Distinct tumor microenvironmental landscapes define molecular subtypes in non-muscle-invasive bladder cancer

Background: Non-muscle-invasive bladder cancer (NMIBC) are biologically heterogeneous tumors that are classified based on stage/grade alone, which does not provide an accurate separation as it misses underlying molecular phenotypes. The UROMOL molecular classification, derived from large-scale transcriptomic profiling of NMIBC (Lindskrog SV, et al. Nat Commun . 2021;12:2301), stratifies tumors into prognosis-associated biology with distinct molecular subtypes. The relationship between molecular states and tumor microenvironment (TME) composition remains poorly defined. Methods: We performed transcriptomic subtyping and single-cell-level TME profiling of NMIBC tumors (patient n=16, sample n=34, cell n = 108108). Tumors were classified into UROMOL subtypes using NMIBC classifier (Lindskrog SV, et al. Nat Commun . 2021;12:2301) based on aggregated gene expression profiles. We quantified epithelial, stromal, and immune populations, and assessed pathway activity using curated immune and stromal gene module scores. Results: 3 UROMOL subtypes were observed: Class 1 (luminal papillary), Class 2a (basal-inflamed), and Class 2b (luminal-inflamed). Class 1 tumors (n=6, 17.6%) exhibited an epithelial-dominant, immune-cold TME characterized by ~80.1% epithelial cells (95% CI: 74-86.2%) and minimal immune infiltration, including T cells (~1.9%, CI: 0.5-3.3%), macrophages (~1.4%, CI: 0.5-3.3%), and fibroblasts (~8.4%, CI: 4.3-12.5%). This composition is consistent with differentiated luminal biology and may have favorable baseline prognosis. Class 2a tumors (n=9, 26.5%) displayed a distinct TME composition pattern, with an increase in T-cell (~4.8%, CI: 1.7-8.0%) and macrophage (~1.9%, CI: 0.6-3.1%) infiltration compared to Class 1, alongside increased fibroblast (~17.5%, CI: 9.6-25.3%) and endothelial (~8.1%, CI: 4.9-11.4%) content. These tumors showed increased effector and checkpoint activity, reflecting a more active immune microenvironment than class 1. Class 2b tumors (n=19, 55.9%) showed a heterogeneous immune-stromal landscape with pronounced fibroblast enrichment (~27%, CI: 19.7-34.4%) and extracellular matrix remodeling, alongside elevated immune infiltration, including T cells (~17.5%, CI: 15-20.0%) and macrophages (~11.2%, CI: 7.4-14.9%), this reflects an immune-active, fibroblast-rich TME architecture. Conclusions: UROMOL subtypes are defined by distinct TME architectures with differences in cellular composition from immune-cold, epithelial-rich Class 1 tumors to highly immune infiltrated Class 2b tumors enriched with T cells and fibroblasts. Given the limited sample size, these observations should be validated in a larger cohort. This study demonstrates that UROMOL molecular classes of NMIBC are associated with distinct TMEs, revealing biologically meaningful differences in immune infiltration and stromal composition.