An Immune-metabolic signature (IMMETCOLS) identifies three clusters in mCRC with different immune-phenotype distribution and potential clinical implications.

Abstract

e15534 Background: The overall response rate following immune check-point blockade (ICB) is < 5% in microsatellite stable (MSS) advanced colorectal cancer (mCRC), and between 30-35% in MSI mCRC. We here evaluate the relationship between our immune-metabolic signature (IMMETCOLS) distribution (cluster (IMC) 1; mesenchymal glycolytic, IMC 2; epithelial-non glycolytic and IMC 3; epithelial mitochondrial oxidative) with immune phenotypes (inflamed (IF), immune-excluded (IE) and immune-desert (ID)) in MSS and MSI mCRC patients (pts). Methods: We have analyzed IMMETCOLS in a retrospective cohort of 128 mCRC pts enriched with MSI-H and BRAF genotype using the nCounter platform (Nanostring Technologies, US). Immune phenotypes were evaluated by immunohistochemistry using 7 immune markers (CD68, CD163, CD8, CD3, FOXP3, PD-L1 and PD1) in a subset of 47 mCRC pts (27 MSS and 20 MSI), in TMAs containing 4 cores from the center of the tumor (CT) and invasive front (IF). The number of stained immune cells was counted using a computerized image analysis system, and PD-L1 was evaluated using CPS score (< 1 vs > 1). Patients having a CD3 cell count below 572 cells/mm2 and a CD8 density below 175 cells/mm2 in IF or CT were considered as ID. (Halama Nils, 2011). CD3 density was also semi-quantitatively estimated in peritumoral stroma and intra-tumoral (range 1-4), defining as IE those cases with a difference > 1 between the two regions in both IF and CT (Brooks JM, 2019). Results: Cluster distribution was similar between MSI (IMC1-40%, IMC2-20% and IMC-40%) and MSS (IMC1-34%, IMC2-24%, IMC3-43%) pts. Patients with high LDH levels (> 1.5ULN) were enriched with IMC3 phenotype (p < 0.0001). MSI have increased CD3 (p = 0.002), CD68 (p = 0.002) and CD163 (p = 0.001) in CT than MSS patients. IMC1 pts had increased CD163/CD68 ratio in CT (p = 0.004). 75% of MSI pts were IE and 25% IF. 55% of MSS pts were IE, 38% ID and 7% IF. IMMETCOLS/ immune-phenotype distribution was: IMC1 (16% ID/ 84% IE), IMC2 (12% ID/ 63% IE/ 25% IF) and IMC3 (35%ID/ 47% IE/ 18% IF). PD1 immuno-expression > 1% (IF and CT) and PD-L1 positivity was observed in 38%, 30% and 28% of pts respectively. There were no differences in PD-L1 (CPS) and PD1 expression between MSS and MSI tumors. Inflamed tumors show higher PD-L1 expression compared with IE and ID immune-phenotypes (p = 0.014). PD1 > 1% expression at IF was significatively lower in IMC1 vs IMC2 and IMC3 tumors (p = 0.032). Conclusions: IMC1 subtype is associated with low PD1 expression, high CD163/CD68 ratio and IE phenotype even in MSI-H mCRC patients. Therefore, IMMETCOLS testing would be of value to identify MSI mCRC pts that do not benefit to ICB therapy.