Written on July 12, 2018 by Campus Life in Professional

I predict that one day when we’ll go to an hospital, we will enter a small room where a machine will take our blood sample, scan our internal temperature with thermic cameras, identify what’s wrong with us and direct us to a specific department or even prescribe directly a treatment. But even if we are slowly going there, the medical industry collaboration with artificial intelligence is progressing with baby steps. However, it is already amazing to see how the current applications some companies created are already bringing results.

I have argued many times with my father about how artificial intelligence can have an impact in the medical industry and whether it is science fiction to think a computer can actually replace doctors. I don’t blame anyone for being pessimist regarding this, however we need to differentiate internet forums low accuracy in terms of medical advices and artificial intelligence powered medical tools. They are not related as medical forums are just content written by anyone and the actual algorithm behind an artificial intelligence framework is based on medical verified resources.

This is not the future, this is now. Here are three concrete applications within the medical industry that help doctors to increase the efficiency of their diagnosis. Those start-ups and big companies are using the concept of deep learning technology to increase the medical research and medical treatments by submitting a large amount of data to identify patterns.

Google AI & Verily : Identify cardiovascular diseases

Scanning patient’s eyes to predict cardiovascular diseases with the help of artificial intelligence. To achieve this, the companies fed their deep learning framework with an input data of 284000 of retina pictures. From this, the algorithm can assimilate patterns with different input such as :
• Sex
• Age
• Smoker or not
• Blood pressure
• If they had Cardiovascular accident or not

Those information are used by doctors to predict future cardiovascular issues that there patient can face.

Constitute a giant DNA database to detect rare diseases: Sophia Genetics

The Swiss startup Sophia Genetics raised $ 30 million last September. Its niche: to propose a genetic database that health professionals can consult to “search for the specific molecular characteristics of a tumor or a congenital disease”, explained his boss at Le Temps. Its strength is based on its database which grows more and more, analyzing “the DNA of 8000 patients each month”, able to identify recurring patterns and anomalies.

Being able to determine if you will die in five years:

Researchers published a study on May 10, 2017 in Scientific Reports that shows that it is possible to predict the life of a patient by having chest scanners read by an algorithm.

Precision Radiology: Predicting longevity using feature engineering and deep learning methods in a radionics framework

The goal is to save doctors time by letting the machine analyze the scans itself and provide predictions. The longer-term goal: to succeed with a single scan to predict future diseases, so that doctors can intervene as soon as possible.

Identify signs that a patient health is getting worse in hospitals.

DeepMind, the Google lab dedicated to artificial intelligence research, has a new, large-scale project: helping doctors find patients who “sink” before it’s too late to save them. 700,000 patient records will be read and analyzed by the algorithm developed by Google to train it to detect signals that indicate that a person’s health is deteriorating.

The goal: to reduce the percentage of deaths in hospital space that could have been avoided if physicians had been able to observe the patient’s health decline in time (11%). DeepMind gives the example of an operation such as the replacement of a hip that can cause kidney failure, impossible to detect with the naked eye.

Other interesting reads about medical and AI applications:

DeepMind develops AI to diagnose eye diseases
How AI is transforming the future of healthcare


Sorry, the comment form is closed at this time.

We use both our own and third-party cookies to enhance our services and to offer you the content that most suits your preferences by analysing your browsing habits. Your continued use of the site means that you accept these cookies. You may change your settings and obtain more information here. Accept