Computer-pneumologist
Stanford Neural Network Diagnoses Pneumonia on X-ray Better Than Doctors
Maxim Agadzhanov, Geektimes
Researchers from Sanford University have developed a self-learning algorithm that is able to make a diagnosis based on medical images (chest X-ray). This software platform specializes only in pneumonia, but it performs its task better than professional radiologists.
Pneumonia can be different, and the neural network distinguishes 14 of its varieties by X-ray. The scientists published the results of their work in open access on the arXiv. "Analyzing medical images is a difficult task, and we know about it," says one of the developers of the software platform. "We decided to develop self-learning algorithms by conducting a "training" based on hundreds of thousands of medical images."
The work uses a data set provided by the NIH Clinical Center. This is a huge database, which includes more than 112 thousand frontal images of the human chest obtained by X-ray. In these pictures, 14 different pathologies can be distinguished. This information made it possible to "train" the software platform for the diagnosis of these pathologies.
After the training was completed, the scientists decided to check how well the system makes a diagnosis. After the machine made its diagnosis, the researchers asked the doctors to do the same. Doctors analyzed 420 images and made their own diagnoses for each of them. As it turned out, the computer system was able to diagnose pneumonia more accurately than a person.
Pneumonia is a dangerous and common disease. In the USA alone, about 1 million people are admitted to hospitals with pneumonia every year. Some pathologies are extremely difficult to detect on X–rays - too implicit signs in some varieties of this disease. At the same time, machine learning can help doctors cope with this problem.
The main danger for the patient here is that if the doctor makes a wrong diagnosis, it can lead to the appointment of incorrect treatment. As a result, the patient will be treated, but his/her condition will worsen. The situation is aggravated by the fact that doctors have to analyze hundreds of images of this kind per day, as a result of which attention is dulled. And the machine, subject to its high-quality training, could work around the clock, practically without making mistakes. Moreover, the computer system can highlight the smallest details in the photo that are important for establishing a correct diagnosis, but invisible to a person.
The developers "taught" their system to create something like a heat map of the human body when analyzing images. Only instead of demonstrating the temperature, the parts of the lungs where the machine "saw" signs of pneumonia are marked with different colors. After processing, the images are examined by a person, paying attention, first of all, to those areas that the machine marked as the "hottest".
Researchers believe that such systems will soon become generally accepted and widespread. "We are going to continue to create and improve medical algorithms to detect deviations from the norm. We also hope that we will soon be able to make publicly available depersonalized sets of medical data that can be used by other specialists working on similar or other problems," said Jeremy Irwin, a representative of the research group.
X–rays in medicine are the most important source of data on the patient's health status. Many doctors are simply inundated with such pictures. After several hours of working with them, the doctor's ability to concentrate decreases, attention decreases, therefore, the probability of error is high. Automation of the process could help solve the problem, which would reduce the number of medical errors.
The work carried out by scientists is not unique. Now many startups and large corporations like IBM and Google are engaged in similar developments. In addition to pneumonia, computer systems are already able to detect signs of tumors, problems of the cardiovascular system and other deviations from the norm on X-rays and other medical images.
Portal "Eternal youth" http://vechnayamolodost.ru