e22525 Background: Non-small cell lung cancer (NSCLC) is highly lethal as most cases are late stage at diagnosis. Screening efforts with computed tomography aimed at reducing NSCLC mortality is limited to specific sub-populations with limited implementation globally. Thus, rapid scalable screening tools to detect NSCLC continue to be an unmet need. Detection and analysis of volatile organic compounds (VOCs) from exhaled breath with proton transfer reaction-time-of-flight mass spectrometry (PTR-MS) is real-time, non-invasive, and highly sensitive. With the hypothesis that PTR-MS analytics of exhaled VOC can discriminate between patients with lung cancer and healthy volunteers, we performed a pilot study to identify and define a unique exhaled-breath VOC signature representative of newly diagnosed, treatment-naïve NSCLC. Methods: A single-centre, prospective, pilot study was conducted. Treatment-naïve patients with newly diagnosed, histologically confirmed NSCLC of any stage were recruited and the control group comprised age, sex, and smoking history matched healthy volunteers with no known malignancy. Exhaled-breath VOC profiles were analysed with PTR-MS testing. A proprietary machine learning algorithm was employed to analyse breath samples to identify a VOC signature that could distinguish between patients with newly diagnosed, treatment naïve NSCLC and healthy controls. Results: 185 newly diagnosed treatment naïve lung cancer patients were recruited over 18 months with over 500 distinct VOC measured in each breath. Amongst these VOC, 20 were identified that differed significantly in concentration within breath of patients with lung cancer compared to breath of healthy controls. To make a 20 VOC multiplex predictive biomarker for lung cancer, a random forest model with 10-fold cross validation was adopted to build the classifier. The optimised VOC signature distinguished patients with newly diagnosed treatment naïve NSCLC from healthy controls with an area under the curve of 0.9 and 0.87 in the training and validation sets respectively with corresponding sensitivity and specificity of 93% and 86% in the training set, and 89% and 85% in the validation set. Conclusions: This study represents the largest known cohort of newly diagnosed, treatment naïve NSCLC patients with exhaled-breath VOC samples. A high-performance VOC signature unique to newly diagnosed NSCLC was identified. The high sensitivity and specificity demonstrated provides a strong signal for the feasibility of breath-based screening for NSCLC. Further large-scale external validation is ongoing to test the utility of exhaled-breath VOC as a real-time, point-of-care non-invasive screening tool for NSCLC.

Supplementary Table 5. GISTIC2 output of LUAD driver genes in Treatment Naive and Neoadjuvant Gefitinib tumours

e23422 Background: FGFR alterations are established actionable drivers in solid tumors, with particular clinical relevance in hepatobiliary malignancies. The advent of selective and next-generation FGFR inhibitors has expanded therapeutic options; however, durable responses remain inconsistent. Emerging evidence suggests that co-occurring genomic alterations significantly influence sensitivity and resistance to FGFR-directed therapies. We report a large Indian multicentric genomic analysis evaluating FGFR alterations, actionability, and resistance mechanisms with direct relevance to contemporary FGFR inhibitors. Methods: Comprehensive genome-wide next-generation sequencing data from 791 patients with solid tumors across multiple Indian centers were analyzed. Alterations in FGFR1–4 were characterized by type, oncogenicity, and clinical actionability. Known and putative resistance mechanisms—including FGFR gatekeeper mutations and co-alterations in bypass signaling pathways—were systematically evaluated. Results: A total of 848 FGFR alterations were identified in 791 patients. FGFR1 (35.5%), FGFR3 (34.6%), and FGFR2 (23.7%) were most frequently altered. Missense variants predominated (56.5%), followed by amplifications (26.9%), deletions (9.3%), and fusions (2.7%). Only 4.2% of FGFR alterations met Level 1 or 2 evidence for clinical actionability, emphasizing the need for careful variant interpretation when considering FGFR inhibitor therapy. FGFR amplifications are also targetable through panTKI agents which result in variable therapeutic response. Clinically relevant resistance mechanisms were identified in 41.8% of patients. FGFR gatekeeper mutations—including FGFR2 N549K, FGFR3 V555M, and FGFR4 V550M were observed in about 2% cases and may impact sensitivity even to newer-generation FGFR inhibitors. Additionally, frequent co-alterations were detected in key oncogenic pathways, notably PIK3CA (15.5%), KRAS (13.4%), PTEN (6.8%), and NRAS (2.5%), suggesting parallel MAPK and PI3K pathway activation as major contributors to intrinsic or early acquired resistance. Our data indicate that both drug-agnostic (pathway bypass) and drug-specific (FGFR gatekeeper) mechanisms coexist at conspicuous frequencies. Conclusions: In the era of next-generation FGFR inhibitors, FGFR alterations alone are insufficient to guide precision therapy. This Indian multicentric pan cancer real-world dataset demonstrates that a substantial proportion of FGFR-altered tumors harbor concurrent genomic events with potential to attenuate therapeutic benefit. Genome-wide profiling should be considered standard in FGFR-driven solid tumors to refine patient selection, anticipate resistance, and inform rational combination or sequencing strategies in precision oncology practice.

Supplementary Figure 7. Inferred clone phylogenies with tumour purity for treatment naive patients. Phylogenetic trees of detected clones from Treatment Naive patients, as inferred by CITUP. Only known LUAD (n=96) and Cancer Gene Census (n=359) driver genes are annotated for each clone on the tree. The mutation burden of a given clone is denoted by the vertical distance between itself and its parent node, whereas the horizontal distances between the nodes do not correspond to genetic distances.

Supplementary Table 3. List of Treatment Naive and Neoadjuvant Gefitinib DNA and RNA samples (All samples and Purity-matched cohorts)

Supplementary Table 2. Pre- and post-treatment primary tumour size and staging

Analysis of immune checkpoint and cytokine/chemokine responses in purity-matched neoadjuvant gefitinib EGFR-mutated NSCLC tumors. A and B, Boxplots of selected CIBERSORTx cell-type proportions present within purity-matched treatment-naïve vs. neoadjuvant gefitinib tumors. C, Gene expression comparison of immune checkpoint and cytokine/chemokine ligand–receptor pairs. D and E, Boxplot of gene expression levels comparing CD47–SIRPA and CD200–CD200R1 immune checkpoint pairs in treatment-naïve vs. neoadjuvant gefitinib tumors. F, Correlation of immune cell type and immune checkpoint gene expression with clinical response, including CT tumor shrinkage, SUVmax change, and pathologic response.

Supplementary Figure 19. Analysis of glycolytic and oxidative phosphorylation signatures in neoadjuvant gefitinib or treatment naive tumours. (A) Immunofluorescence (IF) co-staining of treatment naive or neoadjuvant gefitinib tumour sections with α-PFK1 (Glycolysis), α-ATP5E (Oxidative Phosphorylation), α-PanCK (Cancer Cells) and DAPI (cell nuclei). (B-C) Co-localization analysis of metabolic IF staining intensities in treatment naive vs. neoadjuvant gefitinib tumours.

Supplementary Figure 5. Associations between mutations detected in neoadjuvant gefitinib treated EGFR mutated tumours and treatment response. (A) CT change from baseline (%), (B) SUVmax change from baseline (%), and (C) pathological response (%) stratified by mutation status in the genes ERBB2, LRBA, MDM2, PIK3CA, PTPN23 or TP53. P-values are from Wilcoxon rank sum tests.

Supplementary Figure 8. Inferred clone phylogenies with pathological response (%) and tumour purity for neoadjuvant gefitinib patients. Phylogenetic trees of detected clones from treatment naive patients, as inferred by CITUP. Only known LUAD (n=96) and Cancer Gene Census (n=359) driver genes are annotated for each clone on the tree. The mutation burden of a given clone is denoted by the vertical distance between itself and its parent node, whereas the horizontal distances between the nodes do not correspond to genetic distances.

Supplementary Figure 2. Genomic landscape of EGFR TKI resistance mutations in neoadjuvant gefitinib treated EGFR mutated tumours. (A) Oncoplot of EGFR TKI resistance genomic alterations in treatment-naive and neoadjuvant gefitinib cohorts. Significantly different gene mutation frequencies between the cohorts denoted by asterisks. (B) Cancer cell fraction estimates from EstimateClonality for EGFR T790M reads in treatment naive, neoadjuvant gefitinib and TKI-Resistant (51) cohorts (****: p<0.0001). (C) Comparison of tumour mutation burdens of treatment naive and neoadjuvant gefitinib cohorts at varying sequencing depths. Dashed line denotes samples with both 100x and 500x sequencing.

Supplementary Figure 10. Correlation plots of the tumour genomic features TMB, tumour purity, and number of tumour subclones against neoadjuvant patient clinical parameters. (A) CT tumour shrinkage (%), (B) Tumour SUVmaxChange (%), (C) Pathological response (%). The Pearson correlation coefficient between each two compared features is as presented with its corresponding p-value within the plot. Each point represents the median molecular feature value across tumour sectors from the same patient.

Purity-matched neoadjuvant gefitinib-treated EGFR-mutated NSCLC tumors undergo a glycolytic to oxidative phosphorylation metabolic switch. A, Top 5 enriched mSigDB “hallmark” gene pathways found in purity-matched neoadjuvant gefitinib or treatment-naïve tumors in GSEA. B, Average glycolytic or oxidative phosphorylation gene expression Z-scores in treatment-naïve vs. neoadjuvant gefitinib tumors. C, Purity-matched metabolic pathway map of differentially expressed glycolysis and oxidative phosphorylation reactions in treatment-naïve vs. neoadjuvant gefitinib tumors. D, GSEA of top 5 enriched mSigDB “hallmark” gene pathways found via DSP ROIs (PanCK+, CD3+, and CD68+ regions) in representative treatment-naïve vs. neoadjuvant gefitinib tumors. E, PanCK+ metabolic pathway map of differentially expressed glycolysis and oxidative phosphorylation reactions in representative treatment-naïve vs. neoadjuvant gefitinib tumors. FC, fold change; TCA, tricarboxylic acid.

Supplementary Figure 6. Measures of clonal diversity in treatment naive and neoadjuvant gefitinib patients. (A-B) Two estimates for clone diversity in purity-matched cohorts, which consider phylogenetic structure (Clonal & Ancestor Diversity Index) and the mutational burden (Clonal & Mutational Diversity Index) for each patient. (C) Number of subclones detected for each patient tumour at matched depth and matched purity.

Supplementary Figure 14. Cytokine/chemokine and immune checkpoint responses in purity-matched neoadjuvant gefitinib NSCLC tumours. (A) Gene expression boxplots of the NK and T-cell ligands PVR and PVRL2 (CD112), as well as the NK and T-cell receptors TIGIT, CD226, CD96 and PVRIG (CD112R). (B) Gene expression boxplot of the OX40L-OX40 immune checkpoint ligand-receptor pair. (C) Boxplot of gene expression levels comparing CCL5-CCR5 and CXCL12-CXCR4 ligand receptor pairs in treatment naïve vs neoadjuvant gefitinib tumours.

Supplementary Figure 18. Correlation analysis metabolic gene expression signatures with neoadjuvant gefitinib patient clinical parameters. (A) Correlation of Glycolysis signature score vs. CT change from baseline (%), SUVmax change from baseline (%), and pathological response (%). (B) Correlation of Oxidative Phosphorylation score vs. CT change from baseline (%), SUVmax change from baseline (%), and pathological response (%).

Supplementary Figure 12. Heatmap of purity-matched immune cell types estimated by CIBERSORTx ranked by the order and direction of significance (adjusted p-value) between treatment naïve vs. neoadjuvant gefitinib cohorts.

Clinical assessments of response and outcomes to neoadjuvant gefitinib-treated EGFR-mutated NSCLC. A, Consort diagram and study workflow. Study includes the baseline PET/CT scan and treatment with gefitinib 250 mg daily for a minimum of 4 weeks, followed by posttreatment PET/CT scan and surgical resection. Multiregion WES and RNA-seq were then performed on tumor sectors for multi-omics analysis. B, Waterfall plot. Dotted line indicates PR or partial metabolic response (PMR; <30% change in baseline). C, Pathologic assessment of response. Dotted line indicates the threshold for a MPR (<10% viable tumor). D and E, Correlations of clinical measures of response. D, CT response compared with pathologic response. E, PET response compared with pathologic response.

Supplementary Table 7. List of differentially expressed genes between Treatment Naive and Neoadjuvant Gefitinib tumours (PanCK+, CD3 and CD68 regions)

Supplementary Figure 16. Correlation analysis of M2 Macrophage and CD8 T-cell tumour proportions (from CIBERSORTx) against neoadjuvant gefitinib patient clinical parameters. (A) Correlation of M2 Macrophage proportion with CT change from baseline (%), SUVmax change from baseline (%), and pathological response (%). (B) Correlation of tumour CD8 T-cell proportion with CT change from baseline (%), SUVmax change from baseline (%), and pathological response (%).