AI possibilities in healthcare: application cases

2019-02-25

AI possibilities in healthcare: application cases

Have you ever thought of the possibilities AI solutions open to medicine? Analysts at Frost & Sullivan are sure that artificial technology and machine learning will be establishing a strong presence in this sphere from year to year. Naveen Jain, Founder at startup Viome, who works with healthcare technologies development expresses the same point of view. In one of the interviews, the businessman told that a real AI tsunami is to crash on the medical sphere. Follow our article to get to know about projects and working solutions.

AI makes a diagnosis

Application of AI-based algorithms is in a full swing. They help doctors identify a disease and prescribe treatment.

Watson Health from IBM is utilized by radiological clinics in the US, India, and Thailand. A cognitive program based on this platform can identify lung cancer and heart conditions. Russian product TeleMD also diagnoses cancer as well as assesses risks of its development. DeepMind Health created by Google is applied in a British ophthalmologic clinic, recognizes some eye diseases, and gives treatment recommendations.

Last year, one of the UK clinics reportedly began testing AI that performs ultrasonography of pregnant women. Software called ScanNav examines a fetus in order to identify pathologies if any. It works along with doctors in a real-time mode.

Microbiologists at BIDMC developed a smart microscope that diagnoses lethal blood infections using AI. Its neural network learned 100 thousand images with hazardous bacteria. The microscope is currently capable of sorting them by visible fingerprints with accuracy of 93%.

AI keeps track

Developers use AI to launch services monitoring patient’s health state. Doctors and scientists investigate results and then carry out clinical trials.

Professors at Duke University created apps Autism & Beyond and mPower tracking symptoms of autism and Parkinson's disease allowing to improve methods of their diagnostics. Later, Apple took the apps as an underlying technology for API software Health Records so that users could share medical data via other applications.

Scientists from Massachusetts Institute of Technology and experts at Massachusetts General Hospital created an AI-powered system for human sleep monitoring. It keeps track of radio signals a person generates, evaluates pulse, breathing rate, and is capable of recognizing any irregularities. The system would help doctors examine patients’ sleep remotely and correct it if needed.

AI makes a forecast

In 2018, American medical journal Anesthesiology published the results of testing of AI helpful for surgical treatment methods. The article tells about the machine learning algorithm for forecasting of hypotonia after surgical intervention. AI analyzed data of more than one thousand patients with total time spent on the operation table of almost 10 thousand hours. It was trained to predict anomalies 15 minutes to their appearance with accuracy of 84%, the same index within 10 minutes, and 87% - within 5 minutes.

Qventus is a monitoring system for clinics created by the eponymous startup. It tracks patient’s actions from the check-in at registration table to the checkout, can predict feeling unwell by analyzing patient’s health state. What is more, Mercy Medical Center was using this system within 4 months and cut down on the number of unnecessary tests by 40% taking into account similar complaints of patients.

Jvion developed a solution based on machine learning to recognize patients with the risk of the return visit to a doctor within 30 days after the release. Besides, it gives recommendations on healthcare and disease prophylaxis.

AI carries out investigations

Pharmaceutical heavyweights such as Sanofi or Novartis cooperate with startups developing medical innovations in order to look for new medicine. Pursuing this idea, manufacturer of biotechnological drugs Roche bought Flatiron Health that applies machine learning for data processing.

To find more effective drug formulas, startup Atomwise has been using neural networks since 2012. Its deep learning system AtomNet investigates 10 m chemical compounds on a daily basis and predicts the best way of their interaction. Biopharmaceutical company Berg Health utilizes the same algorithm.

Discovered connections may be effective to combat a disease but that is not a guarantee that a human organism reacts to them positively. NorthShore Medical Center is also engaged in pharmacogenomics – the study of the effect drugs have on a certain group of people within the MedClueRx project. The system identifies the best medicine for each patient suffering from epilepsy, infectious diseases, depression, and gastrointestinal diseases.

Last year, scientific journal Nature Microbiology published an article dedicated to VarQuest algorithm. It can recognize 10 times more variations of antibiotics than during all of the previous similar tests.

AI as an assistant

AI technologies are applied for the development of a wide range of smart assistants: prom personal doctors to robotic surgeons.

Woebot helps deal with depressive thoughts and conditions. The chatbot was developed by AI experts along with the psychologists at Stanford University. Its underlying component is cognitive behavioral therapy that is capable of identifying behavioral patterns and self-destructive stereotypes. The similar app is being developed by startup Wyse.

DOC+ mobile clinic from the Russian eponymous startup gives an opportunity to have remote doctor advice, call a doctor to home, order drugs at nearby clinics. The app also creates an individual medical card available only to a user who can share it with a doctor.

da Vinci is a well-known surgical robot with embedded AI. It performed operative treatment in hundreds of clinics across the globe, including Russia (Moscow City Hospital № 31). Minimally invasive operations are successfully performed by Senhance from TransEnterix.

Meanwhile, Chinese scientists successfully cloned pigs using AI-fueled robots. Smart micromanipulators collected and transferred DNA of donors to surrogates. In April 2018, two female pigs gave birth to 13 healthy piglets.

Conclusion

The Medical Futurist Bertalan Meskó once said that artificial intelligence is a stethoscope of the 21st century. He meant that such a simple tool was not initially adopted by the medical community. It took several decades to have an established presence in the healthcare segment. The same story can be observed in case of AI: somebody finds ways to use it while someone else is afraid of the cutting-edge technology.

Nonetheless, AI and machine learning technologies as well as neural networks tremendously simplify the life of doctors and patients. Medical innovations help bring more accurate diagnoses, find necessary drugs more quickly, track patient’s health state. These are only some of the possibilities AI opened for the medical industry. 

Innovation news

05.03.19
Selection of health gadgets from CES 2019: part 2

Previous week, M-Health Congress told you about the first top six most impressive health devices...

05.03.19
Global healthcare 2019: development of digital technologies and personalized medicine

Analysts from Deloitte forecast that the global healthcare expenditures may rise to $10.1 trillion by...

01.03.19
Digital medicine as a way to reduce healthcare paperwork

According to Frost&Sullivan, the market of digital healthcare solutions will reach $6 billion by 2021...