Celsus. Mammography

AI-Solution for early detection of breast cancer

The model is trained on20 000verified studies

Pilots in 13 Russian regions

Integrated with Moscow Unified
Radiological Information Service

Winner of the 1st AI-solutions battle in Russia

Analysis

System detects and highlights
11 classes of neoplasms on the image

Interpretation

System identifies breast tissue density
by ACR and assigned the Bi-RADS
category to the study

Speed

4 projections pack (RCC, LCC, RMLO, LMLO) is processed <60 seconds

Accuracy

Bi-RADS interpretation accuracy >95%

Analysis

System detects and highlights
11 classes of neoplasms on the image

Interpretation

System identifies breast tissue density
by ACR and assigned the Bi-RADS
category to the study

Speed

4 projections pack (RCC, LCC, RMLO, LMLO) is processed <60 seconds

Accuracy

Bi-RADS interpretation accuracy >95%

Use cases

Diagnosis and screening on an outpatient basis

Сelsus is adapted to work in diagnostic and screening scenarios based analysis

Possibility of post hoc analysis

The system supports multithreading and allows you to analyze the pack of studies in a given period

Remote diagnosis

Possibility to analyze studies remotely using telecommunication channels

Screening on mobile mammography equipments

The system can be installed locally on the hardware or can be used remotely

How it works

Data transmission

Equipment sends studies to the RIS/PACS. RIS/PACS transmits studies to the Celsus AI Core via the API

Image analysis

Celsus receives images, the algorithm analyzes them and detects signs of abnormal changes

Work results

Upon completion, the results of the analysis are transmitted to the RIS/PACS and are available to the radiologist at his workplace

Integration with PACS/HIS/RIS

After integration, Celsus can be used in the usual doctor's workflow

Protocol support HL7/FHIR

International standards for health care data exchange

Pilot projects results

The results of pilot projects were summed up in the Tambov and Bryansk regions and in the Republic of Dagestan

Time minimization

The time spent by a specialist for analyzing research has been reduced by one third

Development

Up to 10% increase of cancer detection at early stages

Scale

> 22,000 studies have been reviewed

Minimization of the risk of errors

In some cases, AI detected signs of pathologies that were invisible for a radiologist
Request demo access
Request the trial access
* - Please, fill in all required fields
Name
E-mail
I agree with processing of my personal data and receiving newsletters from Medical Screening Systems LLC. Read more...