Evaluation of the clinical cardiac safety of pemigatinib, a fibroblast growth factor receptor inhibitor, in participants with advanced malignancies

Abstract Pemigatinib is a potent inhibitor of the fibroblast growth factor receptor (FGFR) family of receptors that is approved for the treatment of cholangiocarcinoma with FGFR2 fusion or other rearrangements. Data from a first‐in‐human clinical study were used to assess the potential for pemigatinib to produce clinically significant effects on heart rate (HR) and cardiac repolarization (QTc). A central tendency analysis for electrocardiogram (ECG) outliers and a plasma concentration‐QTc analysis were conducted to assess cardiac safety in the first‐in‐human pemigatinib study (FIGHT‐101; NCT02393248). The study included 113 participants who received at least one dose of pemigatinib as monotherapy and had at least one pair of plasma pharmacokinetic (PK) and ECG data points collected. Timed 12‐lead ECGs were performed within 15 min of PK blood draws. The ECG parameters for each dose group in the study varied within expectations for patients with advanced malignancies. Categorical analysis of QT interval corrected for HR by Fridericia's method did not reveal dose dependence in the incidence of outliers, and the results of the central tendency and concentration‐QTc analyses did not suggest a dose‐ or concentration‐dependent drug effect. Least squares mean change from baseline in HR was small and did not indicate a clinically relevant effect on HR, and no effect was observed on cardiac conduction as assessed by PR and QRS intervals. In conclusion, pemigatinib does not exhibit any clinically significant prolongation of QTc or dose‐dependent changes in HR. Clinical trial registration: ClinicalTrials.gov NCT02393248.


Pemigatinib (INCB054828) is a potent and selective inhibitor of fi-
broblast growth factor receptor (FGFR)1, FGFR2, and FGFR3 1 that is under development for the treatment of malignant diseases or other diseases related to FGFR dysregulation. It is primarily metabolized by cytochrome P450 3A4 and exhibits low renal clearance and minimal renal excretion. 2,3 Pemigatinib was recently approved by the US Food and Drug Administration, the European Medicines Agency and the Japan Pharmaceuticals and Medical Devices Agency for the treatment of adults with previously treated, unresectable, locally advanced or metastatic cholangiocarcinoma with an FGFR2 gene fusion or other rearrangement. 2,4,5 The International Council for Harmonisation (ICH) E14 guidance provides recommendations for assessing the potential of non-antiarrhythmic drugs to delay cardiac repolarization. 6 In accordance with these guidelines, the cardiac safety of pemigatinib was investigated during the drug development process. In nonclinical studies, pemigatinib did not have an effect on cardiac repolarization in in vitro and in vivo assays (unpublished data on file, Incyte Corporation). Assessment of human ether-á-go-go channel interactions in clinical study participants showed a half maximal inhibitory concentration of >8 μM (the highest feasible concentration because of solubility), which is >300-fold higher than the free mean maximum plasma drug concentration at the therapeutic dose of 13.5 mg (236 nM [total] or 22.2 nM [free]), 3 suggesting low proarrhythmic activity. Based on these studies, pemigatinib is not expected to have a cardiac effect.
Concentration-response analysis from data acquired during early clinical development can be used to evaluate the potential for clinically relevant effects of a drug on QT prolongation. 7 This analysis of data from the first-in-human study FIGHT-101 (NCT02393248) describes the relationship between pemigatinib plasma concentration and QTc interval, based on data from study participants receiving pemigatinib monotherapy at doses of 1 to 20 mg once daily (QD).
The results of QT interval corrected for heart rate by Fridericia's method (QTcF) categorical and central tendency analyses, as well as the effect of pemigatinib on cardiac conduction (PR and QRS) and heart rate (HR), are also described.

| Ethics
The study protocols, protocol amendments, and consent forms were approved by qualified independent review boards/independent ethics committees before participant enrolment. All participants provided written informed consent before enrolling in the study. The study was conducted at 14 sites (13 in the United States and 1 in Denmark) in accordance with Good Clinical Practice guidelines and the ethical principles outlined in the Declaration of Helsinki.

| Study design and populations
FIGHT-101 (NCT02393248) was a first-in-human, open-label, doseescalation and dose-expansion study of the safety, tolerability, doselimiting toxicities, pharmacokinetics (PK), pharmacodynamics, and preliminary efficacy of pemigatinib in patients with advanced malignancies. The study population included male and female participants, 18 years of age or older, with any advanced solid tumor malignancy.
Participants self-administered doses of pemigatinib QD on either an intermittent schedule (2-weeks-on and 1-week-off) or a continuous administration regimen. For the purpose of analysis, the data for intermittent and continuous QD dosing of pemigatinib monotherapy treatments were combined, as the PK and electrocardiogram (ECG) assessments were collected prior to the first break in intermittent dosing (day 14).
The study was conducted in three parts. Part 1 was a monotherapy dose escalation of pemigatinib, in which doses of 1-20 mg QD (1, 2, 4, 6, 9, 13.5, and 20 mg) were administered on an intermittent or continuous schedule. Part 2 was a monotherapy dose expansion that evaluated doses that were selected based on evaluations in Part 1 (9, 13.5, and 20 mg) in participants with amplification, mutation, or translocation of FGFR 1, 2, or 3, or alteration of FGF 1-23. Part 3 involved dose-finding to determine the recommended phase 2 dose of pemigatinib in combination with chemotherapy. Data collected from participants in Parts 1 and 2 of the study are included in the analyses reported here.

What is already known about this subject
• Pemigatinib is a novel targeted therapy approved for treatment of cholangiocarcinoma with FGFR2 fusion/ rearrangements.
• It is important to understand the risk of cardiac toxicity (in particular, QTc prolongation) associated with nonantiarrhythmic drugs.
• Early-phase clinical studies are used to explore the potential for clinically relevant effects on QTc and heart rate.

What this study adds
• Pemigatinib was not associated with any clinically relevant effects on cardiac repolarization.
• There was no pemigatinib plasma concentrationdependent increase in the change from baseline QT interval corrected for heart rate by Fridericia's method. only on days 2, 8, 15, and 16. Participants fasted for 8 h before taking study drug in the clinic and fasted for 1 additional hour after. The plasma samples were assayed by a validated liquid chromatography-tandem mass spectrometry method with a linear validated range of 1 to 1000 nM in human plasma. Standard noncompartmental PK methods were used to analyze pemigatinib plasma concentration versus time data using Phoenix ® WinNonlin ® version 8.0 (Certara USA Inc.). ECGs were performed before and within 15 min of the PK blood draw at the corresponding time point. All 12-lead ECGs were recorded by centrally provided equipment and analyzed at the central ECG laboratory using a semiautomated technique (i.e., over-readings of ECG intervals). Machine-generated ECG interval readings were analyzed together with over-readings by the central laboratory for the ECG method bias analysis. ECG intervals were measured in a blinded manner by the core laboratory and the ECG database was locked before any statistical analysis was undertaken.

| Electrocardiogram assessments
The baseline ECG was defined as the average of all ECGs measured prior to the first administration of study drug on the first day of the first cycle of treatment.

| Time-matched mapping between pharmacokinetics and electrocardiogram
For the purpose of time-matched concentration-QTcF (C-QTcF) analysis, only the 1:1 data pairs of PK data and QTcF were included.
Mapping between ECG and PK was based on the actual date/time of PK and ECG assessments.

| Statistical methods
All statistical analyses were performed using the statistical software SAS ® version 9.4 (SAS Institute, Inc.).

| Categorical analysis
A participant or time point was determined as an outlier if the ECG intervals fell into the following categories (which were assessed separately): absolute QTcF values of >450 and ≤480 ms, >480 and ≤500 ms, or >500 ms; change from baseline (Δ)QTcF of >30 and ≤60 ms or ΔQTcF >60 ms; ΔPR >25% resulting in PR >200 ms; ΔQRS >25% resulting in QRS >120 ms; and HR changes reflecting either a >25% decrease from baseline to an HR of <50 bpm or a >25% increase from baseline to an HR of >100 bpm. All outliers were summarized for each treatment group on the basis of incidence rates. A participant was counted only once for a particular outlier event if the participant experienced more than one episode of that event.

| Central tendency analysis
The central tendency analysis for ΔQTcF (and other ΔECG intervals such as ΔHR, ΔPR, and ΔQRS) was based on a repeated measures model that had ΔQTc as the dependent variable, composite stage/ time points and dose as categorical fixed effects and baseline QTcF as continuous fixed effects covariates. A compound symmetry covariance matrix was specified for the repeated measures at the composite stage/time points for participants due to the convergence challenges encountered with the otherwise default unstructured covariance matrix. From this analysis, the least squares (LS) mean, standard error, and two-sided 90% confidence interval (CI) were calculated for each dose of pemigatinib and each stage/time point, separately.

| Concentration-QTc analysis
QTcF was defined as: for participant i at time point j, in which the RR interval is in the unit of seconds. Change from baseline QTcF for participant i at time j was derived as: where QTcF i0 stands for the baseline QTcF for participant i. The other ΔECG parameters, such as ΔHR, ΔPR, and ΔQRS, were derived in the same way.
The relationship between pemigatinib plasma concentrations and ∆QTcF was investigated using the prespecified linear mixedeffects modeling approach, 8 which can be written as: with random effects of participants on both intercept and concentration. It was assumed that the random effects are normally distributed as a bivariate normal random variable with mean [0, 0] and a 2 × 2 unstructured covariance matrix G, whereas the residuals are independent and identically normally distributed with mean 0 and variance 2 e . Model parameters were estimated using the restricted maximum likelihood approach. The degrees of freedom estimates were determined by the Kenward-Roger method. 9 An exploratory analysis was performed to assess key assumptions undertaken by the C-ΔQTcF linear model. Change from baseline HR was assessed with mean increases or decreases >10 bpm to be with the addition of a quadratic concentration term, was fitted and the quadratic term was tested at the two-sided 5% significance level.
The following goodness-of-fit plots were assessed for the C- The geometric mean of the steady-state peak plasma concentration (C max,ss ) values of pemigatinib for participants in each of the pemigatinib dose groups was retrieved from the PK analysis of FIGHT-101. 3 The predicted effect and its two-sided 90% CI for ΔQTcF at the geometric mean C max,ss were obtained for each pemigatinib dose separately. The mean and two-sided 90% CI for ΔQTcF were computed using the ESTIMATE statement per SAS ® PROC MIXED.

| Method bias sensitivity analysis
In this analysis, the automatic machine readings of ECGs were compared with the results of cardiologists' over-readings at the ECG central laboratory, as described in Ferber et al. 11 and Gong et al. 12

| Nomenclature of targets and ligands
Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guide topha rmaco logy. org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY, 13 and are permanently archived in the Concise Guide to PHARMACOLOGY 2021/22. 14

| Demographics
One hundred sixteen participants with advanced malignancies were enrolled in the first-in-human study and received at least one dose of pemigatinib as monotherapy. Of these, 113 participants enrolled in Part 1 or Part 2 of this study had at least one pair of PK-ECG data points collected (the ECG or PK/QTc population). Of the participants in this analysis population, 69 (61.1%) were women and the median age was 57 years (Table 1). In general, the ECG parameters at baseline in each dose group varied within expectations for small groups of patients with advanced malignancies. Mean absolute QTcF intervals at baseline ranged from 400 to 427 ms.

| QTcF categorical and central tendency analyses
Assessment of absolute QTcF and ∆QTcF outliers did not reveal dose dependence in the incidence of outliers (Table S1). Across all dose groups, no participant had absolute QTcF >500 ms. Two participants   The LS means of ∆HR were all within ±10 bpm with respect to the nominal stage/time points for all dose groups ( Figure S1B). These results suggest that pemigatinib had no drug effect on HR, and QTcF was the primary QTc endpoint.

| Effect on cardiac conduction
At the studied doses, pemigatinib did not have a relevant effect on

| Concentration-QTcF analysis
The clinical PK of pemigatinib in the first-in-human study have been reported, including mean plasma drug concentration-time profiles of pemigatinib for each dose group. 3 Pemigatinib is absorbed rap-

F I G U R E 2
Model-predicted ΔQTcF (mean and 90% confidence interval) overlaid with observed mean ΔQTcF (mean and 90% confidence interval) across deciles of pemigatinib plasma concentrations. ΔQTcF, change from baseline in QT interval corrected for heart rate by Fridericia's method  Figure S2). Across the observed range of pemigatinib concentrations on day 1 of the first cycle, the upper limit of the twosided 90% CI of the predicted ΔQTcF did not exceed 10 ms; thus, a large effect (i.e., >20 ms) on QTc interval can be excluded based on ICH E14 guidance. 7

| Electrocardiogram bias analysis
Comparison between fully automated machine readings of QTcF and the corresponding cardiologists' over-readings demonstrated minimal bias between the two ECG methods on the same set of ECG charts recorded from participants in Part 1 and Part 2 of the study.
To analyze bias by dose group, the low dose levels of 1, 2, 4, 6, and

| DISCUSS ION
Current ICH E14 guidelines support the use of ECG data from early clinical development to detect small changes in cardiac repolarization as an assessment of the potential proarrhythmic effects of a drug. 6,7 The criteria for demonstrating that a new drug does not cause ECG effects of clinical concern include careful collection of ECGs, data analysis of ECGs with matched plasma concentration (C-QTc) and sufficiently were well below the threshold for large QT effects (>20 ms).
In the C-QTcF analysis, a prespecified linear mixed-effects model was determined to be an appropriate fit to the data and was used to establish the relationship between plasma pemigatinib concentration and ΔQTcF. The estimated slope of the C-QTcF relationship was shallow and not significantly different from 0. Using this C-QTcF model, a large QT/QTc effect exceeding the threshold of concern (20 ms) could be excluded within the observed range of pemigatinib plasma concentrations (up to 549 nM). In a sensitivity analysis, the same exposure-response analysis was conducted using data from day 1 of the first cycle only, which accounts for approximately twothirds of total C-QTcF data. The sensitivity analysis yielded similar results, supporting a lack of large effect on QTc interval at the maximum therapeutic dose level.

ACK N OWLED G M ENTS
The authors wish to thank the study participants and their families, the investigators and the site personnel who conducted this study. Editorial assistance was provided by Envision Pharma Group, Inc.
(Philadelphia, PA, USA), and was funded by Incyte Corporation.

D I SCLOS U R E
X.G., T.J., X.L., X.C., and S.Y. are employees of Incyte Corporation and own stock in Incyte.

AUTH O R CO NTR I B UTI O N S
T.J., X.C., and S.Y were involved in research design. T.J. and X.C were involved in research performance. X.G., T.J., and X.L were involved in data analysis. All authors were involved in drafting the manuscript and critical revision of the manuscript.

E TH I C S A PPROVA L S TATE M E NT
The study protocols, protocol amendments, and consent forms were approved by qualified independent review boards/independent ethics committees before participant enrolment. The study was conducted in accordance with Good Clinical Practice guidelines and the ethical principles outlined in the Declaration of Helsinki.

PATI E NT CO N S E NT S TATE M E NT
Written informed consent was obtained from all individual participants involved in the study.

PE R M I SS I O N TO R E PRO D U CE M ATE R I A L FRO M OTH E R S O U RCE S
Not applicable.