• Tuberculosis and Respiratory Diseases
• /
• v.57 no.1
• /
• pp.19-24
• /
• 2004

Background : It is difficult to differentiate between tuberculous pneumonia and Community Acquired Pneumonia, so the diagnosis and treatment of tuberculous pneumonia can be delayed frequently. In this study, we attempted to retrospectively evaluate the clinical and radiologic characteristics of tuberculous pneumonia. Methods : We conducted a retrospective analysis of clinical characteristics of 58 patients diagnosed with tuberculous pneumonia from Nov. 1997 to May 2001 at Korea university kuro hospital. Result : The male to female ratio was 1:1 and the mean age at diagnosis was $54.5{\pm}18.6$ years. Fifty five patients were confirmed microbiologically and three patients pathologically. There were 20 patients(34.5%) who had diabetes mellitus(8cases), chronic obstructive pulmonary disease(3cases), malignancy(3cases), bronchiectasis(2cases), chronic renal failure(1cases) or long term history of corticosteroid treatment(3cases). Many patients had multilobar infiltration in chest X-ray, dominantly in the lower lobe. thirty two patients(55.2%) had infiltration in more than 2 lobes and 5 patients in more than 4 lobes. The significant correlation between the diabetes mellitus and the infiltrated Rt lower lobe(RLL) was found on the borders of confidence limit.(P=0.07<0.1). There was significant correlation between woman and infiltrated lobe(RML, RLL, LLL) excluding the both upper lobe(P=0.029). Conclusion : We must consider tuberculous pneumonia when lobar pneumonia with consolidation resistant to antibiotics, especially in the patients who have diabetes mellitus, chronic obstructive pulmonary disease, malignancy, bronchiectsis, chronic renal failure or long term history of corticosteroid treatment.

• Journal of Chest Surgery
• /
• v.38 no.12 s.257
• /
• pp.828-834
• /
• 2005

Background: Off-pump coronary artery bypass grafting (OPCAB) has shown better outcome in chronic renal failure (CRF) patients by avoiding the effects of cardiopulmonary bypass. We evaluated renal function after OPCAB in CRF patients. Material and Method: 656 patients underwent OPCAB between January, 2001 and December, 2004. Data were collected in 26 CRF patients (Cr > 1.7 mg/dL). Preoperative/postoperative creatinine (Cr) levels, creatinine clearance and postoperative data were evaluated. We divided the patients into group 1 (Cr < 3 mg/dL) and group 2 (Cr $\geq$ 3 mg/dL). Result: Three patients started dialysis after surgery. Preoperative mean creatinine level (4.19$\pm$3.4 mg/dL) was elevated to 4.36$\pm$2.7 mg/dL at the third postoperative day and decreased below Preoperative level at the fifth postoperative day. In group 1 (mean Cr level=1.87$\pm$0.25 mg/dL), Cr level reached its peak level of 2.19$\pm$0.52 mg/dL at the fourth postoperative day (p=0.017), with subsequent decrease. Patients without pre- or postoperative dialysis (n=15) showed peak Cr elevation on postoperative day four (p=0.017) and subsequent decrease (p=0.01). Postoperative creatinine clearance showed reverse correlation with creatinine level. Conclusion: Creatinine level was elevated at third/fourth postoperative day, but decreased 5 days after surgery. Thus, if urgent dialysis is not indicated, postoperative renal replacement therapy in CRF patients may be better to be considered after four days observation.

• Tuberculosis and Respiratory Diseases
• /
• v.60 no.5
• /
• pp.540-547
• /
• 2006

Background : Vital stability and right side heart failure are major prognostic factors of acute pulmonary thromboembolism. While it is important to recognize right side heart failure, it is often difficult in real practice. Recently, several studies have described early diagnostic tools for detecting right side heart failure including echocardiography and biochemical markers. This study, we evaluated the prognostic role of the B-type natriuretic peptide (BNP) in an acute pulmonary thromboembolism. Methods : Thirty-four patients with a diagnosis of acute pulmonary thromboembolism were enrolled in the study. The BNP levels were measured and echocardiography was performed at the Emergency Department. Data on the prognostic factors including ventilatory support, vital stability, pulmonary artery pressure, degree of tricuspid valve regurgitation, complications and death was collected from the patients' medical records. The patients with an acute pulmonary thromboembolism were divided into two groups based on the vital stability and the BNP level and the cutoff values and prognostic factors of the two groups were compared. Results : The predictors of the vital stability that influence the prognosis of patients with acute pulmonary thromboembolism were the BNP level, ventilatory support and death. The plasma BNP levels showed a strong correlation with the vital stability, ventilatory support, thrombolytic therapy and death. When the BNP cutoff level was set to 377.5 pg/dl in a ROC curve, the sensitivity and the specificity for differentiating between the groups with stable or unstable vital signs was 100% and 90%, respectively. Conclusion : This study indicates that a measurement of the plasma BNP levels may be a useful prognostic marker in patients with an acute pulmonary thrombo-embolism.

• Tuberculosis and Respiratory Diseases
• /
• v.61 no.3
• /
• pp.265-272
• /
• 2006

Background: Chest wall deformities such as kyphoscoliosis, thoracoplasty, and fibrothorax cause ventilatory insufficiency that can lead to chronic respiratory failure, with recurrent fatal acute respiratory failure(ARF). This study evaluated the frequency and outcome of ARF, the physiologic status, and the long-term prognosis of these patients. Methods: Twenty-nine patients with chest wall disorders, who experienced the first requirement of ventilatory support from ARF were examined. The mortality and recurrence rate of ARF, the pulmonary functions with arterial blood gas analysis, the efficacy of home oxygen therapy, and the long-term survival rate were investigated. Results: 1) The mortality of the first ARF was 24.1%. ARF recurred more than once in 72.7% of the remaining 22 patients, and overall rate of successful weaning was 73.2%. 2) Twenty-two patients who recovered from the first ARF showed a restrictive ventilatory impairment with a mean FVC and TLC of 37.2% and 62.4 % of predicted value, respectively, and a mean $PaCO_{2}$ of 57mmHg. Among the parameters of pulmonaty functions. the FVC(p=0.01) and VC(p=0.02) showed a significant correlation with the $PaCO_{2}$ level. 3) There were no significant differences between the patients treated with conservative medical treatment only and those with additional home oxygen therapy due to significant hypoxemia in the patients with recurrent ARF and the mortality. 4) The 1, 3, 5-year survival rates were 75%, 66%, and 57%, respectively, in the 20 patients who had recovered from the first ARF, excluding the two patients managed by non-invasive nocturnal ventilatory support. Conclusion: These results suggest that active ventilatory support should be provided to patients with ARF and chest wall disorders. However, considering recurrent ARF and weak effect of home oxygen therapy, non-invasive domiciliary ventilation is recommended in those patients with these conditions to achieve a better long-term prognosis.

• Radiation Oncology Journal
• /
• v.8 no.2
• /
• pp.207-212
• /
• 1990

To determine the correlation between the response to induction chemotherapy and subsequent radiotherapy we analyzed the clinical records of 60 patients with locally advanced carcinoma of the head and neck retrospectively who had completed a full course ($2\~3$ cycle) of induction chemotherapy and curative radiotherapy in Korea Cancer Center Hospital between 1986 and 1989. Chemotherapy was administeredd with CDDP+Bleomycin (BP) in 20, CDDP+5-FU (FP) in 37, and hybrid of BP and FP in three patients. Radiotherapy was giver conventionally with a dose of 65 to 75 Gy or more over seven to eight weeks according to the size of lesion. Response rates following induction chemotherapy were $80\%$ for the tumors and $879\%$ for the nodes whereas complete reponse rates were $12\%\;and\;13\%$, respectively. Six months after radiotherapy $67\%$ of the tumors and $77\%$ of the nodes achieved a complete response. Among the 48 tumor responders and the 31 nodal responders to chemotherapy,39 ($81\%$) and 28 ($90\%$), respectively, achieved complete response after radiotherapy. Thus, whether or not the tumor and node respond to induction chemotherapy was predictive of the response to subsequent radiotherapy (p<0.0005 in tumor, p<0.0001 in node). By reanalyzing according to disease subsets (i.e. primary site, T-stage, N-stage) this relationship was not observed at T1-T2 disease (p>0.3). Therefore the tumor or node's response to induction chemotherapy is a predictor for subsequent radiotherapy except in T1-T2 tumors, and complete response to radiotherapy can be expected despite the failure of induction chemotherapy in $T_1-T_2$ tumors.

• Tuberculosis and Respiratory Diseases
• /
• v.48 no.5
• /
• pp.773-780
• /
• 2000

Background : pH measurement is an important test in assessing the etiology of pleurisy and in identifying complicated parapneumonic effusion. Although the blood gas analyzer is the gold standard method' for pleural pH measurement, pH meter & pH strip methods are also used for this purpose interchangably. However, the correlation among the pH data measured by the three different methods needs to be evaluated. In this study, we measured the pH of pleural fluid with the three different methods respectively and evaluated the correlation among the measured data. Methods : From August 1999 to March 2000, we measured the pleural fluid pH in 34 clinical samples with three methods-blood gas analyzer, pH meter, and pH strip. In the blood gas analyzer and pH meter methods, the temperature of pleural fluid was maintained around $0^{\circ}C$ in air-tight condition before analysis and measurement was performed within 30 minutes after collection. As for the pH strip method, the pleural fluid pH was checked in the ward immediately after tapping and in the clinical laboratory of our hospital. This part is unclear. Results : The causes of pleural effusion were tuberculosis pleurisy in 16 cases, malignant pleural effusion 5 cases, parapneumonic effusion 9 cases, empyema 3 cases, and congestive heart failure 1 case. The pH of pleural fluid (mean$\pm$SD) was 7.34$\pm$0.12 with blood gas analyser, 7.52$\pm$0.25 with pH meter, 7.37$\pm$0.16 with pH strip of immediate measurement and 6.93$\pm$0.201 with pH strip of delayed measurement. The pH measured by delayed pH strip measurement was lower than those of other methods (p<0.05). The correlation of the results between the blood gas analyzer and pH meter(p=0.002, r=0.518) and the blood gas analyzer and pH strip of immediate measurement(p<0.001, r=0.607). Conclusion : In the determination of pH of pleural fluid, pH strip method could be a simple and reliable method under immediate measurement conditions after pleural fluid tapping.

• Tuberculosis and Respiratory Diseases
• /
• v.41 no.2
• /
• pp.111-119
• /
• 1994

Background: Because of the common etiologic factor, such as smoking, lung cancer and chronic obstructive pulmonary disease are often present in the same patient. The preoperative prediction of remaining pulmonary function after the resectional surgery is very important to prevent serious complication and postoperative respiratory failure. $^{99m}Tc$-MAA perfusion scan has been used for the prediction of postoperative pulmonary function, but it may be inaccurate in case of large V/Q mismatching. We compared $^{99m}Tc$-DTPA radioaerosol inhalation scan with $^{99m}Tc$-MAA perfusion scan in predicting postoperative lung function. Method: Preoperative inhalation scan and/or perfusion scan were performed and pulmonary function test were performed preoperatively and 2 month after operation. We predicted the postoperative pulmonary functions using the following equations. Postpneurnonectomy $FEV_1$=Preop $FEV_1x%$ of total function of lung to remain Postlobectomy $FEV_1$=Preop $FEV_1{\times}$(% of total 1-function of affected lung${\times}$$\frac{Number\;of\;segments\;to\;be\;resected}{Number\;of\;segments\;of\;affected\;lung})$ Results: 1) The inhalation scan showed good correlations between measured and predicted $FEV_1$, FVC and $FEF_{25-75%}$. (correlation coefficiency; 0.94, 0.91, 0.87 respectively). 2) The perfusion scan also showed good correlations between measured and predicted $FEV_1$, FVC and $FEF_{25-75%}$. (correlation coefficiency; 0.86, 0.72, 0.87 respectively). 3) Among three parameters, $FEV_1$ showed the best correlations in the prediction by lung scans. 4) Comparison between inhalation scan and perfusion scan in predicting pulmonary function did not show any significant differneces except FVC. Conclusion: The inhalation scan and perfusion scan are very useful in the prediction of postoperative lung function and don't make a difference in the prediction of pulmonary function a1though the former showed a better correlation in FVC.

• Radiation Oncology Journal
• /
• v.17 no.2
• /
• pp.136-140
• /
• 1999

Purpose : Prostate specific antigen (PSA) is a useful tumor marker, which is widely used as a diagnostic index and predictor of both treatment and follow-up result in prostate cancer. A prospective analysis was carried out to obtain the period of PSA normalization and the half life of PSA and to analyze the factors influencing the period of PSA normalization. The PSA level was checked before and serially after radical radiotherapy. Materials and Method : Twen쇼 patients with clinically localized prostate cancer who underwent radical external beam radiotherapy were enrolled in this study. Accrual period was from April 1993 to May 1998. Median follow-up period was 20 months. Radiotherapy was given to whole pelvis followed by a boost to prostate. Dose range for the whole pelvis was from 45 Gy to 50 Gy and boost dose to prostate, from 14 Gy to 20 Gy. The post-irradiation PSA normal value was under 3.0 ng/ml. The physical examination and serum PSA level evaluation were performed at 3 month interval in the first one year, and then at every 4 to 6 months. Results : PSA value was normalized in nineteen patients (95%) within 12 months. The mean period of PSA normalization was 5.3 (${\pm}$2.7) months. The half life of PSA Of the nonfailing patients was 2.1 (${\pm}$0.9) month. The nadir PSA level Of the nonfailing Patients waS 0.8 (${\pm}$0.5) ng/ml. The period of PSA normalization had the positive correlation with pretreatment PSA level (R$^{2}$=0.468). The nadir PSA level had no definite positive correlation with the pretreatment PSA level (R$^{2}$=0.075). The half life of serum PSA level also had no definite correlation with pretreatment PSA level (R$^{2}$=0.029). Conclusion :The PSA level was mostly normalized within 8 months (85%). If it has not normalized within 12 months, we should consider the residual disease in prostate or distant metastasis. In 2 patients, the PSA level increased 6 months or 20 months before clinical disease was detected. So the serum PSA level can be used as early diagnostic indicator of treatment failure.

• Journal of Intelligence and Information Systems
• /
• v.16 no.4
• /
• pp.99-112
• /
• 2010

Ensemble learning is a method for improving the performance of classification and prediction algorithms. It is a method for finding a highly accurateclassifier on the training set by constructing and combining an ensemble of weak classifiers, each of which needs only to be moderately accurate on the training set. Ensemble learning has received considerable attention from machine learning and artificial intelligence fields because of its remarkable performance improvement and flexible integration with the traditional learning algorithms such as decision tree (DT), neural networks (NN), and SVM, etc. In those researches, all of DT ensemble studies have demonstrated impressive improvements in the generalization behavior of DT, while NN and SVM ensemble studies have not shown remarkable performance as shown in DT ensembles. Recently, several works have reported that the performance of ensemble can be degraded where multiple classifiers of an ensemble are highly correlated with, and thereby result in multicollinearity problem, which leads to performance degradation of the ensemble. They have also proposed the differentiated learning strategies to cope with performance degradation problem. Hansen and Salamon (1990) insisted that it is necessary and sufficient for the performance enhancement of an ensemble that the ensemble should contain diverse classifiers. Breiman (1996) explored that ensemble learning can increase the performance of unstable learning algorithms, but does not show remarkable performance improvement on stable learning algorithms. Unstable learning algorithms such as decision tree learners are sensitive to the change of the training data, and thus small changes in the training data can yield large changes in the generated classifiers. Therefore, ensemble with unstable learning algorithms can guarantee some diversity among the classifiers. To the contrary, stable learning algorithms such as NN and SVM generate similar classifiers in spite of small changes of the training data, and thus the correlation among the resulting classifiers is very high. This high correlation results in multicollinearity problem, which leads to performance degradation of the ensemble. Kim,s work (2009) showedthe performance comparison in bankruptcy prediction on Korea firms using tradition prediction algorithms such as NN, DT, and SVM. It reports that stable learning algorithms such as NN and SVM have higher predictability than the unstable DT. Meanwhile, with respect to their ensemble learning, DT ensemble shows the more improved performance than NN and SVM ensemble. Further analysis with variance inflation factor (VIF) analysis empirically proves that performance degradation of ensemble is due to multicollinearity problem. It also proposes that optimization of ensemble is needed to cope with such a problem. This paper proposes a hybrid system for coverage optimization of NN ensemble (CO-NN) in order to improve the performance of NN ensemble. Coverage optimization is a technique of choosing a sub-ensemble from an original ensemble to guarantee the diversity of classifiers in coverage optimization process. CO-NN uses GA which has been widely used for various optimization problems to deal with the coverage optimization problem. The GA chromosomes for the coverage optimization are encoded into binary strings, each bit of which indicates individual classifier. The fitness function is defined as maximization of error reduction and a constraint of variance inflation factor (VIF), which is one of the generally used methods to measure multicollinearity, is added to insure the diversity of classifiers by removing high correlation among the classifiers. We use Microsoft Excel and the GAs software package called Evolver. Experiments on company failure prediction have shown that CO-NN is effectively applied in the stable performance enhancement of NNensembles through the choice of classifiers by considering the correlations of the ensemble. The classifiers which have the potential multicollinearity problem are removed by the coverage optimization process of CO-NN and thereby CO-NN has shown higher performance than a single NN classifier and NN ensemble at 1% significance level, and DT ensemble at 5% significance level. However, there remain further research issues. First, decision optimization process to find optimal combination function should be considered in further research. Secondly, various learning strategies to deal with data noise should be introduced in more advanced further researches in the future.

• Restorative Dentistry and Endodontics
• /
• v.32 no.3
• /
• pp.208-221
• /
• 2007

The fracture toughness test is believed as a clinically relevant method for assessing the fracture resistance of the dentinal restoratives. The objectives of this study were to measure the fracture toughness $(K_{1C})$ and microtensile bond strength of dentin-resin composite interface and compare their relationship for their use in evaluation of the integrity of the dentin-resin bond. A minimum of six short-rod specimens for fracture toughness test and fifteen specimens for microtensile bond strength test was fabricated for each group of materials used. After all specimens storing for 24 hours in distilled water at $37^{\circ}C$, they were tensile-loaded with an EZ tester universal testing machin. Statistical analysis was performed using ANOVA and Tukey's test at the 95% confidence level, Pearson's coefficient was used to verify the correlation between the mean of fracture toughness and microtensile bond strength. FE-SEM was employed on fractured surface to describe the crack propagation. Fracture toughness value of Clearfil SE Bond (SE) was the highest, followed by Adper Single Bond 2 (SB), OptiBond Solo (OB), ONE-STEP PLUS (0S), ScotchBond Multi-purpose (SM) and there was significant difference between SE and other 4 groups (p < 0.05). There were, however, no significant difference among SB, OB, OS, SM (p > 0.05). Microtensile bond strength of SE was the highest, followed by SB, OB, SM, OS and OS only showed significant lower value (p < 0.05). There was no correlation between fracture toughness and microtensile bond strength values. FE-SEM examination revealed that dentin bonding agent showed different film thickness and different failure pattern according to the film thickness. From the limited results of this study, it was noted that there was statistically no correlation between K1C and ${\mu}TBS$. We can conclude that for obtaining the reliability of bond strength test of dentin bonding agent, we must pay more attention to the test procedure and its profound scrutiny.