Till Örebro universitet

Background: We have previously shown the prognostic value of PD-L1 protein and gene expression in early breast cancer (BC), however, the prognostic role of PD-1 expression remains unclear.

Methods: The prognostic value of PD-1 in early BC was investigated using three different approaches: i) evaluation of PD-1 at the protein (IHC, immunohistochemistry in tissue microarrays) and mRNA levels in a retrospective patient cohort of 586 patients treated for early BC in Stockholm, Sweden between 1997-2005, ii) systematic review and trial-level meta-analysis of studies published in Medline, Embase, Cochrane Library and Web of Science Core Collection libraries on the prognostic value of PD-1 IHC expression, and iii) pooled analysis of transcriptomic data from 39 publicly available datasets for the prognostic capacity of PD-1 gene expression. Univariate and multivariable Cox regression models were used.

Results: In the retrospective study cohort, PD-1 protein was significantly associated with biologically high-risk characteristics. PD-1 protein, but not gene expression, was correlated with improved overall survival (OS) (adjusted HR = 0.73, 95% CI 0.55 – 0.96, p = 0.023 and adjusted HR = 0.88, 95% CI 0.68 – 1.13, p = 0.307, respectively). In the trial-level meta-analysis, 4736 entries were initially identified and 15 studies, including our original cohort, fulfilled the predefined eligibility criteria. PD-1 IHC expression was not prognostic in unselected patients. However, a significant correlation to improved disease-free survival was seen within the triple-negative subtype (pooled multivariate HR = 0.57, 95% CI 0.29 – 0.90, p = 0.02). In the pooled gene expression analysis, PD-1 gene expression was associated with improved OS in the entire population (adjusted HR = 0.89, 95% CI 0.80 – 0.99, p = 0.025) and in basal-like (adjusted HR = 0.77, 95% CI 0.63 – 0.95, p = 0.014) tumors.

Conclusions: PD-1 expression at the RNA and protein levels represent promising prognostic factors, especially in the triple-negative and basal-like subtypes. Standardization and further validation are needed prior to clinical implementation.