BALVERSA - ANNAR Study

SUMMARY

• ANNAR (NCT03955913) is a prospective, non-interventional study to identify adult patients with urothelial cancer (UC) and selected fibroblast growth factor receptor (FGFR) aberrations through molecular testing of their archival tumor tissue and to assess eligibility status for BALVERSA studies. The primary data source for this study will be the patients’ medical records. The study has enrolled 3679 patients from 199 study sites.¹ For complete study details, refer to https://clinicaltrials.gov/study/NCT03955913.
• Matsubara et al (2025)² reported an analysis from the global, prospective, noninterventional ANNAR biomarker study, to assess valid FGFR test results, reasons for test failure, and the prevalence of FGFR alterations (FGFRalt) in patients with metastatic urothelial cancer (mUC) and non-muscle-invasive bladder cancer (NMIBC).
  • Overall, the proportion of valid FGFR test results (positive or negative) was significantly higher in mUC than in NMIBC (86% vs 66%; P<0.001).
  • Among the samples that underwent molecular testing for FGFRalt, the most common causes of FGFR test failure were low tumor content, insufficient RNA (P<0.001), and failure of run controls (P<0.001).
  • Among samples with FGFRalt, FGFR mutations were observed in 83% vs 92% of mUC vs NMIBC samples, whereas fusions were less common (15% vs 6.3%, respectively).

CLINICAL DATA

Matsubara et al (2025)² reported an analysis from the global, prospective, noninterventional ANNAR biomarker study (NCT03955913, N=3757), to assess valid FGFR test results, reasons for test failure, and the prevalence of FGFRalt in patients with mUC (n=2706) and NMIBC (n=962).

Study Design/Methods

• Global, prospective, multicenter, noninterventional biomarker study (ANNAR).
• Archival formalin-fixed paraffin-embedded tumor tissue from patients with mUC or NMIBC was obtained for molecular testing.
• FGFRalt were assessed centrally using the QIAGEN therascreen FGFR real-time polymerase chain reaction (RT-PCR) assay.
• Primary data sources included patient medical records and central laboratory results; no additional interventions or safety assessments were performed.
• Inclusion criteria: age >18 years; a confirmed diagnosis of transitional UC; a diagnosis of one of the following: metastatic or surgically unresectable UC stage IV), localized surgically resectable or resected UC with a T classification of T2 or above, or non-muscle-invasive UC of the bladder (Ta, T1, and carcinoma in situ); and having archival tumor tissue samples available for an FGFR testing.
• Endpoints: proportion of valid FGFR test results (positive or negative), reasons for FGFR test failure, and proportion of samples with FGFRalt.
  • Reasons for FGFR test failure included failure of run controls, low tumor content, and insufficient RNA (determined based on the range predefined in the central laboratory).

Results

Patient Characteristics

• A total of 3757 patients with mUC or NMIBC were enrolled.
• Samples from 2706 patients with mUC and 962 patients with NMIBC underwent molecular testing for FGFRalt.
• Patient characteristics with valid FGFR test results are shown in Table: Characteristics of Patients and Archival Tumor Samples with Valid FGFR Test Results.

FGFR Test Results

• Overall, the proportion of valid FGFR test results (positive or negative) was significantly higher in mUC than in NMIBC (86% vs 66%; P<0.001).
• Valid FGFR test results were influenced by sample-related factors, including archival sample age, sample source, and tumor location as shown in Table: Factors Influencing Valid FGFR Test Results.

• Among the samples that underwent molecular testing for FGFRalt, the most common causes of FGFR test failure were low tumor content (mUC [n=2706] vs NMIBC [n=962]: 3.1% vs 3.7%, P=0.4), insufficient RNA (3.7% vs 8.4%, P<0.001), failure of run controls (5.7% vs 21%; P<0.001), and others (1.1% vs 1.4%, P=0.6).

Prevalence of FGFRalt

• FGFRalt was detected in 431 (19%) mUC samples and 333 (53%) NMIBC samples.
• The frequency of FGFRalt was consistent across subgroups, except age in samples from patients with NMIBC (<75 years: 55%; ≥75 years: 46%), as confirmed by multivariable analysis (Table: Multivariable Analysis of Predictors of FGFR Alterations).

• Among samples with FGFRalt, FGFR mutations were observed in 83% vs 92% of mUC vs NMIBC samples, whereas fusions were less common (15% vs 6.3%, respectively).
  • The most frequently identified alterations included FGFR3 mutations (S249C and Y373C), and FGFR3-TACC3 was the most common fusion. Details of FGFRalt are included in Table: FGFR Alterations in mUC and NMIBC.
  • FGFR3-TACC3_V3 fusion, other fusions, and other FGFR mutations/fusions were captured but numeric data was not reported.

• Multivariable analysis showed no specific signal for predictors of FGFR mutation vs fusion in mUC and NMIBC.

Literature Search

A literature search of MEDLINE^®, Embase^®, BIOSIS Previews^®, and Derwent Drug File (and/or other resources, including internal/external databases) conducted on 08 June 2026.