Hepatocellular carcinoma (HCC) ranks among the most prevalent malignancies and represents a significant contributor to cancer-related fatalities on a global scale. While hepatic resection (HR) is often the primary choice for HCC patients with well-preserved liver function, recurrence of tumors frequently arises following curative HR [1,2]. The ability to precisely anticipate the post-resection prognosis holds the potential to tailor personalized treatment strategies for HCC patients. Consequently, there is a demand for prognostic models that can quantitatively forecast the post-resection prognosis in patients diagnosed with HCC.

Meanwhile, HCC is a well-established reason for liver transplantation (LT) and represents a significant indication for living donor liver transplantation (LDLT) in many Asian nations. Nonetheless, despite active treatment for HCC recurrence post-transplantation, it is frequently linked to reduced patient survival. Hence, careful selection of LT candidates is imperative to mitigate the risk of tumor recurrence [3-5]. While several criteria have been suggested for selecting LT candidates among HCC patients, most of them rely on binary concepts that may be challenging to apply to cases falling within the gray areas of these criteria. Consequently, there is a need to create a prognostic model that can quantitatively predict the post-transplantation prognosis for HCC.

A composite score, known as the ADV score, which is calculated by multiplying the levels of α-fetoprotein (AFP) and des-γ-carboxy prothrombin (DCP), alternatively referred to as proteins induced by vitamin K antagonist or absence-II (PIVKA-II), along with tumor volume (TV), has been demonstrated as an all-encompassing prognostic indicator for HCC subsequent to HR or LT [2-13]. This ADV score serves as a quantifiable metric that reflects the underlying characteristics of the tumor. This study was designed to validate the prognostic impact of ADV score in patients who underwent HR or LT for HCC through a collective review of twelve studies [2-14].

Multiplication of Tumor Volume by Two Tumor Markers is a Post-Resection Prognostic Predictor for Solitary Hepatocellular Carcinoma [2]

Authors hypothesized that microvascular invasion (MVI) and post-resection prognosis in patients with solitary HCC could be predicted using blood tumor markers and TV. Authors intended to identify a simple surrogate marker of HCC via a combination of clinical variables. This retrospective study used the strictly selected development cohort (n=1,176) and validation cohort (n=551) containing patients who underwent curative resection of solitary HCC. In the development cohort study, the median values were 13.7 mL for TV, 24.2 ng/mL for AFP, and 75 mAU/mL for DCP; there was no correlation among these three factors (r2≤0.237, p<0.001). The 1-, 3-, and 5-year rates were 22.4%, 41.7%, and 46.8% for tumor recurrence and 93.6%, 84.0%, and 78.2% for patient survival, respectively. Independent risk factors for both tumor recurrence and patient survival were tumor diameter >5 cm or TV >50 mL, MVI, satellite nodules, and high DCP. ADV score resulted in prediction of MVI at a cutoff of 5log with sensitivity of 73.9% and specificity of 66.7%. Patient stratifications according to ADV score with cutoffs of 5log alone, 6log and 9log, and its combination with MVI showed significant prognostic differences (all p<0.001). These prognostic significances were reliably reproduced in the validation cohort study (all p<0.001). Authors suggest that ADV score is an integrated surrogate marker of HCC prognosis. It can be used to predict MVI and post-resection prognosis for solitary HCC.

Small HCC with Low Tumor Marker Expression Benefits more from Anatomical Resection than Tumors with Aggressive Biology [6]

Authors assessed prognostic advantage of anatomical resection (AR) over non-anatomical resection (NAR) for HCC according to multiplication of ADV scores. Superiority of AR over NAR is debated. ADV score is surrogate marker of post-resection prognosis for solitary HCC. This study included 1,572 patients who underwent curative resection for solitary HCC of 2.0 to 5.0 cm between 2006 and 2014. Preoperative patient profiles were not statistically different between AR and NAR groups. In 1,324 naïve patients without preoperative treatment, AR group showed lower recurrence rates (p=0.003) and higher patient survival rates (p=0.012) than NAR group. AR group showed lower recurrence rates in patients with ADV ≤5log (p≤0.046). ADV scores >4log and >3log were independent risk factors for tumor recurrence and patient survival in treatment-naïve patients, respectively. In treatment-naïve group with preserved hepatic functional reserve, AR group showed lower recurrence rates in patients with ADV ≤4log (p=0.026). Absence of MVI also showed lower recurrence rates (p=0.007) in AR group. In 248 patients with preoperative treatment, AR group showed lower recurrence rates (p=0.001) and higher patient survival rates (p=0.006). AR group showed lower recurrence rates in patients with ADV ≤4log (p<0.001) and higher survival rates in patients with ADV ≤5log (p≤0.043). Authors concluded that prognostic benefit of AR was evident in patients with ADV score ≤4log or absence of MVI. Patients with less aggressive tumor biology benefit more from AR than NAR, thus being reasonably indicated for AR.

Validation of Prognostic Impact of ADV Score for Resection of Hepatocellular Carcinoma: Analysis Using Korea Liver Cancer Registry Database [7]

Authors aimed to validate the prognostic predictive power of ADV score for predicting patient survival after resection of HCC. This study included 1,390 patients with HCC registered in the Korea Liver Cancer Registry. Patients underwent HR between 2008 and 2012 and were followed up until December 2016. They were divided into 4 groups according to the number of tumors and preoperative treatment. There was no significant correlation among AFP, DCP, and TV values (r2≤0.04, p<0.001). In group 1 with single treatment-naïve tumor (n=1,154), patient stratification with postoperative ADV 1log-interval and cutoffs of 5log, 7log, and 10log showed great prognostic contrast (p<0.001). In group 2 with multiple treatment-naïve tumors (n=170), patient stratification with postoperative ADV 1log-interval and above-mentioned 3 cutoffs also showed great prognostic contrast (p<0.001). In group 3 (n=50) and group 4 (n=16) with preoperative-treated tumors, patient stratification with postoperative ADV 1log-interval and above-mentioned 3 cutoffs showed noticeable prognostic contrast (p≤0.031). Preoperative ADV score-based on preoperative findings also showed great prognostic contrast in 1,106 patients preoperatively diagnosed as having single treatment-naïve tumor (p<0.001). Confining patients to tumor-node-metastasis stages I and II (n=1,072) as well as Barcelona Clinic Liver Cancer stage 0 and A (n=862), postoperative ADV cutoffs showed further prognostic stratification. Author concluded that this validation study strongly suggests that ADV score is an integrated surrogate marker for post-resection prognosis in patients with HCC.

Prognostic Accuracy of the ADV Score Following Resection of Hepatocellular Carcinoma with Portal Vein Tumor Thrombosis [8]

Authors assessed the prognostic accuracy of ADV score following resection of HCC with portal vein tumor thrombosis (PVTT). This was a retrospective observational study. This study included 147 patients who underwent HR for HCC with PVTT. They were followed up for ≥66 months or until patient death. The grades of PVTT were Vp1 in 121 (14.3%), Vp2 in 41 (27.9%), Vp3 in 71 (48.3%), and Vp4 in14 (9.5%) cases. Preoperative HCC treatment was performed in 48 (32.7%) patients. R0 and R1 resections were performed in 119 (81.0%) and 28 (19.0%) cases, respectively. The 5-year tumor recurrence, HCC-specific survival, and post-recurrence survival rates were 79.2%, 43.5%, and 25.4%, respectively. Neither PVTT grade nor history of preoperative HCC treatment was a significant prognostic indicator. Stratification in accordance with ADV scores of 1log- and 3log-intervals resulted in high prognostic accuracy in predicting tumor recurrence and patient survival. Following cluster analysis, the cutoff for ADV score was determined at 9log and was more prognostically significant in terms of tumor recurrence and patient survival than surgical curability or MVI. Further comparisons revealed that prognostic prediction with an ADV score cutoff at 9log was more accurate than that using the Eastern Hepatobiliary Surgery Hospital-PVTT score. Authors concluded that ADV score is an integrated surrogate biomarker for post-resection prognosis in HCC with PVTT. The prognostic prediction model (PPM) using ADV scores provides reliable post-resection prognosis for patients with various grades of these tumors.

Prediction of Post-Resection Prognosis Using the ADV Score for Huge Hepatocellular Carcinomas ≥13 cm [9]

ADV score is a surrogate marker for post-resection prognosis of HCC. This study aimed to validate the predictive power of the ADV score-based PPM for patients with solitary huge HCC. Of 3,018 patients, 100 patients who underwent HR for solitary HCC ≥13 cm between 2008 and 2012 were selected. The median tumor diameter and TV were 15.0 cm and 886 mL, respectively. Tumor recurrence and overall survival (OS) rates were 70.7% and 66.0% at one year and 84.9% and 34.0% at five years, respectively. MVI was the only independent risk factor for disease-free survival (DFS) and OS. DFS and OS, stratified by ADV score with 1-log intervals, showed significant prognostic contrasts (p=0.007 and p=0.017, respectively). DFS and OS, stratified by ADV score with a cut-off of 8log, showed significant prognostic contrasts (p=0.014 and p=0.042, respectively). The combination of MVI and ADV score with a cut-off of 8log also showed significant prognostic contrasts in DFS (p<0.001) and OS (p=0.001) considering the number of risk factors. Prognostic contrast was enhanced after combining the MVI and ADV score. Authors concluded that the PPM with the ADV score could reliably predict the risk of tumor recurrence and long-term patient survival outcomes in patients with solitary huge HCC ≥13 cm. The results of this study suggest that our PPMs can be used to guide surgical treatment and post-resection follow-up for patients with huge HCCs.

Validation of Quantitative Prognostic Prediction Using ADV Score for Resection of Hepatocellular Carcinoma: A Korea-Japan Collaborative Study with 9,200 Patients [10]

ADV score has been shown to be prognostic of HCC recurrence following HR or LT. This multi-center, multinational validation study included 9200 patients who underwent HR from 2010 to 2017 at 10 Korean and 73 Japanese centers, and were followed up until 2020. AFP, DCP, and TV showed weak correlations (ρ≤0.463, r≤0.189, p<0.001). DFS, OS, and post-recurrence survival rates were dependent on 1.0log and 2.0log intervals of ADV scores (p<0.001). Receiver operating characteristic curve analysis showed that ADV score cutoffs of 5.0log for DFS and OS yielded the areas under the curve ≥0.577, with both being significantly prognostic of tumor recurrence and patient mortality at 3 years. ADV score cutoffs of ADV 4.0log and 8.0log, derived through K-adaptive partitioning method, showed higher prognostic contrasts in DFS and OS. Receiver operating characteristic curve analysis showed that an ADV score cutoff of 4.2 log was suggestive of MVI, with both MVI and an ADV score cutoff of 4.2 log showing similar DFS rates. Authors concluded that this international validation study demonstrated that ADV score is an integrated surrogate biomarker for post-resection prognosis of HCC. Prognostic prediction using ADV score can provide reliable information that can assist in planning treatment of patients with different stages of HCC and guide individualized post-resection follow-up based on the relative risk of HCC recurrence.

Validation of the OncoHepa Test, a Multigene Expression Profile Test, and the Tumor Marker-Volume Score to Predict Post-Resection Outcome in Small Solitary Hepatocellular Carcinomas [11]

OncoHepa test is a multigene expression profile test developed for assessment of HCC prognosis. ADV score was also developed for assessment of HCC prognosis. The predictive powers of OncoHepa test and ADV score were validated in 35 patients who underwent curative HR for naïve solitary HCCs ≤5 cm. Median tumor diameter was 3.0 cm. Tumor recurrence and patient survival rates were 28.6% and 100% at 1 year, 48.6% and 82.9% at 3 years, and 54.3% and 71.4% at 5 years, respectively. The site of first tumor recurrence was the remnant liver in 18, lung in 1, and the peritoneum in 1. All patients with HCC recurrence received locoregional treatment. OncoHepa test showed marginal prognostic significance for tumor recurrence and patient survival. ADV score at 4log also showed marginal prognostic difference with respect to tumor recurrence and patient survival. Combination of these 2 tests resulted in greater prognostic significance for both tumor recurrence (p=0.046) and patient survival (p=0.048). Authors concluded that both OncoHepa test and ADV score have considerably strong prognostic power, thus individual and combined findings of OncoHepa test and ADV score will be helpful to guide post-resection surveillance in patients with solitary HCCs ≤5 cm.

Prognostic Prediction Models for Resection of Large Hepatocellular Carcinoma: A Korean Multi-Center Study [12]

Authors developed a PPM using 4 factors for HR of large HCC. ADV score is a surrogate marker for post-resection prognosis. This study intended to validate the predictive power of 4-factor PPM and to develop new ADV score-based PPM. A total of 526 patients who underwent HR for solitary HCC ≥8 cm were selected from 9 Korean institutions between 2008 and 2014. Median tumor diameter and TV were 11.0 cm and 398 mL, respectively. Tumor recurrence and patient survival rates were 53.0% and 78.4% at 1 year and 70.2% and 49.3% at 5 years, respectively. Independent risk factors for both tumor recurrence and patient survival included AFP ≥100 ng/mL, hypermetabolic fluorodeoxyglucose (FDG)-positron emission tomography (PET), MVI and satellite nodules, which comprised 4 factors of the PPM. Five subgroups based on the number of involved risk factors exhibited significant differences in tumor recurrence and patient survival. ADV score cutoff was set at 7log (ADV7log) after cluster prognostic analysis. Patient grouping according to combination of ADV7log and FDG-PET findings (ADV7log-PET) exhibited significant differences in tumor recurrence and patient survival, comparable to those of the 4-factor PPM. Authors concluded that two PPMs using 4 risk factors and ADV7log-PET could reliably predict the risk of early HCC recurrence and long-term survival outcomes in patients who underwent HR for large HCC. We believe that these PPMs can guide surgical treatment for large HCCs from preoperative HR planning to post-resection follow-up.

Quantitative Prognostic Prediction Using ADV Score for Hepatocellular Carcinoma Following Living Donor Liver Transplantation [3]

Authors assessed the prognostic impact of the ADV score for predicting HCC recurrence and patient survival after LDLT. This study included 843 HCC patients who underwent LDLT between January 2006 and December 2015 at Asan Medical Center. These cases were divided into treatment-naïve (n=256) and pretransplant-treated (n=587 [69.6%]) groups. Results: There were weak or nearly no correlations among AFP, DCP, and TV. There existed high correlations between the pretransplant and explant findings regarding tumor number, size, and ADV score. Right lobe grafts were implanted in 760 (90.2%) patients. HCC recurrence and all-cause patient death occurred in 182 (15.9%) and 126 (15.0%) respectively during the follow-up period for 75.6±35.5 months. The 5-year tumor recurrence and OS rates were 21.5% and 86.2%, respectively. The pretransplant-treated group showed higher tumor recurrence (p<0.001) and lower OS rates (p<0.001). Tumor recurrence and OS were closely correlated with both pretransplant and explant ADV scores in the treatment-naïve and pretransplant-treated groups. The ADV score enabled further prognostic stratification of the patients within and beyond the Milan, University of California San Francisco, and Asan Medical Center criteria. Compared with the 7 pre-existing selection criteria, ADV score with a cutoff of 5log showed the highest prognostic contrast regarding tumor recurrence and OS. Authors concluded that their PPM using ADV scores is an integrated quantitative surrogate biomarker for posttransplant prognosis in HCC patients and can provide reliable information that assists the decision-making for LDLT.

Salvage Living Donor Liver Transplantation for Hepatocellular Carcinoma Recurrence after Hepatectomy: Quantitative Prediction Using ADV Score [4]

Salvage LT is a definite treatment for recurrent HCC after hepatectomy. Prognostic accuracy of ADV score was assessed in patients undergoing salvage LDLT and their outcomes were compared with patients undergoing primary LDLT. This study was a retrospective, single-center, case-controlled study. Outcomes were compared in 125 patients undergoing salvage LDLT from 2007 to 2018 and in 500 propensity score-matched patients undergoing primary LDLT. In patients undergoing salvage LDLT, median intervals between hepatectomy and tumor recurrence, between first HCC diagnosis and salvage LDLT, and between hepatectomy and salvage LDLT were 12.0, 37.2, and 29.3 months, respectively. DFS (p=0.98) and OS (p=0.44) rates did not differ significantly in patients undergoing salvage and primary LDLT. Pretransplant and explant ADV scores were significantly predictive of DFS and OS in patients undergoing salvage and primary LDLT (p<0.001). DFS after prior hepatectomy (p=0.52) and interval between hepatectomy and LDLT (p=0.82) did not affect DFS after salvage LDLT. Milan criteria and ADV score were independently prognostic of DFS and OS following salvage LDLT, and prognosis of patients within and beyond Milan criteria could be further stratified by ADV score. Authors concluded that risk factors and posttransplant outcomes were similar in patients undergoing salvage and primary LDLT. ADV score is surrogate biomarker for posttransplant prognosis in salvage and primary LDLT recipients. Prognostic model incorporating ADV scores can help determine whether to perform salvage LDLT.

Quantitative Prediction of Posttransplant Hepatocellular Carcinoma Prognosis Using ADV Score: Validation with Korea-Nationwide Transplantation Registry Database [13]

The aim of this study is to validate the prognostic impact of ADV score for predicting prognosis of HCC following LT. ADV score has been reported as a prognostic surrogate biomarker of HCC following LT and hepatectomy. The study patients were 1599 LT recipients selected from the Korean Organ Transplantation Registry database. Deceased-donor and living-donor LTs were performed in 143 and 1456 cases, respectively. Weak correlation was present among AFP, DCP, and TV. The viable HCC group showed ADV score-dependent DFS and OS rates from 1log to 10log (p<0.001). Prognosis of complete pathological response group was comparable to that of ADV score <1log (p≥0.099). ADV score cutoff of 5log (ADV-5log) for DFS and OS was obtained through receiver operating characteristic curve analysis with area under the curve ≥0.705. Both ADV-5log and Milan criteria were independent risk factors for DFS and OS, and their prognostic impacts were comparable to each other. Combination of these two factors resulted in further prognostic stratification, showing hazard ratios for DFS and OS as 2.98 and 2.26 respectively for one risk factor and 7.92 and 8.19 respectively for two risk factors (p<0.001). ABO-incompatible recipients with ADV score ≥ 8log or two risk factors showed higher recurrence rates. Authors concluded that this validation study revealed that ADV score is a reliable surrogate biomarker for posttransplant HCC prognosis, which can be used for selecting LT candidates and guiding risk-based posttransplant follow-up surveillance.

Prognostic Impact of Serum Soluble PD-1 and ADV Score for Living Donor Liver Transplantation in Patients with Previously Untreated Hepatocellular Carcinoma [14]

The programmed death protein 1 (PD-1) pathway is the critical mechanism in development of HCC. The present study analyzed the prognostic impact of pretransplant serum soluble PD-1 (sPD-1) concentration and ADV score in patients with previously untreated HCC undergone LT. This retrospective single-center study enrolled 100 patients with HCC who underwent living donor LT from 2010 to 2016. Concentrations of sPD-1 were measured in stored serum samples. Receiver operating characteristic curve analysis of 2-year tumor recurrence resulted in an sPD-1 cutoff of 177.1 µg/mL, which was associated with higher rates of tumor recurrence (p=0.022), but not with overall patient survival (p=0.460). The derived cutoff for pretransplant ADV score was 5.4log, which was associated with higher tumor recurrence rate (p<0.001) and lower overall patient survival rate (p<0.001). Both sPD-1 of >177.1 µg/mL (hazard ratio=2.26, p=0.020) and pretransplant ADV score of >5.4log (hazard ratio=3.56; p<0.001) were independent risk factors for posttransplant HCC recurrence. The combination of these 2 factors enabled the stratification of patients into 3 groups, with groups having 0, 1, and 2 risk factors differing significantly in the prognosis of tumor recurrence (p<0.001) and overall patient survival (p=0.006). Authors concluded that both sPD-1 concentration and ADV score have prognostic impacts in patients who underwent LT for untreated HCCs. These factors, both individually and combined, can help in predicting posttransplant prognosis.

Accurately predicting the post-resection prognosis of HCC patients poses a challenge due to the heterogeneity of HCC tumor biology, the complexity of tumor load, and variations in background liver diseases [1,2,6,12]. The ADV score was initially devised to amalgamate tumor size and the tumor markers AFP and DCP, based on a single-center, high-volume cohort of treatment-naïve patients, each with a single HCC [2]. Considering the intricate nature of HCC tumor biology, tumor size and number can be viewed as quantitative components, while the expression of tumor markers represents a qualitative component. The integration of these components becomes imperative in the development of dependable surrogate biomarkers that accurately reflect the aggressiveness of the tumor [2-8,12].

The ADV score can be significantly influenced by preoperative locoregional treatments. Even in cases where there is no complete radiological or pathological response, some viable tumor portions may persist, allowing for the calculation of ADV scores both before and after HR. In HCC patients who have undergone transarterial chemoembolization prior to HR, the estimation of TV can be achieved by measuring the volumes of contrast-enhancing tumor regions, while lipidolized zones are classified as non-enhancing lesions according to the modified Response Evaluation Criteria in Solid Tumors (mRECIST) criteria [15]. This approach helps offset the therapeutic effects of locoregional treatments. Remarkably strong correlations have been observed in the sizes of viable tumors between preoperative imaging results and the pathology findings from explant or resection [16-18]. Consequently, preoperative ADV scores based on imaging analyses and postoperative ADV scores derived from pathological findings yield comparable results, both leading to similar prognostic outcomes [3,6].

The attainment of complete radiological or pathological response following preoperative locoregional treatment has been documented to substantially diminish the risk of HCC recurrence subsequent to HR. In a single-center study encompassing 110 patients who achieved complete pathological response following HR, their outcomes were comparable to those of patients with ADV scores below 1.0log [19], implying that the risk of HCC recurrence in individuals with complete pathological response could be arbitrarily represented as an ADV score of 0.7log.

While MVI is recognized as a significant prognostic factor in HCC patients, its preoperative determination poses a challenge due to its reliance on pathological findings. Numerous studies have made efforts to predict MVI using various imaging modalities and/or tumor markers [20-23]. An extensive international study with a substantial dataset revealed that receiver operating characteristic curve analysis for MVI achieved an area under the curve of 0.708, with an ADV score cutoff of 4.2log demonstrating a sensitivity of 68.6% and specificity of 62.0%. Survival analysis indicated that both MVI and an ADV score exceeding 4.2log resulted in comparable DFS and OS outcomes. These findings suggest the practicality of employing a preoperative diagnostic approach utilizing an ADV score cutoff of 4.2log for predicting the presence of MVI [10].

The PPM based on ADV scores has proven effective in quantitatively forecasting the posttransplant prognosis [3]. Traditional selection criteria, which primarily consider tumor size and number, typically involve binary concepts with cutoffs determined through statistical analysis. Patients meeting the Milan and University of California San Francisco criteria generally exhibit favorable outcomes. Nevertheless, subgroup analyses employing ADV scores have revealed that these seemingly homogenous cases can be further stratified into subgroups with varying prognostic outcomes. Conducting subgroup analysis using ADV scores within pre-established selection criteria assists in identifying patients situated in the ambiguous range of various LT selection criteria [3].

The PPM utilizing the ADV score was employed to compare the predictive capabilities of eight pre-existing selection criteria systems for LT in HCC patients, with the aim of assessing the inclusivity and predictive accuracy across a spectrum of patients [3]. Among these eight systems, the model using the ADV score with a cutoff of 5log, in conjunction with the Metroticket 2.0 criteria [5], demonstrated superior performance in terms of predicting tumor recurrence. The prognostic predictive abilities of the model employing the ADV score and the Metroticket 2.0 criteria appeared to be comparable, but the former provided more intricate prognostic insights, particularly for patients with high DCP expression or those with multiple tumors of 10 or more. Additionally, the ADV score-based prediction model features flexible cutoff values, ensuring that detailed prognostic information can be obtained for patients across a broad range of ADV scores. These attributes of the ADV score-based prediction model seem to be beneficial in selecting patients with advanced HCC who do not meet the criteria of existing selection criteria.

The culmination of evidence from these twelve investigations, spanning single-center, multi-center, and international validation studies, affirms the ADV score's role as a comprehensive surrogate biomarker for predicting the prognosis following resection or transplantation in patients diagnosed with HCC. The utilization of the ADV score for prognostic predictions offers a dependable means to inform the management of patients across varying stages of HCC, facilitating the tailoring of individualized postoperative follow-up plans based on the assessed risk of HCC recurrence.

There was no funding related to this study.

All authors have no conflicts of interest to declare.

Conceptualization: SH. Data curation: SH, DHJ. Formal analysis: DHJ, GWS. Investigation: DHJ, GWS. Methodology: All. Project administration: SH. Validation: SH. Writing – original draft: All. Writing – review & editing: All.