3586 Background: Neoadjuvant therapy's effect on vascular, tumor, and immune compartments, and their co-regulation, in colorectal liver metastases (CRLMs) remains poorly understood. We performed multiplex-based tissue profiling to quantify how treatment reshapes stromal compartment and co-occurrence networks, defined by inside-case correlations between features. Methods: We analyzed 76 CRLM cases (27 untreated, 49 neoadjuvant-treated 43 chemotherapy, 6 chemo + VEGF-/EGFR-targeted). Tissue microarrays (1-3 1-mm cores/tumor) underwent multiplex immunofluorescence to quantify immune/tumor subsets (CD3/CD4/CD8/PD-1/FOXP3/TIM-3/Ki67/CDX2), vessels (claudin-5), and stromal perivascular compartments (αSMA/PDGFRβ). Statistical analysis included Mann-Whitney U tests and Spearman correlations. Results: Neoadjuvant-treated CRLMs had 42% fewer proliferating tumor cells (CDX2⁺KI67⁺density, p=0.006) and 33% lower CDX2⁺ density (p=0.031) compared with untreated CRLMs. This was accompanied by a 32% reduction in proliferating cytotoxic T-cells (CD3⁺CD8⁺Ki67⁺ density, p=0.029). A reduction in vessel size was also observed in treated CRLMs (vessel size, -12%, p=0.077). Analyses suggested different effects by type of neoadjuvant treatment. Chemotherapy alone reduced proliferating tumor cells (CDX2⁺KI67⁺ density, -37%, p=0.016) and suppressed cytotoxic T-cell proliferation (CD3⁺CD8⁺Ki67⁺ density, -40%, p=0.017). The addition of VEGF/EGFR-targeted agents produced the most pronounced reduction in tumor burden (CDX2⁺ density, -73%, p=0.012; CDX2⁺KI67⁺ density, -76%, p=0.010) vs. untreated. CRLM co-occurrence networks were also affected by treatment. Treatment abolished positive correlations between tumor cells (CDX2⁺ and CDX2⁺KI67⁺ densities) and αSMA⁺PDGFRβ⁺ perivascular fraction/vessel size (r=+0.42 – +0.60, p<0.05), and negative correlations between tumor cells and αSMA⁺PDGFRβ⁻ perivascular fraction/vessel density (r=-0.47 – -0.40, p<0.05) in untreated (vs +0.05 – +0.27 and +0.09 – +0.21, p=n.s., in treated). Treatment also eliminated positive correlations between vessel size and exhausted (CD8⁺PD-1⁺) and regulatory (CD4⁺FOXP3⁺) T-cell densities (untreated r=+0.40 – +0.47, p<0.05). In parallel, new positive links emerged between αSMA⁻PDGFRβ⁺ perivascular fraction and CD4⁺ and CD8⁺ T cells (r=+0.41 – +0.42, p<0.05). Conclusions: This exploratory study indicates that neoadjuvant treatment alters the CRLM microenvironment, affecting tumor, vascular, and immune compartments. Additionally, analyses suggest that treatment induces a restructuring of their interdependence networks, defined by the loss of specific tumor-stromal-vascular links seen in untreated CRLMs, while establishing new ones. These findings motivate continued mechanistic studies and analyses on how these changes are related to benefits of treatment.
Österlund et al. (Wed,) studied this question.