Automated Methods for Clustering, Clinical Indices Prediction, and Vertebrae and Lung Segmentation

Monday, June 10, 2024

10:00 AM-12:00 PM

BIOMED PhD Thesis Defense

Title:
Automated Methods for Clustering, Clinical Indices Prediction, and Vertebrae and Lung Segmentation in Early Onset Scoliosis (EOS) Patients

Speaker:
Girish Viraraghavan, PhD Candidate
School of Biomedical Engineering, Science and Health Systems
Drexel University

Advisor:
Sriram Balasubramanian, PhD
Associate Professor
School of Biomedical Engineering, Science and Health Systems
Drexel University
 
Details:
Early Onset Scoliosis (EOS) is a complex spinal deformity affecting children under 10 years of age, accounting for approximately 10% of all pediatric scoliosis cases. The etiology of EOS patients varies widely and can be classified into congenital, idiopathic, neuromuscular, and syndromic scoliosis. Progressive spine deformity in EOS patients leads to modified thoracic cage development, resulting in thoracic insufficiency syndrome (TIS) and potential complications such as pulmonary hyperplasia and premature death.
 
Current management strategies for EOS lack consensus among surgeons regarding intervention age, instrumentation type, and surgical location. Additionally, there is a need for structural and functional outcome data to assess the efficacy of surgical interventions and prognosis for EOS patients. Manual radiographic measurement of clinical indices is time-consuming and associated with significant inter-observer errors, particularly in the presence of vertebral, lung, and ribcage abnormalities in EOS patients.
 
This thesis aims to address these challenges by developing automated methods to cluster EOS deformities and guide surgical treatment selection using data-driven algorithms to optimize outcomes. Furthermore, this research seeks to create automated methods for identifying anatomical structures, such as vertebrae and lungs, from patient radiographs using machine learning techniques. These automated methods will aid in making accurate measurements of clinical indices, reducing the variability associated with manual measurements.
 
The specific objectives of this thesis are:
 
1. Introduce automated methods to identify meaningful subgroups based on pre-operative clinical indices of EOS patients.
2. Create data-driven machine learning models to predict post-operative clinical indices.
3. Develop automated methods for identifying anatomical structures, such as vertebrae and lungs, from patient radiographs to aid in accurate measurement of clinical indices.

Contact Information

Natalia Broz
njb33@drexel.edu