Developing an Inflammation Intensity Score based on AI analysis of blood biomarkers- ON-422

Desired discipline(s): Engineering - chemical / biological, Engineering, Engineering - computer / electrical, Computer science, Mathematical Sciences
Company: KeyIntel Medical Inc.
Project Length: 4 to 6 months
Preferred start date: As soon as possible.
Language requirement: English
Location(s): Toronto, ON, Canada
No. of positions: 1
Preferred institutions: McGill University, University of Toronto, University of Waterloo

Search across Mitacs’ international networks - check this box if you’d also like to receive profiles of researchers based outside of Canada: 
No

About the company: 

KMI incorporated in 2018 as a medical diagnostic company with a focus on the development of “novel biomarker signatures in chronic inflammatory diseases”. These signatures may have significant value in future immunotherapy models and patient management.

Our first product (KeyIn-1) is a unique blood diagnostic test that can assess inflammation in locally affected tissues. We have comprehensive data on spondylitis and IBD patients with which we have recently filed a provisional patent application on “Methods for systematically assessing local inflammation and active repair” and submitted a manuscript.

These innovative biomarker signatures have significantly lower cost compare to current diagnostic techniques. Using machine learning and deep learning techniques, KMI will be able to offer an inflammation intensity scoring system with higher accuracy, lower cost, and in a fraction of time.

KMI secured the seed funding and has engaged government agencies such as the industrial research assistance program (IRAP). The initial KOL engagements showed very positive feedback. Current challenges: funding for the next steps in product development and clinical validation; the cost of IP protection and commercialization of products.

Please describe the project.: 

The project: Data sets include a cohort of 286 spondylitis patients followed annually for up to 12 years. Serum biospecimens were collected along with the concurrent clinical assessment parameters. KeyIn-1 levels were analyzed in 1204 serum specimens. Biomarker levels were correlated with MRI scoring and the patient response to tretaments. Precise signatures were identified in spondylitis and inflammatory bowel disease patients, based on KeyIn-1 levels and clinical data analysis. Inflammatory activities in each signature were quantified, and early algorithms were manually established based on these outcomes.

Main goals of the company which could use assistance of AI:

Short term:

  • Build the data structure based on regulatory compliant requirements
  • Build machine learning algorithms to evaluate the potential biases in data sets
  • Build machine learning algorithms to predict the clinical outcomes or patient response to treatment
  • Compare the outcomes of machine learning signatures to manually established algorithms
  • Define AI inflammation intensity scoring system for early diagnosis, identification of disease and subgroups in spondylitis and IBD patients
  • Validate the AI inflammation intensity scoring system
  • Prepare guidelines for physicians on the use of AI inflammation intensity scoring system (personalized disease management).

Long term:

  • Evaluate and inclusion of other serological and clinical parameters in AI inflammatory scoring system to enhance the efficacy and efficiency
  • The use of machine learning methods for early disease diagnosis and personalized disease management

Main tasks to be performed by the candidates:

  • Evaluate the current data sets
  • Data structure
  • Data set relations
  • Datapoint definitions (clinical and analytical)
  • List potential biases
  • Build proper back-end codes to
  • Regulatory compliant data structure
  • Build algorithms to identify potential biases
  • Build machine learning algorithms to evaluate the clinical outcomes based on serological biomarkers
  • Build a deep learning algorithm using clinical and analytical data sets to predict the clinical outcomes or patient response to treatment
  • Develop the AI inflammatory scoring system (clinical meaningful)
  •  
  • Design the structure to validate the AI inflammation intensity scoring model
  • Validate the AI inflammation intensity scoring model
  • Support the development of guidelines for clinicians
  • Prepare final reports

Required expertise/skills: 

  • Work experience as a Python Developer
  • Expertise in at least one popular Python framework (like Django, Flask or Pyramid)
  • Knowledge of object-relational mapping (ORM)
  • A deep understanding and multi-process architecture and the threading limitations of Python
  • Ability to integrate multiple data sources into a single system
  • Familiarity with testing tools
  • Familiarity with front-end technologies (like JavaScript and HTML5)
  • Team spirit and ability to collaborate on projects and work independently when required
  • Good problem-solving skills