Objective: This study aimed to develop and validate a deep learning (DL) algorithm for the quantitative measurement of thoracolumbar (TL) fracture features, and to evaluate its efficacy across varying levels of clinical expertise. Methods: Using the pretrained Mask Region-Based Convolutional Neural Networks model, originally developed for vertebral body segmentation and fracture detection, we fine-tuned the model and added a new module for measuring fracture metrics—compression rate (CR), Cobb angle (CA), Gardner angle (GA), and sagittal index (SI)—from lumbar spine lateral radiographs. These metrics were derived from six-point labeling by 3 radiologists, forming the ground truth (GT). Training utilized 1,000 nonfractured and 318 fractured radiographs, while validations employed 213 internal and 200 external fractured radiographs. The accuracy of the DL algorithm in quantifying fracture features was evaluated against GT using the intraclass correlation coefficient. Additionally, 4 readers with varying expertise levels, including trainees and an attending spine surgeon, performed measurements with and without DL assistance, and their results were compared to GT and the DL model. Results: The DL algorithm demonstrated good to excellent agreement with GT for CR, CA, GA, and SI in both internal (0.860, 0.944, 0.932, and 0.779, respectively) and external (0.836, 0.940, 0.916, and 0.815, respectively) validations. DL-assisted measurements significantly improved most measurement values, particularly for trainees. Conclusion The DL algorithm was validated as an accurate tool for quantifying TL fracture features using radiographs. DL-assisted measurement is expected to expedite the diagnostic process and enhance reliability, particularly benefiting less experienced clinicians.

Objective: This study aimed to develop and validate a deep learning (DL) algorithm for the quantitative measurement of thoracolumbar (TL) fracture features, and to evaluate its efficacy across varying levels of clinical expertise. Methods: Using the pretrained Mask Region-Based Convolutional Neural Networks model, originally developed for vertebral body segmentation and fracture detection, we fine-tuned the model and added a new module for measuring fracture metrics—compression rate (CR), Cobb angle (CA), Gardner angle (GA), and sagittal index (SI)—from lumbar spine lateral radiographs. These metrics were derived from six-point labeling by 3 radiologists, forming the ground truth (GT). Training utilized 1,000 nonfractured and 318 fractured radiographs, while validations employed 213 internal and 200 external fractured radiographs. The accuracy of the DL algorithm in quantifying fracture features was evaluated against GT using the intraclass correlation coefficient. Additionally, 4 readers with varying expertise levels, including trainees and an attending spine surgeon, performed measurements with and without DL assistance, and their results were compared to GT and the DL model. Results: The DL algorithm demonstrated good to excellent agreement with GT for CR, CA, GA, and SI in both internal (0.860, 0.944, 0.932, and 0.779, respectively) and external (0.836, 0.940, 0.916, and 0.815, respectively) validations. DL-assisted measurements significantly improved most measurement values, particularly for trainees. Conclusion The DL algorithm was validated as an accurate tool for quantifying TL fracture features using radiographs. DL-assisted measurement is expected to expedite the diagnostic process and enhance reliability, particularly benefiting less experienced clinicians.

Objective: This study aimed to develop and validate a deep learning (DL) algorithm for the quantitative measurement of thoracolumbar (TL) fracture features, and to evaluate its efficacy across varying levels of clinical expertise. Methods: Using the pretrained Mask Region-Based Convolutional Neural Networks model, originally developed for vertebral body segmentation and fracture detection, we fine-tuned the model and added a new module for measuring fracture metrics—compression rate (CR), Cobb angle (CA), Gardner angle (GA), and sagittal index (SI)—from lumbar spine lateral radiographs. These metrics were derived from six-point labeling by 3 radiologists, forming the ground truth (GT). Training utilized 1,000 nonfractured and 318 fractured radiographs, while validations employed 213 internal and 200 external fractured radiographs. The accuracy of the DL algorithm in quantifying fracture features was evaluated against GT using the intraclass correlation coefficient. Additionally, 4 readers with varying expertise levels, including trainees and an attending spine surgeon, performed measurements with and without DL assistance, and their results were compared to GT and the DL model. Results: The DL algorithm demonstrated good to excellent agreement with GT for CR, CA, GA, and SI in both internal (0.860, 0.944, 0.932, and 0.779, respectively) and external (0.836, 0.940, 0.916, and 0.815, respectively) validations. DL-assisted measurements significantly improved most measurement values, particularly for trainees. Conclusion The DL algorithm was validated as an accurate tool for quantifying TL fracture features using radiographs. DL-assisted measurement is expected to expedite the diagnostic process and enhance reliability, particularly benefiting less experienced clinicians.

Objective: This study aimed to develop and validate a deep learning (DL) algorithm for the quantitative measurement of thoracolumbar (TL) fracture features, and to evaluate its efficacy across varying levels of clinical expertise. Methods: Using the pretrained Mask Region-Based Convolutional Neural Networks model, originally developed for vertebral body segmentation and fracture detection, we fine-tuned the model and added a new module for measuring fracture metrics—compression rate (CR), Cobb angle (CA), Gardner angle (GA), and sagittal index (SI)—from lumbar spine lateral radiographs. These metrics were derived from six-point labeling by 3 radiologists, forming the ground truth (GT). Training utilized 1,000 nonfractured and 318 fractured radiographs, while validations employed 213 internal and 200 external fractured radiographs. The accuracy of the DL algorithm in quantifying fracture features was evaluated against GT using the intraclass correlation coefficient. Additionally, 4 readers with varying expertise levels, including trainees and an attending spine surgeon, performed measurements with and without DL assistance, and their results were compared to GT and the DL model. Results: The DL algorithm demonstrated good to excellent agreement with GT for CR, CA, GA, and SI in both internal (0.860, 0.944, 0.932, and 0.779, respectively) and external (0.836, 0.940, 0.916, and 0.815, respectively) validations. DL-assisted measurements significantly improved most measurement values, particularly for trainees. Conclusion The DL algorithm was validated as an accurate tool for quantifying TL fracture features using radiographs. DL-assisted measurement is expected to expedite the diagnostic process and enhance reliability, particularly benefiting less experienced clinicians.

Objective: This study aimed to develop and validate a deep learning (DL) algorithm for the quantitative measurement of thoracolumbar (TL) fracture features, and to evaluate its efficacy across varying levels of clinical expertise. Methods: Using the pretrained Mask Region-Based Convolutional Neural Networks model, originally developed for vertebral body segmentation and fracture detection, we fine-tuned the model and added a new module for measuring fracture metrics—compression rate (CR), Cobb angle (CA), Gardner angle (GA), and sagittal index (SI)—from lumbar spine lateral radiographs. These metrics were derived from six-point labeling by 3 radiologists, forming the ground truth (GT). Training utilized 1,000 nonfractured and 318 fractured radiographs, while validations employed 213 internal and 200 external fractured radiographs. The accuracy of the DL algorithm in quantifying fracture features was evaluated against GT using the intraclass correlation coefficient. Additionally, 4 readers with varying expertise levels, including trainees and an attending spine surgeon, performed measurements with and without DL assistance, and their results were compared to GT and the DL model. Results: The DL algorithm demonstrated good to excellent agreement with GT for CR, CA, GA, and SI in both internal (0.860, 0.944, 0.932, and 0.779, respectively) and external (0.836, 0.940, 0.916, and 0.815, respectively) validations. DL-assisted measurements significantly improved most measurement values, particularly for trainees. Conclusion The DL algorithm was validated as an accurate tool for quantifying TL fracture features using radiographs. DL-assisted measurement is expected to expedite the diagnostic process and enhance reliability, particularly benefiting less experienced clinicians.

Background: This study aimed to generate a Z score calculation model for coronary artery diameter of normal children and adolescents to be adopted as the standard calculation method with consensus in clinical practice. Methods: This study was a retrospective, multicenter study that collected data from multiple institutions across South Korea. Data were analyzed to determine the model that best fit the relationship between the diameter of coronary arteries and independent demographic parameters. Linear, power, logarithmic, exponential, and square root polynomial models were tested for best fit. Results: Data of 2,030 subjects were collected from 16 institutions. Separate calculation models for each sex were developed because the impact of demographic variables on the diameter of coronary arteries differs according to sex. The final model was the polynomial formula with an exponential relationship between the diameter of coronary arteries and body surface area using the DuBois formula. Conclusion A new coronary artery diameter Z score model was developed and is anticipated to be applicable in clinical practice. The new model will help establish a consensus-based Z score model.

BACKGROUND: AND PURPOSE: The aim of this study was to determine the diagnostic performance and safety of a new ¹⁸F-labeled amyloid tracer, ¹⁸F-FC119S. METHODS: This study prospectively recruited 105 participants, comprising 53 with Alzheimer's disease (AD) patients, 16 patients with dementia other than AD (non-AD), and 36 healthy controls (HCs). In the first screening visit, the Seoul Neuropsychological Screening Battery cognitive function test was given to the dementia group, while HC subjects completed the Korean version of the Mini Mental State Examination. Individuals underwent ¹⁸F-FC119S PET, ¹⁸F-fluorodeoxyglucose (FDG) PET, and brain MRI. The diagnostic performance of ¹⁸F-FC119S PET for AD was compared to a historical control (comprising previously reported and currently used amyloid-beta PET agents), ¹⁸F-FDG PET, and MRI. The standardized uptake value (SUV) ratio (ratio of the cerebral cortical SUV to the cerebellar SUV) was measured for each PET data set to provide semiquantitative analysis. All adverse effects during the clinical trial periods were monitored. RESULTS: Visual assessments of the ¹⁸F-FC119S PET data revealed a sensitivity of 92% and a specificity of 84% in detecting AD. ¹⁸F-FC119S PET demonstrated equivalent or better diagnostic performance for AD detection than the historical control, ¹⁸F-FDG PET (sensitivity of 80.0% and specificity of 76.0%), and MRI (sensitivity of 98.0% and specificity of 50.0%). The SUV ratios differed significantly between AD patients and the other groups, at 1.44±0.17 (mean±SD) for AD, 1.24±0.09 for non-AD, and 1.21±0.08 for HC. No clinically significant adverse effects occurred during the trial periods. CONCLUSIONS ¹⁸F-FC119S PET provides high sensitivity and specificity in detecting AD and therefore may be considered a useful diagnostic tool for AD.

PURPOSE: The optimal treatment for synchronous liver metastasis (LM) from colorectal cancer (CRC) depends on various factors. The present study was intended to investigate the oncologic outcome according to the time of resection of metastatic lesions. METHODS: Data from patients who underwent treatment with curative intent for primary CRC and synchronous LM between 2004 and 2009 from 9 university hospitals in Korea were collected retrospectively. One hundred forty-three patients underwent simultaneous resection for primary CRC and synchronous LM (simultaneous surgery group), and 65 patients were treated by 2-stage operation (staged surgery group). RESULTS: The mean follow-up length was 41.2 ± 24.6 months. In the extent of resection for hepatic metastasis, major hepatectomy was more frequently performed in staged surgery group (33.8% vs. 8.4%, P < 0.001). The rate of severe complications of Clavien-Dindo classification grade III or more was not significantly different between the 2 groups. The 3-year overall survival (OS) rate was 85.0% in staged surgery group and 69.4% in simultaneous surgery group (P = 0.013), and the 3-year recurrence-free survival (RFS) rate was 46.4% in staged surgery group and 30.2% in simultaneous surgery group (P = 0.143). In subgroup analysis based on the location of primary CRC, the benefit of staged surgery for OS and RFS was clearly shown in rectal cancer (P = 0.021 and P = 0.015). CONCLUSION: Based on our results, staged surgery with or without neoadjuvant chemotherapy should be considered for resectable synchronous LM from CRC, especially in rectal cancer, as a safe and fairly promising option.
Classification ; Colorectal Neoplasms ; Drug Therapy ; Follow-Up Studies ; Hepatectomy ; Hospitals, University ; Humans ; Korea ; Liver ; Neoplasm Metastasis ; Rectal Neoplasms ; Retrospective Studies

The majority of malignant melanomas in the small intestine are metastases from primary cutaneous lesions, it can also develop as a primary mucosal tumor in the gastrointestinal tract. In this report, we present rare cases of primary small bowel melanoma and review the current literature. A 78-year-old male presented with abdominal pain and CT enterography identified a ileal mass. A 79-year-old female presented with signs and symptoms of partial small bowel obstruction. Abdominopelvic CT and small bowel series revealed a obstructing mass in the distal jejunum. The masses were confirmed on laparotomy and histologically diagnosed as melanoma. Extensive postoperative clinical examination revealed no cutaneous lesions. A primary small bowel melanoma is an extremely rare neoplasm. A definite diagnosis can only be made after a thorough investigation has been made to exclude the coexistence of a primary lesion. Curative resection of the tumor remains the treatment of choice.
Abdominal Pain ; Aged ; Diagnosis ; Female ; Gastrointestinal Tract ; Humans ; Intestine, Small* ; Jejunum ; Laparotomy ; Male ; Melanoma* ; Neoplasm Metastasis

Endogenous endophthalmitis secondary to group B Streptococcus (GBS) is extremely rare, particularly in healthy adults. However, the visual prognosis is poor. We report the first South Korean case of GBS infective endocarditis presenting as endogenous endophthalmitis and skin and soft tissue infection. Cultures of blood, vitreous humor, and pus from skin aspirates yielded a penicillin-susceptible serotype V strain of Streptococcus agalactiae. After 6 weeks, the patient completely recovered from GBS infective endocarditis. However, despite early antibiotic treatment and early surgical intervention, the patient's right eye developed phthisis bulbi and was a candidate for evisceration.
Adult* ; Endocarditis* ; Endophthalmitis* ; Humans ; Patient Rights ; Prognosis ; Serogroup ; Skin ; Soft Tissue Infections ; Streptococcus agalactiae* ; Streptococcus* ; Suppuration ; Vitreous Body