Developing Robust Machine Learning Models to Predict NSCLC Immune Checkpoint Therapy Outcomes
Date
2024
Authors
Contributor
Advisor
Department
Instructor
Depositor
Speaker
Researcher
Consultant
Interviewer
Narrator
Transcriber
Annotator
Journal Title
Journal ISSN
Volume Title
Publisher
Volume
Number/Issue
Starting Page
Ending Page
Alternative Title
Abstract
Non-small cell lung cancer (NSCLC) remains a leading cause of cancer-related deaths worldwide, despite recent advancements in treatment options. Immune checkpoint inhibitors (ICIs) have shown promise in improving patient outcomes, but their efficacy varies widely among individuals. Predicting patient responses to ICIs is crucial for optimizing treatment strategies and improving overall survival. However, current predictive biomarkers, such as PD-L1 expression and tumor mutational burden (TMB), have limitations in terms of accuracy and invasiveness.This dissertation aims to advance the field of personalized medicine in non-small cell lung cancer (NSCLC) treatment by developing innovative strategies for predicting patient responses to immune checkpoint inhibitors (ICIs). Through a series of interconnected studies, we leverage the potential of transcriptomics, machine learning, and non-invasive biomarkers to create more accurate and precise predictive models.
Our research begins with the development of a robust 11-gene signature derived from the largest dataset of pre-treatment tumor tissue samples compiled for this purpose. This signature demonstrates strong predictive performance, distinguishing responders from non-responders with an AUC of 0.85 in discovery datasets and 0.84 in validation datasets. We then apply advanced machine learning techniques and address class imbalance issues to further enhance the performance and reliability of our predictive models, showing their superiority over established biomarkers.
Recognizing the limitations of invasive tissue-based approaches, we develop a novel 11-gene signature derived from peripheral blood mononuclear cells (PBMCs) and buffy coat samples. This non-invasive biomarker proves effective in predicting ICI responses, with AUCs of 0.87 and 0.84 in PBMC and buffy coat cohorts, respectively. By offering a less invasive approach to monitoring treatment efficacy, this signature has the potential to improve NSCLC management, particularly for the people of Hawaii who face challenges in accessing specialized care.
The findings of this dissertation contribute to the growing body of tools for tailoring immunotherapy to individual patient needs and optimizing healthcare resource allocation. As we continue to validate and refine these predictive biomarkers and machine learning models, we move closer to a future where precision medicine becomes the standard of care for NSCLC treatment, offering improved outcomes for patients in Hawaii and beyond. This dissertation represents a significant step forward in the quest for personalized medicine, laying the foundation for a new era of targeted, effective, and patient-centric care in the face of a major health challenge.
Description
Keywords
Bioinformatics, Immune Checkpoint Therapy, Machine Learning, NSCLC
Citation
Extent
98 pages
Format
Geographic Location
Time Period
Related To
Related To (URI)
Table of Contents
Rights
All UHM dissertations and theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission from the copyright owner.
Rights Holder
Local Contexts
Email libraryada-l@lists.hawaii.edu if you need this content in ADA-compliant format.