AI medical imaging has ushered in the era of commercialization

Introduction

AI+ medical imaging uses deep learning to achieve functions such as lesion identification and annotation, and target area delineation, helping doctors discover hidden lesions faster and complete diagnosis and treatment. Currently, global AI medical care is in a stage of rapid development. Countries are actively launching various support policies. A number of AI medical imaging equipment have been approved for sale, covering many fields such as heart, tumors, and lungs. AI medical imaging has officially ushered in the era of commercialization.

The upstream of the AI medical imaging industry chain is mainly divided into hardware providers, software providers and algorithm and data companies. Hardware providers mainly provide MCU (Microprocessing Sensor), ASIC (Application Specific Integrated Circuit), DSP (Digital Signal Processor), precision resistors, power chips, sensors and other electronic components. Major companies include ABB, TSMC, etc. Software providers mainly provide basic medical equipment software such as image acquisition software, image archiving, output systems and image printing systems. Representative companies include GE General and Siemens. The algorithm platform and data platform mainly provide learning materials for later AI products. The algorithm models of most AI medical imaging companies are derived from studying various papers.

In the middle of the AI medical imaging industry chain, companies such as Siemens and Philips rely on medical imaging equipment or image management and writing systems to enter medical institutions. Internet companies such as Baidu, Huawei, and Tencent have mastered and developed advanced AI technologies due to their capital and algorithm advantages, and have made breakthroughs at the application layer. Tencent’s fundus image-assisted diagnosis software for chronic glaucoma-like optic neuropathy and pneumonia CT imaging The triage and assessment software has been approved as a Class III medical device. Among technology companies, Eagle Eye Technology, Shenrui Medical, etc. have made early arrangements in AI medical imaging, focusing on the construction of the application layer, with their own algorithms, and have received high attention from capital. They are the types of companies that currently mainly occupy the domestic market share.

The downstream application scenarios of the AI medical imaging industry chain include hospitals, community clinics, imaging centers, physical examination centers, etc., which mainly perform image reconstruction, brain disease diagnosis, injury identification, etc. From the perspective of purchasing entities, tertiary hospitals are still the main users, and the penetration rate of terminal application scenarios is insufficient. According to CMAI 2022 Forum data, currently 74% of tertiary hospitals have AI products, 32.5% of secondary hospitals have AI products, and only 32.5% of primary-level institutions have AI products. About 10%.

FDA further speeds up the approval of AI medical products. In order to accelerate the development of AI medical care, the FDA broke the restrictions on AI products by traditional medical device approval policies, established a separate AI and digital medical review department, and reduced some Class III medical AI products to Class II by lowering the threshold for medical AI products. category products for approval to shorten the approval process. Under this circumstance, the number of AI/ML medical products approved by the FDA ranks first in the world. As of October 2022, the FDA has supported a total of 521 AI/ML medical devices.

Global AI medical imaging has entered a period of rapid development. Nova One Advisor predicts that by the end of 2027, the global scale of AI medical imaging will be approximately US$20 billion, with a compound annual growth rate of 35.9%. From January to July 2022 alone, the FDA approved 79 AI/ML medical devices, an increase of 58% month-on-month. Many companies around the world are actively launching new AI medical imaging products. GE launched the software-intensive platform Vivid E95/E90/E80 in July 2022 for cardiac imaging; Siemens launched the intelligent integrated imaging software syngo.CT for bone diagnosis; Philips launched the AI application MRCAT Head & Neck, with In the treatment of head and neck tumors.

In terms of market share, established medical imaging institutions such as GE, Philips and Siemens occupy dominant positions in the industry, with these three institutions accounting for 80% of the global market share. Although the competitive landscape in the field has been basically determined, there are still start-up companies in various countries entering the field from segmented tracks and emerging in the field. Viz.ai and Subtle Medical in the United States are exploring the launch of Viz SDH and SubtlePET for stroke monitoring and machine image quality improvement. BriefCase and HealthCCSng developed by Israeli company Aidoc Medical have both been approved by the FDA for use in large vessel occlusion and cardiac imaging.

The global AI medical imaging track capital market is active. In the first half of 2022, 329 financing events occurred in the U.S. digital health track, with a total financing amount of US$10.3 billion. Viz.ai completed US$100 million in Series D financing. The company's independently developed Viz ANEURYSM system was approved by the FDA to detect stroke-related indicators. RedBrick AI completed a US$4.6 million seed round of financing. The company mainly accelerates the development and application of artificial intelligence in clinical environments by quickly annotating data on medical images. Subtle Technology has completed tens of millions of US dollars in Series B financing. The company previously received US$12.2 million in Series A financing. Its self-developed SubtlePET and SubtleMR systems were approved by the FDA to optimize MRI image quality. South Korea's Lunit (Lunying Medical) successfully IPOed on the country's GEM in August 2022. Lunit specializes in the field of tumor diagnosis and treatment. Its core products are Lunit INSIGHT, an AI image analysis solution for cancer screening, and Lunit SCOPE, an AI biomarker platform.

Foreign medical imaging giants and R&D companies are actively joining forces. Siemens has cooperation with RaySearch, Varian Medical Systems, and elekta. At the same time, the company will acquire Varian Medical in 2021 to enhance its competitiveness in radiotherapy. GE acquired Elekta in 2022 and cooperated with RaySearch to increase research and development in radiotherapy. In terms of application implementation, Subtle Medical has reached application implementation business agreements with RadNet, the American imaging center alliance, Capstone, the medical procurement alliance, and China Panorama Medical. Previously, Shentou Medical had reached business agreements with European Affidea and South American DASA.

China's AI medical imaging industry situation

China is accelerating the improvement of the AI medical imaging supervision system, and AI medical imaging products have entered the era of commercialization. In 2017, the CFDA released a new version of the "Medical Device Classification Catalog", adding new categories corresponding to artificial intelligence-assisted diagnosis, and the approval channel for medical imaging AI products has been initially established. In July 2019, the CFDA released the "Key Points for Review of Deep Learning-Assisted Decision-Making Medical Device Software", marking the official opening of a green channel for the approval of three types of medical imaging AI products. In January 2020, Keya Medical's coronary blood flow reserve fraction calculation software became the first AI medical imaging device to be approved as a Class III device. Subsequently, a number of AI medical imaging devices such as Infer Medical's pulmonary nodule CT image-assisted detection software and Yukun Technology's coronary CT angiography image vascular stenosis auxiliary triage software have received approval notices. This marks that my country's AI medical imaging products have officially entered commercialization, and the domestic AI medical imaging track is also expected to usher in a new wave of development.

Competition in my country's AI medical imaging field is fierce, and a circle of high-quality companies has emerged. In 2021, my country's medical imaging equipment market size is 95.4 billion yuan, and is expected to increase to 148.6 billion yuan in 2026, with an average annual compound growth rate of 9.3%. Among them, the AI medical imaging market size is 820 million yuan, which is expected to increase to 13.76 billion yuan in 2025, with a compound growth rate of 102.4%, higher than the global AI medical imaging growth level. Although it entered this field late, the development level of my country's AI medical imaging technology has advanced by leaps and bounds. Up to now, my country has approved 25 domestic Class III medical device companies, with 45 approved products, mainly used in lungs, new blood vessels, skull, bones and other fields.

Shukun Technology, Eagle Eye Technology, Infer Medical, Shenrui Medical and Shanghai United Imaging are representatives of outstanding domestic AI medical imaging companies. The five major manufacturers collectively account for 53% of the domestic AI medical imaging market share. Eagle Eye Technology is the first domestic AI medical imaging listed company focusing on the field of ophthalmology. Its independently developed diabetic retinopathy diagnostic product Airdoc-AIFUNDUS has reached the world's advanced level. Shanghai United Imaging operates in multiple fields such as craniology, orthopedics, and lungs, and has a number of world-first and China-first innovative products with completely independent intellectual property rights, including 2-meter Total-body PET-CT and whole-body 5.0T magnetic resonance uMR Jupiter. The company's 6 products and technologies, including uMR 780 and uCT ATLAS, have been approved by the FDA.

Due to its advantages in the fields of AI and big data, major Internet companies are the most active in the domestic AI medical imaging capital market. Since 2020, a total of 89 financing events have occurred in the domestic AI+ medical field, with a total financing amount of 17.5 billion yuan, of which medical imaging accounts for about 1/3 of the total financing. Major Internet companies are showing positive performance. Alibaba Health invested 225 million yuan in Wanli Cloud to deploy the medical imaging platform "Doctor You". Tencent led a 100 million yuan investment in Voxel Technology to explore the integration of AI medical imaging products and hospital scenes. Baidu made a strategic investment of 20 million in Neusoft Medical and became the company's 10th largest shareholder. At the same time, the AI medical imaging track has achieved remarkable results, and many companies in the field have successfully gone public, including Keya Ark, Shanghai United Imaging, Shukun Technology, Infer Technology, and Eagle Eye Technology.

Summary

The current global AI medical imaging field is still in a period of rapid development and exploration. As a new business format combined with manufacturing and artificial intelligence, AI medical imaging has the characteristics of multi-discipline and high added value. It is one of the important emerging industries in the future. The development level of AI medical imaging technology in my country has advanced by leaps and bounds. United Imaging Medical, Mindray Medical, GE Healthcare, and Philips are in the first echelon, and Lepu Medical and Canon are in the second echelon. As the demand for medical imaging in the medical market continues to grow, the shortage of imaging doctors in the country has become increasingly apparent. The application of artificial intelligence can greatly improve the problem of low efficiency in medical image analysis. As an emerging industry, competition in the domestic AI medical imaging track is still fierce, and the "domestic substitution" of medical equipment also provides fertile ground for development for more Chinese companies and equipment manufacturers with self-research capabilities.