Ophthalmoscope for cardiologist

The Google neural network determines gender, age, body mass index and blood pressure by the fundus

Anatoly Alizar, Geektimes

Scientists from Google and its subsidiary Verily, which specializes in the development of medical technologies, have developed a new way to determine risk factors for cardiovascular diseases, such as coronary heart disease and stroke. The trained neural network calculates these factors fairly accurately from the image of the fundus.

The photo in the upper left corner shows a color sample of the fundus scan from the UK Biobank database. The other images show the same image, but in black and white. Each of them has a green heat map corresponding to each of the learned signs: age, gender, smoking (yes/no), average blood sugar HbA1c, body mass index BMI, arterial systolic pressure SBP, arterial diastolic pressure DBP. The real data from the database for each parameter and predicted by the neural network are indicated.

Knowing these factors, you can fairly accurately calculate the probability of developing cardiovascular diseases, which are the main cause of death worldwide (about 31% of deaths are caused by this cause).

With the help of the new system, doctors can save a lot of time, because instead of several tests, preliminary diagnosis is now performed in a few minutes. Moreover, theoretically, the algorithm allows such diagnostics to be carried out remotely. You only need an ophthalmoscope and a specialist who can take a picture.

Mirror and electronic ophthalmoscopes

Of course, while the accuracy of the neural network is not so high as to replace a full-fledged diagnosis, but it shows promising results. Here, AI does not replace the doctor, but expands his capabilities.

To train the neural network, Google and Verily scientists used medical records with photos of the fundus of approximately 300,000 patients. The lion's share of the data set was obtained from the EyePACS database (236,234 patients, 1,682,938 images). The rest of the information was taken from the UK Biobank database. Although there is less data here, but for each patient there was information on body mass index, blood pressure and the fact of smoking, which is not in the EyePACS database.

The idea of detecting human diseases by the retina is not new. Back in the Soviet Union, such studies were conducted and software for analyzing retinal images was created. But then there were no machine learning systems, so the possibilities of programmers were limited.

If the Google neural network receives photos of the fundus of two patients for processing, one of whom has suffered from cardiovascular disease over the past five years, and the other has not, then it correctly determines whether the photo belongs to the patient in 70% of cases. This is slightly worse than the accuracy of the SCORE algorithm currently used in medicine. It has 72% accuracy.

Experts say that Google's approach to using a neural network in this particular diagnostic task inspires confidence, because it has long been known that the retina of the eye predicts the risk of developing cardiovascular diseases well. So Artificial Intelligence can significantly speed up, and potentially improve the accuracy of such diagnostics. Of course, before the actual application in clinics, the program must undergo thorough testing so that doctors begin to trust it.

This discovery was another proof that neural networks can be widely used in modern medicine, especially in diagnostics. We are only groping for the most obvious applications of AI in this area: diagnosis of arrhythmia by cardiogram, diagnosis of pneumonia by X-rays, diagnosis of skin cancer, etc.

The amazing possibilities of using AI to diagnose diseases is one of the reasons why Google launched the Baseline project to collect detailed medical records of 10,000 people over four years.

The scientific article was published on February 19, 2018 in the journal Nature Biomedical Engineering (Poplin et al., Prediction of cardiovascular risk factors from retinal fundus photographs via deep learning).

Portal "Eternal youth" http://vechnayamolodost.ru