Direct Estimation of Patient Attributes based on MRI Brain Atlases

Case ID:
C14008
Disclosure Date:
2/11/2016
Description:
This invention describes a novel way for diagnostic estimation of the patient attributes based on a unique atlas-based information-retrieval technology. The use of MRI atlases facilitate automated image segmentation and region-of-interest based feature analysis. The recent development of multi-atlas techniques, in particular, offers superior segmentation accuracy. In previous atlas-based approaches, segmentation is usually the single end-goal. We realized that the multi-atlas repository provides not only the pre-segmented labels for segmentation, but also a knowledge database associated with various subject attributes, such as clinical and diagnostic information. In this study, we proposed a context based image retrieval (CBIR) approach based on the multi-atlas framework. We developed an image similarity measurement to weigh the atlases according to anatomical feature matching between the target and co-registered atlas images, on a structure-by-structure basis; and then the atlas weightings were incorporated with known diagnostics of the atlases, to obtain diagnostic estimation of the target. We performed CBIR-based diagnosis in Alzheimer’s disease patients, which showed the estimated disease probability and cognitive scores closely matched the clinical measure in several core regions of the disease. The proposed CBIR framework provides a direct estimation of clinically relevant patient attributes based on the multi-atlas knowledge database, independent of the image segmentation, which potentially opens a new perspective of atlas-based image analysis.
Patent Information:
Title App Type Country Serial No. Patent No. File Date Issued Date Expire Date Patent Status
Direct Estimation of Patient Attributes based on MRI Brain Atlases ORD: Ordinary Utility United States 15/602,578 5/23/2017     Pending
Inventors:
Category(s):
For Information, Contact:
Mark Maloney
dmalon11@jhu.edu
410-614-0300
Save This Technology:
2017 © Johns Hopkins Technology Ventures. All Rights Reserved. Powered by Inteum