940PcfDNA is an acceptable but insufficient means of characterizing FGFR3 mutation in patients with metastatic urothelial cancer (mUC)

Abstract Background Previous studies indicate that genomic alterations (GAs) in cell-free (cf)DNA are found in > 90% of patients (pts) with mUC [Agarwal et al Cancer 2018]. The ease of collection of cfDNA makes it an attractive alternative to tumor tissue-based screening, but the equivalency of cfDNA and tumor tissue for biomarker testing has yet to be defined in a prospective trial in mUC. We examine this in a phase Ib trial of infigratinib (BGJ398), a potent and selective FGFR1–3 inhibitor, in pts with mUC bearing FGFR3 alterations [Pal et al Cancer Discovery 2018]. Methods Eligible pts had mUC with activating FGFR3 mutations/fusions and prior platinum-based chemotherapy, unless contraindicated. Pts received infigratinib 125 mg orally daily (3 wks on/1 wk off). Comprehensive genomic profiling (CGP) was performed on FFPE tissues in a CLIA-certified lab (Foundation Medicine; Cambridge, MA), and blood was collected for cfDNA analysis using a 600-gene panel on an Illumina HiSeq 2500 sequencer. Patients were monitored for their best overall response rate (ORR) based on radiological imaging. Results 67 pts were enrolled; ORR was 25.4% and disease control rate was 64.2%. Blood for cfDNA was collected in 53 pts at the time of screening; collection occurred at a mean of 8.8 days (range 0–37 days) prior to therapy initiation. In contrast, tumor tissue was collected a mean of 520 days (range 28–2315 days) prior. GAs in cfDNA were found in all pts; 38 pts (68%) had detectable FGFR3 mutation. FGFR3 alterations were concordant in 30/38 (79%) of tumors with both tumor tissue and cfDNA at screening. Other commonly encountered alterations were found in KDM6A (30%), KMT2D (20%) and TP53 (20%). In this cohort, all 11 pts with progressive disease (PD) as a best response had detectable FGFR3 in cfDNA at screening, while this was found in only 7/12 responders (58%). Conclusions cfDNA identified FGFR3 mutations in 79% of pts whose mutations were previously identified in tumor tissue, suggesting that cfDNA is a secondary screening option for trials assessing FGFR3-directed therapies. The paradoxically higher rate of PD in pts with detectable FGFR3 mutations in cfDNA warrants further study. Clinical trial identification NCT01004224. Editorial acknowledgement Lee Miller (Miller Medical Communications Ltd). Legal entity responsible for the study QED Therapeutics. Funding QED Therapeutics. Disclosure S.K. Pal: Honoraria (self), Advisory / Consultancy: Novartis; Honoraria (self), Re