Search

Introduction In the PURE-01 study, a proportion of patients (pts) with muscle-invasive bladder cancer (MIBC; T2-4N0M0) had residual invasive disease (ypT2-4) at radical cystectomy (RC) after neoadjuvant pembrolizumab (pembro). The pembro-induced alterations in immunotherapy-resistant tumors remain largely unstudied. We aimed to investigate the biological characteristics of pembro-resistant tumors in comparison to neoadjuvant chemotherapy (NAC)-resistant tumors. Methods Gene expression profiling was performed on 26 RC samples from pts with ypT2-4 disease post-pembro, of which 22 had matched pre-pembro transurethral resection of the bladder tumor (TURBT) samples. Matched cases were analyzed by differential gene expression, hallmark signatures, and molecular subtyping (Decipher Bladder, TCGA, Consensus and Lund classifiers). Unsupervised consensus clustering (CC) was used to compare the 26 post-pembro samples with 133 post-NAC samples and to 21 samples collected from the former tumor bed of NAC-treated patients (scar tissue), all derived from RC specimens. Results Molecular subtyping of the pre- and post-pembro samples revealed significant ‘subtype switching’ with only 41% of samples having concordant subtypes using the Decipher Bladder classifier. The post-pembro samples did not cluster with scar tissues on clustering but were highly associated with immune-infiltrated cases from the post-NAC cohort. Two major groups of post-pembro tumors were identified, the first defined by markedly higher immune, stromal, angiogenesis and epithelial-to-mesenchymal signature scores and the second by higher tumor purity, KRAS and reactive oxygen species signature scores. Checkpoint inhibitor genes (CD274, PDCD1LG2) were consistently expressed in both groups. Conclusions This study expands our knowledge of pembro-resistant tumors finding the molecular characteristics of these tumors is strikingly different from NAC-resistant tumors. Two distinct clusters of tumors were identified post-pembro, neither of which were characterized as having a scar-like character. A larger cohort will be required to further understand the clinical implications of these findings.

Purpose:Neuroendocrine (NE) bladder carcinoma is a rare and aggressive variant. Molecular subtyping studies have found that 5% to 15% of muscle-invasive bladder cancer (MIBC) have transcriptomic patterns consistent with NE bladder cancer in the absence of NE histology. The clinical implications of this NE-like subtype have not been explored in depth.Experimental Design:Transcriptome-wide expression profiles were generated for MIBC collected from 7 institutions and clinical-use of Decipher Bladder. Using unsupervised clustering, we generated a clustering solution on a prospective training cohort (PTC; n = 175), developed single-sample classifiers to predict NE tumors, and evaluated the resultant models on a testing radical cystectomy (RC) cohort (n = 225). A random forest model was finalized and applied to 5 validation cohorts (n = 1302). Uni- and multivariable survival analyses were used to characterize clinical outcomes.Results:In the training cohort (PTC), hierarchical clustering using an 84-gene panel showed a cluster of 8 patients (4.6%) with highly heterogeneous expression of NE markers in the absence of basal or luminal marker expression. NE-like tumors were identified in 1% to 6.6% of cases in validation cohorts. Patients with NE-like tumors had significantly worse 1-year progression-free survival (65% NE-like vs. 82% overall; P = 0.046) and, after adjusting for clinical and pathologic factors, had a 6.4-fold increased risk of all-cause mortality (P = 0.001). IHC confirmed the neuronal character of these tumors.Conclusions:A single-patient classifier was developed that identifies patients with histologic urothelial cancer harboring a NE transcriptomic profile. These tumors represent a high-risk subgroup of MIBC, which may require different treatment.

Objectives To determine the potential impact of the presence of secondary variant histology on the gene expression profiles of muscle-invasive bladder cancer (MIBC) tumors. Methods For six tumors, revised samples were collected from urothelial and secondary variant components (cohort A). The commercial cohort (cohort B) consisted of the anonymized gene expression profiles of 173 patients with MIBC. Samples were obtained from the clinical use of the Decipher Bladder test that were available as part of the Decipher GRID prospective registry (NCT02609269). Secondary variant presence in cohort B was abstracted from institutional pathology reports. For the commercial cohort, only the urothelial carcinoma component was profiled. Results Molecular subtyping of both urothelial and variant components found micropapillary and nested cases were classified as a luminal subtype. Conversely, the sarcomatoid and small cell cases were classified as basal/squamous or neuroendocrine-like, respectively. For cohort B, 50 (29%) of 173 cases had reported secondary variant histology. Cases with squamous variant had basal profiles, small cell cases expressed neuronal markers, and micropapillary cases were classified as luminal. Sarcomatoid tumors had robust epithelial-mesenchymal transition marker expression. Conclusions Our data suggest that in MIBC with secondary variant, the urothelial component can demonstrate an expression profile that closely resembles the variant component.

Purpose: Neuroendocrine (NE) bladder carcinoma is a rare and aggressive variant. Molecular subtyping studies have found that 5% to 15% of muscle-invasive bladder cancer (MIBC) have transcriptomic patterns consistent with NE bladder cancer in the absence of NE histology. The clinical implications of this NE-like subtype have not been explored in depth. Experimental Design: Transcriptome-wide expression profiles were generated for MIBC collected from 7 institutions and clinical-use of Decipher Bladder. Using unsupervised clustering, we generated a clustering solution on a prospective training cohort (PTC; n = 175), developed single-sample classifiers to predict NE tumors, and evaluated the resultant models on a testing radical cystectomy (RC) cohort (n = 225). A random forest model was finalized and applied to 5 validation cohorts (n = 1302). Uni- and multivariable survival analyses were used to characterize clinical outcomes. Results: In the training cohort (PTC), hierarchical clustering using an 84-gene panel showed a cluster of 8 patients (4.6%) with highly heterogeneous expression of NE markers in the absence of basal or luminal marker expression. NE-like tumors were identified in 1% to 6.6% of cases in validation cohorts. Patients with NE-like tumors had significantly worse 1-year progression-free survival (65% NE-like vs. 82% overall; P = 0.046) and, after adjusting for clinical and pathologic factors, had a 6.4-fold increased risk of all-cause mortality (P = 0.001). IHC confirmed the neuronal character of these tumors. Conclusions: A single-patient classifier was developed that identifies patients with histologic urothelial cancer harboring a NE transcriptomic profile. These tumors represent a high-risk subgroup of MIBC, which may require different treatment.

Background In PURE01 study (NCT02736266), neoadjuvant pembro resulted in 42% pT0 in patients (pts) with MIBC. pT0 pts had features suggesting pre-existing immunity or higher tumor mutational burden (TMB) may promote response. In this study, we investigated potential mechanisms for resistance to pembro, including FGFR3 genomic alterations (GA). Methods Pts enrolled in the PURE-01, which is still recruiting pts in its amended design, had predominant urothelial carcinoma histology and stage cT ≤ 4N0 MIBC. Biomarker analyses in the expanded cohort of 96 pts included PD-L1 combined positive score and comprehensive genomic profiling (FoundationCDx assay). In addition, TCGA MIBC (n = 405) and a prospective commercial cohort (PCC) of 415 MIBC pts from the clinical use of the Decipher Bladder TURB test from the GRID registry (NCT02609269) were analyzed. A single-sample genomic classifier (GC) was trained to identify FGFR3-active tumors (FGFR3+). Results In PURE01 cohort, despite a linear association of TMB with ypT0 in multivariable models (p = 0.017), there was no association between FGFR3 GA and pathological response nor with PD-L1 expression. FGFR3 GA were found in 17.6% ypT0 vs 20.7% in ypT3-4 disease (corrected p = 0.82). Applying the GC to the TCGA MIBC cohort, we found 45% of FGFR3+ cases had FGFR3 mutations vs 9% for the rest of the cohort. Applying the GC to the PCC, we found that the FGFR3+ tumors were Luminal (n = 14), Luminal Infiltrated (n = 10) or Luminal Papillary (n = 30). FGFR3+ pts showed significantly lower PD-L1 (-0.03 vs. 0.109, p Conclusions In our combined cohorts, FGFR3 GA did not correlate with response to pembro. FGFR3 GA are enriched in FGFR3+ tumors, which in turn tend to have lower immune activity, PD-L1 and PD-L2, suggesting FGFR3 activity may provide a potential tool for further discriminating the mechanisms underlying response and resistance to neoadjuvant pembro in MIBC. Clinical trial identification NCT02736266 (PURE01) NCT02609269 (GRID Registry). Legal entity responsible for the study The authors. Funding Decipher Bioscience. Disclosure A. Necchi: Honoraria (institution), Advisory / Consultancy, Research grant / Funding (institution): Merck; Honoraria (institution), Advisory / Consultancy: Roche; Advisory / Consultancy: AstraZeneca; Advisory / Consultancy: BMS; Advisory / Consultancy, Research grant / Funding (institution): Incyte; Advisory / Consultancy: Clovis; Advisory / Consultancy: Janssen. R. Madison: Full / Part-time employment: Foundation Medicine. J. Chung: Full / Part-time employment: Foundation Medicine. Y. Liu: Full / Part-time employment: Decipher Bioscience. J. Chung: Full / Part-time employment: Foundation Medicine. J.S. Ross: Full / Part-time employment: Foundation Medicine. S.M. Ali: Full / Part-time employment: Foundation Medicine. E. Davicioni: Full / Part-time employment: Decipher Bioscience. E.A. Gibb: Full / Part-time employment: Decipher Bioscience. All other authors have declared no conflicts of interest.