FGFR2 fusion detection in plasma: A new era in the clinical monitoring of iCCA.

4085 Background: Actionable genetic alterations can be identified in over 50% of intrahepatic cholangiocarcinoma (iCCA). Since EMA approved Pemigatinib, a selective Fibroblast growth factor receptor 1-3 (FGFR1-3) inhibitor for the treatment of CCA with FGFR2 fusions or rearrangements, the screening of patients who may benefit from such targeted therapies is especially relevant. In addition, novel FGFR inhibitors that are effective for treatment of resistant mutant clones are under development. In patients with no available tissue for genomic profiling or at different timepoints after targeted therapy, NGS testing of circulating cell-free DNA (ccfDNA) would be the most convenient option. Hence, two important issues must be addressed: i) technical set-up and validation of detection of FGFR2 rearrangements in plasma, and ii) patient shedding in iCCA, in order to consider liquid biopsy for routine clinical care. Methods: We conducted a retrospective study in a cohort of 18 iCCA patients with known FGFR2 fusion or rearrangement events previously identified in tumor FFPE by NGS (FoundationOne CDx test). A custom-designed capture-based NGS panel for use in plasma or tissue (VHIO-iCCA test) was developed to detect FGFR2 rearrangements and other common altered genes in iCCA. After validating our VHIO-iCCA panel with fusion positive FFPE samples, a concordance study was conducted to evaluate the detection of FGFR2 fusion and rearrangements in matched-to-tissue timepoint plasmas. Finally, additional serial plasma samples taken during FGFR inhibitor treatment were also analyzed. Results: From the 18 rearrangement events previously detected with FoundationOne CDx, we were able to identify all 18 in FFPE samples and 15 in the paired plasma using our VHIO-iCCA panel, representing a 100% and 83% concordance, respectively. In the 3 discordant cases, no additional alterations were detected in plasma, indicating a lack of circulating tumor DNA (ctDNA) shedding. Serial sampling of these patients indicated persistent non-shedding. The analysis of fusion allele fraction (FAF) in serial plasma samples revealed that detection of fusion FGFR2 changed during treatment and correlated with best response. In general, patients with a stable FAF showed a SD, while patients which reduced FAF presented a PR. Conclusions: This extremely valuable set of cases has allowed us to validate our VHIO-iCCA panel to be used in tissue and plasma, and to determine that the sensitivity in plasma is >80%, making this a feasible option to avoid tissue biopsies, whenever patients cannot undergo the procedure and even to aid in cancer monitoring. Patient shedding is high in iCCA, yet a fraction of patients may not find a useful resource in liquid biopsy. For those who shed ctDNA, monitoring through the FAF may guide clinical management of iCCA.