A deep learning approach to identify the fetal head position using transperineal ultrasound during labor DOI Creative Commons
Ruben Ramirez Zegarra, Francesco Conversano, A. Dall’Asta

et al.

European Journal of Obstetrics & Gynecology and Reproductive Biology, Journal Year: 2024, Volume and Issue: 301, P. 147 - 153

Published: Aug. 9, 2024

To develop a deep learning (DL)-model using convolutional neural networks (CNN) to automatically identify the fetal head position at transperineal ultrasound in second stage of labor.

Language: Английский

Assessing Rapid Visual Perception in Medical Students Trained in Ultrasound DOI Creative Commons

O. Mescher,

Christopher T. Musick,

Brianna C. Landis

et al.

Medical Science Educator, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 27, 2025

Language: Английский

Citations

0
Platform for Automated Assessment of Obstetric Ultrasounds, Using Machine Learning DOI
Jonathan Sánchez Luna, Arnulfo Alanís, Efrain Patiñό

et al.

Smart innovation, systems and technologies, Journal Year: 2025, Volume and Issue: unknown, P. 247 - 257

Published: Jan. 1, 2025

Language: Английский

Citations

0
Artificial Intelligence–Assisted Automation of Fetal Anomaly Ultrasound Scanning DOI
F. Sessions Cole

NEJM AI, Journal Year: 2025, Volume and Issue: 2(4)

Published: March 27, 2025

Language: Английский

Citations

0
Effectiveness and clinical impact of using deep learning for first-trimester fetal ultrasound image quality auditing DOI Creative Commons

Xiaoyan Cao,

Binghan Li,

Yongsong Zhou

et al.

BMC Pregnancy and Childbirth, Journal Year: 2025, Volume and Issue: 25(1)

Published: March 31, 2025

Regular auditing of ultrasound images is required to maintain quality; however, manual time-consuming and can be inconsistent. We therefore aimed develop validate an artificial intelligence-based image quality audit (AI-IQA) system from the four key planes used in first-trimester scanning. The AI-IQA was developed based on YOLOv7 structure detection network a multi-branch regression using large multicenter internal dataset. Clinical validation performed 567 cases scanned by radiologists with different experience levels, which 349 were without feedback (clinical test set 1) 218 after 2–3 rounds 2). proportion standard obtained detailed expert results compared verify whether could objectively accurately provide deficiencies nonstandard assist at levels improving quality. In set, achieved high average accuracy precision, recall F1-score overall plane (0.881, 0.833, 0.842 0.837, respectively) (0.906, 0.861, 0.857 0.859, respectively). clinical sets 1 2, showed strong consistency assessment results, Cohen's Kappa coefficient exceeding 0.8 for all planes. addition, following feedback, junior mid-level increased 7.7% 5.1%, respectively. takes only 0.05 s assess each image, while experts require more than 20 (p < 0.001). proposed proved highly accurate efficient method automatically scanning quality, providing precise rapid control. This tool also improve

Language: Английский

Citations

0
Artificial intelligence based automatic classification, annotation, and measurement of the fetal heart using HeartAssist DOI
Rina Kim, Mi‐Young Lee, Yoo Jin Lee

et al.

Research Square (Research Square), Journal Year: 2025, Volume and Issue: unknown

Published: April 4, 2025

Abstract Objectives: This study evaluated the feasibility of HeartAssist, a novel automated tool designed for classification fetal cardiac views, annotation structures, and measurement parameters. Unlike previous AI tools that primarily focused on classification, HeartAssist integrates capabilities, enabling more comprehensive assessment. Methods: Cardiac images from fetuses (gestational ages 20–40 weeks) were collected at Asan Medical Center between January 2016 October 2018. was developed using convolutional neural networks to classify 10 annotate 26 measure 43 One expert performed manual classifications, annotations, measurements, which then compared outputs assess feasibility. Results: A total 65,324 2,985 analyzed. achieved 99.4% accuracy, with recall, precision, F1-score 0.93, 0.95, 0.94, respectively. Annotation accuracy 98.4%, while automatic success rate 97.6%, an error 7.62% caliper similarity 0.613. Conclusions: HeartAssist is reliable screening, demonstrating high in classifying views annotating comparable outcomes measuring could enhance prenatal detection congenital heart disease improve perinatal outcomes.

Language: Английский

Citations

0
Artificial intelligence based automatic classification, annotation, and measurement of the fetal heart using HeartAssist DOI Creative Commons
Rina Kim, Mi‐Young Lee, Yoo Jin Lee

et al.

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: April 16, 2025

This study evaluated the feasibility of HeartAssist, a novel automated tool designed for classification fetal cardiac views, annotation structures, and measurement parameters. Unlike previous AI tools that primarily focused on classification, HeartAssist integrates capabilities, enabling more comprehensive assessment.Cardiac images from fetuses (gestational ages 20-40 weeks) were collected at Asan Medical Center between January 2016 October 2018. was developed using convolutional neural networks to classify 10 annotate 26 measure 43 One expert performed manual classifications, annotations, measurements, which then compared outputs assess feasibility. A total 65,324 2,985 analyzed. achieved 99.4% accuracy, with recall, precision, F1-score 0.93, 0.95, 0.94, respectively. Annotation accuracy 98.4%, while automatic success rate 97.6%, an error 7.62% caliper similarity 0.613. is reliable screening, demonstrating high in classifying views annotating comparable outcomes measuring could enhance prenatal detection congenital heart disease improve perinatal outcomes.

Language: Английский

Citations

0
Fetal cardiac diagnostics in Indonesia: a study of screening and echocardiography DOI
Muhammad Adrianes Bachnas, Wiku Andonotopo, Adhi Pribadi

et al.

Journal of Perinatal Medicine, Journal Year: 2025, Volume and Issue: unknown

Published: April 15, 2025

Abstract Introduction Congenital heart defects (CHDs) are a leading cause of neonatal morbidity and mortality globally. Accurate prenatal detection is crucial to improving outcomes. In Indonesia, two primary methods used: fetal cardiac screening (FCS), which accessible but limited in sensitivity (40–60 %), echocardiography (FE), the gold standard with over 90 % access due infrastructural financial challenges. Content This review analyzes Indonesia’s diagnostic disparities, highlighting how rural regions rely heavily on FCS, while FE remains restricted urban centers. Emerging technologies, such as AI-enhanced diagnostics telemedicine, show promise bridging gaps by increasing FCS accuracy extending through remote consultations. Summary AI has potential boost up 30 %, making it an effective preliminary tool, telemedicine platforms connect practitioners specialists. However, barriers like insufficient infrastructure, regulatory issues, training hinder widespread adoption. Outlook Addressing these requires standardized national protocols, capacity-building initiatives, public-private partnerships finance infrastructure reduce costs. With technology integration systemic reforms, Indonesia can achieve equitable CHD diagnostics, maternal outcomes aligning global standards.

Language: Английский

Citations

0
Research advancements in the Use of artificial intelligence for prenatal diagnosis of neural tube defects DOI Creative Commons
Maryam Yeganegi,

Mahsa Danaei,

Sepideh Azizi

et al.

Frontiers in Pediatrics, Journal Year: 2025, Volume and Issue: 13

Published: April 17, 2025

Artificial Intelligence is revolutionizing prenatal diagnostics by enhancing the accuracy and efficiency of procedures. This review explores AI machine learning (ML) in early detection, prediction, assessment neural tube defects (NTDs) through ultrasound imaging. Recent studies highlight effectiveness techniques, such as convolutional networks (CNNs) support vector machines (SVMs), achieving detection rates up to 95% across various datasets, including fetal images, genetic data, maternal health records. SVM models have demonstrated 71.50% on training datasets 68.57% testing for NTD classification, while advanced deep (DL) methods report patient-level prediction 94.5% an area under receiver operating characteristic curve (AUROC) 99.3%. integration with genomic analysis has identified key biomarkers associated NTDs, Growth Associated Protein 43 (GAP43) Glial Fibrillary Acidic (GFAP), logistic regression 86.67% accuracy. Current AI-assisted technologies improved diagnostic accuracy, yielding sensitivity specificity 88.9% 98.0%, respectively, compared traditional 81.5% 92.2% specificity. systems also streamlined workflows, reducing median scan times from 19.7 min 11.4 min, allowing sonographers prioritize critical patient care. Advancements DL algorithms, Oct-U-Net PAICS, achieved recall precision 0.93 0.96, identifying abnormalities. Moreover, AI's evolving role research supports personalized prevention strategies enhances public awareness AI-generated messages. In conclusion, significantly improves leading greater As continues advance, it potential further enhance healthcare raise about ultimately contributing better outcomes.

Language: Английский

Citations

0
The performance of sonographic antenatal birth weight assessment assisted with artificial intelligence compared to that of manual examiners at term DOI Creative Commons

Alex Horky,

Marita Wasenitz,

C. Iacovella

et al.

Archives of Gynecology and Obstetrics, Journal Year: 2025, Volume and Issue: unknown

Published: April 29, 2025

Language: Английский

Citations

0
Development and validation of ultrasound-based radiomics deep learning model to identify bone erosion in rheumatoid arthritis DOI
Lei Yan, Jing Xu, Xiaojian Ye

et al.

Clinical Rheumatology, Journal Year: 2025, Volume and Issue: unknown

Published: May 19, 2025

Language: Английский

Citations

0