Many companies are aware that artificial intelligence is the future of medicine and are marketing surprising and futuristic applications.
From startups to large corporations, companies often have the ability to predict the future and get one step ahead of it. They know very well that the twenty-first century has great expectations from artificial intelligence in every sector, including medicine. Artificial intelligence has already brought many transformations to the health system and has more on the way. Doctors will be assisted even more in gathering, analyzing and organizing clinical data, performing early diagnosis, planning treatments and finding better solutions for patients. Robot doctors and medical record management software are already being used on a regular basis and, more recently, vanguard projects have been launched and cutting-edge products have been placed on the market.
The future is now
The diagnosis of cancer, heart failure, diabetes and adverse reactions to medicines are just some of the health sectors large IT companies have been investing more in as a result of progress made in artificial intelligence.
Google, in fact, has launched the project DeepMind Health. DeepMind is able to process millions of pieces of medical information in just a few minutes, speeding up many clinical, health procedures such as filing medical records or diagnostics. DeepMind researchers are also developing models that emulate the ability to imagine the consequences of an action before carrying it out in an effort to understand what intelligence and imagination are and turn it into an algorithm. Even Verily, the life sciences branch of Google, is working on a project called Baseline Study, which includes gathering genetic data with the objective of adopting some of Google’s algorithms to analyze what allows people to enjoy good health. For this project, researchers even use technology to monitor disease. For example, intelligent contact lenses can even measure blood sugar levels.
IBM is also active in the field of healthcare with the Watson system, which has become an established unit in hospital wards as it seems to be able to diagnose heart failure two years earlier than more traditional methods. The algorithm is based on data gathered during hospital visits. Jianying Hu, one of the researchers involved, explains, “We have discovered that the diagnosis of other diseases, prescribed medications and medical records from hospitalizations- in this order- can provide signs capable of predicting the disease. Then we add information obtained from the physician’s notes using natural language”.
Another large corporation, Intel, has entered the healthcare sector focusing instead on lung cancer. Intel, in collaboration with the giant Alibaba and other partners, has recently launched a contest to develop an algorithm capable of reading x-rays and other patient medical data to obtain early diagnosis and follow the growth of the tumor.
Artificial intelligence can also be used to predict the eventual side effects of a drug, as demonstrated in a study carried out by Stanford University and published by ACS Central Science. With just a little information about the chemical structure of the potential drug, the algorithm is able to formulate predictions about potential toxicity and the instability of the molecule, considerably speeding up the synthesis of drugs. Stanford University also demonstrated that deep learning algorithms (automatic learning and artificial intelligence research based on neural networks) are capable of diagnosing different types of skin cancer with the same accuracy of the best dermatologists. This can be done simply by taking a photo of the skin with a smartphone.
Startups are also getting involved, like Zephyr Health, a company that helps healthcare companies achieve more improved and faster data analysis through specific databases, learning algorithms and a presentation of the most usable data.
Startup Atomwise, instead, launched a virtual research project to find safe medication that already exists and can potentially be re-designed to treat Ebola. The study revealed two drugs that could significantly reduce infectivity.
Monitoring and predictive apps
The world of apps has responded to the call of artificial intelligence as well.
To monitor whether or not patients are following the treatment prescribed, for example, is the AiCure app (funded by the National Institutes of Health). This app uses artificial intelligence and a smartphone’s webcam to confirm medication has been ingested. This is very useful for patients with serious conditions who are reluctant to follow their physician’s advice.
Not feeling well and not having the time to immediately see a doctor is a very common reality. Many apps are trying to resolve this problem by using artificial intelligence to offer responses based on personal medical records and more general notions. Interaction is quite simple. The user inserts his or her symptoms and the software compares them with a database of diseases. After having considered the patient’s history and circumstances, the app provides a series of possible diagnoses. WebMD, for example, has announced the launch of its own technology for all devices infused with Amazon’s Alexa.
There is also Molly, the first virtual nurse in the world, developed by the startup Sensely. In addition to her smile and friendly expression, she helps people monitor their conditions and the treatment prescribed. This interface uses machine learning methods to care for patients with chronic diseases, providing monitoring and follow-up care.
The risk of these apps, however, is a high number of self-diagnosing patients that never receive confirmation from an expert in the medical field.
Radiology is yet another sector in the field of medicine that is using artificial intelligence. Enlitic is a pioneer in this sector with software that does not substitute radiologists, but makes their work faster and more precise. The first thing robot radiologists must do, for example, is make sure the x-ray marked by a radiologist as a left hand is not a right hand. Afterwards, the robot analyzes the x-ray, looking for abnormalities.
Based on these findings, robot radiologists mark each x-ray given to the specialized radiologist with a level of priority. If no abnormalities are found, it is given low priority and vice-versa. When the radiologist receives the image, he or she studies it and draws up a report.
This system speeds up the analytical and diagnostic processes of an x-ray immensely.
Humans cannot be replaced
Proof that artificial intelligence is now a commercial reality in the field of medicine can be found with the Food and Drug Administration (FDA), which granted a clearance to market the first deep learning technique applied to medicine, “Deep Ventricle”, an algorithm developed by Arterys that can calculate in just 30 seconds, using magnetic resonance, how many liters of blood our heart pumps per minute.
To better determine the future of artificial intelligence and its relationship with humankind, at the last Health Innovation Summit in Brussels, Microsoft (and several partners) launched the AI in Health Alliance, which proposes the adoption of these technologies on a large scale and the identification of common standards that unite the scientific community, from startups to universities.
While taking full advantage of artificial intelligence, we must not forget that these technologies cannot replace people. They offer immense support in analyzing great quantities of data which would be impossible for the human brain, but it is the human touch that gives sense to the data that is analyzed.
Humankind is responsible for using AI wisely, both from a scientific and moral point of view, without forgetting the ethical aspects of relationships with patients.
© Domedica s.r.l.