Tumour Associated Neutrophils as a Potential Biomarker for Anti-PD-1 Immunotherapy in Non-Small Cell Lung Carcinoma: An Automated Approach

Background: Despite recent advancements in immunotherapy and clinical biomarkers for the treatment of lung adenocarcinoma (LUAD), objective response rates are still low at around 20%. Current biomarkers are limited in their ability to stratify patients for predicting response, highlighting the need for novel biomarkers to better guide therapeutic strategies. Tumour associated neutrophils (TANs) can influence non-response to immunotherapy and have been explored using manual scoring criteria. Improvements in digital pathology provide an opportunity to automate this process. Hypothesis: LUAD patients that are TANhigh harbour a consistent, therapeutically targetable molecular signature that sets them apart from TANlow/(-) tumours. This phenotype can be detected accurately and more efficiently with an AI-driven image-analysis pipeline. There were two main aims for this project: Aim 1: Dissect the molecular foundation of the TANhigh phenotype. Aim 2: Develop and validate an automated TAN scoring approach. Methods: Manual TAN scoring was performed on the Clinical Proteomic Tumour Analysis Consortium (CPTAC) LUAD cohort (n=85). Mass spectrometry proteomic and transcriptomic analyses were conducted to characterize the molecular landscape of TANhigh tumours. Quantitative pathology features, such as granulocyte density in the tumour, were used to develop a classifier for automated TAN status prediction. Results: Manual visual scoring identified a TANhigh phenotype in approximately 18% of LUAD cases. This phenotype was enriched in males, patients with smoking history, STK11 mutations, and a near absence of EGFR mutations. Proteomic analysis revealed upregulation of potentially targetable neutrophil proteins (MMP9, ARG1, ELANE, PADI4, S100A9). The automated TAN classifier achieved high accuracy (92.8%) and an AUC of 0.931 in internal CPTAC 5-fold cross-validation and identified a biologically relevant neutrophil-rich subgroup in an external cohort with consistent EGFR associations. Significance: This study enhances our understanding of TAN biology in LUAD, revealing a distinct molecular phenotype associated with TANhigh status and identifies several therapeutic targets. The development of an accurate, automated TAN classifier offers a scalable tool for future research and potential clinical application. TAN status is an accessible metric with the potential to improve patient stratification and guide novel therapeutic interventions in LUAD, improving response to immunotherapy and patient outcomes.

Department dc:contributor.department

Biomedical and Molecular Sciences

Year dc:date.issued

2025

Author dc:creator

• Smith, Jaxen Carter

Advisors dc:contributor.supervisor

• Simpson, Amber
• Cottrell, Tricia

• NSCLC
• Lung cancer
• Neutrophils
• Immunotherapy
• Artificial intelligence
• Pathology
• Digital pathology
• Immunology

Statement dc:rights

• Attribution-NonCommercial-NoDerivatives 4.0 International

Licence dc:rights.uri

• http://creativecommons.org/licenses/by-nc-nd/4.0/

Language dc:language.iso

eng

Handle dc:identifier.uri

https://hdl.handle.net/1974/34642

OAI identifier oai:identifier

oai:queensu.scholaris.ca:1974/34642