Background: This study aimed to evaluate the applicability of deep learning technology to thyroid ultrasound images for classification of thyroid nodules. Methods: This retrospective analysis included ultrasound images of patients with thyroid nodules investigated by fine-needle aspiration at the thyroid clinic of a single center from April 2010 to September 2012. Thyroid nodules with cytopathologic results of Bethesda category V (suspicious for malignancy) or VI (malignant) were defined as thyroid cancer. Multiple deep learning algorithms based on convolutional neural networks (CNNs) —ResNet, DenseNet, and EfficientNet—were utilized, and Siamese neural networks facilitated multi-view analysis of paired transverse and longitudinal ultrasound images. Results: Among 1,048 analyzed thyroid nodules from 943 patients, 306 (29%) were identified as thyroid cancer. In a subgroup analysis of transverse and longitudinal images, longitudinal images showed superior prediction ability. Multi-view modeling, based on paired transverse and longitudinal images, significantly improved the model performance; with an accuracy of 0.82 (95% confidence intervals [CI], 0.80 to 0.86) with ResNet50, 0.83 (95% CI, 0.83 to 0.88) with DenseNet201, and 0.81 (95% CI, 0.79 to 0.84) with EfficientNetv2_ s. Training with high-resolution images obtained using the latest equipment tended to improve model performance in association with increased sensitivity. Conclusion CNN algorithms applied to ultrasound images demonstrated substantial accuracy in thyroid nodule classification, indicating their potential as valuable tools for diagnosing thyroid cancer. However, in real-world clinical settings, it is important to aware that model performance may vary depending on the quality of images acquired by different physicians and imaging devices.

Background: This study aimed to evaluate the applicability of deep learning technology to thyroid ultrasound images for classification of thyroid nodules. Methods: This retrospective analysis included ultrasound images of patients with thyroid nodules investigated by fine-needle aspiration at the thyroid clinic of a single center from April 2010 to September 2012. Thyroid nodules with cytopathologic results of Bethesda category V (suspicious for malignancy) or VI (malignant) were defined as thyroid cancer. Multiple deep learning algorithms based on convolutional neural networks (CNNs) —ResNet, DenseNet, and EfficientNet—were utilized, and Siamese neural networks facilitated multi-view analysis of paired transverse and longitudinal ultrasound images. Results: Among 1,048 analyzed thyroid nodules from 943 patients, 306 (29%) were identified as thyroid cancer. In a subgroup analysis of transverse and longitudinal images, longitudinal images showed superior prediction ability. Multi-view modeling, based on paired transverse and longitudinal images, significantly improved the model performance; with an accuracy of 0.82 (95% confidence intervals [CI], 0.80 to 0.86) with ResNet50, 0.83 (95% CI, 0.83 to 0.88) with DenseNet201, and 0.81 (95% CI, 0.79 to 0.84) with EfficientNetv2_ s. Training with high-resolution images obtained using the latest equipment tended to improve model performance in association with increased sensitivity. Conclusion CNN algorithms applied to ultrasound images demonstrated substantial accuracy in thyroid nodule classification, indicating their potential as valuable tools for diagnosing thyroid cancer. However, in real-world clinical settings, it is important to aware that model performance may vary depending on the quality of images acquired by different physicians and imaging devices.

Background: This study aimed to evaluate the applicability of deep learning technology to thyroid ultrasound images for classification of thyroid nodules. Methods: This retrospective analysis included ultrasound images of patients with thyroid nodules investigated by fine-needle aspiration at the thyroid clinic of a single center from April 2010 to September 2012. Thyroid nodules with cytopathologic results of Bethesda category V (suspicious for malignancy) or VI (malignant) were defined as thyroid cancer. Multiple deep learning algorithms based on convolutional neural networks (CNNs) —ResNet, DenseNet, and EfficientNet—were utilized, and Siamese neural networks facilitated multi-view analysis of paired transverse and longitudinal ultrasound images. Results: Among 1,048 analyzed thyroid nodules from 943 patients, 306 (29%) were identified as thyroid cancer. In a subgroup analysis of transverse and longitudinal images, longitudinal images showed superior prediction ability. Multi-view modeling, based on paired transverse and longitudinal images, significantly improved the model performance; with an accuracy of 0.82 (95% confidence intervals [CI], 0.80 to 0.86) with ResNet50, 0.83 (95% CI, 0.83 to 0.88) with DenseNet201, and 0.81 (95% CI, 0.79 to 0.84) with EfficientNetv2_ s. Training with high-resolution images obtained using the latest equipment tended to improve model performance in association with increased sensitivity. Conclusion CNN algorithms applied to ultrasound images demonstrated substantial accuracy in thyroid nodule classification, indicating their potential as valuable tools for diagnosing thyroid cancer. However, in real-world clinical settings, it is important to aware that model performance may vary depending on the quality of images acquired by different physicians and imaging devices.

Background: This study aimed to evaluate the applicability of deep learning technology to thyroid ultrasound images for classification of thyroid nodules. Methods: This retrospective analysis included ultrasound images of patients with thyroid nodules investigated by fine-needle aspiration at the thyroid clinic of a single center from April 2010 to September 2012. Thyroid nodules with cytopathologic results of Bethesda category V (suspicious for malignancy) or VI (malignant) were defined as thyroid cancer. Multiple deep learning algorithms based on convolutional neural networks (CNNs) —ResNet, DenseNet, and EfficientNet—were utilized, and Siamese neural networks facilitated multi-view analysis of paired transverse and longitudinal ultrasound images. Results: Among 1,048 analyzed thyroid nodules from 943 patients, 306 (29%) were identified as thyroid cancer. In a subgroup analysis of transverse and longitudinal images, longitudinal images showed superior prediction ability. Multi-view modeling, based on paired transverse and longitudinal images, significantly improved the model performance; with an accuracy of 0.82 (95% confidence intervals [CI], 0.80 to 0.86) with ResNet50, 0.83 (95% CI, 0.83 to 0.88) with DenseNet201, and 0.81 (95% CI, 0.79 to 0.84) with EfficientNetv2_ s. Training with high-resolution images obtained using the latest equipment tended to improve model performance in association with increased sensitivity. Conclusion CNN algorithms applied to ultrasound images demonstrated substantial accuracy in thyroid nodule classification, indicating their potential as valuable tools for diagnosing thyroid cancer. However, in real-world clinical settings, it is important to aware that model performance may vary depending on the quality of images acquired by different physicians and imaging devices.

Purpose: Biliary atresia (BA) is a rare but progressive cholangiopathy and the leading cause of pediatric liver transplantation worldwide. The Korean Association of Pediatric Surgeons (KAPS) has conducted three national surveys (2001, 2011, and 2023) to assess long-term trends in the diagnosis, treatment, and outcomes of BA. This study provides a comparative analysis of the 2nd and 3rd national surveys, with reference to selected findings from the 1st survey. Methods: This study included 453 patients from the 3rd national survey (2011–2021) and 435 patients from the 2nd survey (2001–2010), all of whom underwent Kasai portoenterostomy. Data were collected via electronic case report forms from pediatric surgical centers nationwide. Comparisons were made regarding demographics, clinical features, diagnostic patterns, operative details, follow-up outcomes, and survival. Kaplan–Meier analysis was used to evaluate long-term survival. Results: The mean number of BA patients per year remained stable between surveys (43.5 in the 2nd, 41.18 in the 3rd), though centralization of care increased, with 61.5% of cases managed by two major institutions in the 3rd survey. The median age at surgery decreased, and the use of preoperative imaging (especially magnetic resonance cholangiopancreatography) increased. The 10-year native liver survival rate declined from 59.8% to 53.7%, while overall 10-year survival improved slightly (92.9% to 93.2%). Postoperative complications, such as cholangitis and liver failure, persisted but were better categorized. The 3rd survey also reported improved mortality (4.9%) and reduced follow-up loss (11.5%) compared to the 2nd survey. Conclusion While overall survival after Kasai operation has remained high and even improved, native liver survival has slightly declined. The findings reflect earlier diagnosis, more consistent diagnostic imaging, and increasing centralization of care. These trends underscore the importance of long-term nationwide data collection in guiding future strategies for BA management in Korea.

Purpose: This study aims to investigate and compare the incidence, demographic characteristics, clinical manifestations, preoperative diagnostic methods, anatomical classifications, associated anomalies, operative treatments, and postoperative outcomes of patients with intestinal atresia treated by the members of the Korean Association of Pediatric Surgeons (KAPS) through three nationwide surveys. Methods: KAPS conducted 3 national surveys in 1998, 2010, and 2024 to examine the patients diagnosed with intestinal atresia. In preparation for the survey, we developed a customized case registration form to obtain data on patient sex, birth weight, gestational age, clinical manifestations, preoperative diagnostic methods, anatomical types, associated anomalies, operative treatments, and postoperative outcomes. Authorized KAPS members completed the case registration form. Results: The first, second, and third national surveys included 218, 222, and 236 individuals diagnosed with intestinal atresia, respectively. The male-to-female ratios were 1.5:1, 1.1:1, and 1.1:1, respectively. The first, second, and third national surveys revealed that 34.3%, 43.3%, and 53.4% of patients were born before 37 weeks of gestation, respectively. Additionally, 28.7%, 32.0%, and 40.7% of patients had a birth weight under 2,500 g. In the third national survey, duodenoduodenostomy was the most common procedure, performed in 70 out of 82 patients diagnosed with duodenal atresia. Resection and anastomosis were the main surgical procedures conducted in 47 out of 54 cases of jejunal atresia and 74 out of 92 cases of ileal atresia. The mortality rates in the first, second, and third national surveys were 13.8%, 3.6%, and 1.3% respectively, with the lowest rate observed in the third national survey. Conclusion These national surveys offer valuable insights into the current state of intestinal atresia, including specific surgical interventions and postoperative outcomes in South Korea. For pediatric surgeons aiming to enhance their understanding of intestinal atresia and its treatment options, these surveys could be an indispensable resource and guide.

Purpose: This study provides insights into the prevalence at birth, clinical characteristics, and outcomes of gastroschisis and omphalocele in Korea over the past decade, addressing the lack of localized data despite advanced healthcare capabilities. Methods: The study retrospectively analyzed data from 20 pediatric surgical centers in Korea from January 2012 to December 2021, including 269 patients diagnosed with gastroschisis or omphalocele. Data variables included gender, gestational age, birth weight, associated anomalies, type of defect, surgical interventions, and outcomes. Results: The study covered 269 patients, with 80 gastroschisis and 189 omphalocele cases. Gastroschisis prevalence at birth remained stable at 2.15 per 100,000 live births, while omphalocele increased to 5.08 per 100,000. Both conditions had similar gender ratios (0.95). Gastroschisis patients had lower birth weights (2,463.90±505.50 g) and smaller head circumferences (31.97±1.86 cm) compared to omphalocele patients (2,757.65±761.24 g, 32.78±2.64 cm). Omphalocele cases had more associated anomalies, especially cardiovascular issues. Prenatal diagnosis rates were high: 93.7% for gastroschisis and 86.4% for omphalocele. About 96.3% of gastroschisis and 84.1% of omphalocele patients were born in their treatment hospitals. Gastroschisis patients underwent surgery sooner (average 3.5 days) and started feeding later (16.5 days) than omphalocele patients (average 56.5 days to surgery, 6.6 days to start feeding). Hospital stays and follow-up durations were similar, averaging around 782.6 days for gastroschisis and 800.3 days for omphalocele patients. Survival rates were 89.7% for gastroschisis and 87.1% for omphalocele. Conclusion The study highlights the need for early diagnosis, centralized care, and specialized surgical approaches to optimize outcomes for gastroschisis and omphalocele patients in Korea. Enhanced prenatal screening and surgical protocols are recommended to improve these patients' prognosis.

Purpose: The Korean Association of Pediatric Surgeons (KAPS) conducts annual nationwide surveys on various aspects of pediatric surgical diseases, with the results being discussed during KAPS’s annual spring meetings. Methods: KAPS conducted two national surveys, in 1995 and 2016, to investigate esophageal atresia (EA) with or without tracheoesophageal fistula (TEF). The authors analyzed data from these surveys to identify differences or changes in the annual occurrence, demographic characteristics, clinical presentation, preoperative diagnostic methods, anatomical type, associated anomalies, surgical treatment, and postoperative outcomes among patients with EA/TEF treated by KAPS members. Results: The first and second national surveys included 148 and 211 patients with EA/TEF, respectively. Excessive salivation was the most prevalent clinical symptom in both surveys. Type C was the most common form of EA/TEF in both surveys. The first survey included 126 patients, all of whom underwent open surgery. In the second survey, 152 (78.4%) of 194 patients underwent open surgery, while 34 (17.5%) underwent thoracoscopic surgery. Primary esophageal repair was performed on 96 (76.2%) of 126 patients in the first survey and on 160 (82.5%) of 194 patients in the second survey. Anastomotic strictures developed in 21.4% and 32.5% of patients, anastomotic leakage in 22.2% and 10.3%, recurrent fistula in 2.4% and 4.2% during the first and second surveys, respectively. The respective survival rates for group A were 90.2% and 98.3% in the first and second surveys. For group B, the rates were 73.9% and 98.1%, and for group C, they were 34.5% and 68.1%, respectively, according to the Waterston classifications. Conclusion These nationwide surveys provide comprehensive information on the status, detailed treatment, and outcomes for Korean pediatric patients with EA/TEF. They are anticipated to be an invaluable resource and guide for pediatric surgeons seeking to expand their knowledge on EA/TEF and its treatment options.

A 58-year-old woman visited the hospital complaining of fatigue and indigestion lasting for more than 3 months. She had no medical history other than taking a calcium plus vitamin D supplement for osteopenia. The initial blood test showed a high calcium level of 14.0 mg/dL. Additional tests were performed to differentially diagnose hypercalcemia. The blood test results were as follows: serum parathyroid hormone (PTH)=247.0 pg/mL, PTH-related peptide <1.0 pg/mL, phosphorous=2.6 mg/dL, 25-hydroxy-vitamin D=14.5 pg/mL, creatinine=1.09 mg/dL, and 24 hr urine calcium=215 mg/dL. A 4.5 cm sized cystic lesion on the intra-thyroidal space was confirmed on neck sonography and 4-dimensional parathyroid computed tomography, but technetium-99m methoxyisobutylisonitrile parathyroid scintigraphy showed equivocal results. After removal of the cystic lesion, serum calcium and PTH were normalized, and parathyroid lipoadenoma was confirmed in the postoperative pathology. Clinical features of parathyroid lipoadenoma are known to be similar to common parathyroid adenoma, but imaging studies often report negative findings. Therefore, it is necessary to better understand this rare disease for the differential diagnosis. For the final diagnosis and treatment of this disease, parathyroidectomy with intraoperative PTH measurement may be required.

Tixagevimab/cilgavimab is a monoclonal antibody used to prevent coronavirus disease 2019 among immunocompromised hosts and maintained neutralizing activity against early omicron variants. Omicron BN.1 became a dominant circulating strain in Korea early 2023, but its susceptibility to tixagevimab/cilgavimab is unclear. We conducted plaque reduction neutralization test (PRNT) against BN.1 in a prospective cohort (14 patients and 30 specimens). BN.1 PRNT was conducted for one- and three-months after tixagevimab/ cilgavimab administration and the average PRNT ND 50 of each point was lower than the positive cut-off value of 20 (12.9 ± 4.5 and 13.2 ± 4.2, respectively, P = 0.825). In the paired analyses, tixagevimab/cilgavimab-administered sera could not actively neutralize BN.1 (PRNT ND 50 11.5 ± 2.9, P = 0.001), compared with the reserved activity against BA.5 (ND 50 310.5 ± 180.4). Unlike virus-like particle assay, tixagevimab/cilgavimab was not active against BN.1 in neutralizing assay, and would not be effective in the present predominance of BA.2.75 sublineages.