Based on proprietary medical knowledge map, intelligent disease database can be constructed through deep mining, extraction and structurization of data with artificial intelligence.
With Yitu Healthcare's intelligent disease database, "AI + big data" research platform, regional key research data center, and intelligent assisted diagnosis system and other solutions are available to facilitate clinics, teaching, research, and management.
Yitu's care.ai Intelligent Integrated Solution for Child Growth and Development Evaluation is an industry-leading solution that combines artificial intelligence technology and hardware device. The solution thoroughly integrates the ultra-low dose X-ray full-shield bone filming machine with the AI system for growth and development evaluation, optimizing and accelerating the entire evaluation process. With its ultra low-dose characteristics, the solution further eliminates the need for a shielded room and minimizes the radiation risk for children. Through these means, the solution is able to facilitate early screening of childhood diseases, empower grassroots diagnosis and treatment, and foster care for the health of children.
Based on 3D RetinaNet and other algorithms and patented technologies, whole-chest multi-task intelligent diagnosis can be achieved with the AI-assisted diagnostic system for chest CT developed by Yitu Healthcare. The system can identify most of the common lesions in lung, such as nodules, masses, patches, stripes, cystic shadows, etc., followed by quantified intelligent analysis and diagnosis; meanwhile, mediastinal and pleural lesions can be identified to provide imaging evidences for clinical staging of malignant lesions. The intelligent 4D technology adds accurate time dimension on 3D images, and provides more abundant decision-making information for clinical imaging diagnosis.
Based on deep learning and NLP technologies, the system can create a disease risk prediction model according to genetic factors, living habits and environmental risks. It can integrate the test and examination results from clinical medical records, multimodality imaging, tumor markers and genetic tests, and generate structured reports. It can perform comprehensive diagnosis on patients with medical guidelines to provide practical and effective diagnostic assistance for doctors.
Based on computer vision and deep convolution neural network technologies, standardized gland classification, intelligent lesion detection, and full-characteristic recognition of masses, asymmetries, calcifications and structural distortions can be performed automatically; lesions can be graded accurately according to BI-RADS standard, and structured report can be generated, which provides a one-stop high-efficient AI solution for mammography.
Based on 3D imaging reconstruction and volume registration algorithm, the system can perform intelligent detection and classification on mammary glands, detect and identify lesions. The lesions are automatically graded and diagnosed according to the BI-RADS standard. Structured reports are generated based on clinical guidelines. The system can dramatically improve doctors' efficiency and quality of image reading and diagnosis, and reduce the risks of missed diagnosis and misdiagnosis for breast ultrasound.
With Yitu's proprietary and well optimized AI model, the system uses the BonNet intelligent anatomical locating engine and 3 main-stream bone age standards, including TW3, GP Method or CHINA 05, to provide accurate bone age evaluation result in 1 second. Customized structured report of comprehensive growth and development based on physiological and biochemical indicators is generated to bring great convenience to clinical diagnosis and treatment, teaching and scientific research in pediatrics.
Based on NLP technology, the system applies in structurizing medical records, disease consultation, disease prediction, critical disease identification, and some other scenario models. It can provide a series of assistant tools, such as AI guidance, AI pre-inquiry, AI clinical decision support, AI follow-up, etc., to establish a total AI solution which covers process before, during and after clinic, optimizes the outpatient process and serves patients, doctors and hospitals.
Based on computer vision and deep convolutional neutral network technologies, the system can intelligently detect acute ischemic lesions in brain CT/DWI-MR multimodality imaging, determine their locations, characters, quantify imaging indicators, locate offending blood vessels and provide assistant clinical decisions based on clinical information. It can automatically perform imaging evaluation in seconds, which was originally demanding for healthcare providers. It can also assist doctors for accurate evaluation of treatment risks, thus enabling patients to access immediate and optimal treatment.
Based on computer vision and deep learning technologies, the system can perform automatic delineation of lesions in thyroid ultrasound imaging, identify lesions as diffusive or local, cystic or solid, single or multiple. It can also intelligently grade the lesions and generate structured reports, which will dramatically improve doctors' efficiency and quality of image reading.
Based on deep convolutional neural network and multiple optimization technologies, the system can conduct intelligent quality evaluation on multimodality imaging data, calculate quality control score in seconds, and support real-time statistical analysis of quality control results. It will dramatically improve the operation and management efficiency of hospital and radiology department by networking, automation and standardization of medical imaging quality control.
Based on medical record data and evidence-based database, the Intelligent Internet medical platform utilizes core AI technologies and diagnosis and treatment modules to assist in the construction of intelligent hospital. Total solution covering the whole clinical workflow are provided for outpatient optimization, standardized triage, clinical decision support and efficient allocation of medical resources, so as to improve the subject development and revenue of central hospitals, to enhance the medical service capability of primary institutions, and to facilitate the establishment of feasible model for tiered diagnosis and treatment.
AI is improving the medical productivity and expanding the healthcare frontiers
YITU has performed a joint study with Guangzhou Women and Children's Medical Center on the applications of NLP technologies in pediatric diagnosis, with the research letter being published in Nature Medicine (IF 32.621). This was the first time in the world that research findings of NLP technology on intelligent clinical diagnosis based on text-based Chinese electronic health records (EHR) had been published in a top international medical journal.
Yitu Healthcare cooperates with hundreds of top medical institutions in China to use AI applications to improve their capabilities in providing medical services and to provide scientific evidence for early screening, diagnosis and treatment against high-risk cancers, scientific research and national public health decision, and to contribute to Plan of Healthy China 2030. YITU's "AI Map of Cancer Screening" for high-risk cancers can reduce the doctors' workload, eliminate misdiagnosis and missed diagnosis, provide strong technical support for large-scale early screening of diseases, and drive cancer screening in China towards the "AI+" age.
Yitu Healthcare has become strategic partner of iKang Healthcare Group and will cooperate substantively in child growth and development monitoring, screening and physical examinations of multiple high-risk cancers. Medical AI application will be expanded into healthcare management, driving the national healthcare management into the "AI+" age.
Yitu Healthcare cooperated with multiple departments of West China Hospital, such as Information Center, Radiology, Pathology, Respiratory Medicine, etc., integrated HIS, LIC, PACS and RIS systems, and created the AI scientific research database of lung cancer, which includes multi-dimensional clinical indicators covering clinical text data, imaging and pathological data.