AI in Medical Diagnostics

Medical Diagnostics Development Workflow

Finding the right patient group is key in precision medicine because it allows us to evaluate the most effective approach to disease treatment. AI can also replace routine decision-making processes such as diagnosing cancer, image analysis, and management of chronic diseases. Diagnostics development takes multiple steps including prototyping, manufacturing, validation, and regulatory. AI can help in selecting prototype biomarkers, optimizing assays, evaluating manufacturability, validating analytics, designing clinical trials, and much more!

PoC

Prototype format
- Blood-test based assays
- Cancer diagnostics
- COVID diagnosis
- Prostate cancer diagnosis from imaging
- Brain cancer diagnosis from imaging
Target product profile
IP filing

Prototype Assay

Optimize the assay
Product specifications
Material selection

Manufacturing

Manufacturability
Equipment design
Process development and validation

Analytical Validation

Analytical performance validation
Linearity of the assay
Method comparison

Clinical Validation

Clinical trial design and execution
Data analysis
On-site testing

Regulatory

In Vitro Diagnostics (IVD) submission
Conformity assessment
Application in another country

Why AI in Medical Diagnostics?

Companion diagnostics to select right patients

Every patient is different. Some drugs work for certain patients but not others. AI can create diagnostic models to find the right patient groups for every drug.

Multi-angle diagnostics

AI can build models from not only structural data (blood test results etc) but also video, sound, and image data. These allow building diagnosis from multi-angle examinations.

Performing Routine Tasks Better Than Humans

From radiology to IPS cell sorting to spectrography – the ability to process volumes of data without human involvement can both accelerate processes and reduce costs and errors.

Our latest research work with Stanford University

Deep learning detection of prostate cancer recurrence with 18F-FACBC (fluciclovine, Axumin®) positron emission tomography