Editor’s Note: Immune checkpoint inhibitors (ICIs) have brought new hope to the treatment of non-small cell lung cancer (NSCLC). However, their effectiveness is limited to only a subset of patients. Current biomarkers, such as tumor mutational burden (TMB) and PD-L1, show limited predictive power. According to Dr. Kenneth O’Byrne’s report(Icon Cancer Centre Greenslope) at the 2024 World Conference on Lung Cancer (WCLC), tests revealing high TMB or PD-L1 positivity only have a positive predictive value of around 40%.

In his oral presentation titled “Predicting the Future: Novel Pathological Assessments and Imaging Biomarkers” at the 2024 WCLC, Dr. O’Byrne introduced a key concept: using spatial metabolic mapping to decode the characteristics of the tumor microenvironment revealed through single-cell analysis. He highlighted findings crucial to accelerating the development of predictive biomarkers for ICI therapy. “To better predict treatment response, we need a deeper understanding of the tumor microenvironment,” he stated. “The complexity of the microenvironment has a significant impact on patients’ responses to ICIs.”

Traditionally, tumors with high concentrations of inflammatory cells (hot tumors) are thought to respond more effectively to ICIs. However, single-cell analysis shows that the mere presence of inflammatory cells isn’t enough to predict a patient’s response to immunotherapy. For instance, hot tumors with extensive infiltration of inflammatory cells may respond to ICIs, while hot tumors where inflammatory cells are confined to the periphery and excluded from the tumor microenvironment (TME) may not. Even with similar numbers of inflammatory cells, their biological activities could differ vastly.

Researchers employed a multimodal approach to retrospectively analyze tissue samples from 45 NSCLC patients treated with ICIs. The goal was to identify characteristics unique to responders and non-responders by using AI-driven unsupervised clustering to classify different cell types.

Unsupervised AI analysis identified 43 distinct cell subpopulations, differentiated primarily by their metabolic and activation states. Key differentially expressed proteins included oxidative phosphorylation markers such as CS, SDHA, and ATPAS, alongside metabolic enzymes like HK1, GLUT1, and LDHA.

The study found a direct link between metabolic states, effector functions, and tissue localization. Metabolically active lymphocytes exhibited higher levels of PD-1, MHC-I, MHC-II, and CD44 expression. Regulatory T cells (Tregs) emerged as predictors of resistance to ICI therapy, consistent with findings from other studies.

Analysis of cellular neighborhoods revealed ten distinct zones, including a macrophage-mixed tumor phenotype associated with treatment response. Dr. O’Byrne noted that this association wasn’t surprising, given the established role of macrophages in differentiating into anti-tumor states once they penetrate the tumor tissue.

Unsupervised clustering analysis of tumor cells revealed three main metabolic states: OXPHOS-positive, OXPHOS-negative, and PPP-positive. Tumors with a PPP-positive state, characterized by upregulation of ASCT2, pNRF2, and G6PD, demonstrated higher proliferation rates and elevated CD44 expression—a marker of cancer stem cells. Tumors with over 40% PPP-positive cells exhibited resistance to PD-1 inhibitors, leading to decreased overall survival (OS) rates. “When PPP levels are high, treatment response is poor; when PPP levels are low, response is good,” Dr. O’Byrne summarized.

Dr. O’Byrne emphasized that while these findings are promising, they must be interpreted with caution. The development of this technology presents intriguing signals for the future of ICI biomarker research.

About Prof Kenneth O’Byrne

Prof Kenneth O’Byrne is a medical oncologist and clinical scientist consulting at Icon Cancer Centre Greenslopes. He is the founder of the British Thoracic Oncology Group (BTOG), and founder member of the European thoracic oncology platform (ETOP), a member of the ESMO chest tumours faculty, is on the education and fellowship committees of the International Association for the Study of Lung Cancer (IASLC) and on the board of the Thoracic Alliance for Cancer Trials (TACT).

He is co-chair of the advanced lung cancer sub-committee of the Australian Lung cancer Trials Group (ALTG), and a member of the Medical Oncology Group of Australasia (MOGA), Australian Genomics Health Alliance (AGHA) and Queensland Genomics Health Alliance (QGHA).

Prof O’Byrne has been involved in more than 200 Clinical Trials in his career including designing investigator-initiated, grant-funded Phase II and III studies. He has established collaborative studies and networks for biomarker detection and validation and has been involved in biomarker driven clinical trials including the LUX-Lung (Boehringer-Ingelheim), checkpoint immunotherapy (BMS, MSD) and the ETOP programs.

He is clinical lead for the ‘Cancer and Ageing Research Program’ at the Queensland University of Technology Translational Research Institute laboratories on the Princess Alexandra Hospital campus, his research including biomarker identification in solid and liquid biopsies, drug resistance linked to cancer ‘stemness’ and genomic instability. His collaborations are global with ongoing projects on drug resistance, DNA repair and genomic instability, and whole genome and exome sequencing linked to groups in Europe and the US.