While the term “artificial intelligence” might conjure images of futuristic robots, the field is actually the new frontier of health care.
The covid pandemic exposed critical challenges within the health care system — such as health care worker shortages. The World Health Organization estimates a shortage of 15m health care workers by 2030.
Doctors, researchers, and entrepreneurs alike have been focused on how AI can improve the health care system. AI-enabled health startups raised nearly $10B in funding in 2021 and more than $3B in the first half of 2022.
From new body imaging devices to oncology platforms, some trailblazing AI companies are already changing how millions of people receive care.
How AI helps health care
The possible applications of AI in health care are extensive, reaching from drug development and imaging to insurance and patient care.
Dr. Marinka Zitnik, an assistant professor of biomedical informatics at Harvard Medical School, says two uses — precision medicine and therapeutic science — stand out above the crowd.
Precision medicine takes into consideration a person’s genetic makeup, environment, and lifestyle to deliver personalized care and treatment. Therapeutic science repurposes already available drugs for other uses, and comes into play in scenarios such as public health emergencies where there is no time to develop a new drug from scratch.
Furthermore, machine learning can carefully process more data than human health care workers, giving AI technologies a leg up at early detection of diseases.
For example, a new system created by Johns Hopkins University detects sepsis hours earlier than traditional methods and reduces the chance of patient death by 20%. Sepsis, usually caused by infection, is easy to miss in a crowded emergency room as it presents with common, nonspecific symptoms such as fever and confusion.
Johns Hopkins’s AI system processes a patient’s medical history, current symptoms, and most recent lab results to warn providers when a patient is at risk for sepsis.
Outside of the emergency room, AI can help with drug development and finding cures for rare diseases. UK-based startup Healx uses AI to match drugs that have passed clinical trials with rare diseases they could possibly treat.
With the cost of developing a prescription drug rising 145% since 2003 to a staggering $2.6B, using machine learning to repurpose existing drug trials could save time, money, and, most importantly, lives.
In addition, AI technologies can help with early diagnosis, imaging, emergency call triage, and much more.
Top AI health care companies
While there are 100+ of AI health care startups, these 11 companies are paving the way with groundbreaking research and technology.
In 2017, Arterys became the first to receive Food and Drug Administration’s clearance for leveraging cloud computing and deep learning in a clinical setting.
The company started with a specific problem: It wanted to make the diagnosis of heart defects in newborns and children simpler for clinicians.
4D Flow MRI, an imaging technology that shows blood flow in the heart, solves the issues with diagnosis, but image-archiving servers in hospitals couldn’t read files as large as 4D Flow’s output.
To solve for that, Arterys used cloud computing to bring 4D Flow’s images to hospital radiologists via a web browser, allowing them to make lifesaving treatment decisions.
When the company saw providers were still manually calculating the size of heart ventricles, it combined deep learning with cloud computing GPUs to automatically measure ventricles.
The company is now working on providing radiologists with automatic, accurate measurements across specialties, applying the technology to cancers, liver, and the brain, among other applications.
Ultrasounds are used to diagnose a wide range of conditions such as blood clots, gallstones, and cancerous tumors. With high-range ultrasound machines costing a staggering $100k or more, underserved communities around the world lack medical imaging.
Butterfly Network is meeting the need by creating the first hand-held whole-body imager. The pocket-sized probes, which attach to smartphones to show imaging, can be taken to the most remote locations. It pairs with AI technology that can interpret results as accurately as a human clinician.
Caption Health is improving patient care by making ultrasound technologies more accessible, with a focus on early disease detection.
Ultrasound technology can be difficult to master and requires specific training. Poor-quality images can result in misdiagnosis and missed treatment opportunities. Caption Health uses AI to guide clinicians through the imaging process and to automatically assess image quality.
This allows any health care provider, regardless of their training, to conduct an ultrasound and interpret results.The technology is currently focused on cardiac ultrasounds with work underway to develop AI guidance for lung ultrasounds.
This digital health care company’s AI-based platform works with coronary computed tomography angiography (CCTA) imaging to help providers diagnose and treat heart disease earlier.
Cleerly measures the plaque buildup on the heart’s arteries rather than indirect markers such as risk factors and symptoms. The technology’s ability to analyze and characterize different types of plaque can help clinicians more accurately determine a patient’s risk of heart attack.
CloudMedX is a computing platform that streamlines clinical processes and improves patient outcomes using predictive analytics.
It uses AI to collect data and build holistic pictures of individuals and communities. The single, unified data platform has functions ranging from operational to clinical and financial, meaning health care providers can find everything they need in one place.
The company’s predictive health care models can predict disease progression and determine the likelihoods that patients may have complications by processing medical data and providing risk assessment scores.
Corti’s voice-based digital assistant offers clinicians support with quality assurance.
The AI was trained using thousands of hours of real patient calls and consultations. It listens alongside health care workers during patient visits through proprietary speech recognition and language processing tech.
The AI tools can be used to better sort and direct incoming emergency calls at hospitals and can help with patient flow and triage on nursing call lines. It can also be used for call support and post-call quality assurance.
DeepMind, acquired by Google Health in 2014, aims to create intelligent AI systems across industries, but has made notable strides in health care specifically.
The company partnered with a hospital to treat patients with diabetic retinopathy (an eye condition that can cause vision loss) and found its AI system could recommend patient referrals as accurately as doctors for 50+ types of eye diseases.
The system was also able to predict whether a patient will develop a more severe stage of eye degeneration months before it happens, which could one day mean sight-loss prevention.
Enlitic uses deep learning to streamline workflows for radiologists and improve diagnoses.
The manual process of reading imaging slows down clinicians trying to provide care and diagnosis to patients. Enlitic’s software reads clinical content and then, based on the knowledge it has gathered through deep learning, analyzes and interprets the data to save radiologists time.
GE Healthcare and Enlitic have partnered to embed Enlitic’s AI into GE radiologists’ workflows to improve data standardization, efficiency, and capacity.
Healthtech company Komodo compiles clinical encounters of 330m patients who have gone through the health care system. This map of data allows Komodo to offer its customers — which span from disease advocacy groups to pharmaceutical companies — clinical insights gathered through analytics.
The company’s technology can identify potential patients for clinical trials, track the adoption of new treatments, and identify disparities in care and outcomes. It is even tackling understanding the short- and long-term health impacts of covid.
10. Oncora Medical
According to Oncora, oncologists log in to at least six different software systems when treating their patients.
Oncora aims to provide a single platform where clinicians can find data about cancer patients, clinical outcomes, and treatments so that providers can use collected data to improve future care.
By using only one platform, clinicians can save time by only making a single data input. Oncora’s analytics software pulls in data from separate systems, allowing providers to access everything in one place.
Owkin uses AI to find better treatments for unmet medical needs, beginning with cancer.
The company takes a “federated learning” approach to data, meaning it partners with organizations to access their data and applies machine-learning algorithms to learn about diseases.
The algorithms comb through past clinical data to help predict disease progression upon diagnosis.
Public AI health care companies
Some prominent public companies with AI health care focuses include:
- Alphabet - which includes Google Health and its AI arm, DeepMind
- Butterfly Network
- Medtronic - launched the first AI-powered endoscopy device
- Merative - a health data company owned by IBM
- Sanofi - the pharmaceutical giant that owns AI company Owkin
- Stryker - a leading manufacturer of orthopedic robots
AI medical imaging companies
Some AI companies working specifically on medical imagine include:
- Aidoc - AI company offering a range of imaging services
- Behold.ai - uses AI to support intelligent CT scans and X-rays
- Caption Health
- Gleamer - makes medical-grade AI for radiology
- Sirona - offers AI tools to assist radiologists in interpreting imaging
- Qure.ai - maker of AI for chest X-rays, head CTs, and ultrasounds, among others
Cost of AI in health care
There is a wide range in the cost of AI: A custom AI solution can range from $20k to $1m.
How much AI software costs depends on factors such as the level of intelligence needed, the data set size it needs to process, and what output is needed from the algorithms.
It will also depend on how AI companies choose to package and sell their technologies to future customers.
“Integration of AI into business models to make profit is not yet fully understood,” says Zitnik. “Most of the existing tools that are currently implemented are used in research studies or collaborations… It’s not clear yet how to monetize AI systems.”
Zitnik explains AI companies will have to make decisions when it comes to monetization: Will they charge based on the number of queries a customer runs? Or pay a single fee upfront for an AI tool that they continually feed new information to?
While AI can be costly, spending on the technology shows no signs of slowing down: The AI health care market is projected to expand to $36.1B in 2025, a growth rate of 50.2%.
What did you think of this article?