Modern technologies are gradually being introduced into medicine. Of course, they cannot replace the doctor, but they can help doctors in their work. Diagnostics is perhaps the most obvious direction for artificial intelligence and neural networks.
Draper is developing a system that analyzes lung images and finds signs of cancer. The problem with this disease is that most of the symptoms appear only in the late stages, so it is difficult to treat.
The neural network is able to process the results of computed tomography and concentrate on suspicious areas. This will not only save doctors time, but also make it possible to make a preliminary diagnosis remotely – for example, patients who live in small towns and do not have access to the best specialists.
Another group of researchers told Scientific Report about their project, which measures glucose using an electrocardiogram. This eliminates the need for a blood test and allows patients with diabetes to control their condition.
The neural network scans ECG waves, comparing data with individual indicators for each person and general statistics. ECG sensors are now even integrated into smart watches. This approach works with an accuracy of 82% – this is enough to be used as an auxiliary diagnostic method.
Back in 2010, IBM introduced its AI-based system. Since then, 600 thousand medical reports and 2 million pages of text have been added to the database. The neural network is able to analyze the patient’s diagnostic data, comparing his case with similar ones, make a preliminary diagnosis and offer a treatment plan. An AI takes only 10 minutes per person, while a doctor would take hours.
In 2019, the same company conducted a study in India, selecting 1,000 patients with lung and breast cancer. Each received an individual treatment plan, which in most cases coincided with the prescriptions of doctors. In 13.6% of cases, after reviewing the artificial intelligence suggestions, doctors adjusted their plans, recognizing that the program offered a more personalized approach.