A team of Russian-based researchers has created an AI software system for diagnosing lung cancer that can read CT images in as little as 20 seconds.
The system, named Doctor Alzimov, after the Russian-born science fiction writer, can be installed on any computer and provides images with clearly marked findings for easy interpretation. It was developed by researchers at Peter the Great St. Petersburg Polytechnic University (SPbPU) in St. Petersburg, Russia along with radiologists from the St. Petersburg Oncological Center.
In late 2018, Doctor Alzimov analyzed its first CT images taken from 60 patients at the Oncological Center and found nodules on the lungs as small as 2 millimeters.
“Initially, we set up an algorithm to search for nodules starting from 6 millimeters, because radiologists themselves start the treatment of tumors of this size,” said Lev Utkin, head of the SPbPU Research Laboratory of Neural Network Technologies and Artificial Intelligence, in a news release. “But the system is so smart that it was able to find nodules of even smaller size.”
Using the “chord system,” Doctor Alzimov utilizes segmented CTs to classify lung nodules. Chords connect points across the nodule that reflect its shape and structure to produce a histogram that Doctor Alzimov then analyzes.
According to the group, their system was trained on 1,000 CT scans. Utkin and colleagues also have been generating their own dataset consisting of 250 patients and plan to have four-times that amount of data by the middle of 2019, according to the statement.
The group plans to eventually transfer CT images to a supercomputer, which could reduce the diagnostic testing time to two seconds.