AI in Health
Artificial intelligence (AI) is part of our daily lives. In about ten years, it has become omnipresent (GPS, social networks, connected objects, etc). It has many applications in health care. We most often benefit from it without knowing it.
Artificial intelligence is the implementation of several techniques to allow machines to simulate a form of human intelligence. It has made itself indispensable in medicine. Indeed, since the decoding of the genome and the explosion of health data (BIG DATA), the human brain is no longer able to store, manage and mobilize all this information. Everything from surgery, treatment, diagnosis and diagnosis is being radically transformed by advances in artificial intelligence.
We will present to you some revolutions made by AI in medicine.
Skin cancer: identifying a melanoma without risk of error
Until now, a dermatologist had to rely on his eyes to differentiate a simple mole from a melanoma. From now on, he can rely on a medical imaging system equipped with artificial intelligence (AI). The SkinUp application developed by the French start-up Anapix Medical allows the specialist to take a picture of the lesion. It is then analyzed by an AI engine capable of determining whether or not it is benign.
Medical imaging: radiologist computers
Thousands of radiological images (scanner, MRI…) have been digitized and stored. Thanks to highly sophisticated image recognition algorithms, this “deep learning” or “machine learning” process allows the software to analyze a new image and make the diagnosis as a radiologist would do. Thus, it augments the doctor’s eye and is more efficient than him thanks to the millions of bits of comparative data it has memorized. In the coming years, the analysis of medical images will be entrusted to machines for increasingly accurate and reliable assessments.
Surgery: robotic surgeons
The cyber-revolution also concerns operating theaters. Once the robot and surgical tools have been programmed by the surgeon, they replace his hand, with unbeatable gesture finesse and safer results. Research is even underway to develop “intelligent” probes capable of recognizing the tissues being operated on very accurately. This will lead to more targeted interventions, making it possible to perfectly remove the diseased area, particularly in cancer surgery, while preserving the surrounding healthy tissues as much as possible. In the long run, machines can be entrusted with the responsibility of increasingly complex procedures in areas of the body where the surgeon did not always dare to venture.
Cancer: predicting the effectiveness of immunotherapy
Immunotherapy is the great advance of the last 10 years in oncology. But there is still an issue : some patients respond very well to this therapy and others do not respond at all. Thanks to new AI software, researchers at the Gustave Roussy Institute and INSERM have been able to predict the response to treatment, taking into account the immunological environment of a tumor.
Assistance in choosing the best treatment
To give the right treatment to a patient, the doctor has a lot of information. He must know precisely the pathology, the patient’s personal and family history, his risk factors, his results of radiological or biological examinations, his behavioral data, the various existing treatments, their possible side effects, etc. This is where therapeutic decision support tools such as Watson, developed by IBM, come in. This machine is able to analyze all the patient’s data and put at his disposal all the scientific knowledge available on his pathology at his disposal. He can then engage in a “discussion” with the doctor in order to find the best therapy. It can even go beyond that and study a particular patient by taking into account all people suffering from the same pathology.