Professor Tang Chengwei, Huaxi Hospital: The latest research and development results, China's first hardware equipment-based digestive endoscopic AI products have been confirmed.

The application of artificial intelligence (AI) in medicine has attracted wide attention, and AI-assisted digestive endoscopy diagnosis has become a hot topic in the industry.
clinical endoscopic operation is completed by the endoscope physician, the diagnosis is completed by the physician's physical eye combined pathology biopsy, its essence is a process of accumulating experience and gradually perfecting.
AI-assisted endoscopy diagnosis is expected to improve the detection rate of endoscopic physicians on digestive tract lesions, effectively reducing misdiagnosis and omission.
after several years of development, AI-Medical has gradually entered the industry shuffle stage, from the initial "fresh" to "landing", at the moment, the real landing ai medical projects have more market and voice.
September 11-13, the 2020 Chinese Digestive Endoscopy Annual Meeting, hosted by the Chinese Medical Association and the Chinese Medical Association Digestive Endoscopy Credit Association, was held online.
meeting, professor Tang Chengwei, Chairman-designation of the Digestive Medicine Branch of the Chinese Medical Association and Digestive Medicine Department of Huaxi Hospital, delivered a wonderful speech on the report "Artificial Intelligence Assisted Digestive Endoscopy" and shared the latest developments in the heterogeneous digestive endoscopy AI project of Huaxi Hospital and Heath, attracting strong attention from the industry.
Digestive endoscopic AI research and development status quo and pain point: papers more products less, difficult to play a role in unknown lesions According to the 2015 cancer statistics published online in the Journal of Clinical Cancer (CA Cancer J Clin), China's digestive tract tumors alone, the annual incidence of stomach cancer as high as 6791 million cases, about 500,000 deaths, accounting for 17.7% of the total cancer death rate.
As the most direct examination equipment for digestive tract diseases, digestive endoscopy has been very popular in China, but China's digestive tract diseases, especially malignant tumors of early diagnosis and treatment work, not satisfactory, the main reason is that endoscopic doctor training cycle is longer, an experienced doctor needs at least ten years of training cycle, talent training depends on the stacking of time, talent is relatively scarce, restricting the level of diagnosis and treatment of digestive tract diseases.
Tang Chengwei pointed out that the high demand in the field of digestive endoscopy, the heavy workload of medical care, digestive endoscopy artificial intelligence technology and equipment research and development, will hopefully change this phenomenon.
the goal of digestive endoscopy AI is to improve diagnostic efficiency, improve diagnostic accuracy in an all-round way, build a new endoscopy system through AI technology, play the role of quality control of regional medical centers, and help implement China's hierarchical diagnosis and treatment policy.
" domestic digestive endoscopy treatment level is uneven, many patients therefore go through multiple endoscopic examination, increase the risk of related examination, waste of medical resources.
The development of digestive endoscopic AI technology, for high-grade hospitals, so that care to give patients more care, so that doctors free up more energy to do minimally invasive treatment, for low-grade hospitals, can overcome the weakness of doctors trained and inexperienced.
, " says Prof Tang.
meeting, Professor Tang Chengwei made a relevant interpretation of the current situation and pain points of digestive endoscopic AI.
pointed out: First, digestive endoscopic AI field papers, fewer products, really put into practice of the project is not many.
, endoscopic clinical diagnosis is still insufficient.
doctors often face unknown endoscopic lesions, only the detection of certain known lesions of AI aids, it is difficult to face a variety of unknown lesions to play an auxiliary role.
, AI testing standards temporarily failed to surpass endoscopy doctors.
bulging lesions are one of the main lesions observed in endoscopy, which may be emoticons, inflammatory lesions, early or progressive cancers, vascular lesions, smooth fibroids, etc.
further screening of bulging lesions under endoscopy is not difficult for endoscopists, and AI may not be better than endoscopy doctors, even if it is difficult to identify a small number of bulge lesions.
" endoscopy doctors most need AI assistance is to assist in the examination process to detect lesions in a timely manner, to avoid missing lesions, which is particularly useful for inexperienced, heavy workload or fatigue of doctors.
," Professor Tang Chengwei points out.
China's first digestive endoscopic AI products have been confirmed, the second generation of product development in March 2017, Huaxi Hospital and Heterogeneous Medical Technology Co., Ltd. jointly established the Huaxi-Heath Medical Artificial Intelligence Research and Development Center, with Professor Tang Chengwei's team as the core, the two sides in the digestive endoscopic artificial intelligence product development in-depth cooperation.
At present, Heath's research and development achievements in the field of endoscopic AI cover the whole digestive tract, including tumors, polygam, atrophy lesions, ulcer lesions, erosion lesions, vascular lesions and other multi-category more than 100 diseases.
Tang Chengwei, after 3 years of exploration and research, the center in the field of digestive endoscopic AI has achieved good results.
an artificial neural network based on the CNNs principle, which converts a large number of endoscopic bulge lesions images into complex data information, conducts endoscopic CAD system model training, and successfully creates the first real-time assisted diagnostic device for endoscopic imaging (endoscopic AI device).
After completing nearly 500 clinical trials, this digestive endoscopic image real-time auxiliary diagnostic device was approved by the pharmaceutical regulatory department in April 2020 as a medical device approved for clinical application, which is the first digestive endoscopic image assisted diagnostic device approved so far in China.
as the first domestic hardware equipment digestion AI products, what are the characteristics of this product? First of all, the product meets the design standards, the recognition speed is not inferior to the endoscopy doctor, can be found with the endoscopy doctor bulging lesions.
5mm bulging lesions were ≥ 96.81% more sensitive to 5mm bulging lesions under the intestinal glasses.
, endoscopic AI devices perform better than design expectations in the role of "assistant", providing a useful basis for the future development of industry standards for endoscopic AI devices.
, the product helps doctors detect lesions in real time and efficiently, and the series enables rapid and efficient identification of a wide range of lesions in the whole digestive tract (upper and lower digestive tracts) to effectively address clinical pain points.
from the disease coverage, clinical productization and other perspectives, the product is in the international leading position.
Professor Tang Chengwei introduced that digestive endoscopic image-assisted diagnostic equipment is only an initial product, artificial intelligence technology in the field of digestive endoscopy applications, not limited to one or several lesions of real-time detection, more applications will cover the different aspects of digestive endoscopy.
the latest research results have matured, and the second generation of products developed in cooperation with Heath iso-type has been basically stereotyped.
She pointed out that the second generation of products will be more closely integrated with the clinical, penetration into the digestive endoscopy more links, on the one hand in real-time video lesions identification has a new breakthrough, on the other hand, in the qualitative, quantitative and quality control of lesions, more technology introduced, to achieve in the digestive endoscopy front and back end to do a holistic layout.
Since 2017, AI, as a national strategy, has played an important role in skin cancer and diabetic retinal lesions, and we believe that future AI-based endoscopic image diagnostic recognition is expected to play an important role in the screening of early gastroscopic cancer in the endoscopy.
As the opening statement, medical AI after several years of development, its value has long been recognized by doctors, but the concept has long passed the "fresh period", only those who really apply in the medical scene, to solve practical problems of the project can have a real market voice.
Xu Wenxuan Source: Arterial Network Copyright Notice: All text, images and audio and video materials on this website that indicate "Source: Mets Medicine" or "Source: MedSci Originals" are owned by Mets Medical and may not be reproduced by any media, website or individual without authorization, and shall be reproduced with the words "Source: Mets Medicine".
all reprinted articles on this website are for the purpose of transmitting more information and clearly indicate the source and author, and media or individuals who do not wish to be reproduced may contact us and we will delete them immediately.
reproduce content at the same time does not represent the position of this site.
leave a message here.