Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Jul 10.
Published in final edited form as: Int J Clin Oncol. 2026 Mar 8;31(5):868–877. doi: 10.1007/s10147-026-03005-2

Comparative genomic landscape of primary and metastatic bladder urothelial carcinoma in a large-scale cohort

Akira Ohtsu 1,2,3, Yusuke Otani 3,4,5, Seiji Arai 1, Anna Rogachevskaya 3,4, Vanessa D Chin 6, Shinichi Toyooka 5, Kazuhiro Suzuki 1, Wenyi Wei 3,4, Atsushi Tanaka 3,4,7
PMCID: PMC13347192  NIHMSID: NIHMS2188977  PMID: 41795757

Abstract

Background

Metastatic bladder urothelial carcinoma has poor survival, and large comparative genomic studies using uniform targeted sequencing of paired primary and metastatic lesions remain limited. We compared gene- and pathway-level alterations between primary and metastatic tumors

Methods

We analyzed 2,880 bladder urothelial carcinoma samples (2,305 primary; 575 metastatic) from 2,343 patients profiled with MSK-IMPACT. Somatic mutations and copy number alterations were integrated per gene and compared between primary and metastatic samples in the full cohort and in a paired subset using standard statistical tests.

Results

Primary and metastatic samples showed broadly similar driver landscapes. In the full cohort, KDM6A, FGFR3, STAG2, and ERCC2 were more frequently altered in primary tumors, whereas no individual genes were enriched in metastases; these differences were not significant in paired analyses. At the pathway level, TP53 pathway alterations were relatively more frequent in metastases, while DNA damage response alterations were enriched in primary tumors; other pathways showed comparable alteration rates. Apoptosis-focused analyses identified no significant gene-level differences, but suggested a trend toward higher alteration rates in the TP53 pathway and apoptosis regulators in metastases.

Conclusion

Primary and metastatic lesions of bladder urothelial carcinoma show broadly similar gene- and pathway-level alteration profiles on targeted DNA sequencing. TP53 pathway and apoptosis-related alterations are modestly more frequent in metastases, consistent with impaired stress responses and apoptosis evasion.

Keywords: Bladder urothelial carcinoma, Primary, Metastasis, Targeted next-generation sequencing, TP53 pathway, DNA damage response

Introduction

Urothelial carcinoma of the bladder is a common malignancy worldwide and remains a major cause of cancer-related mortality despite advances in surgery, systemic therapy, and immunotherapy [1]. Standard treatment for muscle-invasive bladder cancer includes cisplatin-based perioperative chemotherapy and radical cystectomy, while metastatic urothelial carcinoma is treated with platinum-based chemotherapy, immune checkpoint inhibitors, antibody–drug conjugates, and, more recently, targeted agents directed against fibroblast growth factor receptor 3 (FGFR3) and other actionable alterations [1-5]. Nonetheless, the prognosis of patients with metastatic disease remains poor, and a better understanding of the genomic events that underpin disease progression and metastasis is needed to refine risk stratification and to identify new therapeutic vulnerabilities.

Large-scale genomic studies have defined the landscape of driver alterations in bladder cancer. The Cancer Genome Atlas project and subsequent efforts have shown that urothelial carcinoma harbors frequent mutations in TP53, chromatin regulatory genes such as KDM6A, ARID1A, KMT2D, CREBBP, and EP300, along with recurrent alterations in cell cycle, receptor tyrosine kinase, and PI3K pathway genes including RB1, CDKN2A/B, FGFR3, ERBB2, and PIK3CA [6, 7]. These landmark studies have typically included several hundred bladder cancer cases, for example, The Cancer Genome Atlas (TCGA) muscle-invasive cohort of 412 samples, and an institutional series of approximately 500 profiled bladder tumor samples. More recently, integrative genomic analyses have also characterized non-muscle-invasive disease, showing that alterations in FGFR3, KDM6A, and STAG2 are particularly common in earlier-stage tumors [8-10]. However, comprehensive analyses of very large cohorts that include both primary and metastatic urothelial carcinoma of the bladder, profiled on uniform targeted sequencing panels used in routine practice, remain limited [11-13].

Changes in DNA damage response genes, including ERCC2 and other elements of the nucleotide excision repair pathway, have been shown to influence how tumors respond to cisplatin-based chemotherapy. These alterations are also associated with distinctive mutational signatures observed in urothelial carcinoma [14-16]. At the same time, loss of TP53 function and perturbation of downstream apoptotic regulators, such as MDM2, BCL2 family members, and caspases, are central mechanisms by which cancer cells evade stress-induced cell death [17, 18]. These apoptotic pathways are therefore attractive targets for therapeutic intervention and may be selectively perturbed in tumors that have acquired the capacity to metastasize. However, systematic analyses of how gene- and pathway-level alterations differ between primary and metastatic urothelial carcinoma of the bladder, particularly in large, clinically annotated cohorts profiled with targeted next-generation sequencing, remain limited [11, 12].

In the present study, we leveraged the American Association for Cancer Research Genomics Evidence Neoplasia Information Exchange (AACR GENIE) consortium dataset to analyze urothelial carcinoma of the bladder profiled by three Memorial Sloan Kettering targeted sequencing panels. We examined 2,305 primary and 575 metastatic samples, representing, to our knowledge, the largest uniformly profiled cohort of bladder urothelial carcinoma reported to date and the largest series including both primary and metastatic lesions analyzed on the same institutional platforms [6, 12, 19]. We first characterized the overall genomic landscape of the top 30 most frequently altered genes. We then compared gene-level alteration frequencies between primary and metastatic samples in both unpaired and paired cohorts, examined pathway-level differences across canonical cancer signaling pathways, and, finally, focused on alterations in apoptosis-related genes. By integrating gene-level and pathway-level analyses in this large cohort, we aimed to identify genomic features associated with metastatic disease and to clarify the extent to which primary tumors recapitulate the genomic profile of metastatic lesions in urothelial carcinoma of the bladder.

Materials and methods

Dataset acquisition

AACR Project GENIE public data (v18.0), including clinical information, somatic mutation calls, and copy number calls, were downloaded from cBioPortal [20-23]. We selected samples with CANCER_TYPE_DETAILED = = "Bladder Urothelial Carcinoma" and excluded samples with “Unspecified” in SAMPLE_TYPE and “Local recurrence” in SAMPLE_TYPE_DETAILED in the metadata. We then selected only MSK-IMPACT assays with 410, 468, or 505 gene panels, resulting in a total of 2,880 samples (2,305 primary bladder urothelial carcinoma samples and 575 metastatic lesion bladder urothelial carcinoma samples from 2,343 patients). The GENIE sample IDs used in this study are listed in Table S1.

Processing and coding of genomic alterations

Somatic mutation data were obtained from the data_mutations_extended file in the GENIE v18.0 release. We used the “Hugo_Symbol” and “Tumor_Sample_Barcode” to map each variant to a specific sample–gene pair. Variant_Classification entries were grouped per gene and summarized at the sample level. For frequency and pathway analyses, a gene was considered altered if it harbored at least one somatic variant (missense, nonsense, frameshift insertion or deletion, in-frame insertion or deletion, splice site, 5′ flank) and/or a copy number alteration. Gene-level copy number data were obtained from the data_CNA file. Copy number calls were provided as integer values for each gene–sample pair. We treated − 2 as a deep deletion and + 2 as an amplification; all other integer values (− 1, 0, + 1) were regarded as copy number neutral for the purpose of defining alteration events. For each sample–gene pair, we therefore assigned a categorical copy number class of “Deep deletion”, “Amplification”, or “Neutral”. Fraction genome altered (FGA) values were taken from cBioPortal [20-23] and used without additional processing as a per-sample measure of global copy number alteration burden. These values were displayed as bar plots above the oncoprint panels for primary and metastatic samples.

Definition of primary, metastatic, and paired cohorts

Tumor samples were classified as primary or metastatic based on the SAMPLE_TYPE annotation in the clinical sample file. Samples annotated as primary tumors were grouped into the “Primary” cohort, and samples annotated as metastases into the “Metastatic” cohort, after exclusion of “Unspecified” and “Local recurrence” as described above. For paired analyses, we identified patients who had at least one primary and one metastatic sample. Among these patients, one primary and one metastatic sample per patient were selected to construct a paired primary–metastatic cohort (primary n = 222, metastatic n = 222), which was used to compare gene-level alteration frequencies between matched disease sites.

Pathway definition

Assignment of genes to specific pathways (Table S2) was performed to provide a landscape of frequently mutated biological processes. Pathway definitions were based on previously published templates for canonical oncogenic signaling pathways [24].

Apoptosis-related gene set

In addition to the pathway groupings described above, we examined a focused set of apoptosis-related genes. This set comprised TP53, CDKN2A, EP300, CDKN1A, MDM2, ATM, MCL1, MYC, NF2, PTEN, RUNX1, CASP8, BIRC3, BCL2L1, PARP1, FOXO1, BCL6, CYLD, STK11, XIAP, BCL10, BCL2L11, BBC3, MDM4, TNFRSF14, GATA1, BCL2, and PMAIP1. Genes were selected a priori based on their established roles as key regulators of p53 signaling and intrinsic or extrinsic apoptosis in cancer, as reported in previous reviews, rather than through any data driven feature selection [17, 18, 25, 26].

Calculation of gene- and pathway-level alteration frequencies

For each gene, alteration frequency was defined as the proportion of samples with at least one qualifying event (mutation and/or copy number alteration) among all evaluable samples in each cohort. In samples harboring multiple event types within the same gene (for example, mutation plus amplification), each sample was assigned a single prioritized event class (copy number alteration events taking precedence over mutations) for display in stacked bar plots. For pathway-level analyses, each pathway was defined as the set of genes listed in Table S2. A pathway was considered altered in a given sample if at least one constituent gene in that pathway was altered. Pathway alteration rates were calculated as the proportion of samples with at least one altered gene in the pathway, separately for primary and metastatic cohorts.

Statistical analysis

All statistical analyses were performed in R (version 4.5.1) [27]. Heatmaps were generated using the ComplexHeatmap package (version 2.24.1), and bar plots were generated using ggplot2. For comparisons of gene-level alteration frequencies between primary and metastatic samples, two-sided Fisher’s exact tests were used, and p-values were adjusted for multiple comparisons using the Benjamini–Hochberg method.

Results

Genomic landscape of primary and metastatic bladder urothelial carcinoma

From the AACR GENIE cohort, we first selected urothelial carcinoma of the bladder profiled with one of three Memorial Sloan Kettering targeted sequencing panels (MSK-IMPACT410, MSK-IMPACT468, or MSK-IMPACT505), yielding 2,305 primary and 575 metastatic samples. The overall genomic landscape of the top 30 most frequently altered genes, ordered by alteration frequency in primary samples and including both somatic mutations and copy number alterations, is shown as an oncoprint (Fig. 1). TP53 was the most frequently altered gene in both primary and metastatic samples (primary 49.3%, metastatic 53.4%), followed by TERT (primary 42.8%, metastatic 40.3%), KDM6A (primary 33.4%, metastatic 26.3%), and ARID1A (primary 28.6%, metastatic 28.3%). Other recurrently altered genes included KMT2D, CDKN2A/B, FGFR3, RB1, PIK3CA, ERBB2, ELF3, STAG2, KMT2C, CREBBP, EP300, ERCC2, and others. Overall, this genomic overview is consistent with previous large-scale reports of bladder urothelial carcinoma and shows that the rank order of the most frequently altered genes was broadly similar between primary and metastatic samples (Fig. 1).

Fig. 1.

Fig. 1

Open in a new tab

Genomic landscape of primary and metastatic bladder urothelial carcinoma. Oncoprint of the 30 most frequently altered genes in primary (n = 2,305) and metastatic (n = 575) bladder urothelial carcinoma samples profiled with MSK-IMPACT 410, 468, or 505 gene panels in AACR Project GENIE v18.0. Each column represents a tumor sample, and each row represents a gene, with genes ordered from top to bottom according to their overall alteration frequency in primary samples. Colored boxes indicate somatic mutations, amplifications, or deep deletions, and black boxes (“Non_exist”) indicate genes that were not targeted by the specific MSK-IMPACT panel used for that sample and were therefore not evaluable for alterations. Mutation types (missense, nonsense, frameshift insertion or deletion, in-frame insertion or deletion, splice site, and 5′ flank), aggregated mutation classes (“Mutation_others” and “Multi_hit”), and copy number events (amplification and deep deletion) are color coded as shown in the legend. The “Mutation_others” category includes intronic and noncoding/protein-silent variant classes (Intron/intron, Nonstop_Mutation, Splice_Region/splice_region, Translation_Start_Site/translation_start_site, 5′UTR, 3′UTR, 3′Flank, and Silent), and “Multi_Hit” is assigned when ≥ 2 mutation classes co-occur in the same gene–sample pair. The bar plots above the heatmap show the fraction genome altered (FGA) for each sample

Gene-level differences in alteration frequency between primary and metastatic samples

We next compared the alteration frequencies of these 30 recurrently altered genes between primary and metastatic samples in the full cohort. Four genes showed significantly higher alteration rates in primary samples than in metastatic samples: KDM6A, FGFR3, STAG2, and ERCC2 (P < 0.05, Fisher’s exact test). In contrast, no genes showed significantly higher alteration frequencies in metastatic samples in this unpaired comparison. These findings are in line with prior reports that KDM6A, FGFR3, STAG2, and ERCC2 alterations are enriched in non-muscle-invasive or earlier-stage bladder cancers and are less common in more advanced disease (Fig. 2, left panel).

Fig. 2.

Fig. 2

Open in a new tab

Comparison of gene-level alteration frequencies between primary and metastatic bladder urothelial carcinoma in all samples and in a paired-sample cohort. Per-gene alteration frequencies of the 30 most frequently altered genes in primary and metastatic bladder urothelial carcinoma. The left panel summarizes all evaluable samples (primary n = 2,305, metastatic n = 575), and the right panel shows the subset of patients with paired primary and metastatic biopsies (primary n = 222, metastatic n = 222). Genes are ordered according to alteration frequency in primary tumors in the full cohort, and the same order is used for the paired-sample analysis. For each gene, alteration frequencies were calculated using only samples in which that gene was actually assayed, excluding “Non_exist” calls from panels that did not include that target. We compared primary and metastatic tumors using two-sided Fisher’s exact tests and adjusted for multiple comparisons with the Benjamini–Hochberg method. Significance is indicated above the metastatic bars: * for FDR < 0.05, ** for FDR < 0.01, and *** for FDR < 0.001. For each gene, the upper bar represents primary samples, and the lower bar represents metastatic samples

To distinguish between differences driven by cohort composition and those arising during tumor evolution, we performed the same analysis in a subset of 222 patients with matched primary and metastatic samples. In this paired analysis, the differences in alteration frequency for KDM6A, FGFR3, STAG2, and ERCC2 were no longer statistically significant (Fig. 2, right panel). Consistent with this finding, gene-level concordance between matched primary–metastatic pairs was generally high (> 90% for most genes), whereas TERT (81%), CDKN2A (82%), and CDKN2B (82%) showed comparatively lower concordance (Supplemental Fig. S1). Overall, this suggests that once a tumor has acquired the capacity to metastasize, the alteration status of these genes tends to be maintained between the primary and metastatic sites, and that the primary versus metastatic differences observed in the full cohort are largely driven by the overrepresentation of these alterations in tumors that do not go on to develop metastases.

Pathway-level comparison of genomic alterations between primary and metastatic samples

We then grouped genes into canonical signaling and cancer-related pathways and compared the proportion of samples with at least one alteration in each pathway between primary and metastatic samples (Fig. 3). Among the pathways examined, the TP53 pathway (primary 59%, metastatic 64%) showed a higher alteration rate in metastatic samples than in primary samples, whereas alterations in DNA damage response genes (primary 50%, metastatic 43%) were more frequent in primary samples. Other pathways, including receptor tyrosine kinase (RTK) (primary 69%, metastatic 69%), cell cycle (primary 66%, metastatic 66%), PI3K (primary 50%, metastatic 48%), NOTCH (primary 48%, metastatic 46%), SWI/SNF (primary 46%, metastatic 44%), RAS (primary 33%, metastatic 32%), immune (primary 27%, metastatic 25%), Hippo (primary 23%, metastatic 24%), Myc (primary 17%, metastatic 16%), telomere (primary 14%, metastatic 15%), WNT/β-catenin (primary 18%, metastatic 15%), and NRF2 signaling (primary 9%, metastatic 9%) showed broadly similar alteration rates between primary and metastatic samples, with only modest differences in frequency. These data suggest that acquisition or selection of TP53 pathway defects may be particularly important for the development or maintenance of metastatic disease, while tumors with prominent DNA damage pathway alterations may be relatively enriched among non-metastatic primary cases.

Fig. 3.

Fig. 3

Open in a new tab

Pathway-level alteration rates in primary and metastatic bladder urothelial carcinoma. Pathway-level alteration rates in primary and metastatic bladder urothelial carcinoma. Pathways were defined from curated gene sets (Table S2) and include receptor tyrosine kinase (RTK), cell cycle, TP53 pathway, PI3K, NOTCH, SWI/SNF, DNA damage response, RAS, immune-related, Hippo, Myc, telomere, WNT/β-catenin, and NRF2-related pathways. For each pathway, bars show the proportion of primary and metastatic tumors with at least one constituent gene altered, calculated among evaluable samples for genes present on the corresponding MSK-IMPACT panel

Alteration patterns in apoptosis-related genes

Because TP53 plays a central role in stress sensing and apoptosis regulation, we next focused on a curated apoptosis-related gene set (see Methods) to examine the distribution of alterations in key regulators of p53 signaling and programmed cell death (Fig. 4). Overall, there were no statistically significant differences in the alteration frequencies of individual apoptosis-related genes between primary and metastatic samples. However, several genes showed numerically higher alteration rates in metastatic samples, including TP53 and its negative regulator MDM2; the anti-apoptotic factor MCL1; the oncogene MYC; the initiator caspase CASP8; and the apoptosis inhibitors XIAP and BCL10; and BCL2 family members. These trends support the notion that metastatic bladder tumors may more frequently harbor alterations in genes that dampen apoptotic responses or enhance survival signaling, although the effect sizes were modest and did not reach statistical significance in this dataset.

Fig. 4.

Fig. 4

Open in a new tab

Alteration frequencies of apoptosis-related genes in primary and metastatic bladder urothelial carcinoma. Per-gene alteration frequencies of selected apoptosis-related genes in primary (n = 2,305) and metastatic (n = 575) bladder urothelial carcinoma samples. Stacked vertical bar plots compare the percentage of primary and metastatic tumors harboring each event type for genes involved in apoptotic regulation

Discussion

In this study, we used a large, uniformly profiled cohort from the AACR GENIE project to compare the genomic landscape of primary and metastatic urothelial carcinoma of the bladder. The overall pattern of recurrent alterations in the full cohort was highly concordant with previous large-scale genomic studies of bladder cancer, with frequent mutations and copy number alterations in TP53, TERT, KDM6A, ARID1A, KMT2D, CDKN2A/B, FGFR3, RB1, PIK3CA, and other established driver genes [6, 7]. The rank order of the most frequently altered genes was broadly similar between primary and metastatic samples, indicating that the same core set of drivers underpins both local and disseminated disease. This observation is consistent with prior reports suggesting that metastatic urothelial carcinoma shares the major driver landscape seen in primary cohorts, even though the clinical behavior of metastatic disease is markedly more aggressive [12]. By virtue of its size and uniform profiling, our cohort extends these observations and, to our knowledge, provides the most comprehensive targeted sequencing-based comparison of primary and metastatic bladder urothelial carcinoma to date.

At the gene level, our comparison of 2,305 primary and 575 metastatic samples identified four genes with significantly higher alteration frequencies in primary samples: KDM6A, FGFR3, STAG2, and ERCC2. KDM6A, FGFR3, and STAG2 have been repeatedly implicated in non-muscle-invasive and earlier-stage bladder cancer and are often associated with more differentiated or luminal phenotypes [9, 10]. In contrast, ERCC2 and other DNA damage response genes have been linked to distinct mutational patterns and improved response to cisplatin-based chemotherapy in muscle-invasive disease [14, 16], suggesting that ERCC2-mutant tumors may be selectively depleted from the population of patients who eventually develop metastatic, biopsy-accessible lesions. Thus, while the enrichment of KDM6A-, FGFR3-, and STAG2-altered tumors in the primary cohort is consistent with the overrepresentation of less dissemination-prone, non-metastatic cases, the higher primary frequency of ERCC2 likely reflects treatment- and sampling-related selection rather than a simple reduction in metastatic potential. In contrast, no genes were significantly more frequently altered in metastatic samples in this unpaired comparison, underscoring that metastatic competence is unlikely to be driven by a small number of metastasis-specific mutations in the genes targeted by the panels used here.

A key strength of our analysis is the inclusion of 222 patients with matched primary and metastatic samples profiled on the same platform. In this paired cohort, the differences in alteration frequency for KDM6A, FGFR3, STAG2, and ERCC2 were no longer statistically significant. These findings suggest that these cohort-level differences do not consistently occur within patients, although we observed gene-level discordance for selected genes. Together with previous exome-based studies showing substantial overlap in driver alterations between metastatic lesions and primary tumors [12], and with the prospective MSK-IMPACT study by Clinton and colleagues, which used a smaller but richly annotated cohort and likewise reported broadly similar frequencies of recurrent driver alterations between muscle-invasive primary and metastatic urothelial carcinomas despite lesion-to-lesion heterogeneity [28], our findings are consistent with a model in which the major driver events that define the overall driver landscape of a tumor are established early during carcinogenesis and are carried forward into metastatic disease. From a practical standpoint, this implies that, particularly in patients who develop metastatic urothelial carcinoma, genomic profiling of primary tumor specimens may provide a reasonable proxy for many recurrent driver and pathway-level alterations captured by current clinical panels [19, 28]. However, clinically meaningful site-specific differences may still occur at the individual patient level.

Pathway-level analysis provided additional insight into the biology of metastatic disease. When genes were grouped into canonical pathways, TP53 pathway alterations were more frequent in metastatic samples, whereas DNA damage response gene alterations were more common in primary samples. In contrast, the alteration frequencies of other pathways, including receptor tyrosine kinase, cell cycle, PI3K, NOTCH, SWI/SNF, RAS, immune, Hippo, Myc, telomere, WNT/β-catenin, and NRF2 signaling, were broadly similar between primary and metastatic samples. The relative enrichment of TP53 pathway alterations in metastatic tumors is consistent with TP53’s central role in integrating stress signals and enforcing cell cycle arrest and apoptosis. Loss of TP53 function, or dysregulation of its downstream network, may facilitate survival under the selective pressures associated with invasion, dissemination, and therapeutic exposure [14, 16].

Given the importance of TP53 in regulating apoptosis, we further examined an apoptosis-related gene set that included TP53 pathway components and additional apoptotic regulators. Although there were no statistically significant differences in alteration frequencies for individual apoptosis-related genes between primary and metastatic samples, several genes showed numerically higher alteration rates in metastatic tumors, including TP53, MDM2, MCL1, MYC, CASP8, XIAP, and BCL10, as well as other BCL2 family members. These trends, though modest, suggest that metastatic bladder tumors may more frequently harbor alterations that either blunt pro-apoptotic signaling or enhance survival and proliferation. Such alterations could contribute to resistance to therapy-induced cell death and may represent rational targets for combination strategies that incorporate inhibitors of the BCL2 family or other apoptosis-modulating agents.

From a clinical perspective, our findings highlight both the continuity and the subtle evolution of the genomic landscape from primary to metastatic bladder cancer. The broad similarity in the rank order of recurrently altered genes and in many pathway-level alteration frequencies supports the use of targeted sequencing of primary tumors to identify actionable alterations and to characterize key driver patterns. In particular, in patients with established metastatic disease, our paired-sample analysis suggests that sequencing of archival or contemporaneous primary tumor tissue often captures many of the key driver and pathway-level alterations present in metastatic lesions, which is important when metastatic sites are difficult or risky to biopsy. However, clinically meaningful site-specific gene alterations may still occur in individual patients. At the same time, our data do not exclude the possibility of therapy-induced clonal evolution or site-specific heterogeneity, and rebiopsy or liquid biopsy may still be warranted in selected clinical scenarios, especially after multiple lines of systemic therapy [28-31].

Several limitations of this study should be acknowledged. First, the analysis was retrospective and based on targeted sequencing panels that interrogate predefined gene sets. We were therefore unable to assess the contribution of alterations outside the panel genes, noncoding variants, structural rearrangements, or epigenetic changes. Second, although the AACR GENIE cohort is large and clinically annotated, detailed treatment information, the timing of sample acquisition, and specific metastatic sites were not available for all cases, which limits the ability to link particular genomic patterns with therapeutic exposures or the organotropism of metastases. Third, because only a single specimen was typically analyzed from the primary lesion, intratumoral heterogeneity within the primary tumors may be underestimated, potentially underrepresenting spatially distinct genomic alterations at the primary site. In addition, because transcriptomic data were unavailable, we could not assign expression-defined molecular subtypes or perform subtype-stratified site comparisons, and our conclusions should therefore not be generalized across these subtypes. Our apoptosis-related gene set, while capturing key regulators, does not capture the full complexity of cell death pathways, including intrinsic and extrinsic apoptosis, necroptosis, ferroptosis, and other stress-response programs. Finally, the statistical power to detect modest differences in alteration frequency for less commonly altered genes is limited, especially in the paired subset.

Future work integrating targeted sequencing or whole exome sequencing with transcriptomic, epigenomic, and proteomic profiling of matched primary and metastatic urothelial carcinoma, as well as longitudinal sampling before and after therapy, will be important to fully delineate the evolutionary trajectories that lead to metastatic competence and treatment resistance [8, 12, 16]. Nonetheless, by lever-aging a very large real-world cohort profiled on clinically relevant sequencing panels, our study provides a comprehensive view of gene- and pathway-level alterations in primary and metastatic bladder urothelial carcinoma and highlights the potential importance of the TP53 pathway and apoptosis-related alterations in the biology of metastatic disease, while supporting the practical use of primary tumor sequencing as a surrogate for metastatic profiling in many patients.

Conclusion

Primary and metastatic bladder urothelial carcinomas exhibit a remarkably similar pattern of recurrent gene and pathway alterations on targeted sequencing. This overlap suggests that primary tumors often reflect the key genomic signals found in metastatic disease, making them useful proxies when metastatic samples are scarce. The similarity also indicates a stable core genomic architecture that endures throughout tumor development. Apoptosis- and TP53 pathway-related alterations were modestly more frequent in metastatic samples, suggesting a role for impaired stress responses and apoptosis evasion in metastatic progression.

Supplementary Material

Table S2
Table S1
Figure S1

Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s10147-026-03005-2.

Acknowledgements

We thank the investigators of the AACR Project GENIE program for generating and sharing the data, as well as the cBioPortal team for providing data access and infrastructure. The AACR-GENIE investigators have no responsibility for the analyses or interpretation presented here

Funding

This work was supported by the Japan Society for the Promotion of Science (JSPS) KAKENHI (Grant Number 23K08773)

Footnotes

Conflict of interest The authors declare that they have no competing interests.

Ethical approval Not applicable. This study used only de-identified, publicly available data and did not involve direct interaction with human participants.

Data availability

The datasets analyzed in this study are publicly available from the AACR Project GENIE repository and cBioPortal

References

  • 1.Patel VG, Oh WK, Galsky MD (2020) Treatment of muscle-invasive and advanced bladder cancer in 2020. CA Cancer J Clin 70(5):404–423 [DOI] [PubMed] [Google Scholar]
  • 2.Bellmunt J, de Wit R, Vaughn DJ et al. (2017) Pembrolizumab as second-line therapy for advanced urothelial carcinoma. N Engl J Med 376(11):1015–1026 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Loriot Y, Necchi A, Park SH et al. (2019) Erdafitinib in locally advanced or metastatic urothelial carcinoma. N Engl J Med 381(4):338–348 [DOI] [PubMed] [Google Scholar]
  • 4.Powles T, Park SH, Voog E et al. (2020) Avelumab maintenance therapy for advanced or metastatic urothelial carcinoma. N Engl J Med 383(13):1218–1230 [DOI] [PubMed] [Google Scholar]
  • 5.Powles T, Valderrama BP, Gupta S et al. (2024) Enfortumab vedotin and pembrolizumab in untreated advanced urothelial cancer. N Engl J Med 390(10):875–888 [DOI] [PubMed] [Google Scholar]
  • 6.Cancer Genome Atlas Research Network (2014) Comprehensive molecular characterization of urothelial bladder carcinoma. Nature 507(7492):315–322 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Robertson AG, Kim J, Al-Ahmadie H et al. (2017) Comprehensive molecular characterization of muscle-invasive bladder cancer. Cell 171(3):540–56.e25 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Lindskrog SV, Prip F, Lamy P et al. (2021) An integrated multi-omics analysis identifies prognostic molecular subtypes of non-muscle-invasive bladder cancer. Nat Commun 12(1):2301. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Hurst CD, Alder O, Platt FM et al. (2017) Genomic subtypes of non-invasive bladder cancer with distinct metabolic profile and female gender bias in KDM6A mutation frequency. Cancer Cell 32(5):701–15.e7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Hurst CD, Knowles MA (2022) Mutational landscape of non-muscle-invasive bladder cancer. Urol Oncol 40(7):295–303 [DOI] [PubMed] [Google Scholar]
  • 11.Damrauer JS, Hoadley KA, Chism DD et al. (2014) Intrinsic subtypes of high-grade bladder cancer reflect the hallmarks of breast cancer biology. Proc Natl Acad Sci U S A 111(8):3110–3115 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Loriot Y, Kamal M, Syx L et al. (2024) The genomic and transcriptomic landscape of metastastic urothelial cancer. Nat Commun 15(1):8603. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Damrauer JS, Beckabir W, Klomp J et al. (2022) Collaborative study from the Bladder Cancer Advocacy Network for the genomic analysis of metastatic urothelial cancer. Nat Commun 13(1):6658. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Van Allen EM, Mouw KW, Kim P et al. (2014) Somatic ERCC2 mutations correlate with cisplatin sensitivity in muscle-invasive urothelial carcinoma. Cancer Discov 4(10):1140–1153 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Liu D, Plimack ER, Hoffman-Censits J et al. (2016) Clinical validation of chemotherapy response biomarker ERCC2 in muscle-invasive urothelial bladder carcinoma. JAMA Oncol 2(8):1094–1096 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Taber A, Christensen E, Lamy P et al. (2020) Molecular correlates of cisplatin-based chemotherapy response in muscle invasive bladder cancer by integrated multi-omics analysis. Nat Commun 11(1):4858. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Aubrey BJ, Kelly GL, Janic A (2018) How does p53 induce apoptosis and how does this relate to p53-mediated tumour suppression? Cell Death Differ 25(1):104–113 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Campbell KJ, Tait SWG (2018) Targeting BCL-2 regulated apoptosis in cancer. Open Biol. 10.1098/rsob.180002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Zehir A, Benayed R, Shah RH et al. (2017) Mutational landscape of metastatic cancer revealed from prospective clinical sequencing of 10,000 patients. Nat Med 23(6):703–713 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.AACR Project GENIE Consortium (2017) AACR Project GENIE: powering precision medicine through an international consortium. Cancer Discov 7(8):818–831 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Pugh TJ, Bell JL, Bruce JP et al. (2022) AACR Project GENIE: 100,000 cases and beyond. Cancer Discov 12(9):2044–2057 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Cerami E, Gao J, Dogrusoz U et al. (2012) The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Discov 2(5):401–404 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Gao J, Aksoy BA, Dogrusoz U et al. (2013) Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal. Sci Signal 6(269):11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Sanchez-Vega F, Mina M, Armenia J et al. (2018) Oncogenic signaling pathways in the cancer genome atlas. Cell 173(2):321–37. e10 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Carneiro BA, El-Deiry WS (2020) Targeting apoptosis in cancer therapy. Nat Rev Clin Oncol 17(7):395–417 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Hanahan D (2022) Hallmarks of cancer: new dimensions. Cancer Discov 12(1):31–46 [DOI] [PubMed] [Google Scholar]
  • 27.Team RC. R: A language and environment for statistical computing. Available from: https://www.R-project.org
  • 28.Clinton TN, Chen Z, Wise H et al. (2022) Genomic heterogeneity as a barrier to precision oncology in urothelial cancer. Cell Rep 41(12):111859. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Helal C, Pobel C, Bayle A et al. (2023) Clinical utility of plasma ctDNA sequencing in metastatic urothelial cancer. Eur J Cancer 195(113368):113368. [DOI] [PubMed] [Google Scholar]
  • 30.Jia E, Agarwal G (2024) Clinical utility of liquid biopsy in bladder cancer: the beginning of a new era. J Liq Biopsy 6(100271):100271. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Guercio BJ, Whiting K, Shah RH et al. (2025) Circulating tumor DNA and response to cisplatin-based chemotherapy in patients with metastatic urothelial carcinoma enrolled in CALGB 90601 (Alliance). Eur Urol Open Sci 75:80–88 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Table S2
NIHMS2188977-supplement-Table_S2.xlsx (9.9KB, xlsx)
Table S1
NIHMS2188977-supplement-Table_S1.xlsx (120.9KB, xlsx)
Figure S1
NIHMS2188977-supplement-Figure_S1.pdf (225.4KB, pdf)

Data Availability Statement

The datasets analyzed in this study are publicly available from the AACR Project GENIE repository and cBioPortal

RESOURCES