MP16-13: Analysis of the tumor microenvironment (TME) in High-Grade Ta (HG Ta) Bladder Cancer reveals an immune and stromal enriched subtype which demonstrates poor prognosis with concurrent carcinoma-in situ (CIS)
Friday, May 3, 2024 1:00 PM to 3:00 PM · 2 hr. (US/Central)
221B
Abstract
Information
Full Abstract and Figures
Author Block
Mark Farha*, Jacob Tallman, New York , NY, Eugene Pietzak, New York, NY
Introduction
The survival rate for most patients with non-muscle invasive bladder cancer (NMIBC) is favorable, but recurrence and progression are common, with 15-year follow-up data demonstrating a rate of progression in Ta patients similar to T1 tumors at 40%. Higher fidelity identification of aggressive disease is needed to personalize therapy and avoid overtreatment. Here we perform an analysis of the TME in HG Ta disease to identify a subtype at increased risk of early recurrence.
Methods
Transcriptomic data from the UROMOL project, a European multicenter prospective study of NMIBC, was used to identify 93 patients with HG Ta NMIBC. TME deconvolution was performed using MCPCounter. Unsupervised hierarchical clustering divided the HG Ta cases into 6 clusters, with optimal cluster number determined by scree plot analysis. Recurrence free survival (RFS) was assessed by cluster for the entire cohort, and then was compared for a cluster of interest (CL5) against all other cases. A multi-variable cox hazards analysis adjusting for clinicopathologic and molecular variables was performed, and gene set enrichment analysis (GSEA) was conducted using GENAVi, a shiny web application, to identify enriched pathways.
Results
Comparison of the TME between HG Ta patients with and without CIS reveals that CIS cases are significantly enriched with cytotoxic lymphocytes (mean Z-score 0.59 vs. -0.14, p=0.03) and monocytes (mean Z-score 0.52 vs. -0.01, p=0.045). Unsupervised hierarchical clustering identified 6 TME clusters. A cluster that was pan-enriched in both immune and stromal content (CL5) emerged (n=15). This cluster had significantly more co-incident CIS (8/15, 53% vs. 9/80, 11%, p=0.0001). CL5 patients with CIS had worse RFS (4.5 vs. 17 months, p<0.0001). On multivariable cox hazards analysis, CL5+CIS was an independent poor prognostic factor for recurrence of HG Ta disease (HR 3.08, 95% CI 1.02 – 9.2, p=0.045) when considering molecular factors including variants in FGFR3, PIK3CA, ARID1A, TP53, ERBB2 and clinicopathologic features such as age, gender, tumor size and EORTC risk. Furthermore, CL5+CIS cases demonstrated an enrichment in cell cycle progression and epithelial to mesenchymal transition hallmark genes.
Conclusions
Our analyses reveals the heterogeneous composition and prognostic implications of the TME in HG Ta disease with CIS. Further studies are needed to explore the use of this classification scheme to guide prognostication and therapeutic decisions.
Source Of Funding
None
Author Block
Mark Farha*, Jacob Tallman, New York , NY, Eugene Pietzak, New York, NY
Introduction
The survival rate for most patients with non-muscle invasive bladder cancer (NMIBC) is favorable, but recurrence and progression are common, with 15-year follow-up data demonstrating a rate of progression in Ta patients similar to T1 tumors at 40%. Higher fidelity identification of aggressive disease is needed to personalize therapy and avoid overtreatment. Here we perform an analysis of the TME in HG Ta disease to identify a subtype at increased risk of early recurrence.
Methods
Transcriptomic data from the UROMOL project, a European multicenter prospective study of NMIBC, was used to identify 93 patients with HG Ta NMIBC. TME deconvolution was performed using MCPCounter. Unsupervised hierarchical clustering divided the HG Ta cases into 6 clusters, with optimal cluster number determined by scree plot analysis. Recurrence free survival (RFS) was assessed by cluster for the entire cohort, and then was compared for a cluster of interest (CL5) against all other cases. A multi-variable cox hazards analysis adjusting for clinicopathologic and molecular variables was performed, and gene set enrichment analysis (GSEA) was conducted using GENAVi, a shiny web application, to identify enriched pathways.
Results
Comparison of the TME between HG Ta patients with and without CIS reveals that CIS cases are significantly enriched with cytotoxic lymphocytes (mean Z-score 0.59 vs. -0.14, p=0.03) and monocytes (mean Z-score 0.52 vs. -0.01, p=0.045). Unsupervised hierarchical clustering identified 6 TME clusters. A cluster that was pan-enriched in both immune and stromal content (CL5) emerged (n=15). This cluster had significantly more co-incident CIS (8/15, 53% vs. 9/80, 11%, p=0.0001). CL5 patients with CIS had worse RFS (4.5 vs. 17 months, p<0.0001). On multivariable cox hazards analysis, CL5+CIS was an independent poor prognostic factor for recurrence of HG Ta disease (HR 3.08, 95% CI 1.02 – 9.2, p=0.045) when considering molecular factors including variants in FGFR3, PIK3CA, ARID1A, TP53, ERBB2 and clinicopathologic features such as age, gender, tumor size and EORTC risk. Furthermore, CL5+CIS cases demonstrated an enrichment in cell cycle progression and epithelial to mesenchymal transition hallmark genes.
Conclusions
Our analyses reveals the heterogeneous composition and prognostic implications of the TME in HG Ta disease with CIS. Further studies are needed to explore the use of this classification scheme to guide prognostication and therapeutic decisions.
Source Of Funding
None