Abstract ： Hua Wei Yun EI Innovation Incubation Lab United with the school of telecommunications, Huazhong University of science and Technology 、 It is one of the top international journals in the field of radiology in Peking Union Medical College Affiliated to Huazhong University of science and technology Radiology（《 Radiology 》） The latest research results have been published on .
recently , Top international journals in the field of radiology Radiology（《 Radiology 》） From China to the cloud EI Innovation Incubation Lab United with the school of telecommunications, Huazhong University of science and Technology 、 The Department of Radiology, Union Medical College Affiliated to Huazhong University of science and technology has jointly published the latest research results ：AI The sensitivity of the algorithm is as high as 97.5%, To help doctors improve their clinical diagnostic sensitivity by about 10 percentage , The rate of missed diagnosis decreased 5 percentage .
Radiology The latest release of Huawei cloud 、 Huazhong University of science and technology latest joint medical research results
This time Huawei cloud EI Innovation Incubation Lab United with the school of telecommunications, Huazhong University of science and Technology 、 This paper was jointly published by the Department of Radiology, Union Hospital, Tongji Medical College, Huazhong University of science and technology 《 Based on deep learning CT Angiography cerebral aneurysm detection algorithm 》（“Deep Learning-Based Algorithm for Detecting Cerebral Aneurysms on CT Angiography Images”）. This paper describes the development of deep learning based on fully automated 、 Highly sensitive CT Algorithm for detecting cerebral aneurysms with angiographic images .
As in the field has been recognized as the latest 、 Authoritative reference to the highest quality research , Top journals in radiology Radiology Last published . Enough to prove that , This is the latest research result of artificial intelligence in the field of medical imaging , It is also a heavy scientific research achievement in the diagnosis of cerebral aneurysms by using artificial intelligence technology , In the field of radiology is a highly professional research results .
AI+ Medical imaging , Break new ground
A cerebral aneurysm is a weakened area of blood vessels in the brain , If not treated , They may leak or break , Sometimes it can be fatal . The risk of aneurysm rupture depends on the size of the aneurysm 、 Shape and location , Therefore, the detection and feature extraction of cerebral aneurysm is the key to guide the treatment .
CT Angiography （CTA） At present, it is the main imaging method to evaluate intracranial aneurysms , But because of the small size of the aneurysm and the complexity of the intracranial vessels , Even a professional radiologist can take a long time to diagnose , Some small aneurysms may also be missed .
“ Artificial intelligence + Medical imaging ”, It is to apply the most advanced artificial intelligence technology to medical imaging diagnosis , Specific application scenarios of artificial intelligence to help doctors diagnose patients' conditions , It can be widely used to identify and label all kinds of lesions , Such as breast cancer 、 Lung cancer 、 Cervical cancer, etc . The use of AI deep learning , It will help doctors diagnose cerebral aneurysms more quickly and accurately .
In order to shorten the diagnosis time 、 Improve diagnostic accuracy , Hua Wei Yun EI Innovation Incubation Lab United with the school of telecommunications, Huazhong University of science and Technology 、 Union Hospital Affiliated to Tongji Medical College of Huazhong University of science and Technology （ Wuhan Union Hospital ） Radiology department has developed a deep learning algorithm , Can identify CT Angiography for the detection of cerebral aneurysm .
Used by Huawei cloud joint team 500 Multi patient CT Angiography , Based on Huawei cloud one stop shop AI Development platform ModelArts To train the model , And then use another one 534 Zhang CT Angiography data set testing , These include 649 Aneurysms .
Whole data display , The algorithm detects 649 Of the intracranial aneurysms 633 individual , The sensitivity is 97.5%. The study also found that 8 New aneurysms .
“ Deep learning algorithm has shown excellent performance in detecting aneurysms . We found that , A very small number of aneurysms that were overlooked in the initial clinical diagnosis report , It was successfully identified by deep learning algorithm .” Dr. Long Xi, a radiologist from Wuhan Union Medical College Hospital who participated in the joint project, said . More Than This , Through the test results show that , Deep learning algorithm has potential in the diagnosis of cerebral aneurysms , It is expected to be a diagnostic tool for the second opinion in clinic .
Help doctors , Instead of replacing
AI There are many advantages , Mainly because it is not subject to the level of experience 、 Working hours and emotions affect human performance . For tertiary hospitals , Artificial intelligence + The introduction of medical imaging can improve the traditional mode of highly relying on doctors to read films manually , To a certain extent, relieve the pressure of medical imaging diagnosis , At the same time, it can also meet the scientific research needs of tertiary hospitals .
For basic hospitals , Compared with the top three hospitals , Its ability to process complex images 、 The judgment ability is weaker , Therefore, the rate of misdiagnosis and missed diagnosis is higher . Artificial intelligence through the early analysis and processing of image diagnosis results , Can increase the number of screening , Reduce the rate of misdiagnosis and missed diagnosis , So as to improve the level of comprehensive medical treatment .
" Cooperate with Huawei cloud for development AI The purpose of deep learning algorithm is to help radiologists , Instead of replacing them . Further collection is needed in the future 、 analysis 、 Validate more heterogeneous high quality data , Further verify the algorithm , This is the key to evaluate its popularization and applicability to daily clinical work , need AI Algorithm expert 、 Imaging experts and other experts to further cooperation ." Said Dr. loncy .
To be a booster of the medical industry
In the medical industry , Huawei cloud EIHealth The platform is committed to building artificial intelligence in the medical industry “ Black land ”, Supporting biotechnology companies 、 The hospital 、 Pharmaceutical companies, etc AI And digital transformation , Break into the deep water area of the industry , Find the second curve of business development . He also said , at present ,EIHealth It mainly covers the genome 、 Three directions of clinical research and drug development .
stay 2020 Fight against novel coronavirus pneumonia in 2006 ,EIHealth Many achievements have been made in the above three fields , Help fight the epidemic in many ways , Realize the maximum value of the platform itself .
for example , In Wuhan , Wuhan hope group and Huawei form a joint team , In the absence of fixation IT Temporary hospitals for infrastructure , The rapid gene detection technology of new coronavirus based on the third generation sequencing technology has been realized and verified .
In guangdong , Hua Wei Yun EI The team assisted the Guangdong Provincial regeneration laboratory , Accelerating the novel coronavirus pneumonia image AI Model development , Novel coronavirus pneumonia image AI model training , And in cooperation with Huazhong University of science and technology 、 Blue net technology jointly launched AI Auxiliary medical imaging screening service . The service integrates the AI+CT Reasoning services , It can be directly given to the hospital for clinicians to use .
More Than This , International Conference on top artificial intelligence medical imaging MICCAI（ International Conference on medical image computing and computer aided intervention ） stay 2019 year 、2020 Huawei cloud artificial intelligence medical image analysis team was successively published in 5 Papers , Covering cervical cancer screening 、 Stroke segmentation 、 Ventricular segmentation and automatic generation of plain film diagnostic reports .
本文为[Huawei cloud developer community]所创，转载请带上原文链接，感谢