Skip to Content


No results


No results


No results

Search Results


No results
Search events using: keywords, sponsors, locations or event type
When / Where
All occurrences of this event have passed.
This listing is displayed for historical purposes.

Biomedical Engineering pres.

Ph.D. Defense: Mac Morris

Gaussian Mixture Model utilization for Classification of Deep Vein Thrombosis

Biomedical Engineering Biomedical Engineering
Biomedical Engineering
Deep vein thrombosis, defined as blood clots forming in veins beneath muscles in the body, is the third most common cardiovascular disease worldwide. The standard treatment of deep vein thrombosis involves using anticoagulants to stop clot progression but is only successful in 2/3 of patients because the composition of the clot changes over time. The standard diagnosis of deep vein thrombosis and determination of the age of the thrombus are based on the patient’s timeline of when these symptoms appeared. With use of magnetic resonance imaging (MRI), thrombus progression can be studied non-invasively from the acute to chronic stage. The aims of this work are to apply a Gaussian mixture model to multiparametric MRI of preclinical thrombi and determine whether spatial correlations exist with composition by histology.

Thrombosis was induced in 30 12-week-old mice divided into three groups and imaged at either acute, sub-chronic, or chronic time points. The multiparametric MRI volumes consisted of T1-, T2-, and T2-weighted images. The MRI volume for each mouse was normalized, and the thrombus was segmented. The T1, T2, and T2 intensities at each location were input into a Gaussian mixture model. Following image acquisition, tissue samples were acquired and stained for Martius scarlet blue trichrome to examine red blood cell, collagen, and fibrin content in each thrombus.

To analyze each Gaussian mixture model region, qualitative and quantitative analyses were performed by comparing the Gaussian mixture model results with the paired histology. 2D correlation values between the Gaussian mixture model tissue class and histological composition revealed that the first cluster class correlated with red blood cells (p < 0.05), and the second cluster class correlated with fibrin and collagen (p < 0.05). This study demonstrates that spatial correlations exist between the classification of multiparametric MRI data and corresponding histology. With the aforementioned methods and results from each study, we are able to move one step closer with assisting healthcare providers in optimizing treatments for patients with deep vein thrombosis.

Chair: Joan Greve
Biomedical Engineering Biomedical Engineering
Biomedical Engineering
Report Event As Inappropriate Contact Event Organizers
Back to Main Content