Risk Model With Predictive Capability Through the Fusion of Experimental/clinical, Computational, and Imaging Data

Technology #ua16-220

Questions about this technology? Ask a Technology Manager

Download Printable PDF

Categories
Researchers
Samy Missoum
Associate Professor, Aerospace & Mechanical Engr
Managed By
Robert Sleeper
Licensing Manager (520) 626-4604

Title: Risk Model with Predictive Capability Through the Fusion of Experimental/Clinical, Computational, and Imaging Data

 

Invention: The technology is a novel use of several data sets to improve the predictive ability of risk models. The algorithm analyzes the experimental data set and related outcome as predicted by the computer model, propagates uncertainties to improve prediction, and adds optional imaging information.

 

Background: Risk prediction models are important tools in clinical decision-making. However, risk prediction models can be inaccurate when there are few events compared to predictors, resulting in underestimation in low risk patients and overestimation in high-risk patients.

 

Applications:

  • Cancer diagnostics
  • Hospitals/clinics to identify areas to improve outcomes
  • Evidenced based clinical software
  • Research organizations/laboratories
  • Medical imaging software
  • Clinical research trials

 

Advantages:

  • Convenient use via phone for optional image analysis
  • Can account for the correlation between factors by limiting the size of the search space with a one-class support vector machine
  • Can refine the risk model through metamodeling techniques and Monte-Carlo simulations of uncertainties
  • Can refine the model further through geometric feature recognition

 

Licensing Manager:

Bob Sleeper

RobertS@tla.arizona.edu

(520) 626-4604