AI-based platform for medical images processing and analysis
- Mathematical models for medical image analysis using artificial intelligence technologies
- Multimodal DICOM Viewer
- AI-based tools for pathology analysis results’ visualization
- Images markup tools
- Customizable interaction processes between radiologists and artificial intelligence tools
Concept of Use
Usage of the Platform example
Task: Automatic revision of diagnostic images from PACS server of healthcare provider or Central Archive of Medical Images
Goal: Reduced risk of skipping pathologies, especially oncological diseases in the early stages
For Healthcare authorities:
- Implementation of national healthcare programmes
- Mortality decrease, including a one-year mortality of patients with malignant neoplasms
- Healthcare equipment usage improvement
For Healthcare providers and radiologists:
- Medical errors probability reduction
- Workload of radiologists reduction
- Possibility of using the product for scientific work and additional training
Key Elements of the Platform
DICOM Viewer Botkin Oncore
Professional DICOM Viewer with AI functions
Own-developed Gateway for integration with Central Archive of Medical Images and PACS. It is deployed in medical institution network. With Botkin Gateway it is possible to configure the interaction between the PACS server of medical institution and the Cloud Platform for medical image analysis.
- The Gateway has following functions:
- Depersonification of medical images based on selected DICOM tags
- Filtering PACS server studies by tags content for processing transmission
- PACS server schedule for information pull out
- Integration with any number of PACS servers
Botkin Workflow Engine
Technology for studies flow received from customers via Botkin Gateway management.
- Research processing scenarios setting, depending on tasks given, content of DICOM-tags, types of studies or pathologies, gateway settings
- Determination of processing status for each step of research scenario
- Logging of all research processing stages
- Visual representation of processing statuses to increase the work efficiency with a large flow of studies
Key Elements of the Technology
Botkin Hybrid Intelligence
Medical imaging technology combining the artificial intelligence usage tools and radiologists’ expertise.
- Adaptable interaction processes for both artificial intelligence tools and radiologists (Botkin Workflows)
- Radiologists’ validation process analysis for artificial intelligence performance results assessment
- Option for cross-validation of the results by a group of radiologists
Own technology that significantly improves the efficiency of using artificial intelligence technologies for analyzing images with various pathologies.
- Consists of:
- Automated machine learning pipelines, which operate when new data is received. They also produce, validate, and integrate updated models.
- A meta algorithm for model architectures testing to determine the best model options for a particular dataset
- Subsystem for monitoring models’ performance
Hardware and software complex for medical image analysis.
- Key Features:
- Possibility of integration with the clinic
- Connection to PACS or directly to a data source (CT, PET / CT, MRI, X-ray, Fluorography)
- Flexible configuration (depending on image and diagnostics types)