PURPOSE: To analyze the causes, success rate, and the changes of Jones tube length in endoscopic conjunctivodacryocystorhinostomy (CDCR) reoperation cases. METHODS: The medical records of 40 patients (41 eyes, a total of 52 cases) who underwent reoperation of CDCR with Jones tube reinsertion using an endoscope from January 2013 to December 2018 were retrospectively reviewed. We analyzed the success rate, causes of reoperation, changes in lengths of used tubes, and the average interval times between operations. RESULTS: The most common cause of reoperation was medial tube migration (53.8%). As in other causes, tube loss (17.3%), obstructions related with conjunctival overgrowth or granulation (13.46%), lateral tube migration (11.53%), iatrogenic removal (1.92%), and acute dacryocystitis (1.92%) followed. In patients with medial tube migration, the average tube length used in reoperations decreased by approximately 1.11 mm compared to prior operations. In cases of lateral tube migration, the average tube length increased approximately 1.00 mm after the reoperation. The success rate of reoperations was 78.04%. The average interval time between the initial operation and the first reoperation was 52 months. In cases with several reoperations, the interval time decreased as the number of reoperations increased. CONCLUSIONS Because medial tube migration was found to be the most common cause of reoperations, it should be considered as a potential problem when performing surgery. The changes in the lengths of inserted Jones tubes were related to certain types of complication, which affected the prognoses. In endoscopic CDCR reoperations, the success rate was favorable. In recurrent cases, the average interval time between reoperations decreased as the number of operations increased.

No abstract available.
Bartonella Infections* ; Bartonella* ; Retinal Vessels* ; Retinaldehyde*

Objectives: ZZThe aim was to find effective vectorization and classification models to predict a psychiatric diagnosis from text-basedmedical records. Methods: ZZElectronic medical records (n = 494) of present illness were collected retrospectively in inpatient admission notes withthree diagnoses of major depressive disorder, type 1 bipolar disorder, and schizophrenia. Data were split into 400 training data and 94 independentvalidation data. Data were vectorized by two different models such as term frequency-inverse document frequency (TF-IDF)and Doc2vec. Machine learning models for classification including stochastic gradient descent, logistic regression, support vectorclassification, and deep learning (DL) were applied to predict three psychiatric diagnoses. Five-fold cross-validation was used to find aneffective model. Metrics such as accuracy, precision, recall, and F1-score were measured for comparison between the models. Results: ZZFive-fold cross-validation in training data showed DL model with Doc2vec was the most effective model to predict the diagnosis(accuracy = 0.87, F1-score = 0.87). However, these metrics have been reduced in independent test data set with final workingDL models (accuracy = 0.79, F1-score = 0.79), while the model of logistic regression and support vector machine with Doc2vec showedslightly better performance (accuracy = 0.80, F1-score = 0.80) than the DL models with Doc2vec and others with TF-IDF. Conclusions ZZThe current results suggest that the vectorization may have more impact on the performance of classification thanthe machine learning model. However, data set had a number of limitations including small sample size, imbalance among the category,and its generalizability. With this regard, the need for research with multi-sites and large samples is suggested to improve the machinelearning models.

Objectives: ZZThe aim was to find effective vectorization and classification models to predict a psychiatric diagnosis from text-basedmedical records. Methods: ZZElectronic medical records (n = 494) of present illness were collected retrospectively in inpatient admission notes withthree diagnoses of major depressive disorder, type 1 bipolar disorder, and schizophrenia. Data were split into 400 training data and 94 independentvalidation data. Data were vectorized by two different models such as term frequency-inverse document frequency (TF-IDF)and Doc2vec. Machine learning models for classification including stochastic gradient descent, logistic regression, support vectorclassification, and deep learning (DL) were applied to predict three psychiatric diagnoses. Five-fold cross-validation was used to find aneffective model. Metrics such as accuracy, precision, recall, and F1-score were measured for comparison between the models. Results: ZZFive-fold cross-validation in training data showed DL model with Doc2vec was the most effective model to predict the diagnosis(accuracy = 0.87, F1-score = 0.87). However, these metrics have been reduced in independent test data set with final workingDL models (accuracy = 0.79, F1-score = 0.79), while the model of logistic regression and support vector machine with Doc2vec showedslightly better performance (accuracy = 0.80, F1-score = 0.80) than the DL models with Doc2vec and others with TF-IDF. Conclusions ZZThe current results suggest that the vectorization may have more impact on the performance of classification thanthe machine learning model. However, data set had a number of limitations including small sample size, imbalance among the category,and its generalizability. With this regard, the need for research with multi-sites and large samples is suggested to improve the machinelearning models.

PURPOSE: To investigate the patterns of locoregional recurrence of pathologic T3N0 (pT3N0) lower rectal cancer omitting postoperative radiotherapy (RT) and explore the potential of modification of a RT field. MATERIALS AND METHODS: From Jan 2003 to Nov 2011, 35 patients omitting preoperative or postoperative RT for pT3N0 lower rectal cancer were included. We defined the lower rectal cancer as the tumor with the inferior margin located below the virtual line-a convergent level between rectal wall and levator ani muscle. All patients had radiologic examinations for recurrence evaluation during the follow-up duration. RESULTS: The median follow-up duration was 66.4 months (range, 1.4 to 126.1 months). Eight (22.9%) of the 35 patients had recurrence. Three (8.6%) was local recurrence (LR) only, 3 (8.6%) was distant metastasis (DM) only, and 2 (5.7%) was LR with DM. All LR were located at primary tumor sites. The overall survival rate, LR-free survival rate, and DM-free survival rate at 5 years was 79.8%, 83%, and 87%, respectively. All LR developed from tumors over 5 cm. However, there was no statistical significance (p = 0.065). There was no other risk factor for LR. CONCLUSION: Even though the patients included in this study had pathologically favorable pT3N0 rectal cancer, LR developed in 14.3% of patients. Most of the LR was located at primary tumor sites prior to surgery. Based on these findings, it might seem reasonable to consider postoperative RT with a smaller radiation field to the primary tumor site rather than the conventional whole pelvic irradiation.
Follow-Up Studies ; Humans ; Muscles ; Neoplasm Metastasis ; Neoplasm Recurrence, Local ; Radiotherapy, Adjuvant ; Rectal Neoplasms ; Recurrence ; Risk Factors ; Survival Rate

PURPOSE: The prevalence rates of metabolic syndrome (MetS) and percentile distribution of insulin resistance (IR) among Korean children and adolescents were investigated. The cutoff values of IR were calculated to identify high-risk MetS groups. METHODS: Data from 3,313 Korean subjects (1,756 boys and 1,557 girls, aged 10–18 years) were included from the Korean National Health and Nutrition Examination Survey conducted during 2007–2010. Three different sets of criteria for MetS were used. Indirect measures of IR were homeostasis model assessment (HOMA-IR) and triglyceride and glucose (TyG) index. The cutoff values of the HOMA-IR and TyG index were obtained from the receiver operation characteristic curves. RESULTS: According to the MetS criteria of de Ferranti el al., Cook et al., and the International Diabetes Federation, the prevalence rates in males and females were 13.9% and 12.3%, 4.6% and 3.6%, and 1.4% and 1.8%, respectively. Uses these 3 criteria, the cutoff values of the HOMA-IR and TyG index were 2.94 and 8.41, 3.29 and 8.38, and 3.54 and 8.66, respectively. The cutoff values using each of the 3 criteria approximately corresponds to the 50th–75th, 75th, and 75th–90th percentiles of normal HOMA-IR and TyG index levels. CONCLUSION: This study describes the prevalence rates of MetS in Korean children and adolescents, an index of IR, and the cutoff values for MetS with the aim of detecting high-risk groups. The usefulness of these criteria needs to be verified by further evaluation.
Adolescent* ; Child* ; Female ; Glucose ; Homeostasis ; Humans ; Insulin Resistance* ; Insulin* ; Korea ; Male ; Nutrition Surveys ; Prevalence ; Triglycerides

PURPOSE: The prevalence rates of metabolic syndrome (MetS) and percentile distribution of insulin resistance (IR) among Korean children and adolescents were investigated. The cutoff values of IR were calculated to identify high-risk MetS groups. METHODS: Data from 3,313 Korean subjects (1,756 boys and 1,557 girls, aged 10–18 years) were included from the Korean National Health and Nutrition Examination Survey conducted during 2007–2010. Three different sets of criteria for MetS were used. Indirect measures of IR were homeostasis model assessment (HOMA-IR) and triglyceride and glucose (TyG) index. The cutoff values of the HOMA-IR and TyG index were obtained from the receiver operation characteristic curves. RESULTS: According to the MetS criteria of de Ferranti el al., Cook et al., and the International Diabetes Federation, the prevalence rates in males and females were 13.9% and 12.3%, 4.6% and 3.6%, and 1.4% and 1.8%, respectively. Uses these 3 criteria, the cutoff values of the HOMA-IR and TyG index were 2.94 and 8.41, 3.29 and 8.38, and 3.54 and 8.66, respectively. The cutoff values using each of the 3 criteria approximately corresponds to the 50th–75th, 75th, and 75th–90th percentiles of normal HOMA-IR and TyG index levels. CONCLUSION: This study describes the prevalence rates of MetS in Korean children and adolescents, an index of IR, and the cutoff values for MetS with the aim of detecting high-risk groups. The usefulness of these criteria needs to be verified by further evaluation.
Adolescent* ; Child* ; Female ; Glucose ; Homeostasis ; Humans ; Insulin Resistance* ; Insulin* ; Korea ; Male ; Nutrition Surveys ; Prevalence ; Triglycerides

Background: For gendered innovations to be streamlined in health research, researchers need tools to measure gender variables quantitatively. Because gender is a sociocultural concept which varies by culture, it is important to check feasibility of using tools developed in western countries before using them in Korea. The objective of this study is to conduct a pilot study for adapting a questionnaire measuring gender variables from the literature for the Korean population. Methods: The research was conducted in two steps: 1) modification of questionnaire by an expert group, and 2) an on-line survey of 3,000 adults, in three age groups, balanced in sex. The questionnaire was composed of 7 gender variables with 2 to 6 items, constituting a total of 24 5-point Likert scale items. Gender variable scores were calculated as the average of Z-scores of items pertaining to each variable. Results: Participants were 45.6% men, 43.9% women, and 10.5% other gender. Scores of five gender variables-caregiver tension, work tension, independence, emotional intelligence, and discrimination-were significantly different among sex and gender groups. Among gender variables, work strain, caregiver strain, and discrimination had notable influences on health status and health-related behaviors among survey participants. Conclusions Several gender variables were related to health status and health behaviors in Korean adults. This is the first attempt in Korea to quantify gender characteristics to be used for health research. Further studies are needed to verify the questionnaire and methods developed in this study for use with diagnosed disease outcomes.

Purpose: This study investigated the relationship between the visceral fat area (VFA) and clinico-pathological outcomes in patients with colorectal cancer (CRC). Methods: This retrospective study included 204 patients who underwent anthropometric measurement by bioelectrical impedance analysis (BIA) before surgical treatment for CRC between January 2016 and June 2020. Results: According to the average value of the visceral fat area, 119 (58.3%) patients had a low visceral fat area, and 85 (59.1%) patients had a high visceral fat area. Patients with visceral obesity showed a higher BMI compared to patients without visceral obesity, (21.8±1.9 vs. 25.7±2.5, P＜0.001). There was no significant difference in the overall perioperative outcomes including total operation time, time to gas out, sips of water, soft diet, hospital stay, and morbidity between patients in the low and high VFA groups. We divided patients into two subgroups according to the degree of cancer progression and more advanced cases with low VFA showed significantly more total and positive retrieved lymph nodes (LNs) (20.9±10.3 vs. 16.1±7.1, P=0.021 and 3.3±2.9 vs. 2.2±2.3, P=0.019, respectively) and a higher proportion of more than 12 retrieved LNs compared to patients with a high VFA (95.1% vs. 90.0%, P=0.047). Body composition analysis showed that phase angle, muscle composition, and body fluid composition were not statistically different between the two groups. However, body fat mass was statistically higher in the high VFA group (22.0±4.6 vs. 12.8±3.1, P＜0.001). Conclusion Visceral obesity measured by BIA showed lower total and positive retrieved LNs and was not associated with adverse peri-operative outcomes, inflammatory and nutritional, and pathologic outcomes for CRC.

Purpose: This study investigated the relationship between the visceral fat area (VFA) and clinico-pathological outcomes in patients with colorectal cancer (CRC). Methods: This retrospective study included 204 patients who underwent anthropometric measurement by bioelectrical impedance analysis (BIA) before surgical treatment for CRC between January 2016 and June 2020. Results: According to the average value of the visceral fat area, 119 (58.3%) patients had a low visceral fat area, and 85 (59.1%) patients had a high visceral fat area. Patients with visceral obesity showed a higher BMI compared to patients without visceral obesity, (21.8±1.9 vs. 25.7±2.5, P＜0.001). There was no significant difference in the overall perioperative outcomes including total operation time, time to gas out, sips of water, soft diet, hospital stay, and morbidity between patients in the low and high VFA groups. We divided patients into two subgroups according to the degree of cancer progression and more advanced cases with low VFA showed significantly more total and positive retrieved lymph nodes (LNs) (20.9±10.3 vs. 16.1±7.1, P=0.021 and 3.3±2.9 vs. 2.2±2.3, P=0.019, respectively) and a higher proportion of more than 12 retrieved LNs compared to patients with a high VFA (95.1% vs. 90.0%, P=0.047). Body composition analysis showed that phase angle, muscle composition, and body fluid composition were not statistically different between the two groups. However, body fat mass was statistically higher in the high VFA group (22.0±4.6 vs. 12.8±3.1, P＜0.001). Conclusion Visceral obesity measured by BIA showed lower total and positive retrieved LNs and was not associated with adverse peri-operative outcomes, inflammatory and nutritional, and pathologic outcomes for CRC.