- Current Status, Prospect and Bottleneck of Ultrasound AI Development: A Systemic Review
- Siyi Xun, MA, Wei Ke, PhD, Mingfu Jiang, MA, Huachao Chen, BA, Haoming Chen, BA, Chantong Lam, PhD, Ligang Cui, MD, Tao Tan, PhD
- 2023, 7 (2): 61-72. DOI:10.37015/AUDT.2023.230020
- Abstract ( 244 ) HTML ( 23 ) PDF ( 11417KB ) ( 300 )
In recent years, ultrasound imaging has become an important means of medical diagnosis because of its safety and radiation-free advantages. With the continuous progress of deep learning, Artificial Intelligence (AI) models can process large amounts of ultrasound data quickly and accurately, providing decision support for clinicians in diagnosis. From the perspective of ultrasound image classification, detection and segmentation, this paper systemically introduces the latest progress of AI technology in ultrasound imaging, and summarizes the recent high-level related work. At the same time, we also discuss the development prospect and bottleneck of AI in ultrasound imaging processing, which provides the future research directions for researchers in related fields.
- Intelligent Ultrasonic Diagnosis and Clinical Application: Technical Development and Prospectives
- Rendong Chen, PhD, Xiaoqian Wang, BS, Ping Liang, MD, Xiaoping Ouyang, PhD, Dexing Kong, PhD
- 2023, 7 (2): 73-81. DOI:10.37015/AUDT.2023.230019
- Abstract ( 170 ) HTML ( 21 ) PDF ( 11300KB ) ( 237 )
Ultrasound intelligent diagnosis is an emerging technology that combines artificial intelligence (AI) and medical ultrasonography. It has gained significant attention in recent years due to its potential to improve the accuracy and efficiency of medical diagnosis. The core elements of ultrasound artificial intelligence are the construction of data and algorithm models. Therefore, developing autonomous and controllable models, algorithms, and data platforms is extremely important. In this paper, we provide a comprehensive review of the current state-of-the-art in ultrasound intelligent diagnosis including the aspects of the construction of ultrasonic database, deep learning techniques in ultrasound intelligent diagnosis, and the clinical application of ultrasound-AI products. With continued advancements in AI and ultrasound imaging technology, we believe ultrasound intelligent diagnosis will be a valuable tool in the hands of healthcare professionals, providing them with more accurate and efficient diagnoses and treatment plans in the coming years.
- Artificial Intelligence in Prenatal Ultrasound: Clinical Application and Prospect
- Wenjia Guo, MM, Shengli Li, MM, Xing Yu, MD, Huaxuan Wen, BM, Ying Yuan, MM, Xia Yang, MM
- 2023, 7 (2): 82-90. DOI:10.37015/AUDT.2023.230024
- Abstract ( 154 ) HTML ( 9 ) PDF ( 11800KB ) ( 141 )
Since the 1990s, researchers have been seeking approaches for applying artificial intelligence (AI) to prenatal ultrasound. With the breakthrough of cloud computing technology and the development of deep learning technology, AI in prenatal ultrasound has already entered the clinical application stage in recent years. How does AI combine with clinical prenatal ultrasound? Is the clinical application of AI in prenatal ultrasound effective? What can we expect from AI in prenatal ultrasound? This review introduces the latest developments in this field and explores the challenges and opportunities brought by AI to prenatal ultrasound.
- Deep Learning on Ultrasound Imaging for Breast Cancer Diagnosis and Treatment: Current Applications and Future Perspectives
- Changyan Wang, BS, Haobo Chen, MS, Jieyi Liu, BS, Changchun Li, BS, Weiwei Jiao, BS, Qihui Guo, BS, Qi Zhang, PhD
- 2023, 7 (2): 91-113. DOI:10.37015/AUDT.2023.230012
- Abstract ( 224 ) HTML ( 8 ) PDF ( 11701KB ) ( 169 )
Ultrasound is a commonly used imaging modality for breast cancer diagnosis and prognosis but suffers from false positives, false negatives and interobserver variability. Deep learning (DL), a subset of artificial intelligence, has the potential to improve the efficiency and accuracy of breast ultrasound. This article provides a comprehensive overview of DL applications for breast cancer diagnosis and treatment in ultrasound, including methodological descriptions of various DL models, and clinical applications on noise reduction, lesion localization, risk assessment, diagnosis, response evaluation and outcome prediction. Furthermore, the review highlights the importance of interpretability and small sample size learning of DL-based systems in clinical practice; specific recommendations for further expanding the clinical impact of DL-based systems are also provided.
- Advanced Application of Artificial Intelligence for Pelvic Floor Ultrasound in Diagnosis and Treatment
- Enze Qu, MD, Xinling Zhang, MD
- 2023, 7 (2): 114-121. DOI:10.37015/AUDT.2023.230021
- Abstract ( 168 ) HTML ( 12 ) PDF ( 11516KB ) ( 123 )
Artificial intelligence-based pelvic floor ultrasound helps the diagnosis, preoperative assessment, and postoperative monitoring of female pelvic floor dysfunction (FPFD). The application of artificial intelligence in pelvic floor ultrasound mainly includes automatic segmentation and measurement, the diagnosis of muscle injury, childbirth prediction and postoperational evaluation. It can not only overcome the problem of operator experience dependence but also improve work efficiency and simplify the workflow, which has popularized the application of pelvic floor ultrasound. However, most of the current research is still limited to the automatic segmentation of three-dimensional axial plane levator hiatus (LH). The automatic reconstruction, real-time tracking of 3D/4D images and the imaging navigation of pelvic floor surgery remain major challenges for researchers.
- Advances in the Research of Ultrasound and Artificial Intelligence in Neuromuscular Disease
- Tianxiang Li, BS, Fei Ji, BS, Ruina Zhao, MD, Huazhen Liu, MD, Meng Yang, MD
- 2023, 7 (2): 122-129. DOI:10.37015/AUDT.2023.230025
- Abstract ( 151 ) HTML ( 13 ) PDF ( 11416KB ) ( 108 )
Neuromuscular disease includes a wide range of muscular disorders, but it lacks convenient and effective tools for clinical diagnosis and therapeutic monitoring. As a widely used imaging tool, ultrasound can clearly display muscle structure and create basic conditions for accurate image analysis. At present, many studies have tried to obtain information on muscle function and pathological changes by analyzing the features of muscle ultrasound images, and have shown reliable results. However, the minimal changes in muscle structure and image texture are easy to be neglected, and manual segmentation and data analysis are time-consuming tasks. Artificial intelligence (AI) can accurately identify image changes and improve the efficiency of image analysis, and the muscle ultrasonic image analysis model developed based on AI has shown advantages in a large number of research results. This review summarizes the relevant studies of muscle ultrasound imaging and AI in the field of it, including a variety of research based on traditional AI methods or deep learning methods, as well as discusses the clinical significance of ultrasound analysis assisted by AI and the future exploration directions in this field.
- Application and Prospect of AI and ABVS-based in Breast Ultrasound Diagnosis
- Rui Chen, MM, Fangqi Guo, MM, Jia Guo, MD, Jiaqi Zhao, MD
- 2023, 7 (2): 130-135. DOI:10.37015/AUDT.2023.230017
- Abstract ( 132 ) HTML ( 7 ) PDF ( 11398KB ) ( 112 )
Breast cancer is the most common malignancy and the leading cause of death for women. Ultrasound is the main tool for breast cancer screening, but it can be influenced by the subjective factors of sonographers. With the continuous development of medical technology and artificial intelligence (AI), the application of breast ultrasound imaging technology is becoming increasingly widespread. Among them, the application of AI and automated breast volume scanning (ABVS) brings new opportunities and challenges for ultrasound diagnosis of breast diseases, while making breast ultrasound diagnosis more accurate and efficient. This article explores the application and prospects of AI and ABVS in ultrasound diagnosis of breast diseases.
- Ultrasound Image Generation and Modality Conversion Based on Deep Learning
- Shujun Xia, MD, Jianqiao Zhou, MD
- 2023, 7 (2): 136-139. DOI:10.37015/AUDT.2023.230011
- Abstract ( 158 ) HTML ( 16 ) PDF ( 11339KB ) ( 136 )
Artificial intelligent (AI) based on deep learning has been used in medical imaging analysis for years. Improvements have been made in the diagnosis of various diseases with the help of deep learning. Multimodal medical imaging combines two or more imaging modalities, providing comprehensive diagnostic information of the diseases. However, some modality problems always exist in clinical practice. Recently, AI-based deep learning technologies have realized the modality conversion. Investigations on modality conversion have gradually been reported in order to acquire multimodal information. MRI images could be generated from CT images while ultrasound elastography could be generated from B mode ultrasonography. Continuous researches and development of new technologies around deep learning are still under investigation and provide huge clinical potentials in the future. The purpose of this review is to summarize an overview of the current applications and prospects of deep learning-based modality conversion of medical imaging.
- Advances in Intelligent Segmentation and 3D/4D Reconstruction of Carotid Ultrasound Imaging
- Cancan Cui, MD, Zhaojun Li, PhD, Yanping Lin, PhD
- 2023, 7 (2): 140-151. DOI:10.37015/AUDT.2023.230015
- Abstract ( 136 ) HTML ( 8 ) PDF ( 11662KB ) ( 84 )
Cardiovascular disease (CVD) is one of the ten leading causes of death worldwide. Atherosclerotic disease, which can lead to myocardial infarction and stroke, is the main cause of CVD. The two main ultrasound image phenotypes used to monitor atherosclerotic load are carotid intima-media thickness (IMT) and plaque area (PA). Early segmentation and measurement methods were based on manual or threshold segmentation, snake models, etc. Usually, these methods are semi-automatic and have poor repeatability and accuracy. Segmentation of the carotid intima-media complex (IMC) and plaque in ultrasound based on artificial intelligence can achieve good accuracy. Compared with two-dimensional ultrasound, three-dimensional/four-dimensional ultrasound can provide spatial dynamic vascular information, which is helpful for doctors to evaluate. This study reviews the progress of artificial intelligence (AI) segmentation methods based on machine learning (ML) and deep learning (DL) used in the segmentation of the IMC and plaque as well as the 3D / 4D reconstruction of carotid ultrasound.
- Rapid Screening of Carotid Plaque in Cloud Handheld Ultrasound System Based on 5G and AI Technology
- Wenjun Zhang, MD, Mi Zhou, PhD, Qingguo Meng, MD, Lin Zhang, MS, Xin Liu, MS, Paul Liu, PhD, Dong Liu, PhD
- 2023, 7 (2): 152-157. DOI:10.37015/AUDT.2023.230018
- Abstract ( 124 ) HTML ( 8 ) PDF ( 11411KB ) ( 88 )
Objective: To evaluate the real-time accuracy of cloud handheld ultrasound system using AI technology in screening carotid plaque.
Methods: 2627 ultrasound images of the carotid artery are collected using the cloud handheld system. Bounding boxes of carotid plaques are labeled by qualified sonographers, and the dataset is trained using a lightweight YOLOv3 model. An additional and separate 390 images are collected and tested using the evaluation metrics average recall (AR), average precision (AP), and frames per second (FPS) for quantifying classification performance and time consumption.
Results: We use a plaque grading definition with a thickness of 1.2-1.5 mm defined as small plaque, 1.5-3 mm as medium plaque, and more than 3 mm thick as large plaque. Our model achieves APIoU=0.50 with 96.5%, with APlarge is 79.9%, APmedium is 90.7%, APsmall is 93.5%; ARIoU=0.50 is 64.5%, where ARlarge is 60.6%, ARmedium is 58.3%, ARsmall is 70.8%, and FPS is 33.3.
Conclusion: We establish a framework for data set construction, model selection, training, and testing of carotid ultrasound images and verify the effectiveness of real-time AI technology in the automatic detection of carotid artery plaque.
- ChatGPT Related Technology and Its Applications in the Medical Field
- Tairui Zhang, BS, Linxue Qian, MD
- 2023, 7 (2): 158-171. DOI:10.37015/AUDT.2023.230028
- Abstract ( 216 ) HTML ( 20 ) PDF ( 12139KB ) ( 146 )
ChatGPT is attracting widespread attention from all walks of life with its excellent multi-round dialogue ability and strong user intent understanding ability, triggering a new wave of artificial intelligence. From the perspective of technical analysis, this article sorts out the various related technologies used in the GPT (Generative Pre-training Transformer) series models as well as large-scale multimodal models, which are more powerful and perform better in multiple downstream tasks. Meanwhile, we guide users to use LLM (Large Language Model) along with GPT more scientifically to maximize their potential. Finally, we analyze the application prospect of the GPT as well as the large-scale multimodal models in the medical field, and the problems are discussed from the perspectives of the risks and limitations of large-scale models applied into the medical field.
- Domestic Large Model Technology and Medical Applications Analysis
- Chengwen Zhang, PhD, Xing Yu, MD
- 2023, 7 (2): 172-187. DOI:10.37015/AUDT.2023.230027
- Abstract ( 108 ) HTML ( 6 ) PDF ( 11943KB ) ( 73 )
In 2023, the capabilities of text communication, language translation, image generation, and code writing demonstrated by ChatGPT and GPT-4 have received widespread attention in and out of China. With years of data and technology accumulation, many Chinese companies and research teams continue to make efforts in the field of large models, and their models cover many industries and have a number of characteristic functions. This paper will introduce the type and development trends of large models, and sort out the methods and characteristics of domestic large models. Finally, we will explain the advantages and disadvantages of the domestic models, and analyze the application prospect and challenges of them in the medical field.
- AI-based ChatGPT Impact on Medical Writing and Publication
- Mofan Li, Yongyue Zhang, MM, Yang Sun, MM, Ligang Cui, PhD, Shumin Wang, PhD
- 2023, 7 (2): 188-192. DOI:10.37015/AUDT.2023.230013
- Abstract ( 224 ) HTML ( 7 ) PDF ( 11302KB ) ( 161 )
ChatGPT, an artificial intelligence (AI) software developed by OpenAI, is a powerful language model. ChatGPT is expected to perform a variety of tasks in the field of medical writing and publishing, including writing drafts, extracting article abstracts, and embellishing language. At the same time, ChatGPT has technical shortcomings and ethical challenges that have raised concerns. This review summarizes the issues faced by ChatGPT in the field of medical writing and publishing, and provides a reference for the development of standards and systems for the use of AI products such as ChatGPT.
- Application of the Virtual Reality in the Teaching of Ultrasonography
- Zheng Zhang, MS, Li Liu, MD, Desheng Sun, MD, Dirong Zhang, MD, Fengbei Kong, MS, Yalin Wu, PhD, Yu Shi, MD
- 2023, 7 (2): 193-196. DOI:10.37015/AUDT.2023.230026
- Abstract ( 109 ) HTML ( 6 ) PDF ( 11274KB ) ( 57 )
This article discusses the potential benefits of using virtual reality (VR) technology in the teaching of ultrasonography. VR technology can provide an immersive learning experience, enabling students to interact with simulated environments and practice various tasks. Ultrasonography has the characteristics of convenient, rapid, real-time feedback, and dynamic, and is indispensable in practical clinical disease diagnosis applications. Combining VR and ultrasound technology can provide a unique and effective teaching method for medical students and medical professionals. This article mainly discusses the current situation, advantages, and challenges of virtual reality technology in the teaching of ultrasonography to ensure their successful implementation in an educational environment.
- Development of 5G-based Remote Ultrasound Education: Current Status and Future Trends
- Jiaojiao Ma, MD, Xinying Jia, MD, Guanghan Li, MD, Dandan Guo, MD, Xuehua Xi, MD, Tongtong Zhou, MD, Ji-Bin Liu, MD, Bo Zhang, MD
- 2023, 7 (2): 197-203. DOI:10.37015/AUDT.2023.230022
- Abstract ( 102 ) HTML ( 6 ) PDF ( 11496KB ) ( 49 )
The rapid advancement of 5G technology has opened new possibilities for remote ultrasound education, offering the potential to enhance training, real-time consultation, and quality control for primary ultrasound doctors. The 5G remote ultrasound education has the potential to revolutionize the way primary ultrasound doctors are trained and supported, ultimately leading to improved patient care and outcomes. By understanding the current status and development trends of this cutting-edge educational approach, the medical community can better prepare for and contribute to its ongoing evolution. Looking towards the future, the development trends in 5G remote ultrasound education are expected to revolve around continuous improvement and innovation in educational methods and technologies. This includes the exploration of artificial intelligence and machine learning applications, the expansion of telemedicine and telementoring programs, and the development of personalized learning plans tailored to individual learners' needs. This article aims to offer an overview of the current status and applications of 5G remote ultrasound education, including the development of theoretical courses and network construction within our institutes, and to discuss future trends in this field.
- The Impact of Deep Learning on Ultrasound in Diagnosis and Therapy: Enhancing Clinical Decision Support, Workflow Efficiency, Quantification, Image Registration, and Real-time Assistance
- Won-Chul Bang, PhD, Vice President, Yeong Kyeong Seong, PhD, Jinyong Lee
- 2023, 7 (2): 204-216.
- Abstract ( 85 ) HTML ( 6 ) PDF ( 11598KB ) ( 60 )
This review article introduces the main concepts and architectures of deep learning networks for medical imaging tasks, such as classification, detection, segmentation, and generation. It then surveys how deep learning has been applied to ultrasound imaging for various purposes, such as image processing, diagnosis, and workflow enhancement. It covers different organs and body parts that can be imaged by ultrasound, such as liver, breast, thyroid, heart, kidney, prostate, nerve, muscle, and fetus. It also discusses how deep learning can help with view recognition, registration, and quantification, measurement, image registration for interventional guidance, and real-time assistance while scanning. Moreover, it explores how generative AI can be used in the future medical field by leveraging deep learning for ultrasound imaging, such as generating realistic and diverse images, virtual organs/patients with diseases, synthesizing missing or corrupted data and augmenting existing data for training and testing.