• 한국정보과학회논문지:시스템및이론
• /
• 제34권7호
• /
• pp.276-281
• /
• 2007

본 논문에서는 한국인의 대표질환 중 하나인 만성 간염에 대한 질환 감수성을 예측하기 위해서 Single Nucleotide Polymorphism 데이타와 대표적인 기계학습 기술인 Support Vector Machine을 이용하였다. 실험을 위한 데이타로 만성간염 환자 173명과 정상인 155명의 SNP 데이타를 사용하였으며, 평가를 위한 방법으로는 Leave-One-Out Cross Valication을 사용하였다. 실험결과 SNP 데이터만으로는 67.1%의 예측 결과를 얻었으며 기본적인 건강요소인 나이와 성별을 특징요소로 사용함으로서 74.9%의 예측 결과를 보였다. 향후 보다 많은 SNP 데이타와 건강관련정보 그리고 생활패턴에 대한 요소들을 특징요소로 감수성 예측에 함께 사용한다면, SVM은 만성 간염 예측을 위한 보다 효과적인 도구가 될 것이다.

• International Journal of Computer Science & Network Security
• /
• 제22권1호
• /
• pp.167-174
• /
• 2022

In metaheuristic algorithms such as Genetic Algorithm (GA), initial population has a significant impact as it affects the time such algorithm takes to obtain an optimal solution to the given problem. In addition, it may influence the quality of the solution obtained. In the machine learning field, feature selection is an important process to attaining a good performance model; Genetic algorithm has been utilized for this purpose by scientists. However, the characteristics of Genetic algorithm, namely random initial population generation from a vector of feature elements, may influence solution and execution time. In this paper, the use of a statistical algorithm has been introduced (Chi2) for feature relevant checks where p-values of conditional independence were considered. Features with low p-values were discarded and subject relevant subset of features to Genetic Algorithm. This is to gain a level of certainty of the fitness of features randomly selected. An ensembled-based learning model for Hepatitis has been developed for Hepatitis C stage classification. 1385 samples were used using Egyptian-dataset obtained from UCI repository. The comparative evaluation confirms decreased in execution time and an increase in model performance accuracy from 56% to 63%.

• Molecules and Cells
• /
• 제40권6호
• /
• pp.418-425
• /
• 2017

Interferon-${\gamma}$-inducible protein 10 (IP-10), also known as chemokine C-X-C motif ligand (CXCL) 10, is closely associated with antiviral immunity and the progression of chronic hepatitis B (CHB). However, the value of baseline serological and histological IP-10 expression levels in predicting the efficacy of the antiviral response to nucleoside/nucleotide analogues (NAs) is still unknown. In our research, intrahepatic and peripheral IP-10 expression levels were systemically examined before and after treatment with entecavir (ETV). Baseline serological and histological IP-10 expression levels were significantly increased in patients with CHB, particularly in patients with higher degrees of liver inflammation and liver fibrosis. Moreover, higher baseline intrahepatic IP-10 levels indicated better prognoses in patients with CHB after entecavir therapy. The baseline IP-10 level was also positively associated with several clinical parameters, including baseline levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), hepatitis B virus (HBV) DNA, and hepatitis B surface antigen (HBsAg), and with the decrease in HBsAg levels after treatment. In addition, monocyte-derived IP-10 was expressed at higher levels in patients with CHB than in patients with liver cirrhosis (LC) and healthy controls (HC). According to the results of our in vitro experiments, IP-10 directly promoted hepatocyte apoptosis. Based on these findings, baseline serological and histological IP-10 levels might predict CHB severity and the decrease in HBsAg levels after entecavir therapy.

• Asian Pacific Journal of Cancer Prevention
• /
• 제16권1호
• /
• pp.381-385
• /
• 2015

Background:Hepatocellular carcinoma (HCC) is the second most common malignancy in Egypt. Data mining is a method of predictive analysis which can explore tremendous volumes of information to discover hidden patterns and relationships. Our aim here was to develop a non-invasive algorithm for prediction of HCC. Such an algorithm should be economical, reliable, easy to apply and acceptable by domain experts. Methods: This cross-sectional study enrolled 315 patients with hepatitis C virus (HCV) related chronic liver disease (CLD); 135 HCC, 116 cirrhotic patients without HCC and 64 patients with chronic hepatitis C. Using data mining analysis, we constructed a decision tree learning algorithm to predict HCC. Results: The decision tree algorithm was able to predict HCC with recall (sensitivity) of 83.5% and precession (specificity) of 83.3% using only routine data. The correctly classified instances were 259 (82.2%), and the incorrectly classified instances were 56 (17.8%). Out of 29 attributes, serum alpha fetoprotein (AFP), with an optimal cutoff value of ${\geq}50.3ng/ml$ was selected as the best predictor of HCC. To a lesser extent, male sex, presence of cirrhosis, AST>64U/L, and ascites were variables associated with HCC. Conclusion: Data mining analysis allows discovery of hidden patterns and enables the development of models to predict HCC, utilizing routine data as an alternative to CT and liver biopsy. This study has highlighted a new cutoff for AFP (${\geq}50.3ng/ml$). Presence of a score of >2 risk variables (out of 5) can successfully predict HCC with a sensitivity of 96% and specificity of 82%.

• Asian Pacific Journal of Cancer Prevention
• /
• 제14권10호
• /
• pp.5891-5893
• /
• 2013

Incidence/mortality of liver cancer follow logistic curves because there is a limit reflecting the prevalence of hepatitis virus carriers in the cohort. The author fitted logistic curves to incidence/mortality data covering the nine five-year cohorts born in 1911-1955 of both sexes. Goodness-of-fit of logistic curves was sufficiently precise to be used for future predictions. Younger cohorts born in 1936 or later were predicted to show constant decline in incidence/mortality in the future. The male cohort born in 1931-35 showed an elevated incidence/mortality of liver cancer early in their lives supporting the previous claim that this particular cohort had suffered massive HCV infection due to nation-wide drug abuse in the 1950s. Declining case-fatality observed in younger cohorts suggested improved treatment of liver cancer. This study demonstrated that incidence/mortality of liver cancer follow logistic curves and fitted logistic formulae can be used for future prediction. Given the predicted decline of incidence/mortality in younger cohorts, liver cancer is likely to be lost to history in the not-so-distant future.

• Asian Pacific Journal of Cancer Prevention
• /
• 제14권11호
• /
• pp.6721-6726
• /
• 2013

Background: Changes in DNA methylation patterns are believed to be early events in hepatocarcinogenesis. A better understanding of methylation states and how they correlate with disease progression will aid in finding potential strategies for early detection of HCC. The aim of our study was to analyze the methylation frequency of tumor suppressor genes, P14, P15, and P73, and a mismatch repair gene (O6MGMT) in HCV related chronic liver disease and HCC to identify candidate epigenetic biomarkers for HCC prediction. Materials and Methods: 516 Egyptian patients with HCV-related liver disease were recruited from Kasr Alaini multidisciplinary HCC clinic from April 2010 to January 2012. Subjects were divided into 4 different clinically defined groups - HCC group (n=208), liver cirrhosis group (n=108), chronic hepatitis C group (n=100), and control group (n=100) - to analyze the methylation status of the target genes in patient plasma using EpiTect Methyl qPCR Array technology. Methylation was considered to be hypermethylated if >10% and/or intermediately methylated if >60%. Results: In our series, a significant difference in the hypermethylation status of all studied genes was noted within the different stages of chronic liver disease and ultimately HCC. Hypermethylation of the P14 gene was detected in 100/208 (48.1%), 52/108 (48.1%), 16/100 (16%) and 8/100 (8%) among HCC, liver cirrhosis, chronic hepatitis and control groups, respectively, with a statistically significant difference between the studied groups (p-value 0.008). We also detected P15 hypermethylation in 92/208 (44.2%), 36/108 (33.3%), 20/100 (20%) and 4/100 (4%), respectively (p-value 0.006). In addition, hypermethylation of P73 was detected in 136/208 (65.4%), 72/108 (66.7%), 32/100 (32%) and 4/100 (4%) (p-value <0.001). Also, we detected O6MGMT hypermethylation in 84/208 (40.4%), 60/108 (55.3%), 20/100 (20%) and 4/100 (4%), respectively (p value <0.001. Conclusions: The epigenetic changes observed in this study indicate that HCC tumors exhibit specific DNA methylation signatures with potential clinical applications in diagnosis and prognosis. In addition, methylation frequency could be used to monitor whether a patient with chronic hepatitis C is likely to progress to liver cirrhosis or even HCC. We can conclude that methylation processes are not just early events in hepatocarcinogenesis but accumulate with progression to cancer.

• 한국방사선학회논문지
• /
• 제17권7호
• /
• pp.1057-1065
• /
• 2023

비침습적 간섬유화 진단은 만성 간질환 환자들에게 매우 중요한 사항이다. 많은 환자들에게 간조직검사를 시행할 수가 없기 때문에 의미 있는 간섬유화 정도를 조기에 예측함으로써 만성 간질환의 치료와 간세포암 및 간경변증으로 인한 합병증을 감소시킬 수 있다. 본 연구는 비침습적인 간섬유화 검사 방법 중 간초음파 소견과 함께 전단파 탄성 초음파를 이용하여 간섬유화의 정량적 측정과 간탄성도 수치와 연관성을 갖는 혈청학적 인자도 비교 분석하였다(p<0.05). 그 결과 정상 집단의 횡파탄성측정값은 4.55 ± 0.69 kPa였으며, 에코 패턴의 비정상 대조군은 8.27 ± 1.83 kPa로 B형 간염 보균 집단에서 전단파 탄성 초음파 측정값이 높게 나타났으며, 혈청학적 인자들 중 AST, ALT, GGT, PT가 SWE 범주에 따른 섬유화의 정도에 따라 통계학적으로 유의한 양의 연관성을 보였으나(p<0.05) ALP와 TB는 통계적으로 유의미한 차이를 나타내지 않았다(p=0.163, p=0.567). 반면에 Albumin과 PLT는 유의한 음의 연관성을 보였다(p<0.05). 임상적으로 간경변 증후가 없는 만성 B형 간염 환자들의 간섬유화 추적 관찰 및 반복 검사 시 간초음파와 전단파 탄성 초음파 검사를 통한 탄성도 값을 추가로 활용한다면 검사자 간 객관적 진단에 있어 도움이될 것이며, 일차적인 진료에 유용할 것으로 사료된다.

• Pediatric Gastroenterology, Hepatology & Nutrition
• /
• 제21권4호
• /
• pp.289-296
• /
• 2018

Purpose: We aimed to study the pattern of liver-injury in children with dengue fever (DF) and validate serum aminotransferase ${\geq}1,000IU/L$ as a marker of severe DF. Methods: Children admitted with DF were included. DF was defined by presence of clinical criteria and positive serological or antigen tests in absence of other etiology. DF severity was graded as dengue without or with warning signs and severe dengue. Liver-injury was defined as alanine aminotransferase (ALT) more than twice the upper limit of normal (boys, 30 IU/L; girls, 21 IU/L). Results: Of 372 children with DF, 144 (38.7%) had liver-injury. Risk of liver-injury and aminotransferase levels increased with DF severity (p<0.001). Recommended ALT and aspartate aminotransferase (AST) cut-off at ${\geq}1,000IU/L$ had sensitivity 4.8% (5/105), specificity 99.3% (265/267) for detection of severe DF. In children with ALT and AST <1,000 IU/L (n=365), the area under receiver operating curves for prediction for severe DF, were 0.651 (95% confidence interval [CI], 0.588-0.714; p<0.001) for ALT and 0.647 (95% CI, 0.582-0.712; p<0.001) for AST. Serum ALT at 376 IU/L and AST at 635 IU/L had sensitivity and specificity comparable to ${\geq}1,000IU/L$ for defining severe DF. Conclusion: Liver-injury is common in DF. The ALT and AST levels increase with DF severity. ALT and AST levels of ${\geq}1,000IU/L$ could be lowered to 376 IU/L and 635 IU/L respectively for defining severe DF.

• 약학회지
• /
• 제55권4호
• /
• pp.352-360
• /
• 2011

Standard combination chemotherapy including isoniazid, rifampin, pyrazinamide, and ethambutol is very effective against tuberculosis. But, these medicines can cause hepatotoxicity which is the main reason for treatment interruption or change in drug regimen. In order to identify risk factors associated with hepatotoxcity in Koreans and assess elevated baseline LFTs' contributions to hepatotoxicity, a retrospective case control study was performed. The medical records of 277 patients who diagnosed with tuberculosis at a community hospital from January 1st, 2007 to June 30th, 2010 were reviewed. Patients were categorized into 3 groups (non toxic group, patients without increase in LFT levels; mild to moderate hepatotoxic group and severe hepatotoxic group). And the correlation between risk factors and hepatotoxicity was analyzed by using SPSS program. The overall incidence of hepatotoxicity was 18% and 8.7% of patients developed severe toxicity. Patients in the severe toxic group had the longest treatment period among the three groups. In 75% of severe toxic group, hepatotoxicity occurred within 18.3 days after starting medication. Hypoalbuminemia (serum albumin <3 g/dl) was a significant risk factor for development of severe toxicity. Elevated baseline transaminase (except ALT), total bilirubin, and preexisting hepatitis were also risk factors which were more than twice as likely to increase risk of severe hepatotoxicity (p>0.05). In conclusion, hypoalbuminemia (serum albumin level <3 g/dl) was a significant risk factor for anti-tuberculosis druginduced severe toxicity. Therefore, before starting antituberculosis chemotherapy, serum albumin level should be assessed at baseline. In high-risk patients (hypoalbuminemia, elevated LFTs) for hepatotoxicty, liver function should be closely monitored up to at least 21 days after taking medication.

• 수면정신생리
• /
• 제24권1호
• /
• pp.38-45
• /
• 2017

목 적 : 불면 증상을 나타내는 수면 장애는 고령화 현상과 더불어 매우 급격히 증가중인 질환이다. 3년간 추적 관찰한 우리나라 농촌의 60세 이상 노인 코호트를 대상으로 수면 장애의 유병률과 이를 예측하는 요인을 알아보고자 하였다. 방 법 : 한국인의 사회적 삶, 건강과 노화에 대한 2012년과 2014년의 조사를 통해 두 번 데이터를 얻었다. 뇌졸중, 심근경색증, 협심증, 관절염, 폐결핵, 천식, 백내장, 녹내장, B형 간염, 요실금, 전립선 비대, 암, 골다공증, 고혈압, 당뇨, 고지혈증, 대사증후군에 대한 진단 여부를 물었다. 치매선별용 간이정신상태검사를 이용하여 인지 기능을 평가하였고 역학연구 우울척도를 이용하여 우울 증상을 평가하였다. 2015년에는 이 중 235명에게 DSM-IV의 제1축 장애의 구조화된 임상적 면담을 시행하였고 피츠버그 수면 질 척도로 불면증의 정도, 즉 수면 장애를 평가하였다. 또한, 지각된 스트레스 척도와 상태-특성 분노표현 척도를 시행하였다. 수면장애를 예측하기 위하여 성별, 나이, 교육, 첫 번째와 세 번째 주기의 우울 점수, 공존 질환의 개수, 그리고 현재의 분노 억제 점수와 지각된 스트레스 정도를 설명 변인으로 삼아 로지스틱 회귀분석을 시행하였다. 결 과 : 조사 대상의 27%가 수면 장애를 가지고 있었다. 회귀 분석 결과, 3년 전의 공존 질환 수, 1년 전의 우울 점수 및 현재 지각된 스트레스 정도가 수면 장애를 유의하게 예측하였다. 결 론 : 3년 전 측정한 공존 질환과 1년 전 평가한 우울증상이 현재의 수면 장애를 예측할 수 있었다. 공존 질환 및 우울 증상의 치료가 수면 장애를 호전시킬 수 있는지 추가적 연구가 필요하다.