Detection of Elbow OCD in the Ultrasound Image by Artificial Intelligence Using YOLOv8

Applied Sciences, Journal Year: 2023, Volume and Issue: 13(13), P. 7623 - 7623

Published: June 28, 2023

Background: Screening for elbow osteochondritis dissecans (OCD) using ultrasound (US) is essential early detection and successful conservative treatment. The aim of the study to determine diagnostic accuracy YOLOv8, a deep-learning-based artificial intelligence model, US images OCD or normal elbow-joint images. Methods: A total 2430 were used. Using YOLOv8 image classification object performed recognize lesions standard views joints. Results: In binary lesions, values from confusion matrix following: Accuracy = 0.998, Recall 0.9975, Precision 1.000, F-measure 0.9987. mean average precision (mAP) comparing bounding box detected by trained model with true-label was 0.994 in YOLOv8n 0.995 YOLOv8m model. Conclusions: joints lesions. Both tasks able be achieved high may useful mass screening at medical check-ups baseball elbow.

Language: Английский

---

Medical Imaging Applications Developed Using Artificial Intelligence Demonstrate High Internal Validity Yet Are Limited in Scope and Lack External Validation

Arthroscopy The Journal of Arthroscopic and Related Surgery, Journal Year: 2024, Volume and Issue: unknown

Published: Feb. 1, 2024

Language: Английский

---

Performance of Artificial Intelligence in Addressing Questions Regarding Management of Osteochondritis Dissecans

Sports Health A Multidisciplinary Approach, Journal Year: 2025, Volume and Issue: unknown

Published: April 1, 2025

Background: Large language model (LLM)-based artificial intelligence (AI) chatbots, such as ChatGPT and Gemini, have become widespread sources of information. Few studies evaluated LLM responses to questions about orthopaedic conditions, especially osteochondritis dissecans (OCD). Hypothesis: Gemini will generate accurate that align with American Academy Orthopaedic Surgeons (AAOS) clinical practice guidelines. Study Design: Cohort study. Level Evidence: 2. Methods: prompts were created based on AAOS guidelines OCD diagnosis treatment, from collected. Seven fellowship-trained surgeons a 5-point Likert scale, 6 categories: relevance, accuracy, clarity, completeness, evidence-based, consistency. Results: exhibited strong performance across all criteria. mean scores highest for clarity (4.771 ± 0.141 [mean SD]). scored relevance accuracy (4.286 0.296, 4.286 0.273). For both LLMs, the lowest evidence-based (ChatGPT, 3.857 0.352; 3.743 0.353). other categories, higher than scores. The consistency between 2 LLMs was rated at an overall 3.486 0.371. Inter-rater reliability ranged 0.4 0.67 (mean, 0.59) (0.67) in category (0.4) category. Conclusion: emphasizes potential gathering clinically relevant answers regarding treatment suggests may be better this purpose model. Further evaluation information procedures conditions necessary before can recommended source Clinical Relevance: Little is known ability AI provide OCD.

Language: Английский

---

Deep learning-based osteochondritis dissecans detection in ultrasound images with humeral capitellum localization

International Journal of Computer Assisted Radiology and Surgery, Journal Year: 2024, Volume and Issue: unknown

Published: Jan. 17, 2024

Abstract Purpose Osteochondritis dissecans (OCD) of the humeral capitellum is a common cause elbow disorders, particularly among young throwing athletes. Conservative treatment preferred for managing OCD, and early intervention significantly influences possibility complete disease resolution. The purpose this study to develop deep learning-based classification model in ultrasound images computer-aided diagnosis. Methods This paper proposes OCD method images. proposed first detects detection using YOLO then estimates probability detected region VGG16. We hypothesis that performance will be improved by eliminating unnecessary regions. To validate method, it was applied 158 subjects (OCD: 67, Normal: 91) five-fold-cross-validation. Results demonstrated achieved mean average precision (mAP) over 0.95, while estimation an accuracy 0.890, 0.888, recall 0.927, F1 score 0.894, area under curve (AUC) 0.962. On other hand, when constructed entire image, accuracy, precision, recall, score, AUC were 0.806, 0.932, 0.843, 0.928, respectively. findings suggest high-performance potential ultrasonic Conclusion introduces method. experimental results emphasize effectiveness focusing on Future work should involve evaluating employing physicians during medical check-ups OCD.

Language: Английский

---

Dynamic Analysis of the Median Nerve in Carpal Tunnel Syndrome from Ultrasound Images Using the YOLOv5 Object Detection Model

Applied Sciences, Journal Year: 2023, Volume and Issue: 13(24), P. 13256 - 13256

Published: Dec. 14, 2023

Carpal tunnel syndrome (CTS) is caused by subsynovial connective tissue fibrosis, resulting in median nerve (MN) mobility. The standard evaluation method the measurement of MN cross-sectional area using static images, and dynamic images are not widely used. In recent years, remarkable progress has been made field deep learning (DL) medical image processing. aim present study was to evaluate dynamics CTS hands YOLOv5 model, which one object detection models DL. We included 20 normal (control group) (CTS group). obtained ultrasonographic short-axis carpal recorded motion during finger flexion–extension, evaluated displacement velocity. model showed a score 0.953 for precision 0.956 recall. radial–ulnar 3.56 mm control group 2.04 group, velocity 4.22 mm/s 3.14 group. scores were significantly reduced This demonstrates potential DL-based analysis as powerful diagnostic tool CTS.

Language: Английский

---

Deep Learning-Based Computer-Aided Diagnosis of Osteochondritis Dissecans of the Humeral Capitellum Using Ultrasound Images

Journal of Bone and Joint Surgery, Journal Year: 2024, Volume and Issue: unknown

Published: May 14, 2024

Background: Ultrasonography is used to diagnose osteochondritis dissecans (OCD) of the humerus; however, its reliability depends on technical proficiency examiner. Recently, computer-aided diagnosis (CAD) using deep learning has been applied in field medical science, and high diagnostic accuracy reported. We aimed develop a learning-based CAD system for OCD detection ultrasound images evaluate system. Methods: The process comprises 2 steps: humeral capitellum an object-detection algorithm classification image network. Four-directional elbow throwing arm 196 baseball players (mean age, 11.2 years), including 104 with normal findings 92 OCD, were training validation. An external dataset 20 (10 10 OCD) was A confusion matrix area under receiver operating characteristic curve (AUC) Results: Clinical evaluation resulted AUCs all 4 directions: 0.969 anterior long axis, 0.966 short 0.996 posterior 0.993 axis. thus exceeded 0.9 directions. Conclusions: propose detect lesions images. achieved directions elbow. This model may be useful screening during checkups reduce probability missing lesion. Level Evidence: Diagnostic II . See Instructions Authors complete description levels evidence.

Language: Английский

---

Advances in Ultrasound-Guided Surgery and Artificial Intelligence Applications in Musculoskeletal Diseases

Diagnostics, Journal Year: 2024, Volume and Issue: 14(18), P. 2008 - 2008

Published: Sept. 11, 2024

Ultrasound imaging is a vital tool in musculoskeletal medicine, with the number of publications on ultrasound-guided surgery increasing recent years, especially minimally invasive procedures sports, foot and ankle, hand surgery. However, ultrasound has drawbacks, such as operator dependency image obscurity. Artificial intelligence (AI) deep learning (DL), subset AI, can address these issues. AI/DL enhance screening practices for hip dysplasia osteochondritis dissecans (OCD) humeral capitellum, improve diagnostic accuracy carpal tunnel syndrome (CTS), provide physicians better prognostic prediction tools patients knee osteoarthritis. Building advancements, DL methods, including segmentation, detection, localization target tissues medical instruments, also have potential to allow surgeons perform more accurately efficiently. This review summarizes advances diseases provides comprehensive overview utilization particularly focusing

Language: Английский

---

Leveraging AI models for lesion detection in osteonecrosis of the femoral head and T1‐weighted MRI generation from radiographs

Journal of Orthopaedic Research®, Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 23, 2024

Abstract This study emphasizes the importance of early detection osteonecrosis femoral head (ONFH) in young patients on long‐term glucocorticoid therapy, including those with acute lymphoblastic leukemia, lupus, and other diagnoses. While X‐ray magnetic resonance imaging (MRI) are standard methods for staging ONFH, MRI can be costly time‐consuming. The research focuses utilizing artificial intelligence (AI) to enhance evaluation radiographic images ONFH detection. involved analyzing from 102 control hips 104 ONFH‐affected at Association Research Circulation Osseous (ARCO) Stage II IIIa. We employed transfer learning YOLOv8 model object detection, using 80% data training 20% validation, then assessed accuracy through mean average precision (mAP) a precision‐recall curve. Additionally, AI generated synthetic (sMRI) Generative Adversarial Network (GAN) evaluated their similarity original MRI. Results showed that mAP was 0.923 YOLOv8n 0.951 YOLOv8x. GAN‐generated sMRI exhibited lower image quality compared originals but maintained potential lesion assessment. Intrarater reliability among evaluators high. findings indicate techniques, particularly GAN generation, effectively assist screening, despite some limitations quality.

Language: Английский

---

Computational intelligence on medical imaging with artificial neural networks

Elsevier eBooks, Journal Year: 2024, Volume and Issue: unknown, P. 227 - 257

Published: Nov. 22, 2024

Language: Английский

---

Quantification of Empty Lacunae in Tissue Sections of Osteonecrosis of the Femoral Head Using YOLOv8 Artificial Intelligence Model

Journal of Biomedical Materials Research Part B Applied Biomaterials, Journal Year: 2024, Volume and Issue: 112(12)

Published: Dec. 1, 2024

ABSTRACT Histomorphometry is an important technique in the evaluation of non‐traumatic osteonecrosis femoral head (ONFH). Quantification empty lacunae and pyknotic cells on histological images most reliable measure ONFH pathology, yet it time manpower consuming. This study focused application artificial intelligence (AI) technology to tissue image evaluation. The aim this establish automated cell counting platform using YOLOv8 as object detection model evaluate validate its accuracy. From 30 rabbits, 270 were prepared; based evaluations by three researchers, ground truth labels created classify each into two classes (osteocytes lacunae) or (osteocytes, cells, lacunae). Two then annotated image. Transfer learning data (80% for training 20% validation) was performed YOLOv8n YOLOv8x with different parameters. To accuracy model, mean average precision (mAP (50)) precision‐recall curve identified. In addition, reliability relative manual evaluated linear regression analysis five unused previous experiments. mAP (50) 0.868 0.883 YOLOv8x. 0.735 0.750 model. quantification obtained highly correlated data. development AI‐applied will significantly reduce effort analysis.

Language: Английский

---