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.

Background: Postpartum depression (PPD) can negatively affect infant well-being and child development. Although the frequency and risk factors of PPD symptoms might vary depending on the country and culture, there is limited research on these risk factors among Korean women. This study aimed to elucidate the potential risk factors of PPD throughout pregnancy to help improve PPD screening and prevention in Korean women. Methods: The pregnant women at 12 gestational weeks (GW) were enrolled from two obstetric specialized hospitals from March 2013 to November 2017. A questionnaire survey was administered at 12 GW, 24 GW, 36 GW, and 4 weeks postpartum. Depressive symptoms were assessed using the Edinburgh Postnatal Depression Scale, and PPD was defined as a score of ≥ 10. Results: PPD was prevalent in 16.3% (410/2,512) of the participants. Depressive feeling at 12 GW and postpartum factors of stress, relationship with children, depressive feeling, fear, sadness, and neonatal intensive care unit admission of baby were significantly associated with a higher risk of PPD. Meanwhile, high postpartum quality of life and marital satisfaction at postpartum period were significantly associated with a lower risk of PPD. We developed a model for predicting PPD using factors as mentioned above and it had an area under the curve of 0.871. Conclusion Depressive feeling at 12 GW and postpartum stress, fear, sadness, relationship with children, low quality of life, and low marital satisfaction increased the risk of PPD. A risk model that comprises significant factors can effectively predict PPD and can be helpful for its prevention and appropriate treatment.

We report a case of a ruptured triple hormone-secreting adrenal mass with hyperaldosteronism, hypercortisolism, and elevated normetanephrine levels, diagnosed as adrenal cortical carcinoma (ACC) by histology. A 53-year-old male patient who initially presented with abdominal pain was referred to our hospital for angiocoagulation of an adrenal mass rupture. Abdominal computed tomography revealed a heterogeneous 19×11×15 cm right adrenal mass with invasion into the right lobe of the liver, inferior vena cava, retrocaval lymph nodes, and aortocaval lymph nodes. Angiocoagulation was performed. Laboratory evaluation revealed excess cortisol via a positive 1-mg overnight dexamethasone suppression test, primary hyperaldosteronism via a positive saline infusion test, and plasma normetanephrine levels three times higher than normal. An adrenal mass biopsy was performed for pathological confirmation to commence palliative chemotherapy because surgical management was not deemed appropriate considering the extent of the tumor. Pathological examination revealed stage T4N1M1 ACC. The patient started the first cycle of adjuvant mitotane therapy along with adjuvant treatment with doxorubicin, cisplatin, and etoposide, and was discharged. Clinical cases of dual cortisol- and aldosterone-secreting ACCs or ACCs presenting as pheochromocytomas have occasionally been reported; however, both are rare. Moreover, to the best of our knowledge, a triple hormone-secreting ACC has not yet been reported. Here, we report a rare case and its management. This case report underscores the necessity of performing comprehensive clinical and biochemical hormone evaluations in patients with adrenal masses because ACC can present with multiple hormone elevations.

Background/Aims: Coronavirus disease 2019 (COVID-19) is associated with acute respiratory syndrome. The mechanisms underlying the different degrees of pneumonia severity in patients with COVID-19 remain elusive. This study provides evidence that COVID-19 is associated with eosinophil-mediated inflammation. Methods: We performed a retrospective case series of three patients with laboratory and radiologically confirmed COVID-19 pneumonia admitted to Chosun University Hospital. Demographic and clinical data on inflammatory cell lung infiltration and cytokine levels in patients with COVID-19 were collected. Results: Cytological analysis of sputum, tracheal aspirates, and bronchoalveolar lavage fluid (BALF) samples from all three patients revealed massive infiltration of polymorphonuclear cells (PMNs), such as eosinophils and neutrophils. All sputum and BALF specimens contained high levels of eosinophil cationic proteins. The infiltration of PMNs into the lungs, together with elevated levels of natural killer T (NKT) cells in BALF and peripheral blood samples from patients with severe pneumonia in the acute phase was confirmed by flow cytometry. Conclusions These results suggest that the lungs of COVID-19 patients can exhibit eosinophil-mediated inflammation, together with an elevated NKT cell response, which is associated with COVID-19 pneumonia.

목적: 이 연구는 2019학년도 역학조사관 입문교육 과정에 참여한 29명의 수습과정생에게 참여형 자기주도 학습 역학조사관 연수 프로그램(FETP)의 효과와 만족도 등 역량 변화를 분석해 그 결과를 향후 과정 개발의 참고 자료로 활용하고자 하였다. 방법: 교육 프로그램의 만족도와 교육 후 모듈에 대한 역량 변화를 평가하는 연구가 수행되었다. 만족도와 역량의 차이 비교는 크루스칼 왈리스 검정(Kruskal-Wallis test)를 실시하였고, 역량의 차이는 윌콕슨 부호순위검정(Wilcoxon signed rank test)에 의해 이루어 졌다. 결과: 2019년 FETP에 참여한 역학조사관 중 여성은 48.3% 였으며, 40세 미만은 9.4% 였다. 역학조사관 입문교육과정 모듈(역학조사, 보건통계 및 정보통계, 감염병 국가 체계, 감염병 질환 감시 체계, 진단 및 실험실 검사, 생물 안전 및 관리, 주요 감염성질환 관리와 조사, 커뮤니케이션, 협동과 리더십, 일반과정)별 만족도는 실무적 도움, 전문성, 기능, 태도 등에서 4점(5점 만점)을 초과하였고, 전체 4.2±0.21(5점 만점)점으로 높은 수준이였다, 모듈의 교육훈련 전후 평균 점수는 2.25±0.91, 3.68±0.63점 등으로 유의한 향상이 있었으며, 모든 모듈 및 하위 주제들도 유의한 향상이 있었다(p<0.001). 그 중에서 현장역학조사 경험이 가장 높은 변화가 있었고, 표본 수집과 실무가 가장 낮은 역량 변화가 있었다. 결론: 2019년 진행된 입문교육 과정은 수료 후 학생들의 역량은 개선되었고, 만족도는 높은 편이었다. 참여형 자기주도학습의 촉진은 역량을 향상시킬 뿐만 아니라 보건 종사자들의 자신감을 높일 수 있었다.

Although the prevalence of hereditary hemolytic anemia (HHA) is relatively low in Korea, it has been gradually increasing in recent decades due to increment in the proportions of hemoglobinopathies from immigrants of South East Asia, raising awareness of the disease among clinicians, and advances in diagnostic technology. As such, the red blood cell (RBC) Disorder Working Party (WP), previously called HHA WP, of the Korean Society of Hematology (KSH) developed the Korean Standard Operating Procedures (SOPs) for the diagnosis of HHA in 2007. These SOPs have been continuously revised and updated following advances in diagnostic technology [e.g., flow cytometric osmotic fragility test (FOFT) and eosin-5-maleimide (EMA) binding test], current methods for membrane protein or enzyme analysis [e.g., liquid chromatography-tandem mass spectrometry (LC-MS/MS), ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS), high-performance liquid chromatography (HPLC)], and molecular genetic tests using next-generation sequencing (NGS). However, the diagnosis and treatment of HHA remain challenging as they require considerable experience and understanding of the disease. Therefore, in this new Korean Clinical Practice Guidelines for the Diagnosis of HHA, on behalf of the RBC Disorder WP of KSH, updated guidelines to approach patients suspected of HHA are summarized. NGS is proposed to perform prior to membrane protein or enzyme analysis by LC-MS/MS, UPLC-MS/MS or HPLC techniques due to the availability of gene testing in more laboratories in Korea. We hope that this guideline will be helpful for clinicians in making diagnostic decisions for patients with HHA in Korea.

Purpose: To assess patient radiation doses during diagnostic and therapeutic neurointerventional procedures from multiple centers and propose dose reference level (RL). Materials and Methods: Consecutive neurointerventional procedures, performed in 22 hospitals from December 2020 to June 2021, were retrospectively studied. We collected data from a sample of 429 diagnostic and 731 therapeutic procedures. Parameters including dose-area product (DAP), cumulative air kerma (CAK), fluoroscopic time (FT), and total number of image frames (NI) were obtained. RL were calculated as the 3rd quartiles of the distribution. Results: Analysis of 1160 procedures from 22 hospitals confirmed the large variability in patient dose for similar procedures. RLs in terms of DAP, CAK, FT, and NI were 101.6 Gy·cm2, 711.3 mGy, 13.3 minutes, and 637 frames for cerebral angiography, 199.9 Gy·cm2, 3,458.7 mGy, 57.3 minutes, and 1,000 frames for aneurysm coiling, 225.1 Gy·cm2, 1,590 mGy, 44.7 minutes, and 800 frames for stroke thrombolysis, 412.3 Gy·cm2, 4,447.8 mGy, 99.3 minutes, and 1,621.3 frames for arteriovenous malformation (AVM) embolization, respectively. For all procedures, the results were comparable to most of those already published. Statistical analysis showed male and presence of procedural complications were significant factors in aneurysmal coiling. Male, number of passages, and procedural combined technique were significant factors in stroke thrombolysis. In AVM embolization, a significantly higher radiation dose was found in the definitive endovascular cure group. Conclusion Various RLs introduced in this study promote the optimization of patient doses in diagnostic and therapeutic interventional neuroradiology procedures. Proposed 3rd quartile DAP (Gy·cm2) values were 101.6 for diagnostic cerebral angiography, 199.9 for aneurysm coiling, 225.1 for stroke thrombolysis, and 412.3 for AVM embolization. Continual evolution of practices and technologies requires regular updates of RLs.