The Preliminary Study Of Clinical-pathological Factors And Serum MicroRNA In The Prediction Of Postoperative Recurrence Of Hepatocellular Carcinoma

PART I:Clinical-Pathological Factors and Prediction Model of Postoperative Recurrence for Hepatocellular Carcinoma[Background and Objective]:Hepatic resection is still the best choice of therapeutic strategies for hepatocellular carcinoma (HCC) currently, but the short-term and long-term relapse free survival (RFS) rate after hepatectomy is unsatisfactory. There are four main factors affecting the prognosis of HCC:surgery operation, tumor characteristic, liver function and perioperative intervention. The technique of selective and dynamic regional hepatic vascular occlusion has been matured and prevalent in the recent years. It could control surgical related risk factors systemically and effectively to a certain degree. For HCC patients underwent hepatectomy with the aforementioned technique, it is necessary to investigate the clinical-pathological risk factors of recurrence, including tumor characteristic and liver function, quantify the criticality of the risk factors and establish a recurrence prediction model to direct the postoperative adjunctive therapy and follow up proposal.[Materials and Methods]:288consecutive HCC patients underwent hepatectomy by the same surgical team with selective and dynamic regional hepatic vascular occlusion from September2007to November2012. All patients were RO resection, less than800ml bloodloss and had no perioperative blood transfusion. The clinical-pathological risk factors of recurrence were retrospectively analyzed and the recurrence prediction model were established.[Results]:The total RFS rate was74.9%,49.3%and34.3%in1,3and5year respectively. The univariate analysis showed that serum gamma-glutamyl-transferase rise more than55U, AFP rise more than400ng/ml, more than5cm diameter, multi-focal, satellite nodules, poor differentiated, microvascular invasion, envelope invasion, postoperative hypohepatia were significantly associated with poor RFS. The multivariate analyses revealed that tumor size, satellite nodules, poor differentiated, microvascular invasion and postoperative hypohepatia were independent prognostic predictors associated with shorter RFS. The recurrence prediction model was established based on the5independent prognostic predictors aforementioned:accuracy was82.90%(sensitivity was70.83%and specificialty was87.96%).For all patients, perioperative TACE associated with poor RFS. For158cases with at least one independent prognostic predictor, further analysis demonstrated that perioperative TACE was not significantly associated with RFS. According to the recurrence prediction model,77cases with high risk of recurrence were selected and survival analysis showed perioperative TACE was beneficial to RFS.[Conclusions]:Selective and dynamic regional hepatic vascular occlusion could effectively control the surgical related risk factors of recurrence for HCC. Tumor features and liver function were the main risk factors of RFS. Perioperative TACE could not benefit patients received curative resection. Comprehensive therapy including TACE would benefit patients with high risk of recurrence. PART II:Serum microRNA in The Prediction of High Risk Postoperative Recurrence of Hepatocellular Carcinoma[Background and Objective]:Tumor characteristics and severity of the underlying liver disease are the main bases that are used to predict the prognosis of postoperative hepatocellular carcinomas (HCC) and to suggest the appropriate treatment option. To date, no predictive molecular marker exists for therapeutic decision-making and follow-up-monitoring algorithm. MicroRNAs (miRNA) are small noncoding RNAs that regulate the expression of target gene by inhibiting translation and/or promoting degradation at post-transcriptional level. In recent years, miRNAs have been proposed as novel diagnositic and prognostic tools for HCC.Our study aimed to identify whether miRNA can serve as biomarker to classify the grade of malignancy and predict prognosis of HCC more precisely.[Materials and Methods]:HCC patients underwent hepatectomy by the same surgical team with selective and dynamic regional hepatic vascular occlusion from January2012to June2013were enrolled into the study. Preoperative blood sample and tumor tissue specimens were collected. First, TaqMan Real-time PCR microRNA Array was used to identify differentially expressed miRNAs from10tumor tissue samples (5from high risk recurrence group vs.5from non-recurrence group) and the matched serum sample. Five serum samples from each group were pooled together, respectively. The miRNAs satisfying the following criteria were considered as candidates:(1) having a15-35Ct value in at least8of the10samples and the two serum pools by Array,(2) showing a2-fold altered expression,(3) having the same variation tendency in tissue and serum samples,(4) having a P value below0.10between the two group of tissue samples. Subsequently, the expressions of candidate miRNAs were detected by quantitative real-time PCR (qRT-PCR) in116HCC patients’ serum samples. Univariate and multivariate analysis were used to identify the associations between the candidate miRNAs and prognosis. The miRNAs-clinic-pathology combined prediction model was established by using logistic regression analysis.[Results]:We identified65and159differentially expressed miRNAs between the two groups in tissue and in serum, respectively. Using the filtering criterion that mentioned above,4miRNAs were screened as candidates:miR-139-5p, miR-422a, miR-486-5p and miR-125b. The four miRNAs were quantified by quantitative real-time PCR (qRT-PCR) in116HCC patients’ serum. Univariate analysis demonstrated that reduced expression of miR-422a, miR-486-5p, miR-125b were significantly associated with poor RFS. Multivariate analyses revealed that reduced expression of miR-486-5p was an independent prognostic predictor associated with shorter RFS. The probability of the recurrence prediction model of miR-486-5p and certain clinical-pathological factors (including tumor diameter, micro vascular invasion and differentiation) was92.11%(89.74%specificity and84.21%sensitivity). Univariate and multivariate analyses showed that the expression level of miR-139-5p was associated with satellite node. A model for preoperative diagnosis of the satellite node was established by combination of serum miR-139-5p and AFP (81.25%probability,91.67%sensitivity and59.62%specificity).[Conclusions]:The levels of miR-139-5p, miR-422a, miR-486-5p and miR-125b were down-regulated both in tumor tissue and serum in early-recurrence HCC patients. The level of serum miR-486-5p was associated with RFS. Furthermore, miR-139-5p may be a predictive marker of HCC intrahepatic metastasis. Serum miRNAs may serve as biomarkers to accurately predict prognosis and diagnosis for HCC.