Diagnosing Medical Images with Artificial Intelligence

The Healthcare Artificial Intelligence Market size was valued at USD 1.3 billion with North America region holding 50% of the market share in 2018 and is expected to reach USD 7.4 billion witnessing a 41.7% Cumulative Average Growth Rate (CAGR) from 2018 to 2023.

AI will play a key role in enabling radiology departments to cope with the ever-increasing volume of diagnostic imaging procedures amidst the chronic shortage of radiologists in many countries. According to a Signify Research carried out in 2018, the market for AI based medical imaging will cross USD 2 billion mark by 2023 with around 32% CAGR over 2018-23 period.

As per Global Market Insights research, the US Medical imaging and diagnosis market was valued around USD 100 million in 2018 and will surpass USD 700 million mark by 2023. Moreover, the increased demand for AI-based clinical analysis tool is further reinforced by the fact that hospitals are projected to spend some USD 2 billion a year on artificial intelligence for medical imaging by 2023.

Driving Factors

  • Cost Saving: According to a survey conducted by the West Health Institute and NORC at the University of Chicago, in 2017, 44% of Americans refused to go to a doctor due to cost concerns. Through implementing AI enabled medical service, the US health sector alone can save up to $150 billion in annually expenses by 2025, according a report published by Accenture.
  • More Accurate and Fast in Critical Cases: Researchers in Germany, France, and the United States already revealed that, when detecting cancer, AI technology is more accurate than a doctor’s eye. Their studies showed that AI-based image analysis software can detect cancer in 95% of images of cancerous moles and benign spots. In comparison, doctors were able to correctly detect cancer only in 87% of the exact same pictures.
  • The Government Support: Fast-Track government approval from FDA is favorably contributing to the growth of the industry as a whole. Since January 2018, the FDA has started a pilot program for providing pre-certification to health-care software developers. The FDA now hope to expand this program to include AI software, given its great potential. This fast-track regulatory approval could lead to major opportunity for the over 70 AI startups in the healthcare space that have raised capital in the past 5 years.
  • Low Number of Radiologists across the U.S

  • Machine Learning Readiness by Imaging Specialty
Clinical Specialty Key Uses for Machine Learning Machine Learning Readiness
Radiology Workflow efficiency and clinical decision
Dermatology Detect skin cancer and diagnose
malignancy of moles
Ophthalmology Detect and diagnose disease, such as
diabetic retinopathy
Pathology Detect and diagnose disease from tissue
Oncology Cancer diagnosis, prognosis and
treatment planning
Surgical Endoscopic image analysis Low

Source: Signify White Paper

World Market for AI based Medical Imaging by Clinical Application 2023

The 20 Most Funded Medical Imaging AI Companies

Company Funding Size (in Million) Application
HeartFlow 476.6 Cardiovascular Imaging
VoxelCloud 78.5 General Imaging
Infervision 75 General Imaging
Imagen Technologies 60 Interpretation of X-Ray data
Zebra Medical 50 General Imaging
Arterys 43.7 General Imaging
Deepwise 43 General Imaging
Volpara 30.8 Breast Imaging
Viz.ai 30.6 Neurology Imaging
12 SIGMA 30 Lung Imaging
Lpixel 30 Lung Imaging
HealthMyne 26.4 Oncology
Bay Labs 21.7 Cardiovascular Imaging
Lunit 20.5 Breast Imaging
Circle Cardiovascular 20 Cardiovascular Imaging
Ultromics 16.03 Cardiovascular Imaging
Vuno 15 Bone, Chest, Lung and Brain Imaging
Huiyi Hui Ying 15 Medical image screening and tumor diagnosis solutions
Brainomix 13.7 Neurology Imaging
MaxQ-AI 9 Neurology Imaging


Funding Analysis of Companies Developing Machine Learning Solutions for Medical Imaging Worldwide (2009-18)

  • There are over 120 start-ups and growth-phase companies developing machine learning solutions for medical imaging. 75 of these have entered the market since the start of 2015. Only 1 of the ten most funded companies is European. 5 are from the USA and 4 are from Asia.
  • Total investment in 2018 was USD 580M, more than double the 2017 amount (USD 270M). HeartFlow accounted for USD 240M of the 2018 total.
  • In 2018, later-stage (Series B onwards) funding, excluding HeartFlow, more than doubled to a total of USD 237M, compared to 2017.
  • Early-stage (Angel, Seed and Series A) funding peaked in 2017 at 29 deals and slowed to 15 deals in 2018.

Total AI Based Medical Imaging Funding in North America in Million

AI Startups in the US that Received FDA Approval

Startup Application
HeartFlow Inc. In 2014, it has received FDA approval for its FFRCT technology, a non-invasive imaging technique for coronary artery disease that provides doctors with insight on both the extent of the blockage as well as its potential impact on blood flow.
  • Arterys’ system enables a much more efficient visualization and quantification of blood flow inside the heart, alongside a comprehensive diagnosis of cardiovascular disease. It has recently partnered with GE Healthcare to combine its quantification and medical imaging technology with GE Healthcare’s magnetic resonance (MR) cardiac solutions.
  • In 2017, FDA cleared clinical product to use cloud computing and deep learning for assessing ventricular function from MRI scan of the heart and for automating cardiac MRI segmentation and measurement.
  • In 2018, it got approval for AI suite to aid cancer diagnosis in the Liver (CT & MRI) and Lungs (CT).
Quantitative Insights   It is the 1st FDA cleared clinical product for breast MRI to evaluate potential breast cancer.
MedyMatch Technology Its AI analyzes head CT images for brain bleeds and send automatic notification to treating physician.
Viz.ai It got approval for CT stroke diagnosis, the first example of applied artificial intelligence software that seeks to augment the diagnostic and treatment pathway of critically unwell stroke patients.
Imagen Its AI-based X-ray wrist fracture diagnosis got FDA approval. With video representing 60% of the global data, Imagen Ltd. ensures companies can preserve their content for the future and maximize its value. Imagen has stored 1+ million hours of video of different leading brands.
IDX IDX’s software can detect diabetic retinopathy (DR) without a doctors help.
ICO Metrix It got approval for its MRI brain interpretation capability.
Neural Analytics It has developed a FDA-approved device for paramedic stroke diagnosis.
OsteoDetect Its system can detect ML Wrist fracture for adults.
Bay Lab It got approval for detecting echo-cardiogram ejection fraction.
Zebra Medical Its platform can do automated calcium scoring. Zebra is empowering radiologists with its revolutionary AI offering at a flat, transparent USD 1 per scan. Zebra’s technology catches misdiagnosed diseases, early-stage cancers, and other life-threatening ailments through its AI-powered algorithms. Using data from millions of high-quality scans, Zebra created a deep learning engine that can automatically detect various medical issues, such as liver, lung, cardiovascular, and bone disease
iCAD Its AI can calculate breast density via mammography.
AIDOC            Its approved product offers CT brain diagnosis and send automated notification to treating physician. AIDOC is a crypto-enabled medical platform and its worklist widget empowers the radiologist to prioritize incoming cases with suspected findings.
MaxQ AI It can do diagnosis and prioritize CT brain bleeds in stoke and trauma.
Subtle AI It is the 1st AI product cleared for medical image enhancement to improve PET image quality.
Quantib On brain MRI, it measures brain atrophy and white matter changes related to aging, dementia and multiple sclerosis.

Challenges in AI-enabled Medical Imaging Industry

  • Gloomy Regulatory Environment with Less-interested Users: The regulatory process remains challenging and there are few approved and fully commercialized products on the market. Moreover, healthcare providers are reluctant to purchase AI tools from multiple software developers due to the vendor specific integration and implementation challenges and the administrative overhead issues.
  • Non-standardized Procedures: The results from AI-based image analysis tools need to be fully integrated into radiologists’ workflows. Due to the scattered nature of the software, many radiologists find that their models are incompatible with others. This causes more damage than good due to inefficient electronic medical record-keeping and a lack of comprehensiveness. Algorithm developers need to partner with imaging IT vendors to ensure their solutions are tightly integrated.
  • Cumbersome Regulatory Approval Process: In the U.S., regulatory approval allowing machines to do the work of trained radiologists is a major obstacle still unsolved. The amount of testing and effort necessary to secure clearance from the U.S. Food and Drug Administration (FDA) for allowing machines to provide primary interpretations of imaging studies without a radiologist would be overwhelming. The FDA categorizes the medical devices into three classes, according to their uses and risks, and regulates them accordingly. The higher the risk, the stricter the control. Class III is the category which includes the devices involving the greatest risk and AI-guided medical imaging falls into this category.

Leave a Reply

Your email address will not be published. Required fields are marked *