Abstract
Purpose:
We report herein a phase Ib trial to determine safety, tolerability and anti-tumor activity of erdafitinib, a pan-FGFR tyrosine kinase inhibitor, with fulvestrant and palbociclib in patients with HR+/HER2-negative metastatic breast cancers (NCT03238196).
Experimental Design:
Thirteen patients were enrolled on escalation phase in a traditional 3+3 trial design to determine maximum tolerated dose (MTD). Subsequently, twenty-two patients were treated at the established MTD during the expansion phase. All patients had received prior treatment with CDK4/6 inhibitors and endocrine therapy, 29 showed FGFR pathway alterations in their tumors.
Results:
The MTD of erdafitinib was 6 mg taken orally once daily when combined with palbociclib and fulvestrant. The triple combination showed clinically manageable tolerability. Most common adverse events were neutropenia likely attributable to palbociclib, and oral mucositis and hyperphosphatemia, attributable to erdafitinib. Three patients showed a partial response, one of them lasting over 2.5 years, despite lacking detectable FGFR1-4 somatic alterations. FGFR1 amplification was not associated with response to FGFR inhibition, but high FGFR1 protein expression, measured by immunohistochemistry, correlated with longer progression-free survival within the FGFR1 amplified cohort. There was no correlation between FGFR1 copy number and FGFR1 protein levels in specimens from metastatic sites, potentially highlighting the need of a more recent metastatic tumor biopsy for biomarker evaluation.
Conclusions:
The trial endpoint was met establishing the MTD of erdafitinib at 6 mg. While the triplet regimen may pose tolerability challenges, alterative doublets with selective FGFR1 inhibitors in patients with FGFR1 dependent tumors, possibly administered in sequence, are worthy of further investigation.
Statement of translational relevance
This clinical trial investigated the combination of erdafitinib, a pan-FGFR tyrosine kinase inhibitor, with fulvestrant and palbociclib in patients with HR+/HER2-negative metastatic breast cancer under the premise that FGFR inhibition could overcome the resistance commonly observed in this setting. While response rates were modest, this study provides insights into patient selection by challenging the reliability of FGFR1 amplification as a biomarker for predicting response to FGFR inhibitors and highlights the importance of using a more recent metastatic tumor biopsy for biomarker evaluation.
Introduction
Hormone-receptor positive HER2-negative (HR+) breast cancer accounts for approximately 75% of newly diagnosed breast cancers. While the majority of patients diagnosed with early-stage HR+ disease are cured, about 20% of patients develop metastatic recurrences. Despite advances in early detection and therapeutic options, unresectable or metastatic HR+ breast cancer remains incurable and is the leading cause of breast cancer-related mortality. Systemic therapy for patients with HR+ metastatic breast cancer (MBC) includes endocrine therapy in combination with CDK4/6 inhibitors, PI3K/AKT inhibitors, antibody drug conjugates, and/or chemotherapy [1]. However, acquired resistance, and occasionally primary resistance, to antiestrogens and CDK4/6i therapy develops in patients with HR+ MBC [2, 3]. After acquired resistance develops, breast cancers may still depend on ligand-independent ER activity in addition to signaling through oncogenic signaling pathways [2].
FGFR pathway signaling aberrations in breast cancer have been associated with poor prognosis and resistance to endocrine therapy and CDK4/6 inhibitors [4–8]. Approximately 15% of patients with HR+ breast cancer harbor genetic/genomic alterations in the FGFR family, the most common being amplification of the FGFR1 gene (8p11.23). Pre-clinical data have shown that the addition of an FGFR inhibitor restores sensitivity to palbociclib and fulvestrant in FGFR1 amplified/overexpressing drug-resistant cell lines and patient derived xenografts (PDXs) [4, 6]. Early-phase clinical trials assessing the effect of multi-target pan FGFR tyrosine kinase inhibitors have shown limited clinical benefit in patients treated with FGFR inhibition, either alone [9–18] or in combination with antiestrogens [19–22]. We showed that the triple combination of FGFR, ERα and CDK4/6 inhibitors induced complete responses in FGFR1-amplified/ER+ PDXs [4], suggesting this approach was worthy of further clinical testing.
Based on these preclinical data [4, 6], we conducted a phase Ib trial combining erdafitinib, a potent, oral pan-FGFR tyrosine kinase inhibitor, with the ERα degrader fulvestrant and the CDK4/6i palbociclib in patients with HR+/HER2-negative MBC (NCT03238196) to determine safety and tolerability of the combination and anti-tumor activity.
Methods
Study participants and design.
NCT03238196 is an open-label, multi-institution, dose escalation and expansion phase Ib trial that evaluates the safety, tolerability, and anti-tumor activity of the combination of fulvestrant, palbociclib and erdafitinib in patients with HR+/HER2-negative metastatic breast cancer.
Institutional review board was obtained at all participating institutions, and informed written consent was obtained from each subject. This study was conducted in compliance with the protocol and Good Clinical Practice, as described in ICH Harmonized Tripartite Guidelines for Good Clinical Practice 1996, US 21 Code of Federal Regulations dealing with clinical studies and the Declaration of Helsinki concerning medical research in humans.
Patient eligibility criteria included: post-menopausal women with clinical stage IV or inoperable locoregional recurrent histologically documented HR+/HER2− invasive mammary carcinoma without prior use of an FGFR inhibitor, at least one line of systemic therapy in the metastatic setting (no more than 2 lines of chemotherapy in the metastatic setting, no limit on endocrine therapy lines, prior exposure to CDK4/6 inhibitor and fulvestrant acceptable), ECOG performance status ≤1, normal baseline blood counts and chemistry laboratory profile; and no intercurrent uncontrolled illness or symptomatic brain metastases. At least 50% of the patients participating in the expansion cohort were required to have tumors with FGFR1-4 amplification determined by local assessment through targeted capture NGS, circulating-tumor DNA, or FISH. Patients with known FGFR1 amplification by clinical assays (NGS, cfDNA or FISH) required central confirmation of FGFR1 amplification by FISH but did not be required to wait for that result to initiate study treatment. Patients with known FGFR2-4 amplification by clinical assays (NGS, cfDNA or FISH) did not require central confirmation of FGFR amplification by FISH. Patients without known FGFR amplification were screened for FGFR1 amplification by FISH analysis at the Vanderbilt-Ingram Cancer Center. Cases were considered as FGFR1-amplified by FISH under one of the following conditions: FGFR1/CEN8 ratio is ≥2.0; and/or average number of FGFR1 signals per tumor cell nucleus is ≥6.
Primary endpoint was the determination of maximum tolerated dose (MTD), defined as the highest dose at which 20% of the patients experienced a DLT. During the escalation cohort, patients were administered fulvestrant and palbociclib in combination with erdafitinib in a traditional 3+3 trial design, with a cohort of 3 patients receiving a staring dose of 5 mg and evaluated for dose limiting toxicities (DLTs) for 4 weeks before the initiation of the next cohort at the escalating dose of 6 mg and then 8 mg. Patients must have received at least 75% of study drug during cycle 1 to be evaluable for a DLT observation period. No new cohort of patients were treated until the previous cohort was fully evaluated for toxicity. Fulvestrant was administered as an intramuscular injection of 500 mg on day 1 and 15 of the first cycle (loading dose) followed by every 28 days. Palbociclib was taken at a dose of 125mg orally once a day, 3 weeks on, one week off. Erdafitinib was taken orally once daily as film coated tablets of 3, 4 and 5 mg. Eridafitinib was provided by Janssen Research and Development and palbociclib by Pfizer.
Patients in the expansion portion of the study, initiated treatment with erdafitinib at the MTD defined in the escalation component of the trial. Dose reductions were allowed after the 4-week DLT observation period in the escalation and in the expansion components for the 3 drugs. Treatment was given until disease progression, unacceptable toxicity or patient withdrawal/non-compliance with protocol. Patients removed from the study for treatment-related adverse events were followed until resolution or stabilization of the event.
Safety and efficacy assessments
Safety was assessed using Common Terminology Criteria for Adverse Events (version 4.03). As per protocol all patients were evaluable for toxicity from the time of their first administration of study treatment drug. Medical history, physical exam, and routine blood work including a metabolic panel were performed on day 1 of each treatment cycle. Events that occurred during cycle 1 (first 4 weeks) listed on Supplemental Table 1 when classified as possibly, probably or definitively related to investigational study treatment or that lead to a dose interruption of more than 21 consecutive days were considered a DLT. Grading scales correspond to the NCI Common Terminology Criteria for Adverse Events version 4.03.
To assess the anti-tumor effect of therapy, overall tumor burden on CT scans of chest, abdomen and pelvis and bone at baseline was compared to subsequent measurements performed every 8 weeks using the Solid Tumor Response Criteria [RECIST] v1.1. Only those patients who had measurable disease present at baseline, had received at least one cycle of therapy, and had their disease re-evaluated were considered evaluable for response. Archival tumor formalin-fixed paraffin embedded tissue collected prior to the initiation of the clinical trial was gathered for correlative analysis. Biopsy samples from metastatic site close to the initiation of treatment were preferred but archival tissue form primary breast cancer was accepted.
Secondary endpoints were progression-free survival (PFS), overall response rate (ORR) and clinical benefit rate (CBR; complete response + partial response + stable disease without disease progression at 6 months).
Correlatives endpoints included FGFR1 fluorescence in situ hybridization (FISH) and next generation sequencing (NGS) of archival tumor specimen to determine the therapeutic predictive role of FGFR1-4 amplification and FGFR1 amplification levels by FISH on clinical outcome and discover any other potential predictive genetic alteration. Consort diagram (Figure 1) shows patients evaluable for efficacy and sample availability for correlatives. Serial measurements of serum phosphate were obtained to determine pharmacodynamic biomarker of FGFR inhibition.
Figure 1:
Consort diagram
Fluorescent in situ hybridization (FISH).
Central confirmation of FGFR amplification by FISH was performed for those patients with available tissue samples (n=29) at Vanderbilt-Ingram Cancer Center in the CLIA-certified Vanderbilt University Medical Center Cytogenetics Laboratory with the ZytoLight SPEC FGFR1/CEN 8 Dual Color Probe (#Z-2072-200, ZytoVision, Bremerhaven, Germany) on archival biopsies from metastatic site or archival primary tumor if metastatic samples was unavailable. Representative images of tumor cells are captured using Cytovision software. Thirty tumor cell nuclei from hot spots or random areas were individually evaluated with the Olympus BX60 at 100X by counting green FGFR1 and orange centromere 8 (CEN8) signals. The average of FGFR1 copy number and the FGFR1/CEN8 was calculated. Tumors are considered as FGFR1 amplified under one of the following conditions: FGFR1/CEN8 ratio ≥2.0; and/or average number of FGFR1 signals per tumor cell nucleus ≥6 [23]. High FGFR1 amplification was defined as FGFR1:CEP8 ≥5 based on the cut-off set arbitrarily by previous publications [13, 14, 23]. We also explored cut-off of FGFR1:CEP8 ≥10.
Immunohistochemistry (IHC).
IHC was performed (n=29) with the primary rabbit monoclonal antibody FGFR1 D8E4 (#9740, Cell Signaling Technology, Danvers, MA, USA) at 1:200 overnight at 4°C as previously validated and optimized [23]. Tumor sections from an MCF7 (RRID:CVCL_0031) FGFR1 knockout xenograft were used as a negative control); cell/tumor blocks from FGFR1 amplified MDA-MB-134 cells (RRID:CVCL_0617) and a FGFR1 amplified patient-derived xenograft [4] were used as positive controls. Cell lines were obtained from the American Type Culture Collection. Semiquantitative scoring was performed by a pathologist based on percentage of tumor cells presenting membranous staining (0-100%) and staining intensity (0-3+). H-scores were calculated using the following equation: [H-score = 3*(% of 3+ intensity cells) + 2*(% of 2+ intensity cells) + 1*(% of 1+ intensity cells)]. Tumors were classified based on HER2 criteria [24] as previously described [23].
Next generation sequencing (NGS).
Results from NGS sequencing panels (Foundation One, Guardant 360) were collected from the patient medical records (n=25). For those patients with available tissue (n=16) MSK-IMPACT panel was performed as previously described [25]. In total, NGS data was collected for 33 patients.
Statistical considerations.
The primary endpoint was the assessment of DLT and determination of MTD. Secondary endpoints were PFS, ORR, and CBR. MTD was determined from patients treated during dose escalation; subsequently, data were pooled with data from patients in the expansion cohort to assess tolerability, efficacy, and exploratory correlative analysis. Descriptive statistics were used to assess safety and efficacy. Logistic regression, Fisher’s exact test, and log-rank tests were used to associate response to biomarker evaluation of correlative analysis. Statistical analyses were performed in R (RRID:SCR_001905) or Graphpad Prism (RRID:SCR_002798).
Data Availability
The data generated in this study are available within the article and its supplementary data files. Deidentified genomic sequencing data generated by Memorial Sloan Kettering IMPACT are available under restricted access to protect patient privacy in accordance with federal and state law. These data can be requested for research use from the corresponding author via execution of a data transfer agreement with MSKCC. Any additional data are available upon request from the corresponding author.
Results
Patient Demographics
Thirty-five post-menopausal women with clinical stage IV or inoperable locoregional recurrent histologically documented HR+/HER2-negative invasive mammary carcinoma were enrolled between August 2017 and September 2020 from five participating institutions. Thirteen patients participated in the escalation phase and another 22 were enrolled to complete the expansion phase. Table 1 shows baseline characteristics for the study population. Patients had received a median of 2 (1 – 6) prior lines of therapy in the metastatic setting excluding an FGFR inhibitor. All (100%) patients had been previously treated with a CDK4/6 inhibitor and endocrine therapy, 18/35 (51%) with fulvestrant, and 33/35 (94%) with chemotherapy. Thirty-three (94%) had previously received palbociclib, 3 abemaciclib and 2 ribociclib. Three patients had received two CDK4/6 inhibitors on different lines of treatment. Median time on CD4/6 inhibition treatment was 10 months and ranged between 2 and 52 months. Overall, 29 patients had tumors with FGFR1-4 amplification as measured by targeted next-generation sequencing (NGS) and/or fluorescent in situ hybridization (FISH), with one patient showing amplification of FGFR1, 2, and 3. Supplemental table 2 provides background information on the representativeness of the study population
Table 1:
Baseline Patient Characteristics
| Escalation cohort (n=13) | Expansion cohort (n=22) | All patients (n=35) | |
|---|---|---|---|
| Median age (years) | 53 | 54 | 54 |
| Race/ethnicity | |||
| Asian | 1 (7.5%) | 2 (9%) | 3 (9%) |
| Black | 1 (7.5%) | 1 (5%) | 2 (6%) |
| White | 7 (54%) | 17 (77%) | 24 (69%) |
| Unknown | 4 (31%) | 2 (9%) | 6 (17%) |
| Hispanic | 0 | 1 (5%) | 1 (3%) |
| Prior lines of treatment in the metastatic setting (median) | 3 | 2 | 2 |
| Hormone therapy | 13 (100%) | 22 (100%) | 35 (100%) |
| Aromatase inhibitors or tamoxifen | 13 (100%) | 21 (95%) | 34 (97%) |
| Fulvestrant | 9 (69%) | 9 (41%) | 18 (51%) |
| GnRH agonists | 3 (23%) | 10 (45%) | 13 (37%) |
| CDK4/6 inhibitor | 13 (100%) | 22 (100%) | 35 (100%) |
| PI3K pathway inhibitor | 4 (31%) | 4 (18%) | 8 (23%) |
| Chemotherapy | 13 (100%) | 13 (59%) | 23 (66%) |
| Metastatic Sites | |||
| Bone | 13 (100%) | 15 (68%) | 28 (80%) |
| Brain | 2 (15%) | 0 | 2 (6%) |
| Liver | 12 (92%) | 8 (36%) | 20 (57%) |
| Lung | 6 (46%) | 7 (32%) | 13 (37%) |
| Lymph node | 3 (23%) | 6 (27%) | 9 (26%) |
| Pleural / peritoneal | 4 (31%) | 1 (5%) | 5 (14%) |
| FGFR gene alteration | |||
| FGFR1 amp | 9 (69%) | 20 (91%) | 29 (83%) |
| FGFR2 amp | 0 | 1 (5%) | 1 (3%) |
| FGFR3 amp | 0 | 1 (5%) | 1 (3%) |
| No FGFR1-4 amp | 4 (31%) | 2 (9%) | 6 (17%) |
Aromatase inhibitors include letrozole, anastrozole and exemestane. GnRH: Gonadotropin-releasing hormone agonists include leuprolide acetate and goserelin acetate.
Determination of maximum tolerated dose, Adverse Events and Safety Assessment
To evaluate safety of the combination, 500 mg fulvestrant and 125 mg of palbociclib were administered in combination with erdafitinib at a starting dose of 5 mg and escalated to 6, and 8 mg in a traditional 3+3 trial design. Thirteen patients were included on the dose escalation phase. No DLTs were observed for the 3 patients treated at 5 mg. One patient treated at 6 mg experienced a DLT so a second cohort of 3 patients was as initiated at 6 mg. In total, one out of 6 (16.6%) and 3 out of 4 (75%) patients experienced DLTs at 6 mg and 8 mg erdafitinib, respectively. DLTs were grade 3 palmar-plantar erythrodysesthesia syndrome (6 mg), oral mucositis (2 patients treated at 8 mg) and elevation of AST/ALT (8 mg). Maximum tolerated dose (MTD) for erdafitinib was established as 6 mg. All DLTs resolved within two weeks after treatment interruption.
Adverse events potentially associated with the triple combination treatment are listed in Table 2 and Supplemental Table 3. Most common adverse events were neutropenia likely attributable to palbociclib, followed by oral mucositis and hyperphosphatemia, attributable to erdafitinib. Retinal epithelial detachments were not observed, although 15 patients (43%) reported ocular toxicity including dry eyes, keratosis, and redness. Overall, these side effects were manageable and reversible. Sixteen patients (46%) suffered adverse events attributable to erdafitinib that resulted in dose modification (reduction or discontinuation, Supplemental Table 4); these included oral mucositis, hyperphosphatemia, palmar-plantar erythrodysesthesia syndrome, skin ulcerations, and diarrhea. Adverse events attributed to palbociclib resulted in dose reduction in 19 patients (54%), mainly due to grade 3 neutropenia. Seven patients (20%) discontinued treatment due to toxicities (Supplemental Table 4); associated adverse events were grade 2/3 oral mucositis, grade 3/4 elevation of AST/ALT, grade 2/3 palmar-plantar erythrodysesthesia syndrome, prolonged/refractory grade 2 hyperphosphatemia, grade 2 vision alterations, and grade 3/4 neutropenia.
Table 2:
Adverse events
| Toxicity | Events | Patients | ||||
|---|---|---|---|---|---|---|
| Grade 3/4 | Any grade | Grade 3/4 | Any grade | |||
| n | % | n | % | |||
| Most likely related to Erdafitinib | ||||||
| Oral mucositis | 5 | 58 | 5 | 14% | 27 | 77% |
| Hyperphosphatemia | 0 | 56 | 0 | 0% | 29 | 83% |
| Palmar-plantar erythrodysesthesia syndrome | 4 | 36 | 3 | 9% | 18 | 51% |
| Diarrhea | 0 | 30 | 0 | 0% | 14 | 40% |
| Constipation | 0 | 25 | 0 | 0% | 19 | 54% |
| Dysgeusia | 0 | 25 | 0 | 0% | 19 | 54% |
| Dry mouth | 0 | 22 | 0 | 0% | 18 | 51% |
| Nail alterations | 0 | 19 | 0 | 0% | 13 | 37% |
| Sore throat | 1 | 18 | 1 | 3% | 8 | 23% |
| Alopecia | 0 | 16 | 0 | 0% | 13 | 37% |
| Elevated AST | 4 | 14 | 2 | 6% | 7 | 20% |
| Dry skin | 0 | 13 | 0 | 0% | 12 | 34% |
| Epistaxis | 0 | 12 | 0 | 0% | 9 | 26% |
| Anorexia | 0 | 11 | 0 | 0% | 7 | 20% |
| Dry eye | 1 | 11 | 1 | 3% | 10 | 29% |
| Skin ulceration | 9 | 10 | 2 | 6% | 2 | 6% |
| Elevated ALT | 4 | 8 | 2 | 6% | 5 | 14% |
| Vision changes/alterations | 1 | 7 | 1 | 3% | 6 | 17% |
| Abdominal pain | 0 | 5 | 0 | 0% | 5 | 14% |
| Dizziness | 0 | 5 | 0 | 0% | 3 | 9% |
| Gastroesophageal reflux disease | 0 | 5 | 0 | 0% | 5 | 14% |
| Hypotension | 2 | 4 | 1 | 3% | 3 | 9% |
| Colitis | 2 | 2 | 1 | 3% | 1 | 3% |
| Esophagitis | 1 | 2 | 1 | 3% | 2 | 6% |
| Eye keratopathy | 1 | 2 | 1 | 3% | 1 | 3% |
| Hyperkeratosis | 1 | 1 | 1 | 3% | 1 | 3% |
| Syncope | 1 | 1 | 1 | 3% | 1 | 3% |
| Most likely related to Palbociclib | ||||||
| Neutropenia | 44 | 74 | 22 | 63% | 25 | 71% |
| Leucopenia | 9 | 30 | 8 | 23% | 15 | 43% |
| Anemia | 3 | 23 | 3 | 9% | 12 | 34% |
| Thrombocytopenia | 4 | 21 | 12 | 34% | 2 | 6% |
| Lymphopenia | 1 | 7 | 1 | 3% | 4 | 11% |
| Febrile neutropenia | 1 | 1 | 1 | 3% | 1 | 3% |
| Neutrophilia | 1 | 1 | 1 | 3% | 1 | 3% |
| Thromboembolic event | 1 | 1 | 1 | 3% | 1 | 3% |
| Likely related to the combination | ||||||
| Fatigue | 1 | 50 | 1 | 3% | 23 | 66% |
| Nausea | 0 | 7 | 0 | 0% | 6 | 17% |
| Vomiting | 0 | 7 | 0 | 0% | 6 | 17% |
Grade 1, 2, 3 and 4 toxicities were recorded by the treating investigators, no grade 5 toxicities were experienced in this trial.
Clinical Efficacy
Four patients were deemed non-evaluable for antitumor efficacy, three due to treatment discontinuation mainly due to adverse events, which occurred prior to completion of the first cycle of treatment, and one due to lack of re-evaluation. Out of 31 patients with evaluable disease, three (10%) showed partial response (PR) and 17 (55%) showed stable disease (SD) as their best response (Table 3). Disease progression was recorded for 28 patients during the follow-up period; 3 patients were lost to follow-up after treatment discontinuation (Figure 2). The overall response rate (ORR) was 10%, and the clinical benefit rate at 6 months was 23% (Table 3). Among patients who responded, the duration of response (PR to PD) ranged from 18 to 136 weeks. Median progression-free survival (PFS) was 12 weeks. Time on treatment for patients with evaluable disease ranged from 5 to 168 weeks, median time on treatment was 8.7 weeks.
Table 3:
Response to fulvestrant, palbociclib and erdafitinib combination according to FGFR1-4 amplification status
| All patients | FGFR1-4 amp | FGFR1-4 non-amp | |
|---|---|---|---|
| Total number of cases | 35 | 29 | 6 |
|
| |||
| Evaluable for efficacy | 31 | 27 | 4 |
| Best Overall Response | |||
| CR | 0 | 0 | 0 |
| PR | 3 (10%) | 2 (8%) | 1 (25%) |
| SD | 17 (55%) | 16 (59%) | 1 (25%) |
| PD | 11 (35%) | 9 (33%) | 2 (50%) |
| Non evaluable | 4 | 2 | 2 |
| Clinical Benefit Rate at 6 months | 23% | 22% | 25% |
| Duration of overall response range | 18-112 w | 18-39 w | 122 w |
| Median progression free survival (range) | 12 w (5-144w) | 12 w (6-55w) | 13 w (5-144w) |
Figure 2: Response to the combination of fulvestrant, palbociclib and erdafitinib.
Swimmers plot showing time on treatment, start of response (if any), reason for end of treatment. Patients are color coded by FGFR1-4 amplification status. Star (*) indicates patient with FGFR1, 2 and 3 amplification bearing tumor. Arrows correspond to follow-up after treatment discontinuation.
Changes in serum phosphate levels were considered a pharmacodynamic effect of erdafitinib, with an increase from baseline observed in all patients. This change did not correlate with treatment response (Supplemental Figure 1).
Molecular Correlatives
Assessment of FGFR1 amplification was conducted through targeted NGS and/or FISH: both methods were used in 25 patients: FISH only, 4; NGS only, 6. Analysis of FGFR1-4 amplification status (Figure 3A), FGFR1:CEP8 ratio (Figure 3B and supplemental Figure 2), and FGFR1 copy number (p=0.75) revealed no correlation with treatment response, whether considering the intention-to-treat group or the subset with FGFR1-4 amplification. Protein expression was assessed by immunohistochemistry (IHC) on 29 patients with available tumor material. FGRF1 IHC and FISH were performed on the same sample for each patient; 14 on a metastasis biospecimen and 15 on the primary tumor. Notably, patients with high FGFR1 protein expression by IHC, defined as scores 2+ and 3+ assessed using HER2 criteria (Figure 3C), within the FGFR1 amplified cohort, exhibited extended PFS compared to IHC expression of 0 or 1+ (median PFS 21.43 vs 12.21 weeks, Figure 3D). High FGFR1 protein expression (2+/3+) was exclusively observed on tumors with FGFR1 amplification. However, not all highly amplified cases showed high protein expression. Only 39% of FGFR1 amplified and 41% of highly amplified tumors, defined as FGFR1:CEP8 ratio ≥5 [13, 14, 23], showed high FGFR1 protein levels by IHC (2+/3+). Correlation between FGFR1 protein expression and FGFR1:CEP8 ratio (r2=0.36) or gene copy number (r2=0.25) was only modest. Discordance was mostly observed in samples from metastatic sites, among which FGFR1 gene copy number to protein correlation was poor (r2=0.07) while correlation for patients whose primary breast cancer tissue were used for FGFR1 FISH and IHC was better (r2=0.46, Figure 3E).
Figure 3: Response to fulvestrant, palbociclib and erdafitinib combination according to FGFR1 gene status and protein expression.
A) Kaplan-Meier curve stratified by A) FGFR1-4 amplification status (n=31), and B) FGFR1:CEP8 ratio by FISH for the FGFR1 amplified sub-cohort (n=22). High FGFR1 amplification was defined as FGFR1:CEP8 ≥5 based on the cut-off set on previous publications. We also explored cut-off of FGFR1:CEP8 ≥10 on Supplemental Figure 2. PFS: progression free survival. C) Representative images of FGFR1 IHC categorized as 3+ (top), 2+ (middle) and 1+ (bottom). D) Kaplan-Meier curve stratified by FGFR1 protein expression by IHC for the FGFR1 amplified sub-cohort (n=21). E) Correlation between FGFR1 CN by FISH and FGFR1 membranous H-score by IHC (n=28). Coloring represents site of sample evaluated. Primary breast cancer samples (cyan circles n= 17) were used for biomarker analysis when metastatic site samples (pink triangles n=12) were unavailable. Black dotted line represents cut-off for amplification and 2+/3+ IHC categories.
Detailed in Supplemental Table 5 are gene mutations and copy number alterations and their correlation with PFS. The adjusted p-value after multi-testing adjustment was not significant for any association with clinical response. An exceptional response was observed in one heavily pretreated patient (PFS 42 months, duration of response 34 months), whose tumor exhibited no FGFR1-4 amplification but displayed 11q13 (which harbors CCND1 and FGF3, FGF4, and FGF19), CRKL, ZNF217 and AURK amplifications, and a BRCA1 Q687P mutation.
Discussion
This is the first clinical trial to report on the triple combination of an antiestrogen plus a CDK4/6 and an FGFR inhibitor. The primary endpoint of this trial was to determine the MTD for erdafitinib, which was established as 6 mg in combination with standard doses of fulvestrant and palbociclib. Toxicity was similar to that reported in previous pan-FGFR1 inhibitor trials with and without combination with endocrine therapy [9–12, 14–22]. Consistent with erdafitinib being a selective FGFR TKI, toxicities attributable to the blockade of other receptors, most notably VEGFR, such as hypertension, proteinuria, and cardiovascular events were not observed. Retinal epithelial detachments were not observed, although 43% of patients reported some form of ocular toxicity. The most common grade 3-4 adverse events were cytopenias, with neutropenia being the most frequent, and oral mucositis. A total of 7 patients (20%) discontinued treatment due to adverse events, 6 likely associated with erdafitinib (oral mucositis, palmar-plantar erythrodysesthesia syndrome, hyperphosphatemia and hepatic enzyme elevation) and one due to grade 4 neutropenia most likely associated to palbociclib. Toxicity was greater than that observed on the phase III PALOMA 3 that evaluated the combination of palbociclib and fulvestrant [26]. However, patients in PALOMA III had not received prior CDK4/6 inhibitors and overall had fewer prior therapies. Of note, the three patients that achieved a PR underwent dose reductions due to toxicities; they received 4mg of erdafitinib, 75 to 100mg of palbociclib and 500mg of fulvestrant for majority of their treatment, including the period when PR was achieved. This adjusted dose could offer a safer toxicity profile and clinical benefit. Recently, erdafitinib was approved as a single agent for the treatment of locally advanced or metastatic urothelial carcinoma with susceptible FGFR3 genetic alterations, with manageable tolerability.
Although ribociclib and abemaciclib later demonstrated longer overall survival outcomes [26–28], palbociclib was the most commonly prescribed and the only CDK 4/6 inhibitor approved for use in combination with fulvestrant when this trial was implemented.
Despite several FGFR inhibitors being evaluated in both the preclinical and clinical space in patients with breast cancer, none have been approved by the Food and Drug Administration to treat breast cancer. Response rates with the triplet combination were not superior to that reported in previous clinical trials with FGFR inhibitors in breast cancer alone or in combination with endocrine therapy[9–22]. In this study, all patients had previously experienced disease progression while on endocrine therapy and a CDK4/6 inhibitor, the three patients that achieved a partial response to the triplet combination had been treated with palbociclib and endocrine therapy, letrozole or tamoxifen (first line), but had received no prior fulvestrant. Notably, one of these patients exhibited an exceptional response that persisted for over 2.5 years, despite lacking detectable FGFR1-4 genetic alteration in the tumor, including single-nucleotide variants, insertions/deletions, copy number variants and chromosomal rearrangements/gene fusions tested by a commercial panel of 648-gene DNA sequencing panel with whole-transcriptome RNA sequencing. She had previously been treated with palbociclib, letrozole and goserelin acetate as fist line metastatic treatment followed by chemotherapy as second line with progression within 5 and 3 months respectively. The clinical response to FGFR inhibition in patients with tumor not bearing FGFR alterations have shown mixed results in the past. We are unable to draw conclusions given the very small number of patents with non-FGFR1 amplified tumors included in this trial, despite the extraordinary response observed in this patient.
The expectation of patient selection through the FGFR pathway has been clouded by inconsistent results in patients with breast cancer. Initial trials with pan-FGFR inhibitors hinted that patients with high FGFR1 amplification may demonstrate a favorable response [9, 11]. However, exploratory biomarker analysis in the present study suggests that FGFR1 amplification is not a reliable biomarker for predicting response to FGFR inhibition, also implying that only a fraction of FGFR1 amplified tumors are FGFR1 dependent. Although the evaluated population is small, we did observe longer PFS for patients whose tumors showed high FGFR1 protein expression by IHC within the FGFR1 amplified cohort; in accordance with that previously reported by Hui et al [11]. The assumption that FGFR1 amplification leads to protein overexpression, and therefore tumor dependence on FGFR1 signaling, is thus challenged by the weak concordance observed in this trial, alongside studies involving other tumor types [29, 30]. We had previously shown a correlation between FGFR1 gene copy number and protein expression in a cohort of patients with stage I-III HR+ breast cancer [23]. Within this cohort, we observed a better correlation between gene copy number and protein expression on samples from a primary breast cancer compared to those from a metastatic site, with less than half of the highly amplified cases showing high protein expression in the latter. This suggests that changes affecting the transcription and/or translation of the FGFR1 gene product may be occurring in heavily treated patients with metastatic breast cancer. Therefore, it appears to be crucial that biopsies used for biomarker analysis are as contemporary as possible to the start of treatment. This approach may provide a higher likelihood that a current biomarker is valid and will be predictive of response.
Patients with FGFR2 amplification-bearing tumors were reported to experience partial response to erdafitinib [17, 18]. One of the patients included in this study showed amplification of the FGFR1, 2 and 3 genes. Although the patient discontinued treatment after 6 weeks, she experienced stable disease for 18 weeks on follow-up. Additionally, FGFR2 and FGFR3 mutations and rearrangement have been found to be targetable with FGFR inhibitors in other tumor types as cholangiocarcinoma [31, 32], but these alterations are quite rare in breast cancer [17, 33, 34].
In sum, we identified 6 mg of erdafitinib as the MTD in combination with standard doses of fulvestrant and palbociclib. Although tolerable, the relatively high toxicity of the triplet combination suggests it might be of difficult implementation on a wide basis. An alternative could be combinations of FGFR inhibitors, preferably FGFR selective inhibitors, with SERDs. These doublets could be sequenced with doublets of FGFR and CDK4/6 antagonists, thus avoiding potentially toxic triplets as a first line of therapy for advanced disease. In sum, while FGFR inhibitors may still have a role in the treatment of CDK4/6inhibitor-refractory HR+ breast cancers, both FGFR1-selective inhibitors and a surrogate biomarker of FGFR1 dependence remain for now as unmet needs.
Supplementary Material
Acknowledgements
Funding for this work was provided by NIH/NCI SPORE P50CA098131, NCI P30 CA068485, NCI P30 CA142543, and Susan G Komen ASPIRE Grant.
Conflict of Interest Statement
JMB receives research support from Genentech/Roche and Incyte Corporation, has received advisory board payments from AstraZeneca, Eli Lilly, and Mallinckrodt and is an inventor on patents regarding immunotherapy targets and biomarkers in cancer. BHP is a paid scientific advisory board member for Celcuity Inc., and an unpaid consultant for Tempus Inc, and has research contracts with Tempus, Caris and Guardant Health. Under separate licensing agreements between Horizon Discovery, LTD and The Johns Hopkins University, B.H.P. is entitled to a share of royalties received by the University on sales of products. The terms of this arrangement are being managed by the Johns Hopkins University in accordance with its conflict-of-interest policies. IAM is an AstraZeneca employee. CLA has received research funding from Pfizer and Lilly. The other authors declare no potential conflicts of interest.
References
- 1.Loibl S, et al. , Breast cancer. Lancet, 2021. 397(10286): p. 1750–1769. [DOI] [PubMed] [Google Scholar]
- 2.Dowsett M, Nicholson RI, and Pietras RJ, Biological characteristics of the pure antiestrogen fulvestrant: overcoming endocrine resistance. Breast cancer research and treatment, 2005. 93: p. 11–18. [DOI] [PubMed] [Google Scholar]
- 3.Dean J, et al. , Therapeutic CDK4/6 inhibition in breast cancer: key mechanisms of response and failure. Oncogene, 2010. 29(28): p. 4018–4032. [DOI] [PubMed] [Google Scholar]
- 4.Formisano L, et al. , Aberrant FGFR signaling mediates resistance to CDK4/6 inhibitors in ER+ breast cancer. Nature communications, 2019. 10(1): p. 1–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Formisano L, et al. , Association of FGFR1 with ERα maintains ligand-independent ER transcription and mediates resistance to estrogen deprivation in ER+ breast cancer. 2017. 23(20): p. 6138–6150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Mouron S, et al. , FGFR1 amplification or overexpression and hormonal resistance in luminal breast cancer: rationale for a triple blockade of ER, CDK4/6, and FGFR1. Breast Cancer Research, 2021. 23(1): p. 1–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Sobhani N, et al. , Targeting aberrant FGFR signaling to overcome CDK4/6 inhibitor resistance in breast cancer. Cells, 2021. 10(2): p. 293. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Turner N, et al. , FGFR1 amplification drives endocrine therapy resistance and is a therapeutic target in breast cancer. 2010. 70(5): p. 2085–2094. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.André F, et al. , Targeting FGFR with dovitinib (TKI258): preclinical and clinical data in breast cancer. 2013. 19(13): p. 3693–3702. [DOI] [PubMed] [Google Scholar]
- 10.Chae YK, et al. , Phase II Study of AZD4547 in Patients With Tumors Harboring Aberrations in the FGFR Pathway: Results From the NCI-MATCH Trial (EAY131) Subprotocol W. Journal of Clinical Oncology, 2020. 38(21): p. 2407. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hui R, et al. , Lucitanib for the treatment of HR+/HER2− metastatic breast cancer: Results from the multicohort phase II FINESSE study. Clinical Cancer Research, 2020. 26(2): p. 354–363. [DOI] [PubMed] [Google Scholar]
- 12.Nogova L, et al. , 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. 2016. 35(2): p. 157–165. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Pearson A, et al. , High-level clonal FGFR amplification and response to FGFR inhibition in a translational clinical trial. 2016. 6(8): p. 838–851. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Smyth EC, et al. , Phase II multicenter proof of concept study of AZD4547 in FGFR amplified tumours. 2015, American Society of Clinical Oncology. [Google Scholar]
- 15.Soria J-C, et al. , Phase I/IIa study evaluating the safety, efficacy, pharmacokinetics, and pharmacodynamics of lucitanib in advanced solid tumors. 2014. 25(11): p. 2244–2251. [DOI] [PubMed] [Google Scholar]
- 16.Voss MH, et al. , A phase I, open-label, multicenter, dose-escalation study of the oral selective FGFR inhibitor Debio 1347 in patients with advanced solid tumors harboring FGFR gene alterations. Clinical Cancer Research, 2019. 25(9): p. 2699–2707. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Bahleda R, et al. , Multicenter phase I study of erdafitinib (JNJ-42756493), oral pan-fibroblast growth factor receptor inhibitor, in patients with advanced or refractory solid tumors. Clinical Cancer Research, 2019. 25(16): p. 4888–4897. [DOI] [PubMed] [Google Scholar]
- 18.Gong J, et al. , Phase II study of erdafitinib in patients with tumors with FGFR amplifications: Results from the NCI-MATCH ECOG-ACRIN trial (EAY131) subprotocol K1. JCO Precision Oncology, 2024. 8: p. e2300406. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Musolino A, et al. , Phase II, randomized, placebo-controlled study of dovitinib in combination with fulvestrant in postmenopausal patients with HR+, HER2− breast cancer that had progressed during or after prior endocrine therapy. 2017. 19(1): p. 18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Seckl M, et al. , RADICAL trial: A phase Ib/IIa study to assess the safety and efficacy of AZD4547 in combination with either anastrozole or letrozole in ER positive breast cancer patients progressing on these aromatase inhibitors (AIs). 2017, American Society of Clinical Oncology. [Google Scholar]
- 21.Campone M, et al. , A phase Ib dose allocation study of oral administration of lucitanib given in combination with fulvestrant in patients with estrogen receptor-positive and FGFR1-amplified or non-amplified metastatic breast cancer. Cancer Chemotherapy and Pharmacology, 2019. 83: p. 743–753. [DOI] [PubMed] [Google Scholar]
- 22.Coombes R, et al. , Results of the phase IIa RADICAL trial of the FGFR inhibitor AZD4547 in endocrine resistant breast cancer. Nature Communications, 2022. 13(1): p. 1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Gonzalez-Ericsson PI, et al. , FGFR1 antibody validation and characterization of FGFR1 protein expression in ER+ breast cancer. Applied Immunohistochemistry & Molecular Morphology, 2022. 30(9): p. 600–608. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Wolff AC, et al. , Human epidermal growth factor receptor 2 testing in breast cancer: American Society of Clinical Oncology/College of American Pathologists clinical practice guideline focused update. Archives of pathology & laboratory medicine, 2018. 142(11): p. 1364–1382. [DOI] [PubMed] [Google Scholar]
- 25.Cheng DT, et al. , Memorial Sloan Kettering-Integrated Mutation Profiling of Actionable Cancer Targets (MSK-IMPACT): a hybridization capture-based next-generation sequencing clinical assay for solid tumor molecular oncology. The Journal of molecular diagnostics, 2015. 17(3): p. 251–264. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Cristofanilli M, et al. , Fulvestrant plus palbociclib versus fulvestrant plus placebo for treatment of hormone-receptor-positive, HER2-negative metastatic breast cancer that progressed on previous endocrine therapy (PALOMA-3): final analysis of the multicentre, double-blind, phase 3 randomised controlled trial. The Lancet Oncology, 2016. 17(4): p. 425–439. [DOI] [PubMed] [Google Scholar]
- 27.Goetz MP, et al. , MONARCH 3: abemaciclib as initial therapy for advanced breast cancer. Journal of Clinical Oncology, 2017. 35(32): p. 3638–3646. [DOI] [PubMed] [Google Scholar]
- 28.Slamon D, et al. , Ribociclib plus fulvestrant for postmenopausal women with hormone receptor-positive, human epidermal growth factor receptor 2-negative advanced breast cancer in the phase III randomized MONALEESA-3 trial: updated overall survival. Annals of oncology, 2021. 32(8): p. 1015–1024. [DOI] [PubMed] [Google Scholar]
- 29.Ng TL, et al. , Preselection of lung cancer cases using FGFR1 mRNA and gene copy number for treatment with ponatinib. Clinical lung cancer, 2019. 20(1): p. e39–e51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Paik PK and Rudin CM, Missing the mark in FGFR1-amplified squamous cell cancer of the lung. 2016, Wiley Online Library. p. 2938–2940. [DOI] [PubMed] [Google Scholar]
- 31.Loriot Y, et al. , Phase 3 THOR study: Results of erdafitinib (erda) versus chemotherapy (chemo) in patients (pts) with advanced or metastatic urothelial cancer (mUC) with select fibroblast growth factor receptor alterations (FGFRalt). Journal of Clinical Oncology, 2023. [Google Scholar]
- 32.Rodón J, et al. , Pemigatinib in previously treated solid tumors with activating FGFR1–FGFR3 alterations: phase 2 FIGHT-207 basket trial. Nature medicine, 2024: p. 1–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Shaver TM, et al. , Diverse, biologically relevant, and targetable gene rearrangements in triple-negative breast cancer and other malignancies. Cancer research, 2016. 76(16): p. 4850–4860. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Wu Y-M, et al. , Identification of targetable FGFR gene fusions in diverse cancers. Cancer discovery, 2013. 3(6): p. 636–647. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data generated in this study are available within the article and its supplementary data files. Deidentified genomic sequencing data generated by Memorial Sloan Kettering IMPACT are available under restricted access to protect patient privacy in accordance with federal and state law. These data can be requested for research use from the corresponding author via execution of a data transfer agreement with MSKCC. Any additional data are available upon request from the corresponding author.