Health

5 ways AI and machine learning can be useful in medicine

Artificial intelligence is a technology that has come to revolutionize a large number of areas of our world, among which medicine stands out. In recent years there have been many advances that artificial intelligence has brought to the medical industry and this ranges from the improvement in disease diagnosis times, the collection of data for studies and has modernized research in laboratories.

Thanks to artificial intelligence and machine learning, machines could not only collect and analyze medical information, but also interpret it, which would greatly shorten decision times in medical contexts. And this not only in the field of physical health, but also mental health, in which big data and deep learning could detect patterns of unusual behavior that a human could not decipher so easily.

There are so many uses that can be given to artificial intelligence and machine learning in medicine, but this time we will focus on five of them that can be useful in this field.

1. Speed ​​up the diagnosis of rare diseases

Thousands of rare diseases are known in the world, the diagnosis of which is usually extremely complicated due to the ambivalent symptoms that are presented and the lack of knowledge. While many symptoms can manifest during childhood, at least 50% of rare diseases manifest in adulthood, and perhaps well advanced by then.

Artificial intelligence could precisely help identify these symptoms early and save patients a lot of time. A 2019 study conducted in Germany yielded interesting results on the diagnosis of rare diseases using artificial intelligence. To do this, they created a neural network that automatically combined images of patients with medical data and genetic information.

Doctor next to a machine reviewing a patient's diagnosis on a piece of paper.
Image: National Cancer Institute via Unsplash

In this way, patients are not only evaluated by a doctor but by an artificial intelligence that filters genetic factors in the patient’s faces and prioritizes the genes that are the root of the disease, reducing the time of data analysis and increasing the rate diagnosis.

Another study published in Nature highlighted the use of DeepGestalt software that works with deep learning to identify facial phenotypes of genetic disorders. The results were very positive, reaching 90% accuracy in identifying syndromes in more than 500 images.

2. Early detection of breast cancer

Breast cancer is much more likely to be cured when detected early, and although artificial intelligence can be crucial in detecting it early, it is also convenient to rule out false positives. At least 50% of women who have an annual mammogram will have a false positive at some point during a 10-year period. Google knows this and that is why in its Health division it created an artificial intelligence in 2020 that in its tests produced very positive results: Google’s AI produced 5.7% fewer false positive diagnoses, and 9.4% fewer false negatives compared to human experts.

For its part, IBM also has an artificial intelligence developed in conjunction with the University of Zurich that seeks to make diagnoses with greater precision to combat breast cancer.

The implementation of artificial intelligence in this area of ​​medicine is essential, as it allows doctors to reach areas where the tools used until now tend to focus on suspicious areas that are more likely to give false positives and consequently impact the patient’s mood. Another risk factor to consider, in addition to the appearance of tumors, is the density of the breast tissue, whose classification is usually not very accurate. MIT created a machine learning tool that allows for more accurate classification, complementing the doctor’s opinion.

3. Determine the prognosis of patients in a coma

In 2018, a group of Chinese scientists developed an artificial intelligence-powered machine to make it easier for doctors to determine the chances of recovery of a patient in a coma. They collected the data obtained from the magnetic resonance imaging of thousands of patients in a coma to process them with machine learning, which created an algorithm that managed to predict the recovery of patients whose primary prognosis was not promising, since it was estimated that they would not recover.

By then, the first tests with the machine were resoundingly successful, having reached 90% accuracy in forecasts. At the time it was used to evaluate approximately 300 patients and they estimated to help more than 50 thousand patients in a coma.

Female doctor with stethoscope around her neck looking at a patient.
Image: Zach Vessels via Unsplash

4. Patient monitoring

The usual thing for many people is to go to a doctor when there are already marked symptoms or when they do not feel well, so patient monitoring is not constant, since it only exists when there is an identified problem. According to a study published in Nature in January 2021, in the future, artificial intelligence could facilitate monitoring without the need to go to a doctor.

Machine learning and wearables could become great allies of doctors by being able to constantly monitor a patient’s health and send a notification or alert to their doctor. An example may be a case of arrhythmia or high blood pressure, both indicators that many wearables such as the Apple Watch already include.

5. Improve disease research in the laboratory

In that same study they mention the potential that artificial intelligence could have in the investigation of diseases in a laboratory. They specifically mention that it could help “rebuild the underlying mechanisms of a disease”, which would have two very positive effects for medicine —among many more—.

One of them is that by being able to simulate the responses of patients —whose clinical and molecular information would be analyzed by machine learning— in pharmacological trials, that is, they could experiment with novel treatments and drugs without the need to harm living beings, be they humans. or animals.

On the other hand, epidemics and pandemics could also be avoided by studying contagious viruses, since the virus would not have to be manipulated, but could be studied with the help of artificial intelligence or machine learning.

Scientist studying a sample under a microscope.
Image: Lucas Vasques via Unsplash

Conclusions

Although this is a slow process, the use of artificial intelligence and machine learning in medicine looks very promising, not to replace a doctor, but to complement their experience to go beyond what a human it is capable.

His consolidated path in medicine shows a promising future in which everyone’s health is a priority so that we all have the opportunity to improve our well-being and prolong the duration of our lives.

Source link

Related Articles

Leave a Reply

Your email address will not be published.

Back to top button