The researchers from Moscow investigated the diagnostic capabilities of various AI services in radiology by choosing five different algorithms, where three nations were involved Russian, India and South Korea. According to their results, Russian solution outperformed international peers in the detection of pulmonary nodules. The research findings were published in the Chinese journal "Quantitative Imaging in Medicine and Surgery".
According to Yuri Vasiliev, the CEO for the Center of Diagnostics and Telemedicine under the Moscow City Health Department, this milestone was very important. "Our in-house AI technology has proven to be more accurate about identifying lung nodules than close solutions from India and South Korea. This is a major step forward for us in our pursuit of improving the quality of medical care.". The over 50 AI services currently in use by radiologists to interpret medical imagery and over 13 million studies analyzed by neural networks alone in Moscow, the growing market of AI solutions continue to introduce solutions that will ease the burden of healthcare professionals yet do not compromise much on quality standards, believes Vasiliev.
He further elaborated, "Our objective is to provide physicians with products that make their jobs easier as they enhance the quality of care. To this end, we created the maturity matrix, which is a wide-ranging tool to measure and benchmark various AI services. Since the end of last year, our AI service has always been a market leader in chest organ radiography.
The system is specifically designed to automatically analyze chest X-rays-it can detect up to 14 distinct pathology signs-and even computes cardiothoracic ratio while generating detailed radiology reports.
Kirill Arzamasov, Head of the Department of Health Informatics, Radiomics, and Radiogenomics at the Center, explained, "We prepared a unique dataset of 100 X-ray studies, 50 with confirmed pulmonary nodules and 50 without. Of these, 25 cases were reported by radiologists to be 'questionable', but became confirmed on subsequent CT scans. CT results validated all 50 studies with nodules."
The evaluation process of the services was conducted in three stages and was compared with the respective reference standards, validated by CT scans. The test shows that the Russian AI service outscored its foreign counterparts on all metrics for diagnostics. The dataset is open-source online and provides developers an opportunity to independently check the quality of various AI services.
This research falls under a greater initiative begun in 2020, which was initiated by the Moscow Social Development Complex and the Department of Information Technologies to integrate computer vision technologies into the healthcare system of Moscow. Developers are to be funded to catalyze innovation in AI in this project.
Diagnostics and Telemedicine Centre As early as 1996, the Diagnostics and Telemedicine Centre established itself as a leader of Russia in introducing AI technology in medicine, directed towards the development of diagnostic imaging, management of medical departments, research, and education of doctors.
