- Case Studies
- /
- Canadian Diagnostics Network
Canadian Diagnostics Network
AI Breast Cancer Detection Platform improves the accuracy of diagnostics using mammography and ultrasound data.
workflow automation in diagnostic triage
model uptime post-deployment
clinical contributors involved in dataset validation
A digital healthcare platform that incorporates Vision AI for mammography and ultrasound medical imaging analysis for breast cancer detection
Download TestimonialsHealthcare
Industry
Canada
Location
AI/ML Engineering, Cloud Infrastructure, Medical Image Annotation
Services
A radiologist’s diagnosis is subjective and often inaccurate, leading to up to 30% errors. Patients suffer repeated visits to hospitals, anxiety about diagnosis delays, and false radiologist conclusions.
See what we can do for youSolution
Our team designed and developed an AI-powered medical imaging analysis platform that, when paired with a radiologist, increased breast cancer detection accuracy by 20-30% in clinical trials compared to a radiologist alone.
To deliver a full AI-powered diagnostic platform combining mammography and ultrasound workflows, Blackthorn AI applied:
Project duration
01 Month
Finalized system architecture and aligned platform vision with the client; approved detailed implementation roadmap.
02 Month
Deployed the mammogram AI pipeline on Azure and connected public dataset inference models to the pipeline.
03 Month
Designed ultrasound annotation strategy; completed and validated annotation for the first batch of 100 images.
04 Month
Built classification model combining mammogram + ultrasound data; performed testing and initial validation.
05 Month
Deployed models via web interface and APIs for clinical use; integrated frontend and backend components.
06 Month
Ran full QA cycle; delivered final documentation and support for onboarding client’s internal teams.
Team Size
Delivering Impact
85%
automationof diagnostic image triage workflows, reducing manual review load for radiologists.
3x
faster diagnosis pipelinefrom 24–48 hours to under 8 hours with integrated AI inference and reporting tools.
~60%
cost reductionin data annotation by combining targeted manual labeling and semi-automated tools.
20+
clinical contributorstrained to validate the dataset, ensuring medical-grade quality and robustness.
