Cristian Barrera
BME PhD Proposal Presentation

Date: 2024-05-10
Time: 2:00PM-3:00PM
Location / Meeting Link: https://emory.zoom.us/j/98857036474

Committee Members:
Anant Madhabushi, Ph.D. FAIMBE, FIEEE, FNAI (Advisor) Ahmet F. Coskun, Ph.D. Sunil S. Badve, MD, FRCPath Kristin Higgins, MD Gari Clifford, DPhil, FIEEE


Title: Advanced Computational Pathology in Lung Cancer: From Immune Microenvironment Analysis to Predictive Biomarkers in Non-small and Small Cell Lung Cancer

Abstract:
Significant clinical challenges are found in non-small cell lung cancer (NSCLC) and small cell lung cancer (SCLC), related to tumor complexity, rapid progression, and resistance to existing treatment. Multiple approaches focus on interpreting the lung cancer tumor microenvironment (TME), to ease said challenges and inform experts of better treatment strategies. However, the approaches need a more straightforward implementation and tend to be obscure in interpreting the biological aspects of the TME. This work proposes developing and applying sophisticated computational pathomics pipelines, named PhenoTIL and PhenopyCell, designed to improve the precision of histopathological image interpretation and inform treatment strategies. Initially, I focused on conducting a deep phenotyping and biomarker discovery in NSCLC and SCLC, analyzing immune cell dynamics and the tumor microenvironment to identify prognostic markers for overall survival and predictive markers for chemotherapy and immunotherapy efficacy. PhenoTIL focuses on structuring and detailing single-cell tumor-infiltrating lymphocytes (TILs) within the TME, through their local morphology and spatial interactions with tumoral cells. The pipeline focused on exploring the clusterization behavior of TILs in the TME and their correlation with survival in NSCLC cases. PhenopyCell on the other hand, is an extension of the methodologies used in PhenoTIL with a wider region of interest, performing deep phenotyping across the entire tissue sample, including areas of tumor, non-tumor, and peri-tumoral regions. It determines the different immune and tumor density states, hotspots formed by immune and tumor clusters, and the immune phenotyping (immune inflamed, desert, and excluded). These pipelines are used to derive a nuanced understanding of treatment responses in NSCLC, aiming to support treatment personalization by identifying key biomarkers that predict patient outcomes, overall survival, and response therapy. For SCLC, PhenopyCell was exclusively employed, as it is capable of generating single-cell phenotyping of tumor and immune cells across the tissue sample, supporting the characterization of the cancer aggressiveness, and identifying biomarkers predictive of chemotherapy for patients categorized as Extensive and Limited. This research integrates advanced computational techniques with clinical pathology to create a robust pipeline in lung cancer treatment, moving towards AI-enhanced pathomic analysis for precise medicine. It also aims to contribute significantly to lung cancer, enhancing survival rates and treatment efficacy for NSCLC and SCLC patients.