Fibroblast growth factor receptor inhibition for succinate dehydrogenase-deficient gastrointestinal stromal tumors: a phase 2 trial

Inclusion and ethics

All patients provided written informed consent before enrollment. This protocol was conducted according to Good Clinical Practice and was consistent with the principles of the Declaration of Helsinki. The protocol, informed consent and subsequent amendments were reviewed and approved by the Central Institutional Review Board (IRB) of the National Cancer Institute (NCI IRB no. 3—Adult CIRB—Early Phase Emphasis, ID IRB00009430), with additional approvals at each site as required by local requirements. A complete list of IRB names and identifiers for all sites is available in the Supplementary Information. A data safety monitoring board was not involved in oversight of the trial. Trial monitoring and auditing was per NCI-Cancer Therapy Evaluation Program UM1 standard practice with Theradex. The principal investigator provided study oversight over all aspects of trial conduct.

Eligibility criteria

Inclusion criteria were:

1. (1)

   Patients must have locally advanced or metastatic disease that is not amenable to surgery.

2. (2)

   Patients must have measurable disease.

3. (3)

   Participant must be willing to undergo pre-treatment biopsy if disease site is amenable to biopsy and low risk for the biopsy procedure. If biopsy is not possible, eligibility may be approved after discussion with the study chair.

4. (4)

   Age ≥18 years.

5. (5)

   ECOG performance status ≤2.

6. (6)

   Patients must have adequate organ and marrow function as defined below:

   • Hemoglobin ≥ 8.0 g dl⁻¹

   • Absolute neutrophil count ≥1,000 per μl

   • Platelets ≥100,000 per μl

   • Total bilirubin ≤1.5× institutional upper limit of normal (ULN)

   • AST(serum glutamic-oxaloacetic transaminase (SGOT))/ALT(serum glutamic-pyruvic transaminase (SGPT)) ≤ 3.0× institutional ULN (unless liver metastases are present in which case it must be ≤5× ULN)

   • Glomerular filtration rate ≥60 ml per min per 1.73 m².

7. (7)

   Human immunodeficiency virus-infected patients on effective non-CYP3A4 interacting anti-retroviral therapy with undetectable viral load within 6 months are eligible for this trial.

8. (8)

   For patients with evidence of chronic hepatitis B virus infection, the hepatitis B virus viral load must be undetectable on suppressive therapy, if indicated.

9. (9)

   Patients with a history of hepatitis C virus (HCV) infection must have been treated and cured. For patients with HCV infection who are currently on treatment, they are eligible if they have an undetectable HCV viral load.

10. (10)

    Patients with treated brain metastases are eligible if follow-up brain imaging after central nervous system-directed therapy shows no evidence of progression.

11. (11)

    Patients must be disease-free of prior invasive malignancies for >5 years with the exception of curatively treated basal cell or squamous cell carcinoma of the skin or carcinoma in situ of the cervix. Note that if there is a history of prior malignancy, patients must not be receiving other specific treatment for that cancer.

12. (12)

    Patients should have completed prior treatment for their cancer: chemotherapy or radiotherapy must have been completed for >2 weeks (6 weeks for nitrosoureas or mitomycin C) before entering the study.

13. (13)

    Patients should have recovered from AEs due to prior anti-cancer therapy (that is, have residual toxicities > Grade 1) with the exception of alopecia.

14. (14)

    Patients with known history or current symptoms of cardiac disease, or history of treatment with cardiotoxic agents, should have a clinical risk assessment of cardiac function using the New York Heart Association Functional Classification. To be eligible for this trial, patients should be class 2B or better.

15. (15)

    Patients must have a corrected QT (QTc) interval length <450 ms.

16. (16)

    Participant is willing to comply with the protocol for the duration of the study including undergoing treatment and scheduled visits and examinations including follow-up.

17. (17)

    Participant must be able to swallow and maintain pills.

18. (18)

    Women of childbearing potential must have a negative urine or serum pregnancy test within 28 days of initial dose of rogaratinib (BAY 1163877), and again within 7 days before treatment on day 1. If screening occurs within 7 days of day 1, only one pregnancy test is required.

19. (19)

    The effects of rogaratinib (BAY 1163877) on the developing human fetus are unknown. For this reason and because kinase inhibitor agents are known to be teratogenic, women of childbearing potential and men must agree to use adequate contraception (hormonal or barrier method of birth control; abstinence) before study entry, for the duration of study participation, and 4 months after completion of rogaratinib (BAY 1163877). Should a woman become pregnant or suspect she is pregnant while she or her partner is participating in this study, she should inform her treating physician immediately. Men treated or enrolled on this protocol must also agree to use adequate contraception before the study, for the duration of study participation, and 4 months after completion of rogaratinib (BAY 1163877) administration.

20. (20)

    Ability to understand and the willingness to sign a written informed consent document. Participants with impaired decision-making capacity who have a legally authorized representative and/or family member available will also be eligible.

Exclusion criteria included:

1. (1)

   Patients who are receiving any other investigational agents.

2. (2)

   History of allergic reactions attributed to compounds of similar chemical or biologic composition to rogaratinib (BAY 1163877).

3. (3)

   Concomitant administration with sensitive substrates or narrow therapeutic index drugs of CYP3A4, P-glycoprotein BCRP, MATE1 and MATE2K, and strong inhibitors and inducers of CYP3A4 should be avoided. Use caution with strong inhibitors and inducers of P-glycoprotein. Because the lists of these agents are constantly changing, it is important to regularly consult a frequently updated medical reference. As part of the enrollment and informed consent procedures, the patient will be counseled on the risk of interactions with other agents, and what to do if new medications need to be prescribed or if the patient is considering a new over-the-counter medicine or herbal product.

4. (4)

   Concomitant administration of medications that prolong the QT/QTc interval is prohibited in accordance with the published US Food and Drug Administration guidance ‘E14 Clinical Evaluation of QT/QTc Interval Prolongation and Proarrhythmic Potential for Non-Antiarrhythmic Drugs’.

5. (5)

   Patients with disturbed calcium and/or phosphate metabolism are excluded from this study.

6. (6)

   Patients with uncontrolled intercurrent illness.

7. (7)

   Patients with psychiatric illness or social situations that would limit compliance with study requirements.

8. (8)

   Pregnant women are excluded from this study because rogaratinib (BAY 1163877) is kinase inhibitor agent with the potential for teratogenic or abortifacient effects. Because there is an unknown but potential risk for AEs in nursing infants secondary to treatment of the mother with rogaratinib (BAY 1163877), breastfeeding should be discontinued if the mother is treated with rogaratinib (BAY 1163877).

Study design, treatment and endpoints

This is a multicenter, open-label, two-stage phase 2 study to evaluate the efficacy and safety of rogaratinib in patients with sarcoma. Two cohorts were defined: cohort A, patients with advanced sarcoma that harbored an alteration in FGFR1–4; and cohort B, patients with SDHd GIST (Clinicaltrials.gov identifier: NCT04595747). The cohorts were enrolled in parallel with separate analysis. This paper reports cohort B the SDHd GIST cohort. The trial was conducted at 11 centers in the USA participating in the NCI-Cancer Therapy Evaluation Program Experimental Therapeutics Clinical Trials Network (ETCTN). Patients received rogaratinib (BAY 1163877) 800 mg by mouth twice daily continuously in 28-day cycles.

The primary objective was to estimate objective response rate. Secondary objectives were to estimate PFS, and safety and tolerability. Exploratory objectives were to evaluate serial measurements of FGF3 and FGF4 and FGFR in serial biopsies, to perform whole-exome sequencing in serial biopsies, and to explore rogaratinib with pharmacodynamic effects.

Dose reductions were permitted by one dose level to 600 mg twice daily or up to two dose levels to 400 mg twice daily based on toxicities. Patients unable to tolerate the lowest dose level or who required a dose interruption of more than 21 days due to toxicities were taken off protocol therapy.

Patients were followed for 30 days after removal from study until disease progression, unacceptable AE or change in clinical condition, patient noncompliance or withdrawal from the trial. Patients removed from study for unacceptable AE(s) were followed until resolution or stabilization of the AE.

The primary endpoint was radiographic response by RECIST 1.1 to single-agent rogaratinib. Tumor assessments were performed using either computed tomography scan or magnetic resonance imaging at baseline, then every 8 weeks for the first year, then every 12 weeks thereafter. Confirmatory scans were required 8 weeks (at a minimum of 4 weeks) following first documentation of objective response by RECIST 1.1.

Secondary endpoints were PFS and safety and tolerability. AEs were graded using the NCI Common Terminology Criteria for Adverse Events v.5.0.

Exploratory objectives were to evaluate the relationship between genomic, transcriptomic, proteomic or epigenomic features of archival, pre-treatment and any post-progression samples to outcomes.

Across our studies, sex was not considered to be a biological variable impacting SDHd GIST cancer. The study design did not limit participants based on sex. Both male and female participants were included. Sex was reported by the study team into the central database. Data were not disaggregated based on sex and gender. No sex- or gender-based analyses were planned a priori.

Statistical design

The trial had a Simon two-stage optimal design, targeting at least a 20% improvement in objective response rate, from a historical control rate of 5% to 25% or higher, with parameters including a one-sided type 1 error of 10%, and a power of 90%. The two-stage structure required one or more responses in the first nine patients to proceed to the second stage, serving as an early futility checkpoint to halt the study if the treatment showed no early signs of activity. If this initial threshold was met, enrollment would continue to a total accrual of 24 patients in the second stage. The final efficacy endpoint required responses in 3 or more of the total 24 enrolled patients to declare the trial positive. The actual power achieved by this setting was 90.3%, with an actual type 1 error of 9.3% and a probability of stopping at the first stage equal to 63%. Time on treatment was defined as time from enrollment until date off treatment or data cut. All patients initiated study drug within 7 days of the enrollment date, except one patient who started study drug 10 days after enrollment. A standard Kaplan–Meier estimator was calculated to plot the PFS curve, including 95% CI. Correlative and biomarker objectives were considered exploratory.

Protocol amendments

• Amendment 1 (v.02Aug2021): clarified that timing of all visits and assessments tied to C1D1, and added a window for scheduling assessments. Height was removed as a screening parameter.

• Amendment 2 (v.18Jan2022): revised protocol to broaden the eligibility criteria to allow a subset of participants to enroll without a disease site considered accessible and low risk for biopsy.

• Amendment 3 (v.30Jun2022): revised consent form to include additional drug risk of hyperphosphatemia (increased levels of phosphorous in the blood).

• Amendment 4 (v.25Oct2022): eliminated the limit of 24 cycles of treatment and removed the secondary objective to estimate overall survival.

• Amendment 5 (v.07Mar2023): decreased frequency of labs beyond cycle 4.

• Amendment 6 (v.26Jul2023): permitted use of archival tissue if pre-treatment biopsy unavailable and revised consent form to explain the use of archival tissue.

Correlatives studies

A pre-treatment (before C1D1) biopsy, if the disease site was amenable to biopsy, and post-progression (optional) tissue biopsy cores per patient were collected, flash frozen and sent to the Early Phase and Experimental Clinical Trials Biospecimen Bank (EET Biobank) for nucleic acid extraction and subsequent genomic and transcriptomic analysis at the National Clinical Laboratory Network genomics laboratory in the Molecular Characterization and Clinical Assay Development Laboratory, Frederick National Laboratory for Cancer Research. Nine of 13 pre-treatment biopsy samples met the >70% tumor content requirement necessary to proceed with genomic and transcriptomic analysis.

Biospecimen nucleic acid extraction

DNA and RNA were extracted, and quality was assessed at the Biopathology Center at EET Biobank. RNA and DNA were extracted from tumors using a modification of the Qiagen AllPrep DNA/RNA Mini kit (cat. no. 80204). Homogenized and lysed tissue utilizing RLT buffer tissue was passed through the AllPrep DNA spin column. The flow-through from the Qiagen DNA column was processed using the Ambion mirVana microRNA isolation kit (cat. no. AM1560). This latter step generated RNA preparations that included RNA of <200 nucleotides suitable for miRNA analysis. The DNA spin column continued utilizing the Qiagen AllPrep kit for DNA isolation. For germline DNA extraction from 10 ml of whole blood, the Qiagen QIAAmp DNA Blood Maxi kit (cat. no. 51192) was utilized.

RNA samples were quantified by measuring fluorescence by Qubit Fluorometer utilizing the Qubit RNA BR assay kit (cat. no. Q10210), and DNA quantified by Quant-iT PicoGreen assay (cat. no. P7589). DNA specimens were resolved by 1% agarose gel electrophoresis to confirm high molecular weight fragments. RNA was analyzed via the RNA6000 Nano assay (Agilent, cat. no. 5067-1529) for determination of the RNA Integrity Number.

Genomic analysis

Mutations and fusions in FGFR1–4 were examined pre-enrollment by clinical sequencing. Tumor DNA samples from pre-treatment biopsies of nine patients were analyzed by whole-exome sequencing. Local standard of care genomic testing on archival tissue was included in the genomic analysis, where available.

Whole-exome sequencing

Exome analysis used a minimum of 50 ng of DNA from tumor and germline collections at the pre-treatment collection time point. The NLCN whole-exome sequencing assay began with initial library preparation using the Kapa HyperPrep Plus protocol. After a pre-hybridization polymerase chain reaction (PCR) and bead-based library purification, the Agilent SureSelect v6 + CoSMiC exome workflow was used for exome hybridization and capture, followed by enrichment PCR. A custom exome panel was used, which covers ~42 Mb of the human exome and focuses on clinically actionable cancer-associated genes. The custom panel includes higher coverage in selected genes to provide increased sensitivity for single-nucleotide variants and insertions and deletions in actionable genes. The panel detects single-nucleotide variants, insertions and deletions, and copy number variations.

RNA sequencing

Bulk RNA-seq was performed with pre-treatment tumor biopsies from nine patients. RNA-seq libraries were prepared by the NLCN RNA-seq assay with 100 ng of total RNA input using the Illumina TruSeq RNA Exome Dual Index library preparation kit, which covers ~45 Mb of the coding transcriptome. Libraries were quantitated with the Thermo Fisher Quant-iT DNA fluorescence-based quantitation kit before pooling and by Bio-Rad Droplet Digital PCR subsequently before being loaded on to an Illumina NovaSeq S4 flow cell. The resulting data were then demultiplexed and analyzed via the Molecular Characterization and Clinical Assay Development Laboratory RNA-seq bioinformatics pipeline. The RNA-seq pipeline detects RNA fusions and filters out small noncoding RNA and residual ribosomal RNA, leaving only reads from messenger RNA.

Bioinformatics

For bioinformatics data processing, the hg19 reference genome was used. From whole-exome sequencing, BWA was used for sequence alignment, and the SDHA and other mutations were identified by the MuTect2 software⁵⁸. The copy number calls were inferred by Sequenza⁵⁹. For the gene expression analysis, alignment using STAR was performed, and normalized counts were obtained from RSEM⁶⁰ and DESeq2⁶¹ results.

SDH status

SDH subunit IHC was performed pre-enrollment. Pathology reports were analyzed for evidence of SDHB or SDHA expression using IHC. Genomic mutation or copy number loss of SDH complex genes were identified from standard of care genomic sequencing or whole-exome sequencing of pre-treatment biopsies.

Pharmacokinetics

Patients were sampled for EDTA plasma pharmacokinetics of rogaratinib pharmacokinetic before treatment and before their daily dose on cycle 1 day 15, cycle 2 day 1 and day 15, and cycle 3 day 1. Rogaratinib concentrations were quantitated with a validated liquid chromatography coupled with tandem mass spectrometry method over a range of 50–10,000 ng ml⁻¹. Each blood sample was centrifuged at approximately 1,000g, and plasma was stored at −70 °C or colder until analysis.

Rogaratinib was purchased from MedChemExpress and [D4]-rogaratinib was obtained from Bayer AG. Acetonitrile and water (both high-performance liquid chromatography grade) were purchased from Fisher Scientific. Formic acid was purchased from Sigma-Aldrich. Control EDTA human plasma was purchased from BioIVT. Nitrogen for mass spectrometric applications was purified using a nitrogen generator (Parker Balston). The liquid chromatography system consisted of a Waters ACQUITY UPLC system and an Acquity UPLC BEH C18 (2.1 × 50 mm, 1.7 μm) column, kept at ambient temperature. The autosampler temperature was maintained at 5 °C. Mobile phase solvent A was 0.1% formic acid in water, and mobile phase solvent B was 0.1% formic acid in acetonitrile. A flow rate of 0.2 ml min⁻¹ was maintained throughout the run. The initial mobile phase composition was 85% solvent A, which decreased linearly to 0% over 1 min. Between 1 and 2 min, the mobile phase composition was held constant, and at 2.1 min, solvent A was increased to 85%, where it was held constant until 4 min, followed by two injections of the solvent (50% methanol in water (v/v)). The total run time was 4 min, with an injection volume of 5 μl. The approximate retention time of rogaratinib was 1.52 min, with a corresponding capacity factor of 2.37, at a void time of 0.45 min. Mass spectrometric detection was carried out using an AB SCIEX TRIPLE QUAD 5500 tandem mass spectrometer with electrospray ionization in positive multiple reaction monitoring mode. The scanning parameters for the mass spectrometer in positive mode were as follows: curtain gas 20, collision gas 7, IonSpray voltage 5,500 V, probe temperature 450 °C, GS1 30, GS2 50, DP of 60 V, EP of 10 V, and CXP of 16 V. The CE was set at 15 V, except for the qualifier transition of rogaratinib (33 V). The multiple reaction monitoring transitions monitored were: m/z 467.1 > 367.1 for rogaratinib; 471.1 > 367.1 for rogaratinib internal standard (IS). The standard curve of ratio response (analyte peak area to IS peak area) versus concentration was plotted using quadratic regression with 1/x² weighting. A 2 mg ml⁻¹ rogaratinib stock solution was used to prepare a human plasma calibration range for rogaratinib from 0.05 µg ml⁻¹ to 10 µg ml⁻¹. Quality control (QC) samples were prepared at QC low, mid and high concentrations of 0.15, 4 and 8 µg ml⁻¹, respectively. A volume of 25 µl of the standard, QC, or sample plasma was pipetted into a 13 × 100 mm disposable borosilicate glass culture tube. Then, 200 μl of acetonitrile extraction solution with 150 ng ml⁻¹ IS was added to each, followed by vortexing for 30 s. Samples were centrifuged at ~1,000g for 10 min. Supernatants were transferred to autosampler vials, and 5 µl was injected into the liquid chromatography–tandem mass spectrometry system. QC-based accuracies ranged from 94.9% to 114.5%. Intra- and inter-assay precisions were between 99.7% and 105.9%. Incurred sample reanalysis of 20 samples yielded the following results: 5% of samples had a difference larger than 20%, with a median difference of 4.7% and a median absolute difference of 4.7%.

Data were processed with WinNonlin (Certara). For exploration of exposure response relationships, cycle 1 day 15, and patient-level geometric mean concentration values were plotted as a function of the presence of: (1) at least possibly related grade ≥2 hyperphopsphatemia in cycle 1 or 2; (2) at least possibly related nail toxicity through cycle 4; (3) at least possibly related eye toxicity through cycle 4; (4) at least possibly related diarrhea through cycle 3; (5) at least possibly related maximum grade toxicity in cycle 1 or 2; (6) any grade 3 or four times grade 2 toxicities; and (7) best response. Week 3 serum phosphate concentration was defined as the earliest value available from day 20 and onwards. Statistical testing was performed with RStudio v.2024.12.1.

PDX experiments

All animal experiments were conducted at Dana-Farber Cancer Institute with the approval of the Institutional Animal Care and Use Committee in an AAALAC accredited vivarium. Mice were housed in a pathogen-free facility under a 12-h light and 12-h dark cycle, with ambient temperature maintained at 22.2 ± 1.1 °C and relative humidity at 35–55%, with food and water provided ad libitum. The PDX model, PG20, was generated from surgical resection tissue of a patient with SDHd GIST as described previously¹⁵. For efficacy studies, tumor fragments were implanted into 8-week-old female NOD.Cg-Prkdc^scid Il2rg^tm1Wjl/SzJ (NSG) mice purchased from The Jackson Laboratory (stock no. 005557). Tumors were allowed to establish to 176.2 ± 45.3 mm³ in size before randomization into various treatment groups (Studylog software) with n = 8 per group as: vehicle control; 40 mg kg⁻¹ once daily sunitinib (LC Laboratories; 0.1 M citrate buffer, pH 4.5); 30 mg kg⁻¹ once daily regorafenib (LC Laboratories; 34% propylene glycol, 34% polyethylene glycol 400, 12% Fluka F-68 and 20% water); 25 mg kg⁻¹ twice daily rogaratinib (Bayer AG; 10% ethanol, 40% Solutol HS15, 50% water); 20 mg kg⁻¹ once daily infigratinib (LC Laboratories; 50% of acetic acid/acetate buffer, pH 4.6 and 50% polyethylene glycol 300); or 1 mg kg⁻¹ once daily pemigatinib (MedChemExpress, 5% N,N-dimethyl acetamide, 95% of 0.5% methyl cellulose in water). Mice were treated orally for 28–35 days with these agents. Growth rates were evaluated by measuring tumor volumes every 3–4 days. To evaluate whether treatment with an agent led to a significant growth rate difference (P < 0.001) compared to vehicle control, a mixed-effect, repeated-measurement linear model was used, including a quadratic term for time and an interaction term between time and drug. All animal experiments were conducted at Dana-Farber Cancer Institute with the approval of the Institutional Animal Care and Use Committee in an AAALAC accredited vivarium.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.