Breast cancer has become one of the most serious malignant tumors that threaten the health of women in China. On March 23, 2018, the National Cancer Center released the latest data on breast cancer in Chinese women, estimating the incidence and mortality of breast cancer in Chinese women in 2014 (due to The existence of collection and statistical work, the general data will lag 3 years), in 2014 the national female breast cancer new cases about 278,900 cases, accounting for 16.51% of female malignant tumors, ranking first in female malignant tumors.
Early diagnosis and early treatment is the key to breast cancer treatment. At present, the detection rate of stage I in Chinese women with early breast cancer is only 20% to 25%, and the 5-year survival rate of patients is only about 80%. If the standardized treatment is found in the early stage of breast cancer and received early, the 5-year disease-free survival rate of the patient can reach 95%, and the 5-year survival rate of the second-stage patient can reach more than 80%.
Faced with huge screening needs, misdiagnosis and missed diagnosis of doctors' scarcity and manual reading have become an urgent pain point. At present, artificial intelligence has become a technical means to face the efficiency of doctors.
Recently, in the 2018 Future Medical Top 100 Forum, Deshang Yunxing released a breast cancer artificial intelligence-assisted diagnostic product. This is another medical-assisted diagnostic product after accurate preoperative planning, intraoperative navigation, postoperative evaluation, and intelligent auxiliary diagnosis system for ultrasound thyroid nodules.
 High incidence of breast cancer, screening for pain points
In the face of high-risk conditions, non-invasive auxiliary examinations commonly used in breast cancer screening include mammography, ultrasound, and magnetic resonance (MRI).
In the comparison of the three inspection methods, the molybdenum target inspection is inexpensive, mainly using X-ray film. Whether domestic or foreign, molybdenum target has become the main means of screening for female breast cancer. This method has the most The advantage, in turn, can be found in asymptomatic or untouchable tumors, and the diagnostic efficiency is even higher than that of magnetic resonance.
The advantage of ultrasound is that the ultrasound of the breast is not radioactive, and can be repeatedly checked according to the needs; and the ultrasound can clearly distinguish the level, and the accuracy of the cystic (the liquid inside the cystic nodule) and the solid mass can reach 100%, which can be roughly judged. The benign and malignant tumors; ultrasound can guide the biopsy and understand whether the lymph nodes on the axillary and clavicle have metastasis; and the examination is convenient and inexpensive, which has advantages in the diagnosis of dense breast and breast hyperplasia.
Another non-invasive method of examination is magnetic resonance. This type of examination is characterized by high sensitivity and no radiation hazard, but it is expensive, so it cannot be used as a means of screening for screening. Generally, it cannot be confirmed as a molybdenum target and color ultrasound. The patient's means of further diagnosis.
China now screens breast cancer through various means for women aged 30-60 years. It is different from fatty women in Europe and America. More than 50% of Chinese women's breasts are dense and have less fat. Doctors are needed. Careful search for tiny bumps, calcifications, or structural distortions hidden in dense images during the reading process usually takes a long time.
Compared with European and American doctors, domestic doctors have a lot of tasks. DeJuan Yunxing breast product leader Zhang Juan is a senior medical imaging expert who has been engaged in image diagnosis for more than 30 years. She told reporters that the number of doctors in Europe and the United States may be around 20-30 patients a day, while domestic doctors At least one day's film of 50 patients is needed, and reports are written from morning till night, and there will be a state of fatigue.
In the face of huge demand, the number of doctors specializing in breast diagnosis is very scarce, and the training period is too long. This has become a difficult problem in front of the imaging department. From medical education to hospital regulation, a senior breast cancer is cultivated. Diagnostic experts, it takes at least 5-10 years.
Faced with the high incidence of breast cancer, the sooner the discovery, the better the treatment effect, the liberation of doctors from the heavy reading task, reducing misdiagnosis and missed diagnosis, artificial intelligence is the most appropriate auxiliary means.
Artificial intelligence assisted diagnosis of breast cancer, the diagnostic level reached the director level
Deshang Yunxing Medical Technology was established in 2013. It features ultrasound artificial intelligence and uses AI technology for preoperative planning, intraoperative navigation and postoperative evaluation of tumor interventional surgery. The company's R&D team has formed a research and development team led by young doctors with the core of experts and scholars. It has done many internationally renowned academic institutions such as Harvard University, New York University Kelang Mathematical Institute, and German Einstein Institute. Academic research has published nearly two hundred academic papers in international authoritative journals.
Hu Hairong, Chairman and General Manager of Deshang Yunxing, said: "We are not just developing artificial intelligence-assisted diagnostic systems for medical imaging, but combining auxiliary diagnosis and ultimately for treatment, creating an AI precision diagnosis and treatment platform."
In this release of the artificial intelligence breast-assisted diagnosis product, Deshang Yunxing cuts through the means of ultrasound and X-ray, assists with artificial intelligence technology, and the entire research and development process has been cleaned and sorted by massive data, strictly controlled by senior breast experts. According to the international ACR standard, the image features of the breast image are extracted and repeatedly confirmed to ensure high accuracy.
In terms of product function, Deshang Yunxing qualitatively and quantitatively analyzed the lesions according to the national breast cancer guidelines, and classified the lesions according to the ACR standard. At the same time, this product will be designed according to China's national conditions, suitable for different versions of the top three hospitals and the majority of grassroots hospitals, hoping to reduce the medical pressure of large hospitals, and can provide auxiliary training for the majority of grassroots hospitals, truly help graded diagnosis and treatment.
In some specific tasks, such as face recognition in computer vision technology applications, computers have surpassed the human eye. The essence of artificial intelligence is that after the amount of data is increased, the processing speed becomes faster and the processing cost becomes lower. Such engineering advances have led to an inflection point in the application. In the medical field, image-assisted diagnosis technology based on big data and artificial intelligence has become a research hotspot. Hospitals and doctors are also embracing artificial intelligence, and handing over heavy and simple primary work to artificial intelligence.
In the application of ultrasound AI, in August 2018, Shanghai Ruijin Hospital initiated and convened more than 400 hospitals in China. Nearly 1,000 ultrasound doctors established the “Chinese Thyroid and Breast Ultrasound Artificial Intelligence Alliance†to establish a multi-level medical institution covering the whole country. The thyroid and breast ultrasound database, the application of ultrasound AI provides effective help for accurate diagnosis and treatment. Unlike the track of AI products in radiology, the ultrasound AI industry is not yet saturated, and there are not many companies occupying this track. Therefore, this field is also facing huge opportunities.
The product has data and algorithm advantages, and the design meets the needs of graded diagnosis and treatment.
Ultrasound is widely deployed in primary hospitals due to its low price and no radiation, but ultrasound doctors are scarce. Different from the process of collecting and reading images of CT, nuclear magnetic, X-ray, etc., the difficulty of ultrasonic detection is that image acquisition and reading need to be completed at the same time.
In addition, compared with the results of magnetic resonance, CT and electrocardiogram examinations, most of the ultrasound images rely on the dynamic images of different sections collected by doctors for diagnosis. The ultrasonic technician's personal operation skill level is relatively high, and the doctor's scanning technique is different. Patients have individualized differences, doctors' observer differences and other factors, which can easily lead to misdiagnosis or missed diagnosis.
Deshang Yunxing develops artificial intelligence-assisted diagnostic products for breast cancer, aiming at solving graded diagnosis and treatment and reducing the pain points of doctors.
Although there are still some difficulties in the development of ultrasound AI, such as the lack of a large amount of data training. However, the enthusiasm of the major companies for ultrasonic AI is not diminished. In addition to Tencent, Samsung, Siemens, Mindray and other large companies also have a layout in the field of ultrasonic AI. In addition, Imagine Technology, Etu Medical and other startup companies also have related products of ultrasound AI.
Hu Hairong believes that the advantages of Desheng Yunxing's artificial intelligence-assisted diagnostic products come from two aspects, one is data and the other is algorithm.
In terms of data, there are tens of thousands of samples for training. In terms of the quality of data annotation, Deshang Yunxing cooperated with many hospitals and personally marked by a breast expert with more than 30 years of work experience. She emphasized: “by experts The data marked, the accuracy of finding the lesion is much higher than the data labeling made by the average doctor."
At present, the accuracy rate of the breast cancer artificial intelligence diagnosis product identifying the mass is 89%, and the calcification recognition degree is over 90%. With the continuous optimization of the algorithm, the accuracy rate is still increasing. When the product is used in the clinic, the diagnostic level can reach the level of the expert at the director level.
In terms of algorithms, according to the introduction, Desheng Yunxing's algorithm team is applied mathematical background. Therefore, they independently developed a deep learning framework and named it Light3, which means light and flexible meaning for different diseases. The frame can be arbitrarily adjusted as needed to adjust parameters and improve the accuracy of the model. The open source deep learning framework cannot be modified arbitrarily, so the accuracy rate is limited. Hu Hairong said: "We have done our own comparison. In the ultrasound thyroid project, the accuracy of Light3 is about 30~40% higher than that of open source."
In terms of product design, according to the needs of graded diagnosis and treatment, Deshang Yunxing has made different product designs for the top three hospitals and grassroots hospitals. For the research needs of the top three hospitals, the product functions are relatively perfect; for the basic medical treatment needs of the primary hospitals, The system will adopt a certain "slimming" method to remove redundant functions and provide them to the primary hospital.
In addition, the product can be trained for young doctors at different levels of hospitals. Because the entire team is lacking, it is impossible for these doctors to return to school for professional training, so the use of artificial intelligence to assist diagnosis is also an effective training method.
On a not-so-congested track, Desheng Yunxing's ultrasound AI products are exploring their own business model. For most AI companies, the ability to pass the certification of relevant qualifications is a "deaf" in front of them.
Yan Yen, general manager of Deshang Yunxing Marketing, said that the current medical AI products on the market are more concentrated on the radiation track, and the AI ​​products of Desheng Yunxing on the ultrasonic track have all-round advantages. Due to regulatory restrictions, there is currently no mature business model in the entire AI industry. This limitation has actually increased the barriers to entry for medical artificial intelligence products.
He stressed: "Deshang Yunxing will do more investment in the market, including finding upstream and downstream partners. Not only the top three hospitals, but also the grassroots hospitals and other experienced products on the product line, can help data development, products Promoted hospitals. The focus of the current investment is on improving product quality and accuracy. Once the threshold is released, we can take advantage of our own advantages in the entire sales market, and take a different path."
At present, Deshang Yunxing has introduced industrial capital and good shares and funds of Fosun Pharma and Huagai Capital in the A and B round financing respectively. According to Hu Hairong, Deshang Yunxing is introducing the next round of cooperation in capital and resources. He hopes that the capital and partners with rich medical resources in the industry will work together to bring artificial intelligence products to clinical diagnosis and treatment.
To Reduce Fat Powder,Reduce Fat Powder,Lose Fat Extract,Natural Green Coffee Bean Weight Loss
Shaanxi Zhongyi Kangjian Biotechnology Co.,Ltd , https://www.zhongyibiology.com