基于K近邻-随机森林集成算法的肝病预测研究

为了提高肝病预测准确率,提出一种基于K近邻-随机森林算法的肝病预测集成模型。首先对UCI数据集中的印度肝病数据集进行数据预处理;然后分别采用K近邻和随机森林算法构建出肝病预测的弱分类器;最后将两个弱分类器利用voting策略进行集成以获得集成肝病预测模型。同时分析了特征对模型的贡献程度。实验结果表明模型的性能指标F1-分数取得了84%的良好表现。因此利用该集成模型可为医生的临床诊断提供支持。

Non-alcoholic fatty liver disease:An early mediator predicting metabolic syndrome in obese children?

AIM:To investigate if non-alcoholic fatty liver disease (NAFLD) is an early mediator for prediction of metabolic syndrome,and if liver B-ultrasound can be used for its diagnosis.METHODS:We classified 861 obese children (6-16 years old) into three subgroups:group 0 (normal liver in ultrasound and normal transaminases);group 1 (fatty liver in ultrasound and normal transaminases);and group 2 (fatty liver in ultrasound and elevated transaminases).We measured the body mass index,waist and hip circumference,blood pressure,fasting blood glucose,insulin,homeostasis model assessment of insulin resistance (HOMA-IR),whole-body insulin sensitivity index (WBISI),lipid profile and transaminases in all the participants.The risk of developing metabolic syndrome (MS) was assessed according to the degree of liver fatty infiltration based on the B-ultrasound examination.RESULTS:Among the 861 obese children,587 (68.18%) were classified as having NAFLD,and 221 (25.67%) as having MS.The prevalence of MS in NAFLD children (groups 1 and 2) was 37.64% (221/587),which was much higher than that in non-NAFLD group (group 0,12.04%) (P < 0.01).There were significantly higher incidences concerning every component of MS in group 2 compared with group 0 (P < 0.05).The incidence of NAFLD in MS patients was 84.61% (187/221),which was significantly higher than that of hypertension (57.46%,127/221) and glucose metabolic anomalies (22.62%,50/221),and almost equal to the prevalence of dyslipidemia (89.14%,197/221).Based on the B-ultrasound scales,the presence of moderate and severe liver fatty infiltration carried a high risk of hypertension [odds ratio (OR):2.18,95% confidence interval (95% CI):1.27-3.75],dyslipidemia (OR:7.99,95% CI:4.34-14.73),impaired fasting glucose (OR:3.65,95% CI:1.04-12.85),and whole MS (OR:3.77;95% CI:1.90-7.47,P < 0.01).The state of insulin resistance (calculated by HOMA-IR and WBISI) deteriorated as the degree of fatty infiltration increased.CONCLUSION:NAFLD is not only a liver disease,but also an early mediator that reflects metabolic disorder,and liver B-ultrasound can be a useful tool for MS screening.

Overexpression of metastasis-associated in colon cancer 1 predicts a poor outcome of hepatitis B virus-related hepatocellular carcinoma

AIM: To investigate the intratumoral expression of metastasis-associated in colon cancer 1 (MACC1) and c-Met and determine their clinical values associated with hepatitis B virus (HBV)-related hepatocellular carcinoma (HCC). METHODS: A retrospective study admitted three hundred fifty-four patients with HBV-related HCC. The expression and distribution of MACC1 and c-Met were assessed by quantitative real-time polymerase chain reaction and immunohistochemistry staining. Prognostic factors influencing survival, metastasis and recurrence were assessed. RESULTS: Intratumoral MACC1 level was found to be associated with HCC disease progression. Both median tumor-free survival (TFS) and overall survival (OS) were significantly shorter in the postoperative HCC patients with high intratumoral MACC1 expression, as compared to those with low intratumoral MACC1 levels (TFS: 34 mo vs 48.0 mo, P < 0.001; OS: 40 mo vs 48 mo, P < 0.01). Multivariable analysis indicated that high MACC1 expression or co-expression with c-Met were independent predictors for HCC clinic outcome (P < 0.001). CONCLUSION: High intratumoral MACC1 expression can be associated with enhanced tumor progression and poor outcome of HBV-related HCC. MACC1 may serve as a prognostic biomarker for postoperative HCC.

Factors predicting occurrence and prognosis of hepatitis-B-virus-related hepatocellular carcinoma

Primary liver cancer is an important cause of cancer death, and hepatocellular carcinoma （HCC） accounts for 70%-85% of total liver cancer worldwide. Chronic hepatitis B virus （HBV） infection contributes to 〉 75% of HCC cases. High serum viral load is the most reliable indicator of viral replication in predicting development of HCC. HBV genotype C is closely associated with HCC in cirrhotic patients aged 〉 50 years, whereas genotype B is associated with development of HCC in non-cirrhotic young patients and postoperative relapse of HCC. Different HBV subgenotypes have distinct patterns of mutations, which are clearly associated with increased risk of HCC. Mutations accumulate during chronic HBV infection and predict occurrence of HCC. Chronic inflammation leads to increased frequency of viral mutation via cellular cytidine deaminase induction. Mutations are negatively selected by host immunity, whereas some immuno-escaped HBV mutants are active in hepatocarcinogenesis. Inflammatory pathways contribute to the inflammation-necrosis-regeneration process, ultimately HCC. Their hallmark molecules can predict malignancy in HBV-infected subjects. Continuing inflammation is involved in hepatocarcinogenesis and closely related to recurrence and metastasis. HBV load, genotype C, viral mutations and expression of inflammatory molecules in HBV-related HCC tissues are significantly associated with poor prognosis. Imbalance between intratumoral CD8^＋T cells and regulatory T cells or Thl and Th2 cytokines in peritumoral tissues can predict prognosis of HBV-related HCC. These factors are important for developing active prevention and surveillance of HBV-infected subjects who are more likely to develop HCC, or for tailoring suitable treatment to improve survival or postpone postoperative recurrence of HCC.

Application of a biochemical and clinical model to predict individual survival in patients with end-stage liver disease

AIM:To investigate the capability of a biochemical and clinical model,BioCliM,in predicting the survival of cirrhotic patients.METHODS:We prospectively evaluated the survival of 172 cirrhotic patients.The model was constructed using clinical(ascites,encephalopathy and variceal bleeding) and biochemical(serum creatinine and serum total bilirubin) variables that were selected from a Cox proportional hazards model.It was applied to estimate 12-,52-and 104-wk survival.The model's calibration using the Hosmer-Lemeshow statistic was computed at 104 wk in a validation dataset.Finally,the model's validity was tested among an independent set of 85 patients who were stratified into 2 risk groups(low risk≤8 and high risk>8).RESULTS:In the validation cohort,all measures of fi t,discrimination and calibration were improved when the biochemical and clinical model was used.The proposed model had better predictive values(c-statistic:0.90,0.91,0.91) than the Model for End-stage Liver Disease(MELD) and Child-Pugh(CP) scores for 12-,52-and 104-wk mortality,respectively.In addition,the Hosmer-Lemeshow(H-L) statistic revealed that the biochemical and clinical model(H-L,4.69) is better calibrated than MELD(H-L,17.06) and CP(H-L,14.23).There were no significant differences between the observed and expected survival curves in the stratified risk groups(low risk,P=0.61;high risk,P=0.77).CONCLUSION:Our data suggest that the proposed model is able to accurately predict survival in cirrhotic patients.

Baseline predictors of virological response for chronic hepatitis B patients

AIM： To determine which baseline factors of chronic hepatitis B patients are predictive of virological response to Peginterferon α-2b therapy. METHODS： A total of 21 HBeAg-positive chronic hepatitis B （CriB） patients treated with Peginterferon α-2b were recruited. They were treated with Peginterferon α-2b （0.5-1.0 μg/kg per week） for 24 wk and followed up for 24 wk. Clinical and laboratory data of the patients were determined at pretreatment and at week 12, at 24 during treatment, and at week 48 during follow up. RESULTS： Ten patients achieved a virological response at the end of treatment. Their baseline serum alanine aminotransferase （ALT）, thyroid-stimulating hormone （TSH）, and total thyroxin （TT4） levels were significantly different from those who failed treatment. The positive predictive values （PPV） and negative predictive values （NPV） of ALT, TSH, and TT4 were 75% and 89 %, 75% and 89 %, and 75% and 75%, respectively. Moreover, combinations of the baseline ALT and TT4, ALT and TSH, TT4 and TSH levels had much higher PPV and NPV （86% and 88%, 89% and 100%, 83% and 100%, respectively）.CONCLUSION： Baseline serum ALT, TSH, and TT4 levels, especially in combination, have high predictive values of virological response to Peginterferon α-2b in HBeAg-positive CriB patients.

Nonalcoholic fatty liver disease is a novel predictor of cardiovascular disease

AIM:To clarify whether nonalcoholic fatty liver disease(NAFLD)increases the risk of cardiovascular disease.METHODS:We carried out a prospective observational study with a total of 1637 apparently healthy Japanese men and women who were recruited from a health check-up program.NAFLD was diagnosed by abdominal ultrasonography.The metabolic syndrome(MS)was defined according to the modified National Cholesterol Education Program(NCEP)ATP Ⅲ criteria.Five years after the baseline evaluations,the incidence of cardiovascular disease was assessed by a self-administered questionnaire.RESULTS:Among 1221 participants available for outcome analyses,the incidence of cardiovascular disease was higher in 231 subjects with NAFLD at baseline(5 coronary heart disease,6 ischemic stroke,and 1 cerebral hemorrhage)than 990 subjects without NAFLD(3 coronary heart disease,6 ischemic stroke,and 1 cerebral hemorrhage).Multivariate analyses indicated that NAFLD was a predictor of cardiovascular disease independent of conventional risk factors(odds ratio 4.12,95% CI,1.58 to 10.75,P = 0.004).MS was alsoindependently associated with cardiovascular events.But simultaneous inclusion of NAFLD and MS in a multivariate model revealed that NAFLD but not MS retained a statistically significant correlation with cardiovascular disease.CONCLUSION:Although both of them were predictors of cardiovascular disease,NAFLD but not MS retained a statistically significant correlation with cardiovascular disease in a multivariate model.NAFLD is a strong predictor of cardiovascular disease and may play a central role in the cardiovascular risk of MS.

Construction and clinical significance of a predictive system for prognosis of hepatocellular carcinoma

AIM: The aims of this study were to explore individualized treatment method for hepatocellular carcinoma (HCC) patients whose maximum tumor size was less than 5 cm to improve prognosis and survival quality. METHODS: Thirty cases of primary HCC patients undergoing tumor resection were retrospectively analyzed (resection group). All the tumors were proved as primary HCC with pathologic examination. The patients were divided into two groups according to follow-up results: group A, with tumor recurrence within 1 year after resection; group B, without tumor recurrence within 1 year. Immunohist-ochemical stainings were performed using 11 kinds of monoclonal antibodies (AFP, c-erbB2, c-met, c-myc, HBsAg, HCV, Ki-67, MMP-2, nm23-H1, P53, and VEGF), and expressing intensities were quantitatively analyzed. Regression equation using factors affecting prognosis of HCC was constructed with binary logistic method. HCC patients undergoing percutaneous microwave coagulation therapy (PMCT) were also retrospectively analyzed (PMCT group). Immunohistochemical stainings of tumor biopsy samples were performed with molecules related to HCC prognosis, staining intensities were quantitatively analyzed, coincidence rate of prediction was calculated. RESULTS: In resection group, the expressing intensities of c-myc, Ki-67, MMP-2 and VEGF in cancer tissue in group A were significantly higher than those in group B (t = 2.97, P= 0.01; t = 2.42, P= 0.03<0.05; t = 2.57, P= 0.02<0.05; t = 3.43, P = 0.004<0.01, respectively); the expressing intensities of 11 kinds of detected molecules in para-cancer tissue in groups A and B were not significantly different (P>0.05). The regression equation predicting prognosis of HCC is as follows: P(1) = 1/[1+e-(3.663-0.412mycc-2.187kl-67c-0.397vegfc)]. It demonstrates that prognosis of HCC in resection group was related with c-myc, Ki-67 and VEGF expressing intensity in cancer tissue. In PMCT group, the expressing intensities of c-myc, Ki-67 and VEGF in cancer tissue in group A were significantly higher than those in group B (t = 4.57, P= 0.000<0.01; t = 2.08, P= 0.04<0.05; t = 2.38, P= 0.02<0.05, respectively); the expressing intensities of c-myc, Ki-67 and VEGF in para-cancer tissue in groups A and B were not significantly different (P>0.05). The coincidence rate of patients undergoing PMCT in group A was 88.00% (22/25), in group B 68.75% (11/16), the total coincidence rate was 80.49% (33/41). CONCLUSION: The regression equation is accurate and feasible and could be used for predicting prognosis of HCC, it helps to select treatment method (resection or PMCT) for HCC patients to realize individualized treatment to improve prognosis.

Globulin-platelet model predicts minimal fibrosis and cirrhosis in chronic hepatitis B virus infected patients

AIM： To establish a simple model consisting of the rou- tine laboratory variables to predict both minimal fibrosis and cirrhosis in chronic hepatitis B virus （HBV）-infected patients. METHODS： We retrospectively investigated 114 chron- ic HBV-infected patients who underwent liver biopsy in two different hospitals. Thirteen parameters were analyzed by step-wise regression analysis and correla- tion analysis. A new fibrosis index [globulin/platelet （GP） model] was developed, including globulin （GLOB） and platelet count （PLT）. GP model = GLOB （g/mL） x 100/PLT （x 109/L）. We evaluated the receiver operating characteristics analysis used to predict minimal fibrosis and compared six other available models. RESULTS： Thirteen clinical biochemical and hemato- logical variables [sex, age, PLT, alanine aminotransfer- ase, aspartate aminotransferase （AST）, albumin, GLOB, total bilirubin （T.bil）, direct bilirubin （D.bil）, glutamyl-transferase, alkaline phosphatase, HBV DNA and pro- thrombin time （PT）] were analyzed according to three stages of liver fibrosis （F0-F1, F2-F3 and F4）. Bivariate Spearman＇s rank correlation analysis showed that six variables, including age, PLT, T.bil, D.bil, GLOB and PT, were correlated with the three fibrosis stages （FS）. Cor- relation coefficients were 0.23, -0.412, 0.208, 0.220, 0.314 and 0.212; and P value was 0.014, 〈 0.001, 0.026, 0.018, 0.001 and 0.024, respectively. Univariate analysis revealed that only PLT and GLOB were signifi- cantly different in the three FS （PLT： F = 11.772, P 〈 0.001; GLOB： F = 6.612, P = 0.002）. Step-wise multiple regression analysis showed that PLT and GLOB were also independently correlated with FS （R2 = 0.237）. By Spearman＇s rank correlation analysis, GP model was significantly correlated with the three FS （r = 0.466, P 〈 0.001）. The median values in F0-F1, F2-F3 and F4 were 1.461, 1.720 and 2.634. Compared with the six available models （fibrosis index, AST-platelet ratio, FIB-4, fibrosis-cirrhosis index and age-AST model and age-PLT ratio）, GP model showed a highest correlation coefficient. The sensitivity and positive predictive value at a cutoff value 〈 1.68 for predicting minimal fibrosis F0-F1 were 72.4% and 71.2%, respectively. The speci- ficity and negative predictive value at a cutoff value 〈 2.53 for the prediction of cirrhosis were 84.5% and 96.7%. The area under the curve （AUC） of GP model for predicting minimal fibrosis and cirrhosis was 0.762 [95% confidence interval （CI）： 0.676-0.848] and 0.781 （95% CI： 0.638-0.924）. Although the differences were not statistically significant between GP model and the other models （P all 〉 0.05）, the AUC of GP model was the largest among the seven models. CONCLUSION： By establishing a simple model using available laboratory variables, chronic HBV-infected patients with minimal fibrosis and cirrhosis can be di- agnosed accurately, and the clinical application of this model may reduce the need for liver biopsy in HBV- infected patients.

Is transient elastography a useful tool for screening liver disease?

Transient elastography(TE)is a new non invasive tool for measuring liver stiffness,which is correlated to the histologic stage of liver fibrosis.Several studies in chronic liver disease(CLD)have determined a good accuracy of TE in predicting significant fibrosis and an optimal accuracy in predicting cirrhosis.Normal liver stiffness ranges between 3.3-7.8 KPa and using a cut off of 7.1 KPa,significant fibrosis and cirrhosis can be excluded with a very high negative predictive value(NPV).Positive predictive value(PPV)for the diagnosis of cirrhosis is lower using just a single scan but increases to 90% if high stiffness values are confirmed by a second independent scan.However the presence of fatty liver and metabolic syndrome slightly increases the readings and may reduce the accuracy of the test.It is uncertain if this increase is related to the presence of steatofibrosis or ifit is caused by steatosis itself.TE can be used in screening patients attending the liver clinics to identify those with signifi cant fi brosis or cirrhosis and may be particularly useful in discriminating HBV inactive carriers from chronic hepatitis B patients.TE,however,is not reliable in predicting the presence of esophageal varices in cirrhotics.Another potential indication for TE is the systematic screening of populations at high risk for CLD,such as intravenous drug users and alcoholics,but further studies are needed to determine its diagnostic accuracy in these settings.

APRI as a predictor of early viral response in chronic hepatitis C patients

AIM:To evaluate the aspartate aminotransferase(AST) to platelet ratio index(APRI) as a predictive factor of early viral response in chronic hepatitis C naive patients.METHODS:We performed an ambispective case-control study.We enrolled chronic hepatitis C naive patients who were evaluated to start therapy with PEGylated interferon α-2b(1.5 μg/kg per week) and ribavirin(>75 kg:1200mg and <75kg:1000mg).Patients were allocated into two groups,group 1:Hepatitis C patients with early viral response(EVR),group 2:Patients without EVR.Odds ratio(OR) and 95% confi dence interval(CI) were calculated to assess the relationship between each risk factor and the EVR in both groups.RESULTS:During the study,80 patients were analyzed,45 retrospectively and 35 prospectively.The mean ± SD age of our subjects was 42.9 ± 12 years;weight 70 kg(±11.19),AST 64.6 IU/mL(±48.74),alanine aminotransferase(ALT) 76.3 IU/mL(±63.08) and platelets 209000 mill/mm3(±84 429).Fifty-five(68.8%) were genotype 1 and 25(31.3%) were genotype 2 or 3;the mean hepatitis C virus RNA viral load was 2 269 061 IU/mL(±7220266).In the univariate analysis,APRI was not associated with EVR [OR 0.61(95% CI 0.229-1.655,P=0.33)],and the absence of EVR was only associated with genotype 1 [OR 0.28(95% CI 0.08-0.94,P=0.034)].After adjustment in a logistic regression model,genotype 1 remains signifi cant.CONCLUSION:APRI was not a predictor of EVR in chronic hepatitis C;Genotype 1 was the only predictive factor associated with the absence of EVR in our patients.

Treatment of chronic hepatitis delta virus with peg-interferon and factors that predict sustained viral response

AIM： To observe the efficacy of peg-interferon in the treatment of hepatitis delta virus （HDV） and to identify the factors that would be predictive of the sustained viral response （SVR）. METHODS： This prospective study was conducted in Medical Unit IV of the Liaquat University of Medi- cal and Health Sciences Hospital Jamshoro from June 2008 to September 2011. This study cohort included all patients of either sex who presented during this time with hepatitis B surface antigen positivity, hepa- titis B virus DNA 〉 20 000 IU/ml, serum glutamic py- ruvic transaminase （SGPT） 〉 2（upper limit of normal）, HDV-RNA positivity with fibrosis stage ≥ 2. Informed consent was obtained from each of these individuals. Patients were diagnosed with hepatitis D on the ba- sis of detectable viral antibodies and the presence of HDV-RNA in their serum. A liver biopsy was performed in all cases and fibrosis staging was performed in ac- cordance with the METAVIR scoring system. All eligible patients were administered peg-interferon at a weekly dosage of 1.5 μg/kg body weight for 48 wk. HDV-RNA was assayed at the end of this treatment period and again at 24 wk later. A biochemical response was de- termined by a normalization of SGPT at the end of the treatment or during follow up. The end of treatment response was defined by a HDV-RNA negative status. A sustained virological response was defined by unde- tectable serum HDV-RNA at six months after the end of treatment. RESULTS： Among the 277 patients enrolled in our present study, 238 completed a course of peg-interfer- on therapy of which 180 （75.6%） were male and 58 （24.4%） female. Biochemical responses were achieved in 122/238 （51.3%） patients. End of treatment re- sponses were achieved in 71/238 （29.8%） cases. A SVR was achieved in 70 of these patients （29.4%）. A strong association was found between the SVR and the end of treatment responses （P = 0.001）, biochemical responses （P = 0.001） and the degree of fibrosis （P = 0.002）. CONCLUSION： Peg-interferon therapy can induce re- mission in nearly one third of patients harboring HDV.

Non-alcoholic fatty liver disease and metabolic syndrome in obese children

I read with great interest the article of Fu et al who investigated whether non-alcoholic fatty liver disease(NAFLD)is an early mediator for prediction of metabolic syndrome,and whether liver B-ultrasound could be used for its diagnosis,in a study involving 861 obese children(6-16 years old).In this study,it was reported that NAFLD is not only a liver disease,but also an early mediator that reflects metabolic disorder,and that liver B-ultrasound can be a useful tool for metabolic syndrome(MS)screening.The authors reported that NAFLD and MS were present in 68.18%and 25.67%of obese children,respectively.Moreover,they observed that the prevalence of MS in NAFLD children was 37.64%,which was much higher than that in the non-NAFLD group.Criteria analogous to those of the Adult Treatment PanelⅢdefinition for MS were used for children in this study.The reported prevalence data on MS in the young has varied markedly,in large part because of disagreement among the variously proposed definitions of MS.Therefore,in my opinion,a study aiming to assess the association between MS components and NAFLD in obese children has to take into account a simple,easy-to-apply clinical definition proposed by the international diabetes federation for MS.Interpretation of the results of the Fu et al study are limited byanother major caveat:that the diagnosis or exclusion of NAFLD was based on liver enzymes and ultrasound imaging,but was not confirmed by liver biopsy.Indeed,it is known that liver enzymes may be within the reference interval in up to 70%of patients with diagnosed NAFLD and that the full histopathological spectrum of NAFLD may be present in patients with normal liver enzymes,which therefore cannot be reliably used to exclude the presence of NAFLD.

Epidemiology of ADRs in Internal Medicine, HCC

We aimed to characterize, and analyze the presence of factors such as polypharmacy and personal medical history that predispose to Adverse Drug Reactions （ADRs） and potential preventability of these. The Civil Hospital of Culiacan （HCC）, in Sinaloa, where the study was conducted has 80 beds in total, in which the Department of Internal Medicine with 24 beds： 12 lbr men and 12 for women, we found the part of the share to contribute to the 200,000 cases of ADRs that according with the WHO each of the countries members report ever）, year to the Uppsala Monitoring Centre. We have not only but also ordered frequency of drugs, pathologies, and analytical tests of the hospitalized patients.