Combined application of information theory on laboratory results with classification and regression tree analysis: analysis of unnecessary biopsy for prostate cancer.

BACKGROUND: The probability of a prostate cancer-positive biopsy result varies with PSA concentration. Thus, we applied information theory on classification and regression tree (CART) analysis for decision making predicting the probability of a biopsy result at various PSA concentrations. METHODS: From 2007 to 2009, prostate biopsies were performed in 664 referred patients in a tertiary hospital. We created 2 CART models based on the information theory: one for moderate uncertainty (PSA concentration: 2.5-10 ng/ml) and the other for high uncertainty (PSA concentration: 10-25 ng/ml). RESULTS: The CART model for moderate uncertainty (n=321) had 3 splits based on PSA density (PSAD), hypoechoic nodules, and age and the other CART for high uncertainty (n=160) had 2 splits based on prostate volume and percent-free PSA. In this validation set, the patients (14.3% and 14.0% for moderate and high uncertainty groups, respectively) could avoid unnecessary biopsies without false-negative results. CONCLUSIONS: Using these CART models based on uncertainty information of PSA, the overall reduction in unnecessary prostate biopsies was 14.0-14.3% and CART models were simplified. Using uncertainty of laboratory results from information theoretic approach can provide additional information for decision analysis such as CART.

[Study on interpretation of quantitative results of prostate-specific antigen using information theory].

BACKGROUND: The prostate-specific antigen (PSA) is considered the most useful among tumor markers currently used. However, its quantitative results are interpreted only qualitatively for the diagnosis of prostate cancer. The recently introduced information theory enables the information of the quantitative results transformed into Shannon's entropy (S) that represents uncertainties and then "1-S" representing diagnostic certainty. METHODS: The 882 urological patients enrolled were categorized into 2 groups: a patient group comprising 233 patients with prostate cancer and a disease control group comprising 649 patients with benign prostate disease. The level of PSA in all the patients was tested and was found to be >or=2 ng/mL. The variables like PSA level and age were modeled on logistic regression analysis to predict the probability of prostate cancer and the diagnostic certainty. RESULTS: The mean (SD) of PSA levels in the patient group and the disease control group were 44.5 ng/mL (37.62 ng/mL) and 5.7 ng/mL (3.70 ng/mL), respectively. The logistic regression model fitted well when the age variable was dichotomized at the age of 55 yr. The diagnostic certainty was lowest at a PSA level of 18.90 ng/mL in the <55-yr age group, and 15.45 ng/mL in the >55-yr age group. CONCLUSIONS: The diagnostic certainty (1-S) of whether to diagnose prostate cancer or not at a certain PSA level could be obtained using the information theory. The methodology used in this study may help interpret the results of other quantitative tests.

Meta-analysis of the association between HLA-DRB1 allele and rheumatoid arthritis susceptibility in Asian populations.

The aims of this study were to summarize results on the association of HLA-DRB1 with rheumatoid arthritis (RA) in Asians and to determine if the shared epitope (SE) hypothesis could explain the meta-analysis results. Among the papers published between January 1987 and July 2006 on RA susceptibility in Asian-Mongoloid populations (Korean, Japanese, Chinese, and Thai), 12 were selected for the metaanalysis. Mongoloid-Asian patients with RA had significantly higher frequencies of HLA-DRB1*0101, *0401, *0410, and *1001 than controls (OR 1.5-2.1, p<0.05 for association). When analyses were restricted to more ethnically homogeneous populations, HLA-DRB1*0405 showed a significant susceptibility to RA in Koreans (OR 5.65, 95% CI 4.32-7.39), whereas the HLA-DRB1*0301, *0403, *0406, *0701, *1301, and *1405 alleles showed protective association with RA (OR 0.32-0.70, p<0.05 for association). In conclusion, it was found that HLA-DRB1 *0101, *0401, *0405, *0410, and *1001 are susceptible, while HLA-DRB1* 0301, *0403, *0406, *0701, *1301, and *1405 are protective in Asian-Mongoloids. All the RA-associated alleles except DRB1*0301 could be explained by the structural model supporting the SE hypothesis that RA susceptibility is determined by the combination of amino acid residues at HLA-DR beta71 and beta74, not by beta71 alone.

Human leukocyte antigen alleles and haplotypes associated with chronicity of hepatitis B virus infection in Koreans.

CONTEXT: Chronic hepatitis B infection is the leading cause of cirrhosis and hepatocellular carcinoma. Human leukocyte antigen may be involved in the chronicity of hepatitis B virus (HBV) infection. OBJECTIVE: To analyze the association between HBV chronicity and human leukocyte antigen alleles and haplotypes of 636 organ donors and recipients. DESIGN: Subjects were categorized into 2 groups according to their clinical and serologic profiles, specifically, an HBV natural convalescent group and an HBV chronic carrier (CC) group. RESULTS: Hepatitis B chronicity was positively associated with A33 (P = .004, odds ratio [OR] = 1.59) and DR7 (P < .001, OR = 2.58), and negatively associated with HLA-DR13 (P < .001, OR = 0.40). Coexpression of A33 and DR7 was significantly higher in the CC group (OR = 3.63), compared with that of either allele alone (OR = 1.76 in A33; OR = 2.53 in DR7). The statistically significant haplotypes were B44-DR7 (P < .001, OR = 5.44), A33-DR7 (P < .001, OR = 4.47), and A33-B44-DR7 (P < .001, OR = 7.31) in the CC group. CONCLUSIONS: Our results indicate that alleles of A33, DR7, and haplotypes containing DR7 are associated with HBV chronicity among Koreans. Moreover, the 2 antigens had an additive effect on chronicity. These findings support the theory that human leukocyte antigen class I-restricted cytotoxic T cells and human leukocyte antigen class II-restricted helper T cells play an important role in HBV chronicity.

Comparison of efficacy and tolerability of amlodipine orotate versus amlodipine besylate in adult patients with mild to moderate hypertension: a multicenter, randomized, double-blind, placebo-controlled, parallel-group, 8-week follow-up, noninferiority trial.

OBJECTIVE: The purpose of this study was to compare the efficacy and tolerability of amlodipine orotate with those of amlodipine besylate in Korean patients with mild to moderate hypertension. METHODS: This multicenter, randomized, double-blind, placebo-controlled, parallel-group study was designed as a noninferiority study. To be included in the study, previously untreated patients had to have a sitting diastolic blood pressure (SiDBP) of 90 to 109 mm Hg. Previously treated patients had to discontinue their current annhypertensive medications and have a baseline SiDBP between 90 and 109 mm Hg after a 2-week washout period. Patients who met the inclusion criteria were randomly assigned to receive 5 mg amlodipine orotate or 5 mg amlodipine besylate for 8 weeks. The medication dose was doubled (10 mg QD for either amlodipine orotate or amlodipine besylate) 4 weeks after enrollment if SiDBP was >or=90 mm Hg. The primary efficacy analysis was noninferiority of the difference in mean trough SiDBP changes from baseline for amlodipin eorotate as compared with amlodipine besylate after 8 weeks of treatment. For the secondary efficacy analysis, 2 other measures were analyzed after 8 weeks of treatment. The SiDBP response rate was defined as an SiDBP measurement<90 mm Hg at the last clinical follow-up visit or an absolute reduction of >or=10 mm Hg in SiDBP from baseline until the last clinical follow-up visit. In addition, noninferiority of the difference in mean trough sitting systolic blood pressure (SiSBP) changes from baseline was analyzed for amlodipine orotate as compared with amlodipine besylate. The drug compliance rate was estimated by pillcount. RESULTS: Eligible patients (n=109; 43 women and 66 men) were randomly assigned to receive amlodipine orotate (n=53) or amlodipine besylate (n=56). No significant differences were found in sex, age, weight, or current smoking between the groups (all, P=NS). The proportion of patients with previous antihypertensive medications was not different between the groups (47.2% [25/53] in the amlodipine orotate group and 50.0% [28/56] in the amlodipine besylate group; P=NS). No significant differences were found in baseline SiDBP (mean [SD], 100 [6] mm Hg [range, 90-109 mm Hg] in the amlodipine orotate group and 100 [6] mm Hg [range, 90-108 mm Hg] in the amlodipine besylate group; P=NS) or in baseline SiSBP (mean [SD], 149 [14] mm Hg [range, 125-179 mm Hg] in the amlodipine orotate group and 146 [10] mm Hg [range, 123-167 mm Hg] in the amlodipine besylate group; p=NS). The mean (SD) changes in SiDBP were -15.6 (6.3) mm Hg for the amlodipine orotate group and -14.5 (5.5) mm Hg for the amlodipine besylate groups was 1.1 (5.9) mm Hg (95% CI, -0.87 to infinity), and because the lower boundary of the 95% CI was greater than -5 mm Hg, amlodipine orotate was considered noninferior to amlodipine besylate. The response rate was 48 of 51 (94.1%) in the amlodipine orotate group compared with 50 (92.6%) of 54 in the amlodipine besylate group after 8 weeks of treatment (P=NS). The mean (SD) compliance rates were 97.6% (3.6%) in the amlodipine orotate group and 96.5% (4.3%) in the amlodipine besylate group (P=NS). The incidence of drug-related adverse events (AEs) was similar between the groups (1/53 [1.9%]) in the amlodipine orotate group vs 4/55 [7.3%] in the amlodipine besylate group; P=NS). The most common drug-related AE overall was peripheral edema (2/55 [3.6%]), and the most common of all the AEs was upper respiratory tract infection (4/55 [7.3%]) in the amlodipine besylate group. The most common drug-related AE was headache (1/53 [1.9%]) in the amlodipine orotate group and peripheral edema (2/55 [3.6%]) in the amlodipine besylate group. No severe AEs were found in either group. CONCLUSION: The reduction in SiDBP after 8 weeks of amlodipine orotate treatment was noninferior to that of amlodipine besylate in these Korean patients with mild to moderate hypertension. The SiDBP response rate and the reduction in SISBP after 8 weeks of amlodipine orotate treatment were not significantly different from those of amlodipine besylate treatment. Both agents were wel tolerated.

[Development of a Web-based Program to Calculate Sample Size for Evaluating the Performance of In Vitro Diagnostic Kits.].

BACKGROUND: Many studies evaluating the performance of in vitro diagnostic kits have been criticized for the lack of reliability. To attain reliability those evaluation studies should be preceded by sample size calculation ensuring statistical power. This study was intended to develop a web-based system to estimate the sample size, which was often neglected because it would require expert knowledge in statistics. METHODS: For sample size calculation, we extracted essential parameters from the performance studies on the 3rd generation anti-hepatitis C virus (HCV) kits reported in the literature. We developed a system with PHP web-script language and MySQL. The statistical models used in this system were as follows; one sample without power consideration (model 1), one sample with power consideration (model 2), and two samples with power consideration (model 3). RESULTS: Among the articles published between 1989 and 2005, 13 articles that evaluated the performance of anti-HCV kits were identified by searching with Medical Subject Headings (MeSH). The diagnostic sensitivity was 83-100% with a median of 145 samples (range; 12-1,091) and the specificity was 97-100% with a median of 1,025 samples (range; 33-4,381). The estimated sample size would be 280 in the model 1, 817 in the model 2, and 1,510 in the model 3, when we set 2% prevalence of HCV infection, 95% sensitivity of a conventional kit, 97% sensitivity of a new kit , 95% significance level (two-sided test), 2% allowable error, and 80% power. CONCLUSIONS: Our study indicates that an insufficient sample size is still a problem in performance evaluation. Our system should be helpful in increasing the reliability of performance evaluation by providing an appropriate sample size.

[Quality Evaluation of the Performance Study of Diagnostic Tests Using STARD Checklist and Meta-Analysis for the Pooled Sensitivity and Specificity of Third Generation Anti-HCV EIA Tests.].

BACKGROUND: The third generation anti-hepatitis C virus (HCV) enzyme immunoassay (EIA) is now in use for screening HCV infection. The aim of this study was to pool the data on the sensitivity and specificity of third generation anti-HCV EIA tests after evaluating the quality of the studies using Standards for Reporting of Diagnostic Accuracy studies (STARD) checklist. METHODS: We searched MEDLINE and PubMed databases using keywords about the accuracy of diagnostic tests for HCV infections. Methodological quality was assessed by two persons with a modified STARD checklist. A heterogeneity test was performed, and in case heterogeneity was present, a sub-group analysis was done. Fixed-effects model was used to obtain pool sensitivity and specificity with 95% confidence intervals (CI). RESULTS: A total of 41 studies from 16 papers were selected. The quality score ranged from 6 to 13 (median 10.5); Inter-observer agreement was 93.62% (k=0.69); and 41 studies revealed heterogeneity. We performed a sub-group analysis with only 28 studies from 13 papers that were evaluated to be of high quality. A subgroup using polymerase chain reaction as the reference test revealed homogeneity and was calculated the pooled sensitivity and specificity of 99.92% (CI 99.77-100.07%) and 99.66% (CI 99.45-99.86%) respectively. Studies on test kits with an increased reactivity to the core region also showed homogeneity in sensitivity and the pooled sensitivity was 99.78% (CI 99.53-100.03%). CONCLUSIONS: For the first time in Korea, the diagnostic accuracy of test kits was evaluated by meta-analysis using STARD checklist. The methodology shown in this study should help extending laboratory medicine to an evidence-based medicine.