It takes the artificial Intelligence 20 seconds to detect even slight changes in a human organ and identify whether the tumor is malignant or benign.
The project developed by the Peter the Great St. Petersburg Polytechnic University has been partly implemented at the Petrov National Research Center of Oncology.
The personal electronic doctor collects and remembers all the data about the health of the patient and warns about the possible reactions of the organism alarming if necessary both the patient and the doctor.
The system can make decisions in an automatic regime without human errors and emotions, says Vladimir Zborovsky, Professor of telematics at the Polytechnic University.
It can detect that what looks like a tumor on the computer tomography scan is not a malignant growth in the lung, but a blood vessel that got scabbed as a result of pneumonia.
The new technology is used in the Petrov National Research Center of Oncology in Pesochny to search for tumors in the lungs of real patients.
The author of the software code is Mikhail Ryabinin, leading developer of the neural networks laboratory of the Polytechnic University.
He says that the program not just compares pixels in the image, but detects all pathologic changes in the lungs, both malignant and benign, as well as tuberculosis.
The precision of the results is 97%, which is equivalent to an experienced radiologist, but the computer works much quicker. It takes the doctor half an hour to analyze more than 250 X-rays and propose a diagnosis, while the robot needs just a moment.
The algorithms of the new software have been patented. In the meanwhile, researchers at the Polytechnic University are working on the reasons of cancer and methods to cure it.
According to Andrey Kozlov, head of the molecular virology and oncology laboratory of the University, the researchers succeeded in identifying the genes that are present in all cancer tumors and are never to be found in normal organs.
This knowledge is expected to help develop an anti-cancer vaccine. Researchers want to ‘teach’ the organism to detect pathologic cells and destroy them.
One of the methods deployed is computer simulation.