IGFBP1(hi)WNT3A(lo) Subtype in Esophageal Cancer Predicts Response and Prolonged Survival with PD-(L)1 Inhibitor

SIMPLE SUMMARY: The lack of precision biomarkers hinders the development of individualized immunotherapy for esophageal cancer (ESCA) patients. The activation of WNT signaling has proved to be associated with the primary resistance to immunotherapy. Therefore, our study aims to develop efficient biomarkers based on WNT signaling to guide ESCA immunotherapy. In the TCGA cohort with 196 cases and our BJCH cohort with 95 cases, we successfully constructed an IGFBP1(hi)WNT3A(lo) signature and validated its correlation with the poor prognosis of ESCA patients. Moreover, in 2 GEO cohorts with a total of 68 cases and our BJIM cohort with 21 cases, we verified that IGFBP1(hi)WNT3A(lo) ESCA patients had a good response and prognosis with immunotherapy. Our findings indicate that the IGFBP1(hi)WNT3A(lo) signature was a potential immunotherapeutic biomarker for ESCA. ABSTRACT: PD-(L)1 inhibitor could improve the survival of locally advanced esophageal cancer (ESCA) patients, but we cannot tailor the treatment to common biomarkers. WNT signaling activation was associated with primary resistance to immunotherapy. In this study, we used our two clinical cohorts (BJCH n = 95, BJIM n = 21) and three public cohorts to evaluate and verify a new immunotherapeutic biomarker based on WNT signaling in ESCA patients. Our findings showed that WNT signaling-related genes stratified TCGA patients into Cluster 1, 2, and 3, among which, Cluster 3 had the worst prognosis. The most up- and down-regulated genes in Cluster 3 were IGFBP1 and WNT3A. Further analysis validated that IGFBP1(hi)WNT3A(lo) ESCA patients had significantly poor RFS and OS in the TCGA and BJCH cohorts. Interestingly, IGFBP1(hi)WNT3A(lo) patients had a good response and prognosis with immunotherapy in three independent cohorts, exhibiting better predictive value than PD-L1 expression (signature AUC = 0.750; PD-L1 AUC = 0.571). Moreover, IGFBP1(hi)WNT3A(lo) patients may benefit more from immunotherapy than standard treatment (p = 0.026). Immune cell infiltration analysis revealed a significant increase in DC infiltration in IGFBP1(hi)WNT3A(lo) patients post-immunotherapy (p = 0.022), which may enhance immune response. The IGFBP1(hi)WNT3A(lo) signature could predict patients who benefited from PD-(L)1 inhibitor treatment and may serve as a biomarker in ESCA.