Pushkar Mutha
BME PhD Proposal Presentation

Date: 2025-10-20
Time: 11:00 AM - 12:00 PM
Location / Meeting Link: https://emory.zoom.us/j/95939356624?pwd=boXtdYqsXRfdP7WjAsWwUsYxfVJFNo.1&from=addon

Committee Members:
Anant Madabhushi, PhD (Advisor); Ahmet Coskun, PhD; Sundaresh Ram, PhD; Lokesh Guglani, MD FAAP; Fatemeh Ardeshir-Larijani, MD, MSc


Title: Chest CT-Based Imaging Biomarkers for Diagnosis, Prognosis and Treatment Response Evaluation in Thoracic Diseases

Abstract:
This work aims to develop and validate novel biology-informed, quantitative CT biomarkers to guide prognosis and treatment selection across thoracic diseases. The core premise is that routine chest CT scans contain sub-visual signals – within tumors, peritumoral vessels, and the airway tree – that reflect underlying biology and can predict clinically meaningful outcomes. Our technical contributions include an optimized Quantitative Vessel Tortuosity (QVT) pipeline that quantifies peritumoral vessel twistedness, curvature, and branching angles in non-small cell lung cancer (NSCLC); a fully automated, lobe- and generation-aware airway phenotyping workflow with novel biology inspired quantitative descriptors tied to inflammation-driven remodeling - tapering indices and cylindricity to detect subtle bronchiectatic airway widening that is often imperceptible on qualitative reads, tortuosity/curvature and branch-angle measurements to capture architectural distortion, and global tree measures (e.g., fractal dimension and branching statistics) to summarize system-level change. Aim 1 delivers a scalable, robust feature extraction pipeline on standard-of-care CT for oncologic phenotyping that extracts intra/peri-tumoral texture and peri-tumoral QVT, a fully automated airway analysis workflow for quantifying airway structure across scales with a test-retest repeatability screen for evaluating feature stability. Aim 2 trains and locks down an interpretable Δ-radiomics + Δ-QVT signature of immune checkpoint inhibitor (ICI) response, ICI Response Imaging Score (IRIS), in NSCLC and evaluates it under blinded conditions in three clinical trials (CP1108, CheckMate 227, INSIGNA). Aims 3 and 4 applies the airway pipeline across non-oncologic settings – COVID-19, fibrotic interstitial lung disease (ILD), and pediatric cystic fibrosis (CF) to relate regionally resolved airway remodeling to acute severity, 63-week mortality, and inflammatory markers of early structural injury, respectively. By leveraging standard-of-care CT, this work aims to deliver practical, generalizable imaging biomarkers that capture host-tumor and airway remodeling dynamics and support timely, evidence-based clinical decisions across thoracic diseases.