compared using the chi-squared test or Fisher's exact test, as appropriate. Recurrence-free survival (RFS) was defined as the time between the curative resection of HCC and confirmation of recurrence. Overall survival (OS) was defined as the time between the operation and all-cause death. Cox proportional hazards models were used to determine the risk factors associated with RFS and OS. Survival analysis based on the Kaplan-Meier method and log-rank tests was also carried out. The level of statistical significance was set at P < 0.05, which was obtained using two-tailed tests.

| Patient characteristics
In the present study, 235 CP class A HCC patients were enrolled, as other classes (B and C) of CP classification were rare in the patients who underwent hepatectomy and the class B cases were hypothesized to have worse prognosis than grade A. 19 Patient demographic and clinical characteristics are shown in Table 1

| Albumin-bilirubin (ALBI) grade and ALBI-TNM score in Child-Pugh class A HCC patients
Among 235 CP class A HCC patients, 142 (60%) patients were classified as ALBI grade 1 and 93 (40%) patients were grade 2 and there was no patient classified as grade 3. Histogram of ALBI grade based on hepatitis virus infection is shown in Figure 1A. In HCC patients with HBV, 47 patients (72%) were classified as grade 1 and the proportion of ALBI grade was significantly different between HBV and HCV patients (P = 0.007). Using ALBI-T score, 231 informative cases with preoperative stage based on the guidelines of the LCSGJ were classified as follows: score 0; 19 (8%), score 1; 79 (34%), score 2; 86 (37%), score 3; 34 (15%), score 4; 13 (6%); and score 5, 0. Histogram of ALBI-T score based on virus infection is shown in Figure 1B. The most dominant ALBI-T score in patients with HBV was score 2 whereas score 1 was most frequently seen in patients with HCV ( Figure 1B). HCC patients without any hepatitis virus had the same tendency of distribution in ALBI and ALBI-T as patients with HCV ( Figure 1A Table S3).

| Association among ALBI grade, ALBI-T score, and ICG-R15
In 154 informative cases with ICG-R15, a significant difference of ICG-R15 value could be identified between ALBI grade 1 and ALBI grade 2 (P < 0.0001, Figure 1C) although there was no significant difference of ICG-R15 between ALBI-T 0,1,2 and ALBI-T 3,4 (P = 0.96, Figure 1D). In addition, correlation analysis showed that the value of a linear predictor for ALBI grade (xb) was moderately correlated with ICG-R15 (r = 0.51, P < 0.0001, Figure 1E). The  Tables 2 and 3). As in our previous report, serosal and vascular invasions that were tumor factors diagnosed with resected specimens were a strong predictor for HCC prognosis. 20 Among the features that can be presurgically obtained and the classifications using presurgical features, ALBI-T score was able to separate CP class A cases into different prognoses both in RFS and OS better than other classifications or clinical features.

| Association of ALBI and ALBI-T with perioperative risk in hepatectomy
We also investigated the impact of ICG-R15, stage, ALBI grade, and ALBI-T score in CP class A patients to operation time, IOBL, and rate of transfusion during hepatectomy ( Figure 4). ALBI-T score 3, 4   (Table S4). According to this analysis, stage was associated with pleural effusion (P < 0.0001); ALBI grade was associated with persistent ascites (P = 0.002); and ALBI-T score was associated with both ascites and pleural effusion (P = 0.04 and 0.0003). Consequently, these results indicate that ALBI-T score has a capability of predicting operation time, IOBL, and postoperative complications preoperatively as well as predicting postoperative prognosis.

| DISCUSSION
In this study, we evaluated perioperative clinical importance of ICG-  The poor prognosis and high frequency of HCC recurrence is associated with both tumor factors and background liver status. 23 To deliver a desired precision treatment to HCC patients, estimating the prognosis of patients planning to undergo hepatectomy is essential.
Furthermore, hepatectomy has a potential risk of massive bleeding requiring blood transfusion that causes severe complications such as hepatic failure after surgery. 24 To avoid perioperative fatality, evaluating background liver status and developing an appropriate strategy are also crucial. Thus, in terms of the unique aspects of prognosis and operative risk of HCC, we compared ICG-R15, stage, ALBI grade, and ALBI-T score and finally shed light on the superiority of ALBI-T comprising both tumor features and background liver status.
Tumor staging is generally needed to determine the patients' survival probability after treatment, decide the most appropriate therapy, and enable an objective comparison among the outcomes of cancer research. Furthermore, it should allow us to predict the prognosis of resected cancer cases and individual treatment risk. For these reasons, staging systems should separate patients into groups with homogeneous prognosis, and serve to select the appropriate treatment. 25 The AJCC/UICC TNM staging system for HCC incorporates tumor size and local invasiveness such as vascular invasion, and number of tumor nodules as well as lymph node and distant metastasis. 8 As for various types of neoplasms, the TNM staging system is a reliable outcome predictor, but prognostic modeling in HCC is more complex than those in other gastrointestinal cancers. There are several classifications of HCC such as the Groupe d'Etude et de Traitement du Carcinome Hépatocellulaire prognostic classification, 26 Cancer of the Liver Italian Program, 27 Barcelona Clinic Liver Cancer staging, 28 the Chinese University Prognostic Index, 29 and staging according to the guidelines of the LCSGJ. 13 However, none of these classifications has received universal acceptance. 30 One of the reasons why HCC staging is difficult is its characteristic recurrence  The present study was able to show a significant association between ALBI-T score and operation time, IOBL, and postoperative prognosis in CP class A patients. However, there are several inherent limitations in this study. First, this study was based on retrospective single-institutional clinical information. The HCC patients enrolled in this study were from Japan only, and it is well known that HCC from different regions has a different etiology and prognosis. In addition, the screening system for HCC in Japan is well established and relatively small HCC can be frequently treated by hepatectomy. Thus, the ALBI-T might only be suitable for evaluating patients from the same country as it comprises the stage defined by guidelines of LCSGJ. Confirmation of the capability of this model as a perioperative risk predictor with HCC patients from other regions is crucial in order to apply this model worldwide in the future. Second, as a result of the limitations of the information availability, we were not able to carry out the analysis specific to each hepatitis etiology. It might be better to assess ALBI and ALBI-T performance in the same background hepatitis, as non-viral hepatitis and associated HCC are becoming more common especially in developed countries. 34,35 In conclusion, the present study showed robust association of ALBI-T score with perioperative risks of hepatectomy and postoperative patient survival in CP class A patients who underwent hepatectomy for HCC. ALBI-T score is a simple and powerful tool for estimating both patient's tumor factor and background liver status simultaneously. With further study, we could use ALBI-T score as a convenient way to assess HCC patients and deliver a more precise treatment to individual HCC patients.