Evaluation of BGJ398, a Fibroblast Growth Factor Receptor 1-3 Kinase Inhibitor, in Patients With Advanced Solid Tumors Harboring Genetic Alterations in Fibroblast Growth Factor Receptors: Results of a Global Phase I, Dose-Escalation and Dose-Expansion Study

Abstract

Purpose

This two-part, first-in-human study was initiated in patients with advanced solid tumors harboring genetic alterations in fibroblast growth factor receptors (FGFRs) to determine the maximum tolerated dose (MTD), the recommended phase II dose (RP2D), and the schedule, safety, pharmacokinetics, pharmacodynamics, and antitumor activity of oral BGJ398, a selective FGFR1-3 tyrosine kinase inhibitor.

Patients and Methods

Adult patients were treated with escalating dosages of BGJ398 5 to 150 mg once daily or 50 mg twice daily continuously in 28-day cycles. During expansion at the MTD, patients with FGFR1-amplified squamous cell non–small-cell lung cancer (sqNSCLC; arm 1) or other solid tumors with FGFR genetic alterations (mutations/amplifications/fusions) received BGJ398 daily on a continuous schedule (arm 2), or on a 3-weeks-on/1-week-off schedule (arm 3).

Results

Data in 132 patients from the escalation and expansion arms are reported (May 15, 2015, cutoff). The MTD, 125 mg daily, was determined on the basis of dose-limiting toxicities in four patients (100 mg, grade 3 aminotransferase elevations [n = 1]; 125 mg, hyperphosphatemia [n = 1]; 150 mg, grade 1 corneal toxicity [n = 1] and grade 3 aminotransferase elevations [n = 1]). Common adverse events in patients treated at the MTD (n = 57) included hyperphosphatemia (82.5%), constipation (50.9%), decreased appetite (45.6%), and stomatitis (45.6%). A similar safety profile was observed using the 3-weeks-on/1-week-off schedule (RP2D). However, adverse event–related dose adjustments/interruptions were less frequent with the 3-weeks-on/1-week-off (50.0%) versus the continuous (73.7%) schedule. Antitumor activity (seven partial responses [six confirmed]) was demonstrated with BGJ398 doses ≥ 100 mg in patients with FGFR1-amplified sqNSCLC and FGFR3-mutant bladder/urothelial cancer.

Conclusion

BGJ398 at the MTD/RP2D had a tolerable and manageable safety profile and showed antitumor activity in several tumor types, including FGFR1-amplified sqNSCLC and FGFR3-mutant bladder/urothelial cancers.

INTRODUCTION

The discovery of targetable genomic aberrations underlying cancer has led to impressive improvements in the treatment of a subset of patients with advanced cancers harboring such oncogenic drivers (eg, lung adenocarcinoma, melanoma).^1-3 However, novel personalized treatment strategies are missing in the vast majority of cancers.

Fibroblast growth factor receptor (FGFR) isoforms 1-4 and their 22 fibroblast growth factor ligands interact in a tissue-specific manner to initiate downstream signaling pathways that regulate cell proliferation, migration, differentiation, and survival—processes required for a variety of cell functions including angiogenesis and calcium/phosphate homeostasis.^4-7 Genomic alterations in FGFR1-3 (eg, gene amplifications, gain-of-function mutations, and chromosomal translocations) that trigger pathway activation^4,5,8 have been identified in bladder cancer,^9,10 squamous cell non–small-cell lung cancer (sqNSCLC),^11,12 squamous cell cancer of the head and neck,¹³ endometrial cancer,^8,14 cholangiocarcinoma,¹⁵ and breast cancer.^4,16-18

Although inhibitors targeting multiple receptor tyrosine kinases, including FGFR, are clinically active in several cancers,⁵ no FGFR-selective tyrosine kinase inhibitor (TKI) has been approved, and no TKI has been approved in a disease with a defined FGFR genetic alteration. BGJ398, an orally bioavailable, selective FGFR1 to 3 inhibitor (half maximal inhibitory concentration values range from 0.9 to 1.4 nM for FGFR1-3 to 60 nM for FGFR4),¹⁹ inhibits proliferation and tumor growth in preclinical cancer models bearing FGFR1-3 genetic alterations.^19,20

On the basis of these preclinical data, we conducted a global, personalized phase I single-agent study to determine the maximum tolerated dose (MTD), recommended phase II dose (RP2D), schedule, safety, pharmacokinetics (PK), pharmacodynamics (PD), and antitumor activity of BGJ398 in patients with solid tumors bearing FGFR alterations (ClinicalTrials.gov identifier NCT01004224).

PATIENTS AND METHODS

Patients

Adults with solid tumors harboring FGFR alterations (eg, amplification, mutation, fusion) for whom no effective standard therapy exists were enrolled.

Patient selection criteria relative to FGFR genetic alterations are defined in the Appendix (online only). FGFR genetic alterations not specified in the protocol were compared with those in public databases (eg, COSMIC and dbSNP) and were adjudicated to determine suitability for enrollment, allowing for continual review and enrollment of patients with newly reported FGFR alterations suggestive of potential sensitivity to FGFR inhibition (Appendix Table A1, online only).

Patients with measurable/evaluable disease per Response Evaluation Criteria in Solid Tumors (RECIST) v1.0²¹ and a WHO performance status ≤ 2²² were eligible (Appendix). Patients with prior FGFR inhibitor (except TKI258) or MEK inhibitor treatment, a history or evidence of endocrine alteration of calcium/phosphate homeostasis or ectopic calcification/mineralization, or evidence of corneal disorder/keratopathy were excluded. Concomitant therapies increasing calcium/phosphate serum levels were not permitted.

Patient screening was conducted in seven countries at 18 study sites, with protocol and amendment approvals granted by each institution’s review board/independent ethics committee. All applicable local regulations and principles of the Declaration of Helsinki were followed.²³ Patients provided written, informed consent before enrollment.

Study Design

The primary objective of this two-part phase I study was to determine the MTD, RP2D, and schedule of oral BGJ398. Secondary objectives included BGJ398 safety, tolerability, PK, PD, and antitumor activity. During dose escalation, sequential patient cohorts received BGJ398 5 (starting dose), 10, 20, 40, 60, 100, 125, and 150 mg once daily in continuous 28-day cycles (Fig 1). A twice daily 50-mg dose was also explored (after MTD/RP2D was established). At least three evaluable patients were treated in each cohort, and at least six patients were treated at the MTD/RP2D. During dose expansion, patients enrolled in one of three arms were treated at the MTD/RP2D: daily treatment of patients with FGFR1-amplified sqNSCLC (arm 1) or other FGFR-altered, advanced solid tumors (arm 2); and daily treatment on a 3-weeks-on/1-week-off schedule of patients with advanced solid tumors (excluding sqNSCLC; arm 3).

Fig 1.

Study scheme. DLT, dose-limiting toxicity; MTD, maximum tolerated dose; PK, pharmacokinetics; RP2D, recommended phase II dose. BID, twice daily; QD, once daily.

Dose-escalation decisions were guided by two adaptive Bayesian logistic regression models using the escalation-with-overdose-control principle, which estimated the rate of dose-limiting toxicities (DLTs), whether specific DLTs (Appendix) or adverse events (AEs)/laboratory abnormalities possibly related to BGJ398, resulting in failure to meet retreatment criteria. These models were reviewed, together with PK, PD, and safety assessments, and site investigator/personnel input to determine subsequent dose cohorts; intrapatient dose escalation was not allowed. The MTD was defined as the highest BGJ398 dose administered for ≥ 21 days resulting in DLTs in ≤ 33% of patients during cycle 1. The probability of the dose with excessive toxicity (> 33%) was < 25%. DLT characterization and MTD determination were based on data from the dose-determining set, including patients who received the planned dose for ≥ 21 days in cycle 1 and were evaluated for safety for ≥ 28 days after the first dose or who experienced a DLT during cycle 1.

Study Evaluations

Investigators assessed tumors at baseline and at every second treatment cycle until discontinuation (Appendix). Tumor evaluations were performed according to RECIST (version 1.0).²¹ A best overall response (BOR) (either a complete response or a partial response [PR]) required a response lasting ≥ 4 weeks. A BOR of stable disease (SD) required a tumor assessment demonstrating SD for ≥ 6 weeks after treatment initiation.

Safety was assessed using Common Terminology Criteria for Adverse Events (version 3),²⁴ with monitoring of AEs until 28 days after the final dose, and periodic physical examinations, laboratory evaluations, and electrocardiograms. FGFRs contribute to phosphate/vitamin D metabolism^6,7; thus, ophthalmologic examinations and assessments of calcium/phosphate homeostasis and renal function were performed (Appendix). PK and PD analyses are described in the Appendix.

Statistical Analyses

Descriptive statistics and/or contingency tables were used to summarize patient characteristics, efficacy and safety measurements, and PK. In a prespecified analysis, data from patients treated during dose escalation were used to determine the MTD/RP2D; subsequently, data were pooled with data from patients in the expansion phase who had the same dose and treatment schedule.

RESULTS

Patient Disposition and Baseline Characteristics

Data are reported from 132 heavily pretreated patients who started treatment between December 21, 2009, and April 7, 2015 (Table 1). At data cutoff (May 15, 2015), most patients (96.2%) had discontinued treatment, primarily because of disease progression (PD; 67.4%), AEs (12.9%), patient decision (10.6%), or death (3.8%; Appendix Table A2, online only).

Table 1.

Patient Demographics and Disease History

Safety

The median duration of BGJ398 exposure was 7.1 weeks (range, 0.6 to 101.0 weeks; N = 132). Treatment-emergent AEs (TEAEs) were reported in 131 patients (99.2%), with most (95.5%) experiencing at least one AE suspected to be treatment related. Hyperphosphatemia (74.2%), constipation (40.2%), and decreased appetite (40.2%) were the most commonly reported TEAEs across all doses (Table 2). The most common AE suspected to be treatment related was asymptomatic hyperphosphatemia (72.7%); other frequent treatment-related AEs included stomatitis (36.4%), decreased appetite (28.8%), diarrhea (27.3%), fatigue (25.0%), alopecia (23.5%), and nausea (22.0%). Grade 3/4 TEAEs were reported in 69 patients (52.3%). Forty-one patients (31.1%) experienced a grade 3/4 AE suspected to be treatment related (Table 2). Seventy-eight patients (59.1%), primarily at ≥ 100-mg doses, experienced an AE requiring dose adjustment or interruption. Eighteen patients (13.6%) had an AE leading to discontinuation (eye related in six patients [4.5%]). Fifteen patients died while receiving treatment or within 28 days of the last BGJ398 dose (Appendix Table A3, online only); 11 were attributed to PD or to related AEs. One death was suspected to be BGJ398 related (Appendix Table A3).

Table 2.

TEAEs Occurring in > 10% of All Patients by Treatment

DLTs, experienced by four of the 34 patients (11.8%) in the dose-determining set, included grade 3 increases in ALT/AST (n = 1 each at 100 and 150 mg), hyperphosphatemia (n = 1, 125 mg), and grade 1 corneal toxicity (n = 1, 150 mg). DLTs were reversible after BGJ398 interruption and/or concomitant medication. The BGJ398 MTD was determined as 125 mg once daily. Common AEs in patients treated at this dose and schedule (n = 57) included hyperphosphatemia (82.5%), constipation (50.9%), decreased appetite (45.6%), and stomatitis (45.6%). Dose adjustments/interruptions and AEs leading to discontinuations occurred in 73.7% and 10.5% of patients, respectively.

The DLT of hyperphosphatemia and the observation that most patients treated with doses ≥ 100 mg experienced AEs of hyperphosphatemia (Table 2) prompted the initiation of additional analyses to evaluate BGJ398 dose/schedule adjustment. Hyperphosphatemia was managed through dietary restrictions, phosphate-lowering therapy, and drug interruptions. Earlier data from 43 patients treated at 125 mg once daily revealed a median time to first dose interruption of 22 days and a median duration of interruption of 7 days. Considering these data and the properties of BGJ398 PK, an intermittent 3-weeks-on/1-week-off schedule of 125 mg once daily was introduced as a dose-expansion arm in a protocol amendment (Appendix Table A1).

The safety profile of BGJ398 125 mg in patients treated on the intermittent schedule was similar to that observed in patients treated continuously. Fewer patients experienced AEs requiring dose adjustment/interruption (50%), but the rate of AEs leading to discontinuation (12.5%) and the most common AEs were similar in the two schedules (Table 2). On the basis of these data, the RP2D for BGJ398 was chosen as 125 mg once daily administered in the 3-weeks-on/1-week-off schedule. With this regimen, the majority of patients achieved exposures above the threshold associated with preclinical evidence of FGFR pathway inhibition and in vivo efficacy (Fig 2).²⁰

Fig 2.

Pharmacokinetics of BGJ398. AUC, area under the plasma concentration-time curve; inf, infinity; QD, once daily; SD, standard deviation.

PK

BGJ398 mean plasma concentration time profiles and area under the plasma concentration-time curve by cohort, starting at dosages ≥ 20 mg once daily, are presented in Fig 2. Estimated PK parameters are reported in Appendix Table A4 (online only). Plasma concentrations for BGJ398 at dosages of 5 mg and 10 mg once daily were frequently below the lower limit of quantification (data not shown).

The median time to reach maximum plasma concentration after a single dose was approximately 2 to 3 hours. The mean area under the plasma concentration-time curve from 0 to 24 hours on day 1 increased by approximately nine-fold, from 20 to 150 mg once daily. Despite the relatively short median terminal elimination half-life on day 1 for these doses (range, 2.69 to 5.90 hours), accumulation was observed with dosing of ≥ 60 mg once daily; mean accumulation ratios ranged from 3.78 to 6.60 (day 15) and from 2.99 to 7.86 (day 28). After 125 mg once daily dosing, the unbound average steady-state BGJ398 concentration on day 28 of cycle 1 was 6.93 nM. Because dose interruptions occurred frequently after continuous once daily dosing, PK parameters on day 28 should be interpreted with caution.

Clinical Activity

Among all the patients treated with BGJ398 (N = 132), 42 had a BOR of SD, six achieved PR, and one achieved an unconfirmed PR (confirmatory scan performed 1 day early). Among the 85 patients treated at ≥ 100 mg with evaluable data, 28 (32.9%) had reduced tumor burden assessed as the best percentage change from baseline in the sum of the longest diameter in the target lesion or lesions (Fig 3).

Fig 3.

Waterfall plots of best change from baseline in the size of target lesions for patients treated at ≥ 100 mg BGJ398.

The disease control rate (DCR: PR + SD) in 36 patients with FGFR1-amplified sqNSCLC treated at doses of ≥ 100 mg continuously was 50%; four patients (11.1%) achieved PRs (100 mg [n = 1], 125 mg [n = 3, 2 confirmed]) and 14 patients had SD. In the subset of 31 patients treated at 125 mg continuously, 12 continued to receive treatment for > 8 weeks, and one half of these patients continued to receive therapy for ≥ 16 weeks (Fig 4).

Fig 4.

Response and duration of exposure in patients treated at BGJ398 doses ≥ 100 mg. Among the patients with other cancer types were five patients with cholangiocarcinoma (green bars). PR, partial response.

Twenty-seven of the 36 patients treated at ≥ 100 mg had pre- and post-treatment target lesion assessments; of these, 11 (40.7%) had reduced tumor burden (Fig 3; Appendix Table A5, online only). The responders remained on study for 39.9 to 76.6 weeks (confirmed PRs) and for 26.3 weeks (unconfirmed PR).

The DCR in eight patients with FGFR3-mutated bladder/urothelial cancer treated at doses ≥ 100 mg was 75%; three patients (37.5%) achieved PRs (125 mg continuously [n = 1], 125 mg 3-weeks-on/1-week-off [n = 2]) and three patients had SD. Five patients (62.5%) had reduced tumor burden and two of three with SD had reductions in tumor measurements nearing PR (27% to 28%; Fig 3). Of the six patients with disease control, the time receiving treatment ranged from 15.1 to ≥ 101 weeks, with one ongoing at data cutoff (Fig 4). One patient withdrew after 2 weeks; this patient had PD in an assessment performed 25 days after the last dose. Of note, one patient with FGFR1-amplified bladder cancer treated at 125 mg progressed rapidly in cycle 1 after 11 days of treatment.

Although no other PRs were observed at doses ≥ 100 mg, 10 of 32 patients (31%) with breast cancer treated at ≥ 100 mg had a best response of SD. Of 26 patients with breast cancer (FGFR1/2 amplified [n = 25]; FGFR3 mutant [n = 1]) with pre- and post-treatment target lesion measurements, four (15.4%) had reduced tumor burden. In addition, all three patients with FGFR2-altered (fusion [n = 2] or mutation [n = 1]) cholangiocarcinoma with pre- and post-treatment target lesion assessments had reduced tumor burden (Fig 3).

PD

BGJ398 treatment led to increased serum phosphate levels (Appendix Fig A1A, online only), as well as to dose- and exposure-related hyperphosphatemia, in most patients treated at doses ≥ 100 mg (Table 2). The median percentage change in FGF23 plasma levels ranged from approximately −25% to 80%, with a trend toward greater increases in patients treated at higher doses (Appendix Fig A1B).

DISCUSSION

This global first-in-human study of the FGFR1-3 inhibitor BGJ398 demonstrated a tolerable safety profile in patients with advanced solid tumors bearing FGFR amplifications, mutations, or fusions. The BGJ398 MTD/RP2D was determined to be 125 mg daily; the recommended schedule at the RP2D was 3 weeks on/1 week off on the basis of DLTs observed in four patients treated at doses ≥ 100 mg, BGJ398 PK, and safety data in patients treated on the intermittent schedule. This BGJ398 dose resulted in an average unbound steady-state concentration of 6.93 nM, which is similar to the half maximal inhibitory concentration of BGJ398 in the RT112 human tumor-derived FGFR gene fusion bladder cancer cell line (4 nM).²⁵ Commonly reported AEs included hyperphosphatemia, constipation, decreased appetite, and stomatitis. A similar safety profile was observed with the pan-FGFR inhibitor JNJ-42756493.²⁶

The FGFR pathway plays an important role in FGF23-mediated phosphate homeostasis.^6,7 Accordingly, hyperphosphatemia, the most common AE, was experienced by most patients receiving ≥ 100 mg BGJ398.²⁷ An intermittent dosing arm (125 mg once daily 3 weeks on/1 week off) was opened during dose expansion to manage hyperphosphatemia-related interruptions, resulting in fewer patients requiring an AE-related dose adjustment/interruption (50.0% v 73.7% [continuous]). Importantly, objective responses and reductions in tumor measurements were observed with both schedules, with similar overall safety profiles. Further exploration of BGJ398 dosing may be performed in the context of ongoing or future trials.

Because the FGF23/FGFR pathway mediates renal tubular phosphate secretion,⁷ FGF23 levels were used initially as a biomarker for FGFR pathway inhibition. Median FGF23 levels increased by ≤ 80% in patients treated with BGJ398, with generally higher levels at higher doses. Elevated serum phosphate levels, consistently observed with BGJ398 doses ≥ 100 mg, proved to be a sensitive, dose-dependent PD indicator of on-target activity and a key biomarker for FGFR pathway inhibition.

BGJ398 treatment provided disease control in 49 of 132 patients across all doses, with PRs observed in patients with FGFR1-amplified sqNSCLC (n = 4) and FGFR3-mutant bladder/urothelial cancer (n = 3). The clinical activity observed in FGFR1-amplified sqNSCLC (11% overall response rate, 50% DCR) is notable given the unfavorable prognosis for patients who relapse after chemotherapy.²⁸ Moreover, the duration on study for responding patients (6.1 to 17.7 months) is comparable to that achieved with nivolumab, an anti–PD-1 antibody approved in metastatic sqNSCLC after failure of platinum-based chemotherapy.^29-32 To date, no oncogenic driver-targeted drug has shown activity as a monotherapy, raising the possibility of a molecularly definable sqNSCLC subtype sensitive to FGFR inhibition. Despite preselecting for FGFR1 amplification, the response rate was lower than expected on the basis of preclinical data, suggesting that FGFR1 amplification may not function as a sole biomarker predicting clinical benefit.¹¹ Although the basis for this discrepancy (eg, tumor heterogeneity, different amplification parameters, and/or presence of required cofactors)^33,34 remains unknown, BGJ398 use as a personalized treatment approach in sqNSCLC warrants further investigation. Future whole genome analyses may help define the exploitable molecular differences between responders and nonresponders,³⁴ providing further insight into additional oncogenic drivers in sqNSCLC.

Responses observed in FGFR3-mutant bladder/urothelial cancer after failure of platinum-based chemotherapy (38% overall response rate, 75% DCR) strongly support a role for FGFR3 mutations as driver alterations in this molecular subgroup and the potent inhibitory function of BGJ398. In light of the < 1-year overall survival for patients with metastatic urothelial/bladder cancer after relapse after first-line chemotherapy,³⁵ BGJ398-targeted treatment warrants further investigation. To this end, a fourth expansion arm was opened to further evaluate BGJ398 activity in patients with urothelial cancer harboring an FGFR3 mutation or fusion.

SD with reduced tumor burden was also observed in patients with cholangiocarcinoma (FGFR2 fusions [n = 2], FGFR2 mutation [n = 1]) and FGFR1-amplified squamous head and neck cancer. The disease control observed in cholangiocarcinoma is notable given the limited treatment options available for patients who progress after chemotherapy.^36,37 Of interest, another patient with cholangiocarcinoma who was enrolled with a presumed FGFR3 mutation progressed rapidly and was later identified to be wild type for FGFR but as having a mutation in KRAS, a negative predictor (preclinically) for BGJ398 sensitivity.²⁰ A phase II study exploring BGJ398 as a second-line or later therapy in patients with FGFR2-altered advanced/metastatic cholangiocarcinoma is ongoing (ClinicalTrials.gov identifier NCT02150967).

The lack of objective responses and the limited disease control observed with BGJ398 in patients with breast cancer challenge the idea of FGFR amplification as a sole oncogenic driver in this disease; however, BGJ398 may prove more effective against advanced breast cancer when combined with other endocrine or targeted agents. Certain FGFR alterations (eg, FGFR3 mutations/gene fusions in bladder/urothelial carcinoma and FGFR2 gene fusions in cholangiocarcinoma) are dominant oncogenic drivers and confer sensitivity to BGJ398-mediated FGFR inhibition, whereas FGFR1 amplification, observed in a number of tumor types including sqNSCLC and breast cancer, may not be sufficient to identify a BGJ398-sensitive population. It is unclear whether alternative or additional biomarkers (eg, FGFR1 protein levels or FGFR pathway activity) would better predict responders, or whether FGFR signaling is less essential for tumor growth in breast cancer.

When this study was initiated in 2009, predictors of FGFR inhibitor sensitivity were limited. As the knowledge of FGFR biology and driver genetic alterations increased and assays for patient selection became available over the 6-year enrollment period, patient inclusion criteria were amended accordingly (Appendix Table A1). This study allowed the clinical evaluation of multiple preclinical hypotheses related to BGJ398-mediated FGFR pathway sensitivity and established which patient populations were likely to benefit from treatment with an FGFR-selective inhibitor.

Taken together, treatment with BGJ398 in patients with advanced solid tumors bearing FGFR alterations was tolerable, with manageable toxicity. The MTD/RP2D was determined to be 125 mg once daily on a 3-weeks-on/1-week-off schedule. BGJ398 demonstrated antitumor activity in FGFR1-amplified sqNSCLC, FGFR3-mutant bladder/urothelial cancer, and FGFR2-gene fusion/mutant cholangiocarcinoma, strongly supporting further biologic and clinical investigation. BGJ398 clinical development is ongoing, including adding a fourth expansion arm to this study for patients with urothelial carcinoma and FGFR3 mutation/gene fusion and a phase II study in cholangiocarcinoma with FGFR2 gene fusion/other FGFR genetic alterations.^38,39

Appendix

Methods

Molecular Prescreening.

Molecular prescreening to assess FGFR genetic alteration status was implemented after a protocol amendment and before enrollment of cohort 4 during the dose-escalation phase (BGJ398 40 mg once daily). Patient eligibility was determined on the basis of FGFR alteration status as assessed centrally or locally using fresh and/or archival tumor samples. Data on the number of patients prescreened versus the number of patients enrolled are not available, because this information was not reported by study sites that performed local assessment of eligibility.

Patients were eligible for enrollment if one of the following genetic alteration criteria was met: (1) FGFR1 or FGFR2 amplification was identified using fluorescence in situ hybridization (defined as a ratio of the respective FGFR to chromosome enumeration probe 8 [FGFR1] or chromosome enumeration probe 10 [FGFR2] of ≥ 2.2 or an average FGFR copy number of six or more signals/nucleus in ≥ 20 contiguous cells from two tumor areas); chromogenic or silver-enhanced in situ hybridization (defined as an average respective FGFR copy number of six or more signals/nucleus or a large gene cluster in ≥ 30% of tumor cells from ≥ 100 contiguous cells from two tumor areas); or quantitative polymerase chain reaction (defined as a respective FGFR copy number of at least six); (2) FGFR3 mutations were detected in exon 7 (R248C, S249C), exon 10 (G372C, A393E, Y375C), or exon 15 (K652M/T, K652E/Q); or (3) other FGFR genetic alterations, including gene fusions, were identified.

Alterations other than those described above were reviewed against public databases (eg, COSMIC and dbSNP), and enrollment was approved by study personnel for appropriate alterations (eg, to exclude known single-nucleotide polymorphisms and synonymous substitutions, or to include more recently identified somatic missense mutations). The intent of including unspecified FGFR alterations, such as fusions, in the prescreening evaluation was to adapt to the rapidly evolving understanding of the role of FGFR in various cancer types and to allow the enrollment of patients with previously unknown activating FGFR alterations (Appendix Table A1).

Additional Patient Enrollment Criteria.

Patients were required to have adequate bone marrow (absolute neutrophil count ≥ 1,500/µL [≥ 1.5 × 10⁹/L]; platelets ≥ 75,000/µL [≥ 75 × 10⁹/L]; and hemoglobin ≥ 10 mg/dL [≥ 100 g/L]) and renal and hepatic function (serum creatinine ≤ 1.5 × the upper limit of normal [ULN]; calculated or measured creatinine clearance > 75% lower limit of normal; and proteinuria grade ≤ 1, total bilirubin ≤ 1.5 × ULN, AST and ALT ≤ 2.5 × ULN, and serum albumin is greater than or equal to the lower limit of normal). Further requirements for inclusion included balanced calcium-phosphate homeostasis and adequate cardiovascular function, including a heart rate–corrected QT interval ≤ 470 ms and blood pressure within the normal range.

Patients with primary CNS tumors or CNS tumor involvement were not eligible for enrollment, unless clinically stable. Concomitant therapies prolonging the QT interval or associated with a risk of torsades de pointes were not permitted.

Study Design Rationale.

During dose escalation, sequential patient cohorts received BGJ398 5 (starting dose), 10, 20, 40, 60, 100, 125, and 150 mg once daily in continuous 28-day cycles. Fifty milligrams twice daily was also investigated after the maximum tolerated dose (MTD)/recommended phase II dose was established. Patients treated at 100 mg once daily had drug exposures approaching the predicted efficacious level (unbound BGJ398 concentration of 70 ng/mL) on the basis of preclinical data²⁰; however, a high degree of variability in exposure among patients was observed, and one patient experienced a dose-limiting toxicity of reversible grade 3 AST/ALT that resulted in dose interruption and modification during cycle 1. This variability in exposure was not observed at 60 mg once daily. These data supported further exploration of alternative dosing schedules for BGJ398, and the twice daily dose cohort was added in Amendment 5 (Jan 2012; Appendix Table A1). Enrollment in the 50 mg twice daily cohort began after the MTD was established, and was completed before patients were treated in expansion arms 1 and 2. The twice daily dosing schedule was not investigated further, because variability in exposure was also observed with this schedule, and suboptimal concentrations were achieved relative to predicted levels associated with preclinical efficacy.

Tumor Evaluations.

Tumors were evaluated, using chest and abdomen computed tomography and, in patients with a history of brain metastases, cranial computed tomography or brain magnetic resonance imaging, at baseline and within 14 days of every second treatment cycle until discontinuation.

Pharmacokinetic Assessments.

Serial blood samples were collected predose and up to 24 hours postdose on days 1, 15, and 28 of cycle 1. Samples were processed, and plasma was frozen at ≤ −60°C until analysis.

BGJ398 plasma concentrations were measured using a validated liquid chromatography–tandem mass spectrometry method with a 1.0 ng/mL lower limit of quantification. Pharmacokinetic (PK) parameters, including maximum plasma concentration, area under the plasma concentration-time curve, time to reach maximum plasma concentration, and half-life, were calculated using noncompartmental methods with Phoenix (Pharsight, Mountain View, CA). Descriptive statistics (mean, median, and standard deviation) were estimated for PK parameters in each cohort. Median values and ranges were provided for half-life and maximum plasma concentration.

Pharmacodynamic Assessments.

Blood samples for determination of fibroblast growth factor (FGF) 23 levels were collected predose on days 1, 2, and 15 of cycle 1 and day 1 on cycles 2 and 3, as well as 4 and 8 hours postdose on day 1. Fibroblast growth factor 23 levels were assessed using standard methods. Median percentage changes in fibroblast growth factor 23 levels from baseline were plotted by dose.

Specific Dose-Limiting Toxicities.

Specific dose-limiting toxicities modeled in the second adaptive Bayesian logistic regression model included corneal opacity, ectopic mineralization/calcification, or elevated serum creatinine (grade 1 for > 7 consecutive days) and serum ionized phosphorus (Pi) > 5.5 mg/dL and/or total serum calcium × Pi > 55 mg²/dL² despite Pi-lowering therapy for ≥ 14 days.

Hyperphosphatemia Grading, Management, and Prevention.

Because grading for hyperphosphatemia is not defined in the Common Terminology Criteria for Adverse Events (version 3) guidelines,²⁴ hyperphosphatemia was defined in this study as any value above the ULN, and severity was based on investigator discretion; severe hyperphosphatemia was considered grade 3. During dose escalation, treatment with phosphate binders was initiated when serum phosphate levels exceeded 5.5 mg/dL, with the goal of maintaining levels at ≤ 7 mg/dL. Prophylactic use of phosphate binders was instituted after the MTD determination and was used routinely in the dose expansion cohorts. Dose interruption was indicated when phosphate levels remained > 7 mg/dL despite the use of phosphate binders for ≥ 14 days.

Results

The BGJ398 125-mg dose level was determined to be the MTD on the basis of clinical safety and PK data and was supported quantitatively by the Bayesian logistic regression models used in this study. The 3-weeks-on/1-week-off intermittent schedule was investigated on the basis of the observation that the majority of patients being treated with the 125 mg once daily dosing schedule required dose interruptions by day 22 to control hyperphosphatemia, with interruptions lasting 7 days (median). The 3-weeks-on/1-week-off schedule was evaluated in an attempt to establish a regimen that allowed control of hyperphosphatemia over the course of the cycle. In addition, although exposure was variable, the majority of patients treated at 125 mg once daily achieved BGJ398 exposures above the threshold associated with preclinical evidence of fibroblast growth factor receptor pathway inhibition and in vivo efficacy.²⁰ Although the data are limited, this was not the case for patients treated at the 100 mg once daily dosage. It was recognized that many patients required dose reduction over the course of their therapy for the management of chronic toxicities; however, given the fact that patients with FGFR alterations are rare and are not easily identified, we chose to initiate treatment with a dose most likely to provide clinical benefit. Moreover, the rate of hyperphosphatemia and the overall safety profile between the 100-mg and 125-mg dose levels was not substantially different; therefore, 125 mg once daily 3 weeks on/1 week off was selected as the optimal BGJ398 dose and schedule for further exploration.

Fig A1.

Pharmacodynamics of BGJ398. Percentage change from baseline in (A) phosphate and (B) FGF23 plasma levels on cycle 1 day 15 by dose. Median values are shown as box plots. BID, twice daily; FGF, fibroblast growth factor; QD, once daily.

Table A1.

Study Protocol Amendments and Patient Molecular Screening

Table A2.

Patient Disposition

Table A3.

Deaths

Table A4.

BGJ398 Pharmacokinetic Parameters (once daily doses)

Table A5.

FGFR Mutations and Fusions

AUTHOR CONTRIBUTIONS

Conception and design: Jose Manuel Perez Garcia, Manuel Hidalgo, Jan H.M. Schellens, D. Ross Camidge, Howard A. Burris, Somesh Choudhury, Francois Ringeisen, Randi Isaacs, Reinhard Buettner, Jürgen Wolf

Administrative support: Martin Schuler

Provision of study materials or patients: Lucia Nogova; Lecia V. Sequist, Manuel Hidalgo, D. Ross Camidge, Martin Schuler, Ulka Vaishampayan, Howard A. Burris, G. Gary Tian, Geoffrey I. Shapiro, Jürgen Wolf

Collection and assembly of data: Lucia Nogova, Lecia V. Sequist, Jean-Pierre Delord, Jan H.M. Schellens, D. Ross Camidge, Martin Schuler, Ulka Vaishampayan, Howard A. Burris, G. Gary Tian, Mario Campone, Wan-Teck Lim, Patricia LoRusso, Geoffrey I. Shapiro, Katie Parker, Somesh Choudhury, Francois Ringeisen, Randi Isaacs, Reinhard Buettner, Jürgen Wolf

Data analysis and interpretation: Lucia Nogova, Lecia V. Sequist, Fabrice Andre, Manuel Hidalgo, Jan H.M. Schellens, Philippe A. Cassier, D. Ross Camidge, Ulka Vaishampayan, Howard A. Burris, Zev A. Wainberg, Wan-Teck Lim, Patricia LoRusso, Geoffrey I. Shapiro, Katie Parker, Xueying Chen, Somesh Choudhury, Francois Ringeisen, Diana Graus-Porta, Dale Porter, Randi Isaacs, Reinhard Buettner, Jürgen Wolf

Manuscript writing: All authors

Final approval of manuscript: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Evaluation of BGJ398, a Fibroblast Growth Factor Receptor 1-3 Kinase Inhibitor, in Patients With Advanced Solid Tumors Harboring Genetic Alterations in Fibroblast Growth Factor Receptors: Results of a Global Phase I, Dose-Escalation and Dose-Expansion Study

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jco/site/ifc.

Lucia Nogova

Honoraria: Pfizer

Consulting or Advisory Role: Boehringer Ingelheim, Roche

Research Funding: Pfizer

Travel, Accommodations, Expenses: Novartis, Pfizer

Lecia V. Sequist

Honoraria: AstraZeneca

Consulting or Advisory Role: Clovis Oncology, Novartis, Merrimack, AstraZeneca, ARIAD Pharmaceuticals, Genentech

Research Funding: Boehringer Ingelheim (Inst), Clovis Oncology (Inst), Genentech (Inst), Merrimack (Inst), ArQule (Inst), Novartis (Inst), AstraZeneca (Inst), Johnson & Johnson (Inst), Eli Lilly (Inst), Merck (Inst), Taiho Pharmaceutical (Inst)

Jose Manuel Perez Garcia

No relationship to disclose

Fabrice Andre

Research Funding: AstraZeneca (Inst), Novartis (Inst), Pfizer (Inst), Eli Lilly (Inst)

Jean-Pierre Delord

No relationship to disclose

Manuel Hidalgo

Stock or Other Ownership: Champions Oncology

Honoraria: Pfizer, Novartis, Merck Sharp & Dohme, Celgene

Consulting or Advisory Role: Novartis, Pfizer, Celgene, Merck, Eli Lilly/ImClone Systems

Research Funding: Pfizer, Celgene, Merck Serono

Travel, Accommodations, Expenses: Pfizer, Celgene, Merck

Jan H.M. Schellens

No relationship to disclose

Philippe A. Cassier

Honoraria: Novartis

Research Funding: Novartis (Inst)

Travel, Accommodations, Expenses: PharmaMar

D. Ross Camidge

Honoraria: Novartis

Martin Schuler

Honoraria: Alexion Pharmaceuticals, Boehringer Ingelheim, Celgene, Eli Lilly, Novartis, Pfizer, Bristol-Myers Squibb

Consulting or Advisory Role: AstraZeneca, Boehringer Ingelheim, Bristol-Myers Squibb, Celgene, Eli Lilly, Novartis

Research Funding: Boehringer Ingelheim (Inst), Bristol-Myers Squibb (Inst), Novartis (Inst)

Patents, Royalties, Other Intellectual Property: University of Duisburg-Essen

Travel, Accommodations, Expenses: Merck Sharp & Dohme

Ulka Vaishampayan

Honoraria: Astellas Pharma, Pfizer, Janssen Pharmaceuticals, Novartis, Bayer AG, Sanofi, Bristol-Myers Squibb

Consulting or Advisory Role: Pfizer, Novartis, Astellas Pharma, Sanofi, Bristol-Myers Squibb

Speakers’ Bureau: Pfizer, Novartis, Astellas Pharma, Bayer AG, Bristol-Myers Squibb

Research Funding: Astellas Pharma, Novartis, Exelixis, Pfizer

Howard A. Burris

Employment: HCA Healthcare/Sarah Cannon

Leadership: HCA Healthcare/Sarah Cannon

Stock and Other Ownership Interests: HCA Healthcare/Sarah Cannon

Consulting or Advisory Role: Mersana Therapeutics (Inst), AstraZeneca (Inst), FORMA Therapeutics (Inst), Janssen (Inst), Novartis (Inst), Roche/Genentech (Inst), TG Therapeutics (Inst), MedImmune (Inst), Bristol-Myers Squibb (Inst)

Research Funding: Roche/Genentech (Inst), Bristol-Myers Squibb (Inst), Incyte (Inst), Tarveda Therapeutics (Inst), Mersana Therapeutics (Inst), AstraZeneca (Inst), MedImmune (Inst), Macrogenics (Inst), Novartis (Inst), Boehringer Ingelheim (Inst), Eli Lilly (Inst), Seattle Genetics (Inst), AbbVie (Inst), Bayer (Inst), Celldex Therapeutics (Inst), Merck (Inst), Celgene (Inst), Agios Pharmaceuticals (Inst), Jounce Therapeutics (Inst), Moderna Therapeutics (Inst), CytomX Therapeutics (Inst), GlaxoSmithKline (Inst), Verastem (Inst), Tesaro (Inst), Immunocore (Inst), Takeda (Inst), Millennium (Inst), BioMed Valley Discoveries (Inst), Pfizer (Inst), PTC Therapeutics (Inst), TG Therapeutics (Inst), Loxo (Inst), Vertex (Inst), eFFECTOR Therapeutics (Inst), Janssen (Inst), Gilead Sciences (Inst), Valent Technologies (Inst), BioAtla (Inst), CicloMed (Inst), Harpoon Therapeutics (Inst), Jiangsu Hengrui Medicine (Inst), Revolution Medicines (Inst), Daiichi Sankyo (Inst), H3 Biomedicine (Inst), Neon Therapeutics (Inst), OncoMed (Inst), Regeneron (Inst), Sanofi (Inst)

Expert Testimony: Novartis

G. Gary Tian

Consulting or Advisory Role: Cardinal Health, Merrimack

Travel, Accommodations, Expenses: Cardinal Health

Mario Campone

Honoraria: Novartis, Servier, Menarini

Consulting or Advisory Role: Novartis, Servier, Menarini, Cellgen

Speakers’ Bureau: Novartis, Amgen

Research Funding: Novartis (Inst)

Travel, Accommodations, Expenses: Novartis

Zev A. Wainberg

Consulting or Advisory Role: Taiho Pharmaceutical, Sirtex Medical, Amgen

Speakers’ Bureau: Genentech

Research Funding: Novartis (Inst), Plexxikon (Inst), Pfizer (Inst), Biomarin (Inst), Merck (Inst)

Travel, Accommodations, Expenses: Genentech, Eli Lilly, Amgen

Wan-Teck Lim

Consulting or Advisory Role: Boehringer Ingelheim, Bristol-Myers Squibb, Novartis

Travel, Accommodations, Expenses: Boehringer Ingelheim, Roche

Patricia LoRusso

Consulting or Advisory Role: Astex Pharmaceuticals, Novartis, Astellas Pharma, Genentech, ARIAD Pharmaceuticals, Alexion Pharmaceuticals, Celgene

Research Funding: Genentech (Inst), Novartis (Inst), Merrimack (Inst), Immunogen (Inst), Tensha Therapeutics (Inst), Taiho Pharmaceutical (Inst), Roche (Inst), Incyte (Inst)

Geoffrey I. Shapiro

Consulting or Advisory Role: Vertex, Chugai Pharmaceutical, G1 Therapeutics, Eli Lilly, EMD Serono, Takeda Pharmaceuticals, Tesaro

Research Funding: Pfizer (Inst), Genentech (Inst), Bayer AG (Inst), Immune Design (Inst), Vertex (Inst), Millennium Pharmaceuticals (Inst), Puma Biotechnology (Inst), Tensha Therapeutics (Inst), Covidien (Inst), Novartis (Inst), Cellceutix (Inst), Sanofi (Inst), Cyclacel (Inst), Mirati Therapeutics (Inst), AstraZeneca (Inst), GlaxoSmithKline (Inst), Eli Lilly, Aileron (Inst), PharmaMar (Inst), PTC Therapeutics (Inst)

Katie Parker

Employment: Novartis

Stock or Other Ownership: Novartis (stock options)

Xueying Chen

Employment: Novartis

Somesh Choudhury

Employment: Novartis

Stock or Other Ownership: Novartis

Travel, Accommodations, Expenses: Novartis

Francois Ringeisen

Employment: Novartis

Stock or Other Ownership: Novartis

Diana Graus-Porta

Employment: Novartis

Stock or Other Ownership: Novartis

Dale Porter

Employment: Novartis

Stock or Other Ownership: Novartis

Patents, Royalties, Other Intellectual Property: Novartis

Randi Isaacs

Employment: Novartis

Stock or Other Ownership: Novartis

Patents, Royalties, Other Intellectual Property: Use patent for a particular drug in a specific indication

Reinhard Buettner

Stock or Other Ownership: Targos Molecular Pathology

Speakers’ Bureau: Roche, Boehringer Ingelheim, Novartis, Pfizer, Merck Sharp & Dohme, Merck, Bristol-Myers Squibb, QIAGEN, Eli Lilly

Jürgen Wolf

Honoraria: AstraZeneca, Bristol-Myers Squibb, Boehringer Ingelheim, Merck Sharp & Dohme, Clovis Oncology, Novartis, Pfizer, Roche

Consulting or Advisory Role: AstraZeneca, Bristol-Myers Squibb, Boehringer Ingelheim, Merck Sharp & Dohme, Clovis Oncology, Novartis, Pfizer, Roche

Research Funding: Novartis (Inst), Pfizer (Inst), Roche (Inst), Bristol-Myers Squibb (Inst)