Prognostic value of tertiary lymphoid structures in triple-negative breast cancer: integrated analysis with the tumor microenvironment and clinicopathological features

Background In triple-negative breast cancer (TNBC), the most immunogenic breast cancer type, tumor-infiltrating lymphocytes (TILs) are an independent prognostic factor. Tertiary lymphoid structures (TLS) are an important TILs source, but they are not integrated in the current prognostic criteria. Methods In this retrospective study, TLS were assessed in hematein-eosin-saffron-stained (HES) histological sections from 397 early, chemotherapy-naive TNBC samples after primary surgical resection. Their association with i) classical clinicopathological features, ii) TILs and CD3+, CD8+, CD20+ lymphoid populations, iii) CD68+, CD163+, CD11b+, CD66b+ myeloid populations, and iv) expression of the PD1/PD-L1 and PVR/TIGIT axis immune checkpoint components and their prognostic significance were evaluated. Results TLS were observed in 88.2% of samples, mainly in peritumoral areas (86.1%). Increased amount of peritumoral TLS (PT-TLS) was significantly associated with younger age (p<0.001), smaller tumor size and higher tumor grade (both, p<0.001), HER2(null) tumors (versus HER2(low) tumors, p<0.002), and non-lobular histological type (p<0.016). TNBC with higher PT-TLS abundance displayed more often a basal-like (p<0.001) and not molecular-apocrine phenotype (p<0.001). TLS abundance was associated with TILs and hot tumor inflammatory pattern (both, p<0.001). Remarkably, PT-TLS abundance was positively associated with the density of the analyzed lymphoid (CD3+, CD8+, CD20+) and myeloid (CD68+, CD163+, CD11b+) cell populations (all p<0.001), with the exception of CD66b+ cells, as well as with expression of the PD1/PD-L1 and TIGIT/PVR immune checkpoint markers. In univariate analysis, beside the classical clinicopathological factors (tumor size, node involvement and adjuvant chemotherapy), TILs, hot tumors and PT-TLS were significantly associated with clinical outcome. Moreover, the risk of relapse was inversely correlated with PT-TLS abundance (Kaplan-Meier analysis). In multivariate analysis, pathological stage, adjuvant chemotherapy and PT-TLS remained correlated with relapse-free survival. Conclusion Our results suggest that TLS are a frequent feature in early TNBC and that their presence, particularly at the tumor periphery, recapitulates the tumor immune microenvironment. In our series, their prognostic value outperformed that of TILs. Therefore, their easy quantification on routine HES sections and their integration into the factors classically analyzed by pathologists could improve the clinical management of TNBC, a breast cancer type whose prognosis remains too poor.