Preoperative prediction score of hepatocellular carcinoma recurrence in living donor liver transplantation: Validation of SNAPP score developed at Asan Medical Center

Abstract

The previously proposed scoring systems are not readily available because of the lack of simplicity for predicting hepatocellular carcinoma (HCC) recurrence. We aimed to develop and validate the new score system, which can predict HCC recurrence after living donor liver transplantation (LDLT) by using morphologic and biologic data. Predictors for HCC recurrence after LDLT were developed (n = 627) and validated (n = 806) in 1433 patients for whom we could collect information to date between 2007 and 2016 at Asan Medical center (AMC) to create the SNAPP score (tumor Size and Number, alpha-fetoprotein [AFP], vitamin K absence-II [PIVKA-II], positron emission tomography [PET]). On logistic regression based on 3-year recurrence-free survival, the SNAPP factors were independently associated with HCC recurrence. The SNAPP score was highly predictive of HCC recurrence (C statistic, 0.920), and 5-year post-LT recurrence rates were significantly different between low, intermediate, and high SNAPP score groups. The performance of the SNAPP score (C-index [95% confidence interval], 0.840 [0.801-0.876]) on predicting tumor recurrence after LDLT was better than that of the New York/California, the Risk Estimation of Tumor Recurrence After Transplant (RETREAT), and the Model of Recurrence After Liver Transplant (MoRAL) score. The SNAPP score provides excellent prognostication after LDLT for HCC patients. Hence, we can help voluntary patients’ decisions about whether to undergo LDLT or not. The previously proposed scoring systems are not readily available because of the lack of simplicity for predicting hepatocellular carcinoma (HCC) recurrence. We aimed to develop and validate the new score system, which can predict HCC recurrence after living donor liver transplantation (LDLT) by using morphologic and biologic data. Predictors for HCC recurrence after LDLT were developed (n = 627) and validated (n = 806) in 1433 patients for whom we could collect information to date between 2007 and 2016 at Asan Medical center (AMC) to create the SNAPP score (tumor Size and Number, alpha-fetoprotein [AFP], vitamin K absence-II [PIVKA-II], positron emission tomography [PET]). On logistic regression based on 3-year recurrence-free survival, the SNAPP factors were independently associated with HCC recurrence. The SNAPP score was highly predictive of HCC recurrence (C statistic, 0.920), and 5-year post-LT recurrence rates were significantly different between low, intermediate, and high SNAPP score groups. The performance of the SNAPP score (C-index [95% confidence interval], 0.840 [0.801-0.876]) on predicting tumor recurrence after LDLT was better than that of the New York/California, the Risk Estimation of Tumor Recurrence After Transplant (RETREAT), and the Model of Recurrence After Liver Transplant (MoRAL) score. The SNAPP score provides excellent prognostication after LDLT for HCC patients. Hence, we can help voluntary patients’ decisions about whether to undergo LDLT or not. In Asia, including South Korea, Japan, Taiwan, Hong Kong, India, and Turkey, living donor liver transplantation (LDLT) has already been established treatment in patients with hepatocellular carcinoma (HCC) satisfying the Milan criteria (MC, 1 lesion of ≤5 cm, 2-3 lesions of ≤3 cm) under the shortage of available deceased donor whole liver donation.1Lee S-G Hwang S Moon D-B et al.Expanded indication criteria of living donor liver transplantation for hepatocellular carcinoma at one large-volume center.Liver Transpl. 2008; 14: 935-945Crossref PubMed Scopus (0) Google Scholar, 2Park GC Song GW Moon DB Lee SG. A review of current status of living donor liver transplantation.Hepatobiliary Surg Nutr. 2016; 5: 107-117PubMed Google Scholar, 3Yoon YI Lee SG. Living donor liver transplantation for hepatocellular carcinoma: an Asian perspective.Dig Dis Sci. 2019; 64: 993-1000Crossref PubMed Scopus (16) Google Scholar, 4Yoon Y-I Song G-W Lee S-G et al.Outcome of ABO-incompatible adult living-donor liver transplantation for patients with hepatocellular carcinoma.J Hepatol. 2018; 68: 1153-1162Abstract Full Text Full Text PDF PubMed Scopus (23) Google Scholar However, the MC has drawbacks that are too conservative, particularly for LDLT, and may exclude some patients having favorable tumor biology among HCC patients beyond the MC.5Halazun KJ Tabrizian P Najjar M et al.Is it time to abandon the Milan criteria? Results of a bicoastal US collaboration to redefine hepatocellular carcinoma liver transplantation selection policies.Ann Surg. 2018; 268: 690-699Crossref PubMed Scopus (0) Google Scholar As a result, many groups offer other criteria and scoring systems that complement the MC. Until now, there has been a lack of a reliable model that can predict HCC recurrence after LDLT.6Lee J-H Cho Y Kim HY et al.Serum tumor markers provide refined prognostication in selecting liver transplantation candidate for hepatocellular carcinoma patients beyond the Milan criteria.Ann Surg. 2016; 263: 842-850Crossref PubMed Scopus (64) Google Scholar, 7Sudan D Chapman WC Cameron JL Agopian V. A novel prognostic nomogram accurately predicts hepatocellular carcinoma recurrence after liver transplantation: analysis of 865 consecutive liver transplant recipients discussion.J Am Coll Surg. 2015; 220: 427-429PubMed Google Scholar, 8Mehta N Heimbach J Harnois DM et al.Validation of a risk estimation of tumor recurrence after transplant (RETREAT) score for hepatocellular carcinoma recurrence after liver transplant.JAMA Oncol. 2017; 3: 493-500Crossref PubMed Scopus (171) Google Scholar, 9Bonadio I Colle I Geerts A et al.Liver transplantation for hepatocellular carcinoma comparing the Milan, UCSF, and Asan criteria: long-term follow-up of a Western single institutional experience.Clin Transplant. 2015; 29: 425-433Crossref PubMed Scopus (19) Google Scholar To create a scoring system that surpasses the ability of the MC to predict HCC recurrence after LDLT, the pathologic outcome, microvascular invasion, should be included in the scoring system as critical factors.10Mazzaferro V Llovet JM Miceli R et al.Predicting survival after liver transplantation in patients with hepatocellular carcinoma beyond the Milan criteria: a retrospective, exploratory analysis.Lancet Oncol. 2009; 10: 35-43Abstract Full Text Full Text PDF PubMed Scopus (1398) Google Scholar,11Kirchner VA Song G-W Mongin S et al.East vs west: predictors of oncologic outcomes following DD and LD liver transplantation for HCC.Transplantation. 2018; 102: S253-S254Crossref Google Scholar The problem is how we can identify whether the microvascular invasion presents or not before the operation. Recently, many study groups have reported the selection criteria of LDLT in HCC patients by using alpha-fetoprotein (AFP), protein induced by vitamin K absence or antagonist-II (PIVKA-II) and positron emission tomography/computed tomography using 18F-fluorodeoxyglucose (FDG PET/CT).12Yaprak O Acar S Ertugrul G et al.Role of pre-transplant 18F-FDG PET/CT in predicting hepatocellular carcinoma recurrence after liver transplantation.World J Gastrointest Oncol. 2018; 10: 336-343Crossref PubMed Google Scholar,13Hong G Suh KS Suh SW et al.Alpha-fetoprotein and F-18-FDG positron emission tomography predict tumor recurrence better than Milan criteria in living donor liver transplantation.J Hepatol. 2016; 64: 852-859Abstract Full Text Full Text PDF PubMed Google Scholar Because they correlate well with the microvascular invasion on explant pathology, therefore, we can derive the closest the outcome to the pathology before LDLT using an elevated level of tumor markers and positive results of FDG PET/CT altogether. However, the concurrent application of those 3 biologic parameters and their significance as a prognostic factor has not been investigated in the prediction of tumor recurrence in LDLT for HCC. In this study, we tried to find out the most effective method to predict HCC recurrence after LDLT using tumor markers with FDG PET-CT together. Subsequently, we aimed to develop and validate a recurrence risk score system for HCC patients undergoing LDLT, a precise and straightforward score system designed at Asan Medical Center, based on preoperative imaging (CT/magnetic resonance imaging [MRI] and FDG PET/CT) and tumor markers at the time of LDLT. The institutional review boards approved this study of Asan Medical Center (AMC; 2014-0831). The study included patients (age ≥ 18 years) with HCC from January 2007 to December 2016. Patients with intrahepatic cholangiocarcinoma or mixed HCC–cholangiocarcinoma on explants were excluded. During the study period, a total of 1581 patients received LDLT for hepatocellular carcinoma, and 1451 had all of the examination results required for the study. Eighteen patients were lost to follow-up. The development and validation group consisted of 627 (2007-2011) and 806 (2012-2016) patients, respectively, who underwent LDLT at AMC. The variables included age, sex, size, number of HCC, and tumor markers (AFP and PIVKA-II) at the time of operation, the cause of the liver disease, Model for End-Stage-Liver Disease (MELD) score. We reviewed explant pathology to evaluate pathologic tumor staging and also to know the presence of a microscopic vascular invasion. Pathologic tumor staging on the explanted liver was based on the size and number of tumors with viability. The following categories of tumor stages were evaluated: within vs beyond University of California San Francisco (UCSF) and AMC criteria. Maintenance immunosuppression was based mainly on a triple regimen of tacrolimus, mycophenolate mofetil, and prednisone. All patients underwent contrast-enhanced CT or MRI and FDG PET/CT within 1 month before LDLT. To evaluate FDG PET/CT uptake, we plot circular regions of interest (ROIs) for the HCC and the normal liver and measured standardized uptake value (SUV) in each ROI. The maximum SUV (SUVmax) was measured in each ROI. For identifying FDG PET/CT positivity, we used tumor SUV-max to normal-liver SUVmax (TSUVmax/LSUVmax) values. From the previous study, a TSUVmax/LSUVmax of 1.15 was used as the most effective prognostic factor on FDG PET in the prediction of tumor recurrence.14Lee JW Paeng JC Kang KW et al.Prediction of tumor recurrence by 18F-FDG PET in liver transplantation for hepatocellular carcinoma.J Nucl Med. 2009; 50: 682-687Crossref PubMed Scopus (132) Google Scholar We grouped tumor size and number in several different ways, including existing criteria, and then analyzed the ability to predict the recurrence of HCC in each group using the receiver operating characteristics (ROC) curve. Among the groups, we choose the one that can predict HCC recurrence after LDLT most precisely. As reported in our previous study, we defined the optimal cutoff value of tumor markers as the point showing the highest C-index among values with specificity ≥ 0.85. The C-index was calculated using the following formula: C – index = (sensitivity + specificity)/2. For the criteria of LDLT for HCC, high specificity is necessary to avoid excluding a large number of patients who would not develop recurrence. The multivariate Cox regression analyses revealed that the combination of tumor markers (AFP and PIVKA-II) were independently significant risk factors for recurrences.15Kim SH Moon DB Kim WJ et al.Preoperative prognostic values of alpha-fetoprotein (AFP) and protein induced by vitamin K absence or antagonist-II (PIVKA-II) in patients with hepatocellular carcinoma for living donor liver transplantation.Hepatobiliary Surg Nutr. 2016; 5: 461-469Crossref PubMed Google Scholar The probabilities of recurrence rate were estimated by the Kaplan-Meier method and compared using the log-rank test. Univariate and multivariable hazard ratios (HRs) for risk factors of HCC recurrence after LDLT were determined by logistic regression based on 3-year recurrence-free survival (RFS) and reported with 95% confidence intervals (CIs).16Sullivan LM Massaro JM D’Agostmo RB. Presentation of multivariate data for clinical use: the Framingham Study risk score functions.Stat Med. 2004; 23: 1631-1660Crossref PubMed Scopus (1087) Google Scholar Risk factors of HCC recurrence with a univariate P < .05 were included in the multivariable analysis, and the final predictors were selected by backward stepwise elimination (P > .05 for removal). The SNAPP (tumor Size and Number, alpha-fetoprotein [AFP], vitamin K absence-II [PIVKA-II], positron emission tomography [PET]) score was then created based on the final multivariable risk factors score. A simple integer scoring system based on the HRs of the independent predictors of HCC recurrence. Each score produced a simplified point scale reflecting the relative impact of model covariables. The integer value for each component was then summed to calculate the SNAPP score. Patients have their risk stratified according to their possibility of recurrence based on the SNAPP score, where patients in the low-risk group had a score of 0-2, the acceptable risk group had a score of 3-4, and the high-risk group had a score of >5 (Table 1).TABLE 1Recurrence risk stratification according to SNAPP scoreSNAPP scoreRisk of recurrenceRecurrence free survival1 y3 y5 yDevelopmentValidationDevelopmentValidationDevelopmentValidation0-2Low98%96%97%94%97%94%3-4Acceptable82%89%71%80%71%77%5 or moreHigh46%55%35%45%31%25%Abbreviations: PET, positron emission tomography; SNAPP, tumor Size and Number, alpha-fetoprotein (AFP), vitamin K absence-II (PIVKA-II). Open table in a new tab Abbreviations: PET, positron emission tomography; SNAPP, tumor Size and Number, alpha-fetoprotein (AFP), vitamin K absence-II (PIVKA-II). The clinical characteristics of the 627 patients in the development group and the 806 patients in the validation group are summarized in Table 2. The mean size and number of HCCs detected on the preoperative CT scan were 2.84 cm (SD, 1.33) and 1.96 nodules (SD, 1.08) in the development group, respectively. Hepatitis B is the main cause of liver disease in development and validation groups (90.0% [n = 564] vs 83.0% [n = 669]). Compared with the developmental group, there were no significantly different preoperative variables of maximal tumor size, number of tumors, the median level of tumor markers, and positivity in FDG PET/CT in the validation group. On explant pathology after LDLT, validation group had no significant difference in tumor stage beyond each criterion (UCSF [15.9% vs 18.7%], Asan [8.6% vs 9.3%]) and microvascular invasion (7.8% [n = 49] vs 10.4% [n = 84]).TABLE 2Clinical characteristics of the patients in this studyClinical characteristicsGroupsP value (SMD)Development (n = 627)Validation (n = 806)Age, mean (SD), y53.02 (6.80)53.15 (6.28).718 (0.02)Male, sex (%)534 (85.2%)709 (88.0%).121 (0.08)DiagnosisHepatitis C (HCV)30 (4.8%)55 (6.8%)Hepatitis B (HBV)564 (90.0%)669 (83.0%)HBV + HCV19 (3.0%)75 (9.3%)Cryptogenic14 (2.2%)7 (0.9%)MELD score, mean (SD)12.60 (5.29)11.00 (4.11)Maximal tumor size (radiologic)≤3 cm423 (67.5%)636 (78.9%)<.001 (0.23)3-6 cm184 (29.3%)163 (20.2%)>6 cm20 (3.2%)7 (0.9%)Number of tumors (radiologic)1293 (46.4%)372 (46.2%).467 (0.04)2-3253 (40.4%)324 (40.2%)≥481 (12.9%)109 (13.6%)AFP, at operation, ng/mL14.5 (0.78-15 400)11.9 (0.72-42 200)PIVKA-II, at operation, mAU/mL29.0 (5.0-6970)25.0 (11.0-19 400)PET-CTIsometabolic476 (75.9%)608 (75.4%).833 (0.01)Hypermetabolic151 (24.1%)198 (24.6%)Milan criteria (radiologic)Within462 (73.7%)615 (76.3%).255 (0.06)Beyond165 (26.3%)191 (23.7%)UCSF criteria (pathologic)Within527 (84.1%)655 (81.3%).169 (0.07)Beyond100 (15.9%)151 (18.7%)Asan criteria (pathologic)Within573 (91.4%)722 (89.6%).649 (0.02)Beyond49 (8.6%)75 (9.3%)Follow-up, mo79 (1-109)43 (5-64)Microvascular invasion (pathologic)No578 (92.2%)722 (89.6%).092 (0.09)Yes49 (7.8%)84 (10.4%)Abbreviations: MELD, model for End-Stage-Liver Disease; PET-CT, positron emission tomography-computed tomography; SNAPP, tumor Size and Number, alpha-fetoprotein (AFP), vitamin K absence-II (PIVKA-II); SMD, standardized mean difference; UCSF, University of California San Francisco. Open table in a new tab Abbreviations: MELD, model for End-Stage-Liver Disease; PET-CT, positron emission tomography-computed tomography; SNAPP, tumor Size and Number, alpha-fetoprotein (AFP), vitamin K absence-II (PIVKA-II); SMD, standardized mean difference; UCSF, University of California San Francisco. Areas under the curve (AUCs) of tumor size for the MC (max tumor size 5 cm) and the development group (max tumor size 6 cm) were 0.728 and 0.741 (P = .000). Also, AUCs of tumor numbers for the MC and the development group were 0.649 and 0.730 (P = .000), respectively (Figure 1). The selected cutoff values were 100 maul/mol for PIVKA-II [(C-index, sensitivity, specificity) = (0.71, 0.55, 0.93)], 150 nag/mol for AFP (0.68, 0.46, 0.90). Using the combination of tumor markers has been shown as a powerful predictor for the expansion of preexisting strict selection criteria of liver transplantation for HCC in the previous study.14Lee JW Paeng JC Kang KW et al.Prediction of tumor recurrence by 18F-FDG PET in liver transplantation for hepatocellular carcinoma.J Nucl Med. 2009; 50: 682-687Crossref PubMed Scopus (132) Google Scholar Mean follow-up after LDLT was 60.7 months (interquartile range [IQR], 25-91 months) in the development group. Because each HCC patient has a different risk for HCC recurrence after LDLT, we employed a 3-category follow-up protocol according to the risk level: within MC (5-year recurrence risk of <20%) vs beyond MC and within Asan criteria (theoretical 5-year recurrence risk of 20%-80%; the actual risk is around 20%), vs beyond Asan criteria (5-year recurrence risk of >80%). Our usual protocol for within-Milan patients included assessment of blood tumor markers (every 1-3 months), routine screening imaging study including dynamic abdomen-pelvis CT scan (or ultrasonography in case of renal dysfunction), chest X-ray (every 3-6 months), and an additional chest CT scan (every 4-12 months) for the first 5 years. In principle, we suggested repeating blood tumor marker tests over 10 years with imaging study for 5 years.17Hwang S Moon D-B Ahn C-S et al.Risk-based long-term screening for hepatocellular carcinoma recurrence after living donor liver transplantation.Transplant Proc. 2013; 45: 3076-3084Crossref PubMed Scopus (0) Google Scholar For patients who are at high risk of recurrence, chest CT and bone scans were also performed (explant tumor stage beyond Milan criteria, or microvascular invasion). Overall HCC recurrence after LDLT at 1, 3, and 5 years was 10.2%, 14.7%, and 18.2% (95% CI, 7.1%-8.9%), respectively (Figure 2). The age, sex, MELD score, cause of the liver disease, and previous treatment for HCC were not predictive of HCC recurrence on univariate analysis. Predictors of HCC recurrence after LDLT in the development group on the multivariate analysis included (1) maximal tumor size, (2) total tumor number, (3) the groups of AFP and PIVKA-II at the time of LDLT as a categorical variable according to cutoffs, and (4) the positivity of FDG PET/CT (Table 3). The HRs of the factors derived from the logistic regression based on 3-year RFS. An individual patient’s SNAPP score is calculated by adding the individual points for each of the 4 variables. The actual statistical model and the coefficients are also summarized in Table 3.TABLE 3Outcomes of univariate and multivariate Cox regression analysis of predictors of HCC recurrence and creation of the SNAPP scorePredictorsUnivariate analysisMultivariate analysisSNAPP score pointsBeta-coefficientPHazard ratio (CI)PHazard ratio (CI)Maximum tumor size, cm0-3NA(Reference)NA(Reference)0NA>3-6<.0015.99 (3.87-9.28).0012.89 (1.51-5.53)11.06>6<.00111.38 (4.46-29.09).0035.787 (1.509-22.183)21.86Total tumor number, n1NA(Reference)NA(Reference)0NA2-3.091.56 (0.93-2.62).0033.06 (1.47-6.36)11.12≥4<.00122.95 (12.39-42.53)<.00137.84 (15.13-94.67)23.63Tumor marker groupAFP ≤ 150 and PIVKA ≤ 100NA(Reference)NA(Reference)0NAAFP ≤ 150 and PIVKA > 100<.0017.99 (4.12-15.52)<.00111.51 (4.80-27.58)12.44AFP > 150 and PIVKA ≤ 100<.00116.41 (9.28-29.02)<.00115.41 (7.40-32.05)22.73AFP > 150 and PIVKA > 100<.00171.79 (28.61-180.14)<.00159.03 (18.50-188.39)34.08PET-CTIsometabolicNA(Reference)NA(Reference)0NAHypermetabolic<.0016.14 (4.02-9.40)<.0014.01 (2.12-7.58)11.39 Open table in a new tab Calculated SNAPP scores ranged from 0 to 7, with the most common scores being 0 (24.2% [n = 152]) and 1 (25.4% [n = 159]). No patient with a SNAPP score of 0 showed recurrence after LDLT. Predicted risk of 1-, 3-, and 5-year HCC recurrence rose with each point scored (Figures 3 and 4), and patients with SNAPP score of 5 or higher (17.7% of the group [n = 111]) had predicted 1-, 3-, and 5-year recurrence risk of 54.0%, 65%, and 69%, respectively. The SNAPP score C statistic was 0.920 in the development group for predicting HCC recurrence compared with 0.675 (95% CI, 0.56-0.79) for Milan criteria by a preoperative imaging study. The C statistics for the UCSF and AMC criteria by explant pathology were 0.632 (95% CI, 0.510-0.754) and 0.637 (95% CI, 0.513-0.762).FIGURE 4Kaplan-Meier probability of hepatocellular carcinoma (HCC) recurrence within 5 y for living donor liver transplant recipients in the development and validation group [Color figure can be viewed at wileyonlinelibrary.com]View Large Image Figure ViewerDownload Hi-res image Download (PPT) In the validation group, the mean follow-up after LDLT was 40.1 months (IQR, 31-54 months). Relatively lower HCC recurrence rates at 1-, 3-, and 5-years after LDLT were observed in the validation group 10.4%, 11.3%, and 13.3% (95% CI, 1.3%-18.0%), respectively, but it was not significant (P = .407) (Figure 2). The SNAPP score performed well in the validation group, with a C statistic of 0.840 (95% CI, 0.801-0.876). As shown in Table 4, the performance of the SNAPP score (C-index [95% CI], 0.840 [0.801-0.945]) on predicting tumor recurrence after LDLT was better than that of New York/California (NYCA; C-index [95% CI], 0.724 [0.682-0.866]) and Model of Recurrence After Liver Transplant (MoRAL) (C-index [95% CI], 0.684 [0.587-0.843]) score. The SNAPP score showed similar c-index with that of the Risk Estimation of Tumor Recurrence After Transplant (RETREAT) score (C-index [95% CI], 0.836 [0.776-0.920]).TABLE 4Comparison between the SNAPP score and other liver transplantation scoring systems for predicting tumor recurrenceDevelopment groupValidation groupPrediction modelRecurrence rate (%)C-index (95% CI)Recurrence rate (%)C-index (95% CI)1 y3 y5 y1 y3 y5 ySNAPP scoreLow2.03.03.00.920 (0.901-0.945)4.06.06.00.840 (0.801-0.876)Moderate16.029.029.011.020.023.0High54.065.069.045.055.075.0New York/California (NYCA) scoreLow4.05.07.00.745 (0.682-0.814)4.05.06.00.724 (0.682-0.866)Acceptable9.012.015.09.013.014.0High46.058.063.046.055.056.0Risk Estimation of Tumor Recurrence After Transplant (RETREAT) score00000.831 (0.781-0.918)0000.836 (0.776-0.920)12.03.06.000027.011.014.06.06.06.0312.020.034.016.022.023.0414.031.040.014.029.029.05 or more34.050.067.034.045.056.0Model of Recurrence After Liver Transplant (MoRAL) score≤314.88.012.015.00.63 (0.592-0.689)6.07.08.00.684 (0.587-0.843)>314.838.050.063.043.057.071.0 Open table in a new tab The SNAPP score correlates well with microvascular invasion on the explant pathology. Almost half the patients (50.0%) in the high-risk group had microvascular invasion compared with 14.3% of the low-risk group patients and 35.7% of acceptable risk group patients (P = .000). Despite recent advances in radiological assessment of liver masses, the risk of underestimating tumor burden remains a significant concern.18Sherman M. The radiological diagnosis of hepatocellular carcinoma.Am J Gastroenterol. 2010; 105: 610-612Crossref PubMed Scopus (38) Google Scholar Even the MC based on preoperative tumor morphologic factors is not accurate to predict HCC recurrence, which may be related to a lack of information about tumor biology.19Costentin CE Bababekov YJ Zhu AX Yeh H. Is it time to reconsider the milan criteria for selecting patients with hepatocellular carcinoma for deceased-donor liver transplantation?.Hepatology. 2019; 69: 1324-1336Crossref PubMed Scopus (25) Google Scholar As a result, the risk rate of recurrence of HCC during 5 years after LDLT is estimated to be 10%-15%, even though the patients satisfying MC have a low risk of HCC recurrence.20Yao FY Mehta N Flemming J et al.Downstaging of hepatocellular cancer before liver transplant: long-term outcome compared to tumors within Milan criteria.Hepatology. 2015; 61: 1968-1977Crossref PubMed Scopus (270) Google Scholar,21Levi DM Tzakis AG Martin P et al.Liver transplantation for hepatocellular carcinoma in the model for end-stage liver disease era.J Am Coll Surg. 2010; 210: 727-734Crossref PubMed Scopus (0) Google Scholar Currently, there are more needs for a refined prognostication system that can more accurately identify explant pathology outcomes and predict HCC recurrence after LDLT thorough preoperative assessment in HCC patients.5Halazun KJ Tabrizian P Najjar M et al.Is it time to abandon the Milan criteria? Results of a bicoastal US collaboration to redefine hepatocellular carcinoma liver transplantation selection policies.Ann Surg. 2018; 268: 690-699Crossref PubMed Scopus (0) Google Scholar,6Lee J-H Cho Y Kim HY et al.Serum tumor markers provide refined prognostication in selecting liver transplantation candidate for hepatocellular carcinoma patients beyond the Milan criteria.Ann Surg. 2016; 263: 842-850Crossref PubMed Scopus (64) Google Scholar,8Mehta N Heimbach J Harnois DM et al.Validation of a risk estimation of tumor recurrence after transplant (RETREAT) score for hepatocellular carcinoma recurrence after liver transplant.JAMA Oncol. 2017; 3: 493-500Crossref PubMed Scopus (171) Google Scholar,10Mazzaferro V Llovet JM Miceli R et al.Predicting survival after liver transplantation in patients with hepatocellular carcinoma beyond the Milan criteria: a retrospective, exploratory analysis.Lancet Oncol. 2009; 10: 35-43Abstract Full Text Full Text PDF PubMed Scopus (1398) Google Scholar Of the various indicators known to date for identifying microvascular invasion of HCC before surgery, serum tumor marker levels are objective and reproducible.22Toso C Asthana S Bigam DL et al.Reassessing selection criteria prior to liver transplantation for hepatocellular carcinoma utilizing the scientific registry of transplant recipients database.Hepatology. 2009; 49: 832-838Crossref PubMed Scopus (272) Google Scholar, 23Merani S Majno P Kneteman NM et al.The impact of waiting list alpha-fetoprotein changes on the outcome of liver transplant for hepatocellular carcinoma.J Hepatol. 2011; 55: 814-819Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar, 24Kim HS Park JW Jang JS et al.Prognostic values of alpha-fetoprotein and protein induced by vitamin K absence or antagonist-II in hepatitis B virus-related hepatocellular carcinoma a prospective study.J Clin Gastroenterol. 2009; 43: 482-488Crossref PubMed Scopus (0) Google Scholar In our previous study, the combination of tumor markers might be used for the expansion of preexisting strict selection criteria of liver transplantation for HCC.15Kim SH Moon DB Kim WJ et al.Preoperative prognostic values of alpha-fetoprotein (AFP) and protein induced by vitamin K absence or antagonist-II (PIVKA-II) in patients with hepatocellular carcinoma for living donor liver transplantation.Hepatobiliary Surg Nutr. 2016; 5: 461-469Crossref PubMed Google Scholar For the first time, to our knowledge, we performed this study to more accurately predict the results of pathologic examinations after LDLT by combining tumor markers and FDG PET-CT results. In this large study reviewing over 1000 HCC patients after LDLT, we developed and validated the SNAPP score using 4 variables that were highly predictive of HCC recurrence: tumor markers, positivity in FDG PET/CT, the maximum tumor diameter, and total tumor number on the pretransplant workup held within 1 month before the operation. Though this study was not a multicenter study but a single-center study, the most significant merit of our research is that the uniformly achieved excellent surgical outcome at AMC during the study period, despite technically complicated LDLT procedures, has minimized surgery-related confounding factors on the posttransplant survival outcome. The SNAPP score showed good discrimination power on selecting optimal candidates for LDLT among HCC patients, and also it was able to stratify 5-year HCC recurrence risks ranging from less than 6% in those with a low risk (SNAPP score; 0-2) to higher than 75% with a high risk (SNAPP score; 5 or higher). As shown in Table 4, the performance of the SNAPP score on predicting tumor recurrence after LDLT was significantly better than those of the scoring system, including NYCA and MoRAL scores. The SNAPP score potentially affects clinical practice on the selection of HCC patients undergoing LDLT. There has been a greater need for LDLT because of a shortage of deceased-donor organs, especially in Asian countries.25Chen CL. Living-donor liver transplantation: the Asian perspective.Transplantation. 2003; 75: S1Crossref PubMed Scopus (1) Google Scholar Furthermore, the benefit of the recipient after LDLT currently outweighs the potential risk for major hepatectomy of the donor with the help of innovation of surgical technique.26Lee SG. A complete treatment of adult living donor liver transplantation: a review of surgical technique and current challenges to expand indication of patients.Am J Transplant. 2015; 15: 17-38Crossref PubMed Scopus (95) Google Scholar As a result, the selection criteria for the HCC recipients can be rationally extended, and the prediction of prognosis is also becoming increasingly important. HCC patients belonging to the “acceptable risk group (SNAPP score 3-4 points)” had a 77% 5-year RFS rate after LDLT. More than 70% of the 5-year RFS rate in LDLT for HCC patients is a satisfactory outcome when compared to other studies.27Clavien P-A Lesurtel M Bossuyt PMM et al.Recommendations for liver transplantation for hepatocellular carcinoma: an International Consensus Conference Report.Lancet Oncol. 2012; 13: E11-E22Abstract Full Text Full Text PDF PubMed Scopus (733) Google Scholar, 28Pomfret EA Washburn K Wald C et al.Report of a National Conference on liver allocation in patients with hepatocellular carcinoma in the United States.Liver Transpl. 2010; 16: 262-278Crossref PubMed Scopus (325) Google Scholar, 29Ioannou GN Perkins JD Carithers RL. Liver transplantation for hepatocellular carcinoma: impact of the MELD allocation system and predictors of survival.Gastroenterology. 2008; 134: 1342-1351Abstract Full Text Full Text PDF PubMed Scopus (214) Google Scholar In our study, HCC patients beyond the MC were 26.2% (n = 164) in the development group and 23.5% (n = 189) in the validation group, but the “high-risk group” of SNAPP score (5 or more) were 56% (n = 91/164) in the development group and 65% (n = 122/189) in the validation group. Totally 60.3% (n = 213) HCC patients who might have been excluded according to the MC criteria during the study period obtained favorable results after LDLT (Table 5). In our experiences, application of the SNAPP score in LDLT for beyond Milan HCC patients definitely help us to identify the good candidates having acceptable future prognosis. Hence, SNAPP score could replace the MC at our institution.TABLE 5Classification of patients with hepatocellular carcinoma within/above the Milan criteria according to SNAPP scoreSNAPP scoreRisk of recurrenceMilan criteriaWithinAboveDevelopment (n = 463)Validation (n = 617)Development (n = 164)Validation (n = 189)0-2Low80%86%17%10%3-4Acceptable16%12%27%25%5 or moreHigh4%1%56%65%Abbreviations: SNAPP, tumor Size and Number, alpha-fetoprotein (AFP), vitamin K absence-II (PIVKA-II), positron emission tomography (PET). Open table in a new tab Abbreviations: SNAPP, tumor Size and Number, alpha-fetoprotein (AFP), vitamin K absence-II (PIVKA-II), positron emission tomography (PET). One of the most important advantages of the SNAPP score is that it is simple, unlike other scoring systems that use complex formulas. For example, MoRAL score is 11xPIVKA-II+2xAFP (median = 108.3, range: 33.7-3928.3).6Lee J-H Cho Y Kim HY et al.Serum tumor markers provide refined prognostication in selecting liver transplantation candidate for hepatocellular carcinoma patients beyond the Milan criteria.Ann Surg. 2016; 263: 842-850Crossref PubMed Scopus (64) Google Scholar They tried to use both tumor markers simultaneously to improve prediction accuracy, but in clinical practice environments, especially in the outpatient’s clinic, it is difficult to apply immediately. However, the SNAPP score based on the results of preoperative evaluation enables us to know intuitively whether the benefit can be obtained after LDLT or not by adding a simple integer score for each item. High correlation rate between SNAPP score and microvascular invasion, which is one of the most important predictors for the posttransplant recurrence, led us to perform specially designed LDLT procedures for the HCC patients with acceptable and high-risk SNAPP scores to minimize the tumor recurrence such as “Hilar dissection first without mobilizing the liver having HCC” and “No-touch en-bloc total hepatectomy with IVC replacement.”30Moon D-B Lee S-G Hwang S et al.No-touch en bloc right lobe living-donor liver transplantation with inferior vena cava replacement for hepatocellular carcinoma close to retrohepatic inferior vena cava: case report.Transplant Proc. 2013; 45: 3135-3139Crossref PubMed Scopus (0) Google Scholar The major limitation of this study is that we did not analyze the effect of previous locoregional treatment on HCC, because our institution is a tertiary medical center and most of the patients included in this study were transferred to us from the other hospitals without sufficient information of HCC treatment. Hence, combined analysis, including previous locoregional treatment, might rather complicate the scoring system, and we had to decide whether to perform LDLT or not to use the SNAPP score assessed at our institution just before the operation. In conclusion, the SNAPP score provides excellent prognostication after LDLT for HCC patients. Hence, we can help the voluntary decision of the patients whether to undergo LDLT or not and can also predict the pertinent post-LDLT HCC surveillance strategies according to the risk groups in advance. In addition, we can make a plan of specially designed LDLT procedures in HCC patients with acceptable and high-risk SNAPP scores for the purpose of minimizing the tumor recurrence.