Doctor of Philosophy in Bioinformatics

in the School of Biological Sciences

Iris Lu

Defends her thesis:
Determination of the accuracy and sensitivity of infrared sensors for anthropometric based lymphedema assessment in clinical environments

Tuesday, April 9, 2019

1:00pm
EBB 1005, Children's Healthcare of Atlanta Seminar Room

Thesis Co-Advisors:

Dr. J. Brandon Dixon
School of Mechanical Engineering
Georgia Institute of Technology

Dr. Rudolph L. Gleason
School of Mechanical Engineering
Georgia Institute of Technology

Committee Members:

Dr. Fredrik Vannberg
School of Biological Sciences
Georgia Institute of Technology

Dr. May Wang

Department of Biomedical Engineering
Georgia Institute of Technology

Dr. Eva Lee

School of Industrial and Systems Engineering
Georgia Institute of Technology

Dr. Sheryl Gabram
School of Medicine
Emory University.

Abstract:

Lymphedema is one of most feared side effects of cancer treatments in the United States. This disease leads to swelling of the affected limb and is associated with physical and psychological distress. Disease onset has no clear timeline. At risk patients may develop lymphedema immediately post-treatment or they may wait decades before developing lymphedema. Current medical care for at risk patients does not provide the continuous surveillance necessary for early lymphedema detection. Therefore, more often than not, patients diagnosed with lymphedema are subjected to a lifetime of maintenance, costing thousands of dollars per year in clinical visits and compression garments. In this dissertation, implementation of infrared sensor systems was explored and evaluated against the standard volume measurement tools used in specialized lymphedema clinics. The infrared sensor with the LymphaTech software resulted in good correlation and agreement with current measurement tools while being easier to use and more cost-effective than commercially available systems. Additionally, the efficacy of utilizing local arm geometries for the detection of arm lymphedema was determined. Anthropometric based features were extracted from a 3D point cloud using custom code and applied to train classification models for lymphedema. These features were shown to detect subtle changes in the arm of lymphedema patients with a sensitivity of 61% compared to the current standard volume difference measurement, which has a sensitivity of 33.3%. Clinics not equipped to detect lymphedema could integrate this infrared system and model as a screening tool to improve referral rates to lymphedema clinics.