MMOOZZGG.com

Fields

AI in radiology: where medical AI has the most practice

2 min read10 sources

AI in radiology gets what it needs: X-ray, mammography, CT and MRI produce large amounts of similar data and have a clear metric. So Russia and the world have built up a lot of practice here. Below are the numbers from regions, what compulsory health insurance pays for, and what a clinic with an X-ray room should take.

On this page · 4 sections

Russia: the scale

According to the Moscow government, Moscow’s AI services process a quarter of all medical images in the country. Labor productivity in radiology grew 5–6 times, errors fell by half, and radiologists have 70 algorithms in 45 areas (September 2026). The Sverdlovsk Region analyzed more than 1 million images in 2024 and found almost 44,000 cases with signs of disease; by February 2025, more than 160 institutions were connected. The Krasnodar Territory ran 78,000 studies with AI in 2024, and the time to analyze an image fell from 15 to 3–5 minutes. Kuzbass has processed more than 100,000 images since the start of 2024, in half the time.

The world: screening and time to treatment

A randomized trial in Sweden with 106,000 women: AI in mammography screening finds 29% more cancers with 44% less radiologist workload. A Stockholm hospital replaced one of the two reading doctors with AI: according to the system’s developer, 15% more cancers were found and reading time fell by 36%. In Scotland, by early results, AI analysis of chest X-rays cut the time from image to the start of lung cancer treatment by 9 days on average. In India, a health technology assessment reported by the developer showed that AI tuberculosis screening by X-ray costs less per person examined than the usual path.

What compulsory health insurance pays for

Russia’s compulsory health insurance pays for AI studies in mammography, radiography, fluorography and chest CT. In September 2026, the fund announced that it adds electrocardiography and colonoscopy, 6 types of studies in total. For regions, this is a basis for purchases; for clinics that work within the insurance system, it is a reason to adopt AI.

What a clinic can take from this

A single image archive comes before the algorithm: in Kuzbass, the Sverdlovsk Region and the Krasnodar Territory, AI is connected to the regional archive or radiology information system. The second reading stays with a doctor in mass screening. Ask the vendor for the registration certificate and risk class. The metric: minutes per report and the share of findings confirmed by a doctor.

Questions and answers

Will AI replace the radiologist?

In Stockholm, AI replaced one of the two reading doctors in screening. In the Sverdlovsk project, the second reading stays with a doctor, and AI does the marking and the priorities.

Sources and fact-checking

Checked by the editors on 29 September 2026.

  1. Digital Platforms, case no. 2780, 2024: Ministry of Health of Kuzbass (platforms.su)
  2. Digital Platforms, case no. 2782: Ministry of Health of the Krasnodar Territory (platforms.su)
  3. Digital Platforms, case no. 2792: Ministry of Health of the Sverdlovsk Region (platforms.su)
  4. Digital Platforms, case no. 3812, 2 September 2026: Moscow Government, Center for Diagnostics and Telemedicine (platforms.su)
  5. Qure.ai, 24 March 2025: Indian Institute of Public Health Gandhinagar (qure.ai)
  6. NHS Grampian, 8 May 2024: NHS Grampian, Scotland, chest X-ray (nhsgrampian.org)
  7. Digital Platforms, case no. 3508, 2 September 2026: Moscow Healthcare Department (platforms.su)
  8. Digital Platforms, case no. 3510, 2 September 2026: Federal Compulsory Health Insurance Fund (platforms.su)
  9. Lund University, 30 January 2026: Lund University, Sweden (lunduniversity.lu.se)
  10. Lunit, 2 December 2024: Capio S:t Göran Hospital, Stockholm (lunit.io)

Found a mistake? Tell us

Publications

Cases on this topic: 6

  • Market case · Russia

    Ministry of Health of Kuzbass more than 100,000 images read by AI since the start of 2024, with processing time cut by half

    The region’s unified radiology system holds X-ray, mammography, CT and MRI images and passes them to AI services. They write a preliminary conclusion and point out abnormalities to the doctor. Since the start of 2024 the system has processed more than 100,000 images, and the time to a result has fallen by half.

    Radiology

    2024 · source: platforms.su

    Read the case
  • Market case · Russia

    Krasnodar Krai Health Ministry 78,000 studies with AI in 2024, image analysis in 3–5 minutes instead of 15

    Since 2023, state medical organizations in the region have used AI for mammography, X-ray and fluorography: 58 mammography machines in 55 organizations connect to the system through a regional radiology network. In 2024, 78,000 studies were done with AI, and by data of the federal compulsory medical insurance fund, the region is among the top three in the country.

    Radiology

    source: platforms.su

    Read the case
  • Market case · Russia

    Sverdlovsk Region Ministry of Health over 1 million chest images in 2024, nearly 44,000 cases with signs of pathology, more than 160 facilities

    Since September 2023 the region has run a pilot: X-ray, mammography, CT and MRI images from the region’s hospitals go into a central archive, where AI marks suspicious areas, writes description protocols and ranks the cases by priority. A rollout across the whole region began in January 2025. In 2024 AI chest X-ray analysis covered more than 1 million chest images and found nearly 44,000 cases with signs of pathology.

    Radiology

    source: platforms.su

    Read the case
  • Market case · Russia

    Moscow Government, Center for Diagnostics and Telemedicine a quarter of all medical images in Russia, productivity up 5–6 times, half as many errors

    According to the Moscow government, the city’s AI services process a quarter of all medical images in Russia. Data presented on 2 September 2026 show labor productivity up 5–6 times and the number of errors down by half.

    Radiology

    2 September 2026 · source: platforms.su

    Read the case
  • Market case · World

    Indian Institute of Public Health Gandhinagar AI tuberculosis screening from an X-ray costs less per person screened, by the model; more than 2,600 sites in 67 countries

    A health technology assessment, ordered by India’s Department of Health Research, used an economic model to compare several regulator-approved AI tools for tuberculosis screening from a chest X-ray with manual reading of images. The developer reports the result: AI tuberculosis screening costs less per person screened than traditional methods and counts as a cost-saving intervention.

    Radiology

    24 March 2025 · source: qure.ai

    Read the case
  • Market case · World

    NHS Grampian, Scotland, chest X-ray by early results, lung cancer treatment starts 9 days sooner on average

    A regional health service deployed AI lung cancer screening on chest X-rays: the software flags suspicious findings and helps send the patient to a CT scan faster. Early results show that the time between the X-ray and the start of lung cancer treatment fell by 9 days on average.

    Radiology

    8 May 2024 · source: nhsgrampian.org

    Read the case
All cases: 62
Contact us