This study investigates the heterogeneity of colorectal cancer (CRC) consensus molecular subtypes (CMS) using spatial transcriptomics (ST) combined with single-cell RNA sequencing (scRNA-seq). The CMS classification is a widely used gene expression-based system for CRC, but its clinical application is limited by intratumoral heterogeneity. The study aims to better understand the spatial composition and molecular features of CMS in CRC. Using ST, the researchers mapped the spatial distribution of cell types and molecular features in CRC samples from seven patients. They identified cell communication events at the tumor-stroma interface, including tumor growth-inhibiting and -activating signals. The study highlights the potential of ST to resolve CRC molecular heterogeneity and support personalized therapy. The study found that CMS2 tumor cells were predominant in several samples, while CMS1 and CMS3 were present in others. The spatial distribution of cell types and molecular features varied between patients, indicating inter-patient heterogeneity. Intra-tumor heterogeneity was also observed, with different regions of the tumor showing distinct molecular features. The study identified key molecular pathways associated with CMS2, including WNT and VEGF pathways, and their potential regulation by ligand-receptor interactions. The study also found that DCN, a proteoglycan secreted by stromal cells, may inhibit tumor progression in CMS2 by modulating the activity of ETV4, MEIS1, and SPI1. Additionally, RNF43, a transmembrane protein, may influence several transcription factors in the TME, including JUN and TEAD4, which are involved in tumor progression and WNT signaling. The study also validated these findings in an independent ST dataset, confirming the presence of CMS2 in liver metastases and the conservation of CMS phenotypes in metastatic tumors. The results demonstrate the potential of ST to reveal inter- and intra-tumor heterogeneity, TME architecture, and spatial patterns of key molecular processes in CRC. The study highlights the importance of integrating ST and scRNA-seq for a comprehensive understanding of CRC and its CMS, providing insights into spatial cellular organization within tumors and their TME. The findings suggest that ST can inform patient-specific treatment strategies and contribute to the development of novel therapies. The study also acknowledges the limitations of the deconvolution-based approach, including the impact of scRNA-seq reference selection and the challenges of region-specific assignment in ST data. Overall, the study underscores the value of ST in CRC characterization beyond bulk- or scRNA-seq, enabling the spatial correlation of morphological and transcriptomic features.This study investigates the heterogeneity of colorectal cancer (CRC) consensus molecular subtypes (CMS) using spatial transcriptomics (ST) combined with single-cell RNA sequencing (scRNA-seq). The CMS classification is a widely used gene expression-based system for CRC, but its clinical application is limited by intratumoral heterogeneity. The study aims to better understand the spatial composition and molecular features of CMS in CRC. Using ST, the researchers mapped the spatial distribution of cell types and molecular features in CRC samples from seven patients. They identified cell communication events at the tumor-stroma interface, including tumor growth-inhibiting and -activating signals. The study highlights the potential of ST to resolve CRC molecular heterogeneity and support personalized therapy. The study found that CMS2 tumor cells were predominant in several samples, while CMS1 and CMS3 were present in others. The spatial distribution of cell types and molecular features varied between patients, indicating inter-patient heterogeneity. Intra-tumor heterogeneity was also observed, with different regions of the tumor showing distinct molecular features. The study identified key molecular pathways associated with CMS2, including WNT and VEGF pathways, and their potential regulation by ligand-receptor interactions. The study also found that DCN, a proteoglycan secreted by stromal cells, may inhibit tumor progression in CMS2 by modulating the activity of ETV4, MEIS1, and SPI1. Additionally, RNF43, a transmembrane protein, may influence several transcription factors in the TME, including JUN and TEAD4, which are involved in tumor progression and WNT signaling. The study also validated these findings in an independent ST dataset, confirming the presence of CMS2 in liver metastases and the conservation of CMS phenotypes in metastatic tumors. The results demonstrate the potential of ST to reveal inter- and intra-tumor heterogeneity, TME architecture, and spatial patterns of key molecular processes in CRC. The study highlights the importance of integrating ST and scRNA-seq for a comprehensive understanding of CRC and its CMS, providing insights into spatial cellular organization within tumors and their TME. The findings suggest that ST can inform patient-specific treatment strategies and contribute to the development of novel therapies. The study also acknowledges the limitations of the deconvolution-based approach, including the impact of scRNA-seq reference selection and the challenges of region-specific assignment in ST data. Overall, the study underscores the value of ST in CRC characterization beyond bulk- or scRNA-seq, enabling the spatial correlation of morphological and transcriptomic features.