Multi‐omics analyses reveal spatial heterogeneity in primary and metastatic oesophageal squamous cell carcinoma

Abstract

AbstractBackgroundBiopsies obtained from primary oesophageal squamous cell carcinoma (ESCC) guide diagnosis and treatment. However, spatial intra‐tumoral heterogeneity (ITH) influences biopsy‐derived information and patient responsiveness to therapy. Here, we aimed to elucidate the spatial ITH of ESCC and matched lymph node metastasis (LNmet).MethodsPrimary tumour superficial (PTsup), deep (PTdeep) and LNmet subregions of patients with locally advanced resectable ESCC were evaluated using whole‐exome sequencing (WES), whole‐transcriptome sequencing and spatially resolved digital spatial profiling (DSP). To validate the findings, immunohistochemistry was conducted and a single‐cell transcriptomic dataset was analysed.ResultsWES revealed 15.72%, 5.02% and 32.00% unique mutations in PTsup, PTdeep and LNmet, respectively. Copy number alterations and phylogenetic trees showed spatial ITH among subregions both within and among patients. Driver mutations had a mixed intra‐tumoral clonal status among subregions. Transcriptome data showed distinct differentially expressed genes among subregions. LNmet exhibited elevated expression of immunomodulatory genes and enriched immune cells, particularly when compared with PTsup (all P < .05). DSP revealed orthogonal support of bulk transcriptome results, with differences in protein and immune cell abundance between subregions in a spatial context. The integrative analysis of multi‐omics data revealed complex heterogeneity in mRNA/protein levels and immune cell abundance within each subregion.ConclusionsThis study comprehensively characterised spatial ITH in ESCC, and the findings highlight the clinical significance of unbiased molecular classification based on multi‐omics data and their potential to improve the understanding and management of ESCC. The current practices for tissue sampling are insufficient for guiding precision medicine for ESCC, and routine profiling of PTdeep and/or LNmet should be systematically performed to obtain a more comprehensive understanding of ESCC and better inform treatment decisions.