Detecting FGFR3 genomicalterations in liquid biopsies from patients with clinically advanced urothelial bladder cancer.

830 Background: Fibroblast growth factor receptor 3 (FGFR3) pathogenic gain of function genomic alterations including activating short variant mutations, rearrangements/fusions and amplifications (FGFR2/FGFR3mut+) are associated with the development and progression of UBC and have emerged as major therapy targets for patients with clinically advanced urothelial bladder cancer (CAUBC). Erdafitinib is FDA-approved pan–FGFR inhibitor for metastatic UC harboring susceptible alterations in FGFR2/FGFR3 e. g. activating mutation or fusion. Conventional genomic testing for FGFR3 typically involves tumor tissue; however, circulating tumor DNA (ctDNA) can be a relevant biospecimen for such testing. We queried whether the tumor fraction (TF) calculation would impact the ability of FGFR3 GA to be detected on liquid biopsies (LBx). Methods: Hybrid capture based comprehensive genomic profiling (CGP) was performed on 10, 531 CAUBC TBx using the FoundationOneCDx assay and on 1, 637 CAUBC LBx using the FoundationOneLiquid CDx assay. The ctDNA tumor fraction (TF) for each LBx sample was determined using assessments of aneuploidy and variant allele frequencies, as previously described. Results: For the TBx group, 1, 917 (18. 2%) CAUBC were FGFR3 mut +. For the LBx group, FGFR3 mut+ detection was increased as the TF increased reaching a peak of 24. 6% when the TF was ≥10% (Table). For CAUBC LBx samples with TF 24% when the TF was ≥ 30%. When the TF was ≥1. 0% the frequency of FGFR3 mut+ was only slightly higher in LBx samples than TBx samples. Conclusions: LBx emerges as a sensitive method for detecting FGFR3 mut+ in CAUBC; but its sensitivity is dependent on TF. Importantly, LBx TF may play a major role in CGP evaluation, as FGFR3mut+ rates using LBx appear to be as frequently detected in LBx as in TBx samples, especially when the LBx TF is ≥1%. However, when TF levels are < 1%, FGFR3 mut+ status may be missed, therefore FGFR3 testing cannot rely on LBx in such cases. Limitations include the retrospective nature, lack of clinical data annotation, possible selection, confounding biases, and lack of matched TBx and LBx samples. Our findings have the potential to increase LBx testing, especially in cases with insufficient tumor tissue available for CGP, and raise the question whether LBx may be complementary to TBx. CAUBC FGFR3 wt FGFR3 mut+ FGFR3 mutfreq TBx 8614 1917 18. 2% LBx TF ≥0% 534 90 14. 4% LBx TF ≥1. 0% 254 68 21. 1% LBx TF ≥5. 0% 174 50 22. 3% LBx TF ≥10. 0% 132 43 24. 6% LBx TF ≥20. 0% 89 29 24. 6% LBx TF ≥30. 0% 63 19 23. 2% LBx TF ≥40. 0% 43 10 18. 9% LBx TF ≥50. 0% 32 7 17. 9%