Quantitative spatial profiling of PD-1/PD-L1 and TIGIT/CD155 interaction indicates poor survival outcome and resistance to adjuvant chemotherapy in pancreatic adenosquamous carcinoma.

BACKGROUND

Pancreatic adenosquamous carcinoma (PASC) is a rare, highly aggressive pancreatic cancer. Its therapeutic response varies, presumably due to tumor microenvironment (TME) complexity.

METHODS

We investigated whether using multiparameter algorithms to measure key immunosuppressive mechanisms in the TME could identify strong predictors of survival and response to adjuvant chemotherapy (ACT). Pre-treatment tumor tissues from 120 patients with PASC were stained using multiplexed quantitative immunofluorescence (training cohort: 92 patients; validation cohort: 28 patients). Novel Automated Quantitative Analysis (AQUA) algorithms were used for spatial profiling of biomarker-positive cells. Biomarker co-localization was determined in pathologist-selected tumor regions. The impact of certain biomarker signatures, including TIGIT/CD155 co-expression and PD-1/PD-L1 interaction scores, on survival outcomes and response to ACT was determined.

RESULTS

In the training cohort, PD-1/PD-L1 colocalization scores were significantly associated with advanced AJCC stage and perineural invasion. A TME with high PD-1/PD-L1 and TIGIT/CD155 co-localization scores represented a unique subtype, characterized by decreased cytotoxic T cell and increased suppressive immune checkpoint molecule levels. High PD-1/PD-L1 and TIGIT/CD155 co-localization scores were independent indicators of worse survival outcomes. Additionally, patients with low TIGIT/CD155 colocalization scores were more likely to benefit from ACT. These findings were reproducible in the external validation cohort.

CONCLUSIONS

Quantitative spatial profiling of the PD-1/PD-L1 and TIGIT/CD155 suppression pathways using novel AQUA algorithms could help predict patient outcomes and ACT responses reliably, guiding development of more effective personalized management for PASC.