Molecular classifications for urothelial bladder cancer appear to be promising in disease prognostication and prediction. This study investigated the novel molecular subtypes of muscle invasive bladder cancer (MIBC). Tumor samples and normal tissues of MIBC patients were submitted for transcriptome sequencing. Expression profiles were clustered using K-means clustering and principal component analysis. The molecular subtypes were also applied to The Cancer Genome Atlas (TCGA) dataset and analyzed for clinical outcome correlation. Three molecular subtypes of MIBC were discovered, clusters A, B, and C. The most differentially upregulated genes in cluster A were BDKRB1, EDNRA, AVPR1A, PDGFRB, and TNC, while the most upregulated genes in cluster C were collagen-related genes, PDGFRB, and PRKG1. For cluster B, COL6A3, COL1A2, COL6A2, tenascin C, and fibroblast growth factor 2 were statistically suppressed. When the centroids of clustering on PCA were applied to TCGA data, the clustering significantly predicted survival outcomes. Cluster B had the best overall survival (OS), and cluster C was associated with poor OS but exhibited the best response to perioperative chemotherapy. Among all groups, cluster B had a better pathologic response to neoadjuvant chemotherapy (40%). Based on the results of the present study, the novel clusters of subtype MIBC appear potentially suitable for integration into clinical practice.
Keywords: molecular subtypes; muscle-invasive bladder cancer; the cancer genome atlas (TCGA); transcriptomic analysis.