No abstract

Lymphoid neogenesis gives rise to tertiary lymphoid structures (TLS) in the periphery of multiple cancer types including muscle invasive bladder cancer (MIBC) where it has positive prognostic and predictive associations. Here, we explored molecular, clinical, and histological data of The Cancer Genome Atlas, as well as the IMvigor210 dataset to study factors associated with TLS development and function in the tumor microenvironment (TME) of MIBC. We also analyzed tumor immune composition including TLS in an independent, retrospective MIBC cohort. We found that the combination of TLS density and tumor mutational burden provides a novel independent prognostic biomarker in MIBC. Gene expression profiles obtained from intratumoral regions that rarely contain TLS in MIBC showed poor correlation with the prognostic TLS density measured in tumor periphery. Tumors with high TLS density showed increased gene signatures as well as infiltration of activated lymphocytes. Intratumoral B-cell and CD8+ T-cell co-infiltration was frequent in TLS-high samples, and such regions harbored the highest proportion of PD-1+TCF1+ progenitor-like T cells, naïve T cells, and activated B cells when compared to regions predominantly infiltrated by either B cells or CD8+ T cells alone. We found four TLS maturation subtypes; however, differences in TLS composition appeared to be dictated by the TME and not by the TLS maturation status. Finally, we identified one downregulated and three upregulated non-immune cell-related genes in TME with high TLS density, which may represent candidates for tumor-intrinsic regulation of lymphoid neogenesis. Our study provides novel insights into TLS-associated gene expression and immune contexture of MIBC and indicates towards the relevance of B-cell and CD8+ T-cell interactions in anti-tumor immunity within and outside TLS.

No abstract

Tertiary lymphoid structures (TLSs) are dense accumulations of lymphocytes in inflamed peripheral tissues, including cancer, and are associated with improved survival and response to immunotherapy in various solid tumors. Histological TLS quantification has been proposed as a novel predictive and prognostic biomarker, but lack of standardized methods of TLS characterization hampers assessment of TLS densities across different patients, diseases, and clinical centers. We introduce a novel approach based on HookNet-TLS, a multi-resolution deep learning model, for automated and unbiased TLS quantification and identification of germinal centers in routine hematoxylin and eosin stained digital pathology slides. We developed a HookNet-TLS model using n=1019 manually annotated TCGA slides from clear cell renal cell carcinoma, muscle-invasive bladder cancer, and lung squamous cell carcinoma. We show that HookNet-TLS automates TLS quantification with a human-level performance and demonstrates prognostic associations similar to visual assessment. We made HookNet-TLS publicly available to aid the adoption of objective TLS assessment in routine pathology.