Background: Patients (pts) with advanced cholangiocarcinoma (CCA) who progressed on or after 1L chemotherapy have few treatment options. FGFR2 fusions or rearrangements occur in 10–16% of pts in global and 6.14% in Chinese pts with intrahepatic cholangiocarcinoma (iCCA), and may predict tumor susceptibility to FGFR inhibitors. HMPL-453 is a novel, selective tyrosine kinase inhibitor against FGFR1, 2, and 3. Here, we report results from an open-label, multi-cohort, single-arm phase 2 study (NCT04353375). Methods: Pts with histologically or cytologically confirmed locally advanced or metastatic iCCA with FGFR2 fusions who were treated at least one previous line of systemic therapy were eligible and included in this analysis. They received 21-day cycles of HMPL-453 orally with 150 mg (QD, cohort 1) or 300 mg (QD, 2w on/1w off, cohort 2) until disease progression or unacceptable toxicity. The primary endpoint was objective response rate (ORR) per RECIST 1.1. Results: At data cutoff (Sep 21, 2022), a total of 25 iCCA pts (12 in cohort 1, 13 in cohort 2) were enrolled and treated with HMPL-453. Median age was 51 yrs (range 28-76) and 12 (48%) were male. All pts had received ≥1 prior treatment line and 21 (84%) had received previous gemcitabine-based chemotherapy. The median follow-up duration was 12 months (mos) and 4.1 mos for cohort 1 and cohort 2, respectively. Of 22 evaluable pts (12 in cohort 1 and 10 in cohort 2) who had at least one post-baseline tumor assessment, the best overall response by investigator assessment was confirmed partial response in seven (31.8%) pts and stable disease in an additional 12 pts (54.5%), resulting in a disease control rate (DCR) of 86.4%. Specifically, in cohort 2 the confirmed ORR was 50% (95% CI, 18.7%–81.3%) and DCR was 90% (95% CI, 55.5%–99.7%). Duration of response was not reached in either cohort. Median progression-free survival was 5.7 (95% CI, 2.6, NR) mos in cohort 1 and not yet mature in cohort 2. Of all pts, 23 (92%) experienced ≥1 treatment-related adverse events (TRAEs). The most common TRAEs (any grade) were diarrhea (56%), dry mouth (44%), and blood phosphorus increased (44%). The common Gr ≥3 TRAEs (≥5% pts) were decreased neutrophil count (8%), nail toxicity (8%) and palmar-plantar erythrodysesthesia syndrome (8%). Compared to cohort 1, pts in cohort 2 experienced a lower incidence of Gr ≥3 TRAEs (23.1% vs 58.3%), less dose interruption/reduction (32.0% vs 41.7%) and less treatment discontinuation (0.0% vs 16.7%) due to TRAE. Based on the better safety profile and preliminary efficacy, 300mg, QD, 2w on/1w off was chosen as the recommended phase 2 dose (RP2D). Conclusions: HMPL-453 showed promising efficacy, particularly in RP2D regimen (300 mg, QD, 2w on/1w off) and acceptable toxicity in pts with previously-treated advanced iCCA and FGFR fusions. These results warrant further study in pts with advanced iCCA. Clinical trial information: NCT04353375.