Prediction Of Hepatocellular Carcinoma Recurrence After Liver Transplantation

Part 1: Preoperative Risk Stratification for Early Recurrence of HBV-related Hepatocellular Carcinoma After Deceased Donor Liver Transplantation: A Five-Eight Model Development and ValidationBackground: Hepatocellular carcinoma(HCC)is the sixth most common malignancy and the third leading cause of cancer-related deaths.In the Pacific region,China accounted for 84.6% of HCC incidence and 86.3% of HCC mortality with Hepatitis B virus(HBV)infection as the most common cause.Although liver transplantation(LT)is an excellent therapeutic choice for HCC as the patients who received LT have the highest chance of cure among all other therapies,the organs shortage is still a main challenge and the early recurrence of hepatocellular carcinoma(HCC)after liver transplantation(LT)is associated with poor surgical outcomes.The objective of this study is to construct a preoperative model to predict individual risk of post-LT HCC recurrence.Methods: The design of this study followed the Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis(TRIPOD)Statement.Inclusion criteria were(1)adult patients with age ? 18(2)pre-operative radiologically diagnosed HCC depending on guidelines of the current guidelines ofAmerican Association for the Study of Liver Diseases(AASLD)(3)no history of previous LT or combined hepatorenal transplantation(4)no evidence of vascular invasion on pre-operative imaging of the liver(mainly CT scan,and MRI)(5)patients who survived at least 3 months after the date of surgery(6)no incidental HCC(7)all the clinical and laboratory data required for the analysis are available.After applying the inclusion criteria,data of 748 adult patients who underwent deceased donor LT for HCC between January 2015,and February 2019 were collected retrospectively from the China Liver Transplant Registry database and randomly divided into training(n=486)and validation(n=262)cohorts.A multivariate analysis was performed,and the five-eight model was developed.Results: A total of 748 patients were included in the study;of them,96% had hepatitis B virus(HBV)and 84 % had cirrhosis.The 486 patients of the training cohort had similar characteristics to the 262 patients of validation cohort without any significant differences.Factors of post-LT HCC recurrence in the training cohort were identified by univariable Cox regression analysis included pre-LT TACE,pre-LT AFP,total tumor diameter(cm),the largest tumor diameter(cm)and the number of nodules at all tested cut-off values.In the multivariate analysis,pre-LT serum alpha-fetoprotein(AFP),tumor number and largest tumor diameter were incorporated to construct the 5-8 model which can stratify patients accurately according to their risk of recurrence into three prognostic subgroups;low-(0-5 points),medium-(6–8 points)and high-risk(>8 points)with 2-year post-LT recurrence rate of(5%,20% and 51%,p<0.001)respectively.The 5-8 model was better than Milan,Hangzhou,and AFP-model for prediction of HCC early recurrence.These findings were confirmed by the results of the validation cohort.Conclusions: The 5-8 model is a preoperative predictive risk score for early recurrence of HCC after LT,was constructed using 3 variables(pre-LT AFP,largest tumor diameter,and the number of nodules).Our model accurately predicts early recurrence of HCC at an individual level in patients with HBV-cirrhosis.It could effectively classify HCC patients into subgroups with a low,medium and high risk of recurrence and can potentially be used to guide therapeutic decisions and facilitate risk communication between surgeons and patients.We believe that our model can play acomplementary role in the selection of HCC patients for LT and prediction of early recurrence alongside the conventional selection criteria.Prospective validation of our model will be an important step to verify its clinical utility.The 5-8 model is a simple validated and accurate tool for preoperative stratification of early recurrence of HCC after LT.Part 2: Prediction of Early Recurrence of Hepatocellular Carcinoma in Patients with Cirrhosis Who Had Received Deceased Donor Liver Transplantation: A Multicenter StudyBackground:liver transplantation(LT)is one of the most effective therapeutic modalities for hepatocellular carcinoma(HCC)as it can cure cancer with its underlying causative disease.Even with the careful selection of HCC candidates,tumor recurrence after LT is still a clinical challenge with a rate of about10%-47%.Since the introduction of Milan criteria in 1996,the recurrence and survival of HCC candidates after LT were improved and it is still recommended as a benchmark for selection of HCC candidates for transplantation.However,Milan criteria was considered to be very restrictive as it depends only on the morphological but not on the biological assessment of the tumor.Thus,many novel criteria have been proposed such as Hangzhou criteria and AFP-model based criteria.This multicenter study aims to evaluate the Milan,Hangzhou and AFP model-based criteria for prediction of early recurrence of HCC in patients with cirrhosis who had undergone LT.Methods:The inclusion criteria for the study subjects were as the following:(1)cirrhotic patients who had undergone LT(2)adult patients ?18 years(3)HCC patients diagnosed preoperatively by radiological assessment depending on guidelines of the American Association for the Study of Liver Diseases(AASLD)(4)no vascular invasion on pre-operative imaging of the liver(mainly CT scan,and MRI)(5)no retransplantation or combined renal transplantation(6)all the essential laboratory and clinical data required for analysis are available.After applying inclusion criteria,we analyzed data of 589 HCC patients who had undergone LT between Jan.2015 and Jan.2019 from the China Liver Transplant Registry(CLTR)database.Imaging data and AFP levels were evaluated immediately before LT.Recurrence and overall survival rates at 2 years were tested using the Kaplan-Meier estimate.Milan,Hangzhou,and AFP model were evaluatedResults : All the patients who included in this study underwent LT for HCC treatment.Of the 589 patients,521 were male(88.5%)and 68 were females(11.5%).The mean age was 52.2± 8.7 years.All subjects of this cohort had cirrhosis and majority of them(n=554)presented with HBV(94.1%).Of the total patients 62.0 %,91.2% and 67.6% were within Milan,Hangzhou and AFP model,respectively.The 2-year recurrence rate was 8.9%,15.8% and 11.8% with corresponding overall survival of 85.3%,82.7% and 86.5%,respectively.The 2-year recurrence rate was different in patients fulfilling and exceeding AFP model among patients who met either Milan criteria(7.9% vs 18.8%,HR=3.83,p=0.006)or Hangzhou criteria(12.0% vs 27.6%,HR = 2.95,p<0.001).However,the 2-year recurrence rate was not statistically different among patients who are beyond either Milan or Hangzhou criteria.Conclusions : For the prediction of early recurrence of HCC in patients with cirrhosis after liver transplantation;Milan criteria,Hangzhou Criteria,and AFP model are effective predictive tools for stratification of patients into low and high-risk groups of recurrence with different prognosis.AFP model can identify a subgroup of patients with high-risk of recurrence among patients who met either Milan or Hangzhou criteria.Part3: Systematic Review: Risk Prediction Models for Recurrence of Hepatocellular Carcinoma After Liver TransplantationBackground : Recurrence of hepatocellular carcinoma(HCC)after liver transplantation(LT)is an important clinical concern associated with poor surgical outcomes.There is an ongoing shift toward evidence-based clinical practice using algorithmic approaches for diagnosis and treatment of diseases and the proliferation of the clinical scoring systems is a result of this trend.The predictive scoring systems are designed to help in clinical decision-making,to determine the best therapeutic option,to stratify patients into risk categories and to facilitate communication between patients and doctors about the prognosis.Numerous studies have attempted to outline the role of these various predictors and to develop predictive scores for HCC recurrence after LT.These predictive models have the potential applicability for individual-based stratification and can be used to develop therapeutic algorithms,guide treatment decisions and advice patients.The aim of this study is to review and analyse the current evidence on risk models for prediction of HCC recurrence after LT.This study aims to summarize the current evidence on risk prediction models of HCC recurrence after LT.Methods:This study was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses(PRISMA)checklist,Checklist for Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies(CHARMS)and Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis(TRIPOD)guidelines where applicable.There was no need for patient informed consent and ethical approval due to all data were extracted from formerly published articles.The systematic literature review was performed according to a priori established study protocol.Pub Med and EMBASE search engines were searched until 25 th May 2019.Studies that originally were designed for development or validation of a risk model for prediction the risk of HCC recurrence after LT were included.Results : From a total of 26 studies,18 original risk prediction models were determined and only 5 prediction models were externally validated.Majority of the original risk prediction models of HCC recurrence after LT originated from western cohorts where hepatitis C virus-related HCC(C-HCC)is a predominant cause(n=11),whereas only 7 risk prediction models originated in eastern country cohorts where hepatitis B virus-related HCC(B-HCC)is epidemic.The AFP model was the most validated prediction model(5 studies),followed by the model of Decan et al(4 studies),then model of Mehta et al(2 studies),and model of Iwatsuki et al(2 studies).Among the 26 included studies,25 were retrospective studies with only MORAL score developed by Halazun et al.based on a prospective cohort.Risk prediction models involved an average number of 3 factors.Most studies showed good discriminatory performance(AUC>0.75).The most frequently employed predictors were alphafetoprotein,tumor size,vascular invasion,tumor number,tumor differentiation,and neutrophil–lymphocyte ratio.Conclusions : In conclusion,we found 18 original risk prediction models intended to predict HCC recurrence in patients who had undergone LT.The quality of studies was generally low according to GRADE criteria.The most common included risk factors were AFP,tumor size,vascular invasion,tumor number,tumor differentiation,and NLR.Only 5 risk prediction models were externally validated.At least 8 prediction models showed perfect discriminatory performance in internal or external validation.Thus,external validation of the current risk prediction models in diverse populations is highly recommended.