Introduction
Liver cancer ranks as the 6
th in incidence and 3
rd in termsof mortality [
1,
2]. Among liver cancers, hepatocellular carcinoma (HCC) is the most common malignancy [
3]. Several high-risk factors can lead to the occurrence of liver cancer, including hepatitis B virus, hepatitis C virus, aflatoxin, alcohol, fatty liver disease, and other conditions [
1,
4]. Currently, there are various treatments for HCC, including surgical treatment, local ablation, transarterial chemoembolization (TACE), radiotherapy, immunotherapy, targeted therapy, and chemotherapy [
5‐
7]. The 5-year survival rate of HCC patients who are candidates for curative therapies has been reported to exceed 70% [
8]. However, with a median survival time of 6 to 12 months, the prognosis of advanced HCC patients is usually poor [
9]. Dramatic progress in the treatment landscape has been made. However, the treatment strategies have certain liver function requirements [
10]. Thus, it is necessary to establish a simple and effective way of assessing liver function to facilitate treatment and predict the clinical efficacy and prognosis of these patients.
Currently, the Child–Pugh score is the most commonly used criteria to evaluate liver function and has been incorporated into the Barcelona Clinic Liver Cancer (BCLC) staging system to guide the clinical treatment and prognosis of HCC patients [
11‐
14]. However, 2 (i.e., ascites and hepatic encephalopathy) of the 5 indicators, which also include albumin, total bilirubin, and prothrombin time, are highly subjective. In addition, albumin and ascites are also commonly affected by other factors, which reduce the objectivity and accuracy of the scoring system [
15,
16]. The Albumin-Bilirubin (ALBI) grade is a new classification method proposed in recent years to evaluate liver function, which only contains two objective indicators (i.e., albumin and bilirubin) [
17]. Previous reports showed that HCC patients with normal liver function have a low ALBI grade, which usually impacts survival rates [
18].
For a long period, sorafenib, a multi-target tyrosine kinase inhibitor, has been recommended as a first-line treatment regimen for advanced HCC patients [
19,
20]. This is exciting time in the field of systemic chemotherapy by introduction of immune checkpoint inhibitors. If first line treatment regimen fails, regorafenib is suggested as the subsequent-line therapy for the treatment of HCC, and it has been shown to have considerable efficacy [
19,
21‐
25]. It has been shown that TACE combined with regorafenib and PD-1 antibody has a higher DCR after failure of TACE combined with regorafenib second-line therapy [
26]. Studies have shown that the ALBI grade can be used to predict the prognosis of HCC patients treated with regorafenib [
27‐
29]. An analysis of 284 regorafenib-treated HCC patients showed that the ALBI grade was an independent predictor of overall survival (OS), and that ALBI grade 1 patients had a 32% lower risk of death than ALBI grade 2 patients [hazard ratio (HR) = 0.68,
P= 0.08] [
30]. However, other study showed no significant correlation between ALBI grade and OS [
31]. The value of the ALBI grade in the prognostic assessment of HCC patients treated with regorafenib remains controversial, and most of the original studies did not address this issue.
Thus, systematic review and meta-analysis are needed to explore the predictive value of the baseline ALBI grade in the long-term prognosis of HCC patients who had been treated with regorafenib.
Methods
Systematic search
Public online databases, such as PubMed, Embase, Cochrane Library, Web of Science, CNKI, Wan Fang Data, and Vip Database were searched to retrieve potential eligible studies. The English search terms were as follows: (hepatocellular carcinoma or HCC or liver cell carcinoma) and (ALBI or Albumin-Bilirubin grade) and (regorafenib). The search was for the period from January 2010 until October 2022. To ensure all eligible studies possible were included, the reference lists of the included studies were also screened, and any relevant articles identified were manually retrieved.
Inclusion and exclusion criteria
Inclusion criteria were based on PICOS principles: (I) patients: include HCC patients whose diagnosis had been confirmed by pathology and/or imaging; (II) Interventions: patients with HCC received regorafenib. (III) Comparisons: ALBI values that were presented as categorical variables and the survival benefit of these patients was compared according to ALBI levels. (IV) Outcomes: use OS and progression-free survival (PFS) as the outcome indicators. (V) have obtainable effect estimates [e.g., the HR, odds ratio (OR) and its 95% confidence interval (CI)]. HR values from multivariate Cox regression analysis were used as preferred. Otherwise, data from univariate analysis will be used for analysis. (VI) Study type is a randomised controlled trial or prospective study or retrospective study.
Articles were excluded from the study if they met any of the following exclusion criteria: (I) concerned a review, animal cell experiment, or conference paper, and/or the full text of the article was not available; (II) had a sample size less than 20 and the 95% CIs could not be obtained.
Data extraction and quality evaluation
The literature screening and data extraction were performed independently by two researchers. If a disagreement arose, consensus was reached via discussion. The following relevant information was extracted from the final included studies: first author, publication year, region, study type, treatment details, sample size, sex, age, time of ALBI acquisition, OS, PFS, objective response rate (ORR), disease control rate (DCR), and their related effect estimates. The ALBI grade was calculated based on the serum albumin and bilirubin levels. The following formula was used [
32]: {ALBI score = [log10 bilirubin (µmol/L) × 0.66] + [albumin (g/L) × − 0.085]}. Classification standard: Grade 1, ALBI ≤ − 2.60; Grade 2, − 2.60 < ALBI ≤ − 1.39; Grade 3, ALBI > − 1.39. The modified ALBI grade was calculated based on the values of ALBI as following: Grade 1, ALBI ≤ − 2.60; Grade 2a, − 2.60 < ALBI ≤ − 2.27; Grade 2b, − 2.27 < ALBI ≤ − 1.39; Grade 3, ALBI > − 1.39.
The Newcastle–Ottawa Scale (NOS) was used to assess the quality of the included studies [
33]. The assessment was done by Cao and Zhou, independently. All studies were evaluated on the basis of their cohort selection, comparability, and outcome measures. The maximum NOS score is 9, Studies scoring 0–3 were rated as low quality; 4–6 as moderate quality; and a study scoring 7–9 was rated as high quality [
34,
35]. The main limitations that reduced the overall quality of the included studies were two: the decline in follow-up rates and the comparability of confounding factors.
Statistical analysis
Review Manager 5.4 software was used for the statistical analysis [
36]. The ORs/HRs and 95% CIs were obtained from univariate analysis of the included studies, and the OR/HR values were combined to calculate the combined estimates. A HR value > 1 indicated that high-grade ALBI was associated with a poor prognosis in HCC. The OR values and their 95% CIs were calculated for the dichotomous variables in the study.
Heterogeneity was assessed using the I2 statistic. An I2 value ≥ 50% and a P value < 0.1 indicated heterogeneity between the studies, and a random-effects model was used. A two-tailed P value < 0.05 indicated a statistical difference. I2 provides an estimate of the proportion of variance in the outcomes not attributable to sampling error alone, thus reflecting the actual heterogeneity in the outcomes. Funnel plots were directly observed to assess whether there was publication bias, and the asymmetric distribution of studies indicated publication bias. Sensitivity analyses were performed by including and excluding studies one by one. A P value < 0.05 indicated a statistically significant difference.
Discussion
In recent years, following its wide clinical application, regorafenib has been shown to have good efficacy and tolerable safety in the treatment of HCC; however, few studies have focused the prognostic biomarkers of regorafenib [
46,
47]. The present study showed that baseline high-grade ALBI in HCC patients is associated with poor OS and PFS. Further, the subgroup analysis showed that the baseline ALBI grade both before sorafenib or regorafenib treatment predicted the efficacy of regorafenib and the prognosis of HCC patients. There was also a significant difference in the DCR between the high-grade ALBI group and the low-grade ALBI group (both
P < 0.05). Further, the ALBI grade can also be used to assist in determining whether HCC patients are suitable for regorafenib treatment.
The ALBI grade is a novel liver function classification system, compared to the Child–Pugh classification system. The ALBI grade has been shown to have good prognostic value in recent studies [
48,
49]. In the prognostic prediction analysis of patients treated with TACE and sorafenib, the ALBI grade was found to be superior to the Child–Pugh grade. In a systematic review, the investigators compared the prognostic predictive value of the ALBI grade and Child–Pugh grade in patients with HCC, and found that ALBI grade had better discriminative ability than Child–Pugh grade in the prognosis of HCC patients [
48]. The prognostic value of the ALBI grade and Child–Pugh grade was also evaluated in patients with HCC treated with TACE. In a study cohort of 303 patients [
49], the survival curves stratified by the ALBI grade were significantly distinct (
P< 0.001), which suggested that the ALBI grade had greater prognostic value than the Child–Pugh grading for HCC patients receiving TACE. Another study showed that the combined application of the ALBI and Child–Pugh grades reduced the heterogeneity of liver cancer patients with different liver function states [
38]. In a study by Kim et al
., only patients with BCLC stage B HCC were included. After regorafenib treatment, the median PFS and OS of the ALBI grade 1/2 and ALBI 3 patients were 2.37
vs. 1.08 months, and 5.55
vs.3.64 months, respectively [
38]. Takada et al [
37]. noted that according to the Child–Pugh classification, patients with a Child–Pugh score of 5 had longer OS than those with a Child–Pugh score of 6 (18.6
vs. 7.9 months,
P < 0.01). In relation to the ALBI grade, patients with ALBI grade 1 had longer OS than those with ALBI grade 2 (21.6
vs. 9.3 months,
P = 0.001). Taken together, these results suggest that patients with good liver function reserve achieve better survival after receiving regorafenib treatment.
Current, evidence suggests that the ALBI grade has predictive value in a variety of tumors, including HCC. Marasco et al
.evaluated the diagnostic value of the ALBI grade in predicting post-hepatectomy liver failure in HCC patients who underwent liver resection [
50]. In their meta-analyses, a total of 7 studies were included, and the results showed that patients with high-grade ALBI at the baseline had increased rates of liver failure compared to those with low-grade ALBI at the baseline [
50]. Another meta-analysis also evaluated the prognostic role of the baseline ALBI grade in HCC patients after surgical resection [
51]. That meta-analysis included 20 studies with 11,365 patients, and the combined results showed that high-grade ALBI was correlated with short OS (HR = 1.64,
P < 0.01) and PFS (HR = 1.42,
P< 0.01) [
51]. In addition, this correlation was not significantly influenced by region or sample size. The clinical utility of ALBI as a prognostic factor was also assessed in HCC patients undergoing TACE [
52]. By extracting data from 8 studies, the authors found that a higher ALBI grade at the baseline was associated with a poor prognosis, and a median OS of 12.0 months in ALBI grade 3 compared to those with a median OS of 33.5 months in ALBI grade 1 [
52]. Recently, changes in ALBI score have been found to be a prognostic factor in patients receiving atezolizumab and bevacizumab (Atez/Bev). Unome et al
., in 62 unresectable HCC patients treated with Atez/Bev, ALBI scores deteriorated significantly within 3 months and were an independent prognostic factor for treatment [
53]. Campani et al
. enrolled 58 HCC patients treated with Atez/Bev, the results showed that the combination of ALBI and AFP early response was significantly associated with poor prognosis in patients with ALBI2-AFP non-responsive patients with OS (
P = 0.046) and PFS (
P= 0.012). These results indicate that ALBI grading combined with early AFP response can improve prognostic differentiation [
54]. In addition, a study using artificial intelligence algorithms to study serum biomarkers and construct survival prediction models, which show that ALBI grading is one of the most important serum biomarkers related to OS [
55]. All the above studies suggest that the baseline ALBI grade has an effective prognostic role in predicting prognosis, and these findings support our results.
However, the clinical value of the ALBI grade in predicting the prognosis of HCC patients treated with regorafenib is not well established. In our study, we assessed the association between the ALBI grade and prognosis of HCC patients who had undergone regorafenib treatment, and found that the baseline ALBI grade was significantly correlated with the survival and disease control outcomes, and its prognostic value was not affected by the definition of the baseline ALBI grade. High-grade ALBI was associated with a poor baseline ALBI prognosis in this population. Additionally, the ALBI grade could also be used to predict which HCC patients would be suitable for regorafenib treatment. Further, the modified ALBI grade was also suggested to be associated with survival outcomes, despite its prognostic value not being significant in the reported studies. The above findings suggest that the prognostic role of the ALBI grade may be applied in clinical practice. Our findings are consistent with the above studies, indicating that the ALBI grade has a certain generality and utility in efficacy prediction.
This study had some limitations. First, as the included studies were all retrospective studies, the baseline data may be inconsistent, and the data may be biased. Second, the included studies differed in terms of the treatment and follow-up, which may increase the heterogeneity of the analysis results. Subgroups, meta-regression and network meta-analysis should be performed. However, we failed to conduct them due to insufficient number of studies, comparisons, and sample sizes. Third, different time points for the ALBI data acquisition may have affected the predictive value of ALBI. Fourth, as some of the included studies only provided PFS data or only survival curves, this limited the studies that could be included in the meta-analysis, which in turn affected the predictive value of the ALBI grade in terms of OS. In addition, since the change in the ALBI score may be prognostic in patients receiving atezolizumab and bevacizumab, it should be studied the clinical value of the ALBI grade in patients that receive tyrosine kinase (TK) inhibitors after novel first-line combinations for HCC including immunotherapies. Therefore, large prospective multicenter study should be conducted to further explore its prognostic value.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.