European Journal of Cancer (2014) 50, 3145–3152
Available at www.sciencedirect.com
ScienceDirect journal homepage: www.ejcancer.com
Clinical Trial
Phase 2 trial of dovitinib in patients with progressive FGFR3-mutated or FGFR3 wild-type advanced urothelial carcinoma Matthew I. Milowsky a,⇑, Christian Dittrich b, Ignacio Dura´n c, Satinder Jagdev d, Frederick E. Millard e, Christopher J. Sweeney f, Dean Bajorin g, Linda Cerbone h, David I. Quinn i, Walter M. Stadler j, Jonathan E. Rosenberg g, Melissa Lochheed k, Paramita Sen k, Matthew Squires l, Michael Shi k, Cora N. Sternberg h a Department of Medicine, Division of Hematology/Oncology, Lineberger Comprehensive Cancer Center, University of North Carolina, Chapel Hill, NC, USA b LBI-ACR VIEnna & ACR-ITR VIEnna, Center for Oncology and Haematology, Kaiser Franz Josef-Spital, Vienna, Austria c Centro Integral Oncologico Clara Campal, Universidad CEU San Pablo, Madrid, Spain d St. James’s Institute of Oncology, Leeds, UK e Department of Medicine, Division of Hematology–Oncology, University of California, San Diego, San Diego, CA, USA f Center for Genitourinary Oncology, Dana-Farber Cancer Institute, Boston, MA, USA g Department of Medicine, Memorial Sloan-Kettering Cancer Center, New York, NY, USA h Department of Medical Oncology, San Camillo and Forlanini Hospitals, Rome, Italy i Division of Medical Oncology, University of Southern California, Norris Comprehensive Cancer Center, Los Angeles, CA, USA j Department of Medicine, University of Chicago, Chicago, IL, USA k Novartis Pharmaceuticals Corporation, East Hanover, NJ, USA l Novartis Pharma AG, Basel, Switzerland
Received 25 June 2014; received in revised form 25 September 2014; accepted 10 October 2014 Available online 30 October 2014
KEYWORDS Urothelial carcinoma Dovitinib Receptor tyrosine kinase Fibroblast growth factor FGFR3 VEGFR
Abstract Background: Second-line treatment options for patients with advanced urothelial carcinoma (UC) are limited. Fibroblast growth factor receptor 3 (FGFR3) is dysregulated in UC by activating mutations or protein overexpression in non-mutant tumours. In this study, the efficacy, pharmacodynamics and safety of dovitinib—a broad-targeted inhibitor of tyrosine kinases, including FGFR3—were evaluated in patients with previously treated advanced UC with and without FGFR3 mutations. Methods: Forty-four adults with advanced UC who had progressed after one to three platinum-based and/or combination chemotherapy regimens were classified as having mutant
⇑ Corresponding author at: University of North Carolina at Chapel Hill, 170 Manning Drive, 3rd Floor, Physician’s Office Building, Chapel Hill, NC 27599-7305, USA. Tel.: +1 919 843 7942; fax: +1 919 966 6735. E-mail address: [email protected] (M.I. Milowsky).
http://dx.doi.org/10.1016/j.ejca.2014.10.013 0959-8049/Ó 2014 Elsevier Ltd. All rights reserved.
3146
M.I. Milowsky et al. / European Journal of Cancer 50 (2014) 3145–3152
(FGFR3MUT; n = 12), wild-type (FGFR3WT; n = 31), or unknown (n = 1) FGFR3 status. Patients received 500 mg dovitinib once daily on a 5-days-on/2-days-off schedule. The primary end-point of this two-stage study was the investigator-assessed overall response rate (ORR). Results: Most of the patients were men (75%) and over half of the patients were aged P65 years (61%). All patients had received P1 prior antineoplastic therapy for UC. The study was terminated at the end of stage 1, when it was determined by investigator review that the ORR of both the FGFR3MUT (0%; 95% confidence interval [CI], 0.0–26.5) and FGFR3WT (3.2%; 95% CI, 0.1–16.7) groups did not meet the criteria to continue to stage 2. The most common grade 3/4 adverse events, suspected to be study-drug related, included thrombocytopenia (9%), fatigue (9%), and asthenia (9%). Conclusion: Although generally well tolerated, dovitinib has very limited single-agent activity in patients with previously treated advanced UC, regardless of FGFR3 mutation status. clinicaltrials.gov NCT00790426. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Urothelial carcinoma (UC) accounts for P90% of cases of urinary bladder cancer [1]. Despite initial sensitivity to standard first-line combination platinum-based chemotherapy in patients with advanced disease, the overall prognosis is poor; the median survival is approximately 15 months in chemotherapy-treated patients [1,2]. Although taxanes are widely used in cisplatinrefractory patients, efficacy is modest and toxicities are limiting [2]. Vinflunine is registered for relapsed/refractory UC in Europe but is not approved in the United States. Thus, there is a significant need for effective and well-tolerated agents for patients with advanced, previously treated UC. A promising target in UC is fibroblast growth factor receptor 3 (FGFR3)—one of four highly conserved FGF receptor tyrosine kinases (RTKs) with known regulatory roles in tumour growth and survival [1]. Bladder cancer is associated with FGFR3 protein overexpression in FGFR3-mutant (85%) and -non-mutant (42%) tumours, and approximately 70% of low-grade non-invasive and 15% of high-grade UC tumours are associated with 11 different FGFR3-activating missense mutations [1,3,4]. Furthermore, transcriptome sequencing of UC tumours revealed recurrent fusion of FGFR3 with sister chromatid cohesion and segregation component TACC3 [5]. FGFRs also regulate angiogenesis (along with vascular endothelial growth factor receptor [VEGFR] and platelet-derived growth factor receptor [PDGFR]) [6]. The modest phase 2 clinical activity of sunitinib, a multitargeted inhibitor of RTKs (including VEGFR and PDGFR, but not FGFR), in previously treated metastatic UC suggests that the VEGFR axis may represent a viable target in advanced disease [7]. Therefore, broader inhibition of angiogenesisassociated RTKs, including FGFR, may provide more potent antitumour effects in patients with advanced UC, regardless of FGFR3 mutation status. Dovitinib (TKI258; Novartis Pharmaceuticals), an inhibitor of FGFR, VEGFR, PDGFRb, CSF-1R,
CKIT, RET, TrkA, and FLT3, has preclinical activity in FGFR3-mutant (FGFR3MUT), FGFR3-fused and FGFR3-overexpressing bladder cancer cell lines and mouse xenografts [8,9]. This phase 2 study evaluated dovitinib in patients with previously treated advanced FGFR3MUT or FGFR3 wild-type (FGFR3WT) UC. 2. Patients and methods 2.1. Patients Patients aged P18 years with histologically confirmed bladder, urethra, ureter or renal pelvis UC who had progressed after one to three platinum-based and/ or combination chemotherapy regimens were eligible for this study. All patients had P1 non-irradiated measurable lesion, archival tumour tissue available for central FGFR3 mutation analysis, adequate blood chemistry, acceptable cardiac and liver function and a World Health Organisation (WHO) performance status of 62. Patients with known or suspected brain metastases, history of another malignancy within 3 years (except adequately-treated cervical, prostate or non-melanomatous skin cancer), uncontrolled diarrhoea, or those receiving anticoagulant therapy were not eligible. All patients provided written informed consent. 2.2. Study design and treatment This phase 2, open-label, two-arm, Simon’s two-stage design, multicenter study evaluated the safety and efficacy of dovitinib in patients with previously treated advanced FGFR3MUT or FGFR3WT UC. The protocol and all amendments were reviewed by the Independent Ethics Committee or Institutional Review Board for each study site, and the study was conducted in accordance with the ethical principles of the Declaration of Helsinki. Patients received 500 mg dovitinib orally once daily on a 5-days-on/2-days-off schedule in 28-day cycles until
M.I. Milowsky et al. / European Journal of Cancer 50 (2014) 3145–3152
disease progression, start of new cancer therapy or unacceptable toxicity. For patients unable to tolerate 500 mg dovitinib, dose adjustments to 400 or 300 mg or interruptions of 621 days were permitted until toxicity resolution. The primary end-point was overall response rate (ORR) in FGFR3MUT and FGFRWT groups by investigator review. Secondary end-points were ORR by central radiology review, PFS and overall survival (OS) by investigator review, disease control rate (DCR), safety and tolerability in both groups. Exploratory objectives included pharmacokinetics and pharmacodynamic effects.
2.3. Patient assessments Tumour assessments were conducted every 8 weeks until disease progression. ORR was defined as the proportion of patients with complete response (CR) or partial response (PR) by investigator or central review using Response Evaluation Criteria in Solid Tumors 1.0. DCR was defined as the proportion of patients with CR or PR, or stable disease (SD) lasting P16 weeks after first dose, by investigator review. PFS was defined as the time from treatment initiation to first-documented progression or death due to any cause. For patients who did not progress or die at the analysis cutoff date, PFS was censored at the last adequate tumour assessment date before data cutoff or at the start date of a new antineoplastic therapy, whichever occurred first. OS, defined as the time from first dose to date of death due to any cause, was censored at the last contact date prior to data cutoff for patients alive at the analysis cutoff date. Adverse events (AEs) were recorded throughout the study. Complete blood counts, blood chemistry, urinalysis, vital signs, weight, WHO performance status, physical condition and cardiovascular function were regularly monitored. AEs, coded using Medical Dictionary for Regulatory Activities v15.0 terminology, were summarised by primary system organ class, preferred term and maximum grade according to the National Cancer Institute’s Common Terminology Criteria for Adverse Events v4.0.
2.4. FGFR3 mutation status SNaPshot analysis (Life Technologies) [10] was initially used to evaluate patients’ archival tissues for nine commonly occurring FGFR3 mutations (R248C, S249C, G372C, Y375C, A393E, K652E, K652M, K652T and K652Q), but was later replaced by Sanger sequencing when two atypical (not found in the literature) and unconfirmable SNaPshot-identified mutations were discovered during screening. SNaPshot-evaluated patient samples were reassessed by Sanger sequencing and FGFR3 mutation status was reclassified, if necessary.
3147
Patients with tissue not analysable by Sanger sequencing were classified as having ‘unknown’ FGFR3 status. 2.5. Pharmacokinetics and biomarker assessments Blood samples were collected on days 1 and 26 of cycle 1, pre- and postdose on day 26 of cycle 2, and on day 26 of cycle 3 for limited-sampling pharmacokinetic central analysis of plasma exposure to dovitinib. For plasma pharmacodynamic analysis, blood samples were collected at baseline, days 1 and 26 of cycle 1, day 26 of cycle 2, day 26 of every other subsequent cycle and at the end of treatment. Circulating growth factors (VEGF, placental growth factor [PlGF] and FGF23) and soluble receptors (sVEGFR1, sVEGFR2 and CKIT) were evaluated as core pharmacodynamic biomarkers and measured by enzyme-linked immunosorbent or multiplex assays (Meso Scale Discovery, Rockville, MD, United States of America (USA)). 2.6. Statistical methods Simon’s two-stage design [11] was used for each group to test the literature-supported null hypothesis ORR 6 0.10 [12] using a one-sided test with 10% significance and 90% power at the alternative ORR = 0.25. Twenty patients were planned for stage 1 and an additional 20 patients for stage 2. Safety and study analysis sets included patients receiving P1 dovitinib dose. Observation of P2 responders from stage 1 was required for stage 2 continuation. Descriptive statistics and frequency counts were used to summarise patient demographics, baseline characteristics and AEs. The Kaplan– Meier product-limit method was used to describe PFS and OS. An analysis of variance model with an estimated-parameter covariance matrix was used to calculate model-adjusted mean biomarker fold-changes from baseline (day 1, cycle 1) at defined time points. 3. Results 3.1. Patient characteristics and disposition Of 259 patients screened (Fig. 1), 231 had FGFR3 mutation data reported. Of these, 206 were FGFR3WT and 25 were FGFR3MUT (14 had S249C, seven had R248C, two had G372C, one had Y375C, and one had A393E). A total of 44 patients from 20 centres in eight countries were enrolled in the study (Table 1) and stratified by FGFR3 mutation status. Eight patients with SNaPshot-identified FGFR3 mutations were reclassified to FGFR3WT status by Sanger sequencing, and one patient with a tissue sample not analysable by Sanger sequencing was reclassified as having unknown FGFR3 status. This reclassification led to an overrecruitment of patients with FGFR3WT (n = 31) relative to
3148
M.I. Milowsky et al. / European Journal of Cancer 50 (2014) 3145–3152
Fig. 1. Flow diagram of progress through enrolment, stratification, follow-up and analysis in the trial. aScreen failure includes closure of the wildtype fibroblast growth factor receptor 3 (FGFR3WT) group.
FGFR3MUT (n = 12) status. The majority of patients were men (75%), aged P65 years (61%), and had UC of the bladder as the primary site (77%). All patients received prior antineoplastic therapy, most commonly surgery (98%) and/or chemotherapy (98%). In general, baseline characteristics were well-balanced between the FGFR3WT and FGFR3MUT groups; however, a WHO performance status P1 (64% versus 33%) and stage IV disease at diagnosis (55% versus 8%) were observed more often in patients with FGFR3WT than FGFR3MUT UC. All patients discontinued study drug (Fig. 1), most commonly due to disease progression (67% and 42%) or AEs (17% and 32%) in the FGFR3MUT and FGFR3WT groups, respectively. Dose reductions or delays were reported for 41% and 48% of patients, respectively, mostly due to AEs, including diarrhoea and asthenia.
3.2. Efficacy Primary efficacy during stage 1, by investigator assessment, showed no responses in the FGFR3MUT group and one PR in the FGFR3WT group (Table 2; Fig. A.1). SD was observed in 42% and 32% of patients in the FGFR3MUT and FGFR3WT groups, respectively; DCRs were similar between groups (25% versus 26%, respectively). Secondary-response analysis by central review showed similar overall results. The number of responders in each group failed to meet the criteria for study continuation to stage 2. Because most patients in the FGFR3MUT group did not receive >6 months of treatment, and meeting the response threshold to proceed to stage 2 was highly unlikely, the study was terminated. The median PFS by investigator review was 3 months (95% confidence interval [CI], 1.6–3.6) for
M.I. Milowsky et al. / European Journal of Cancer 50 (2014) 3145–3152
3149
Table 1 Patient characteristics by FGFR3 mutation status. Demographic variable
FGFR3MUT (n = 12)
FGFR3WT (n = 31)
All patients (N = 44)a
Age, median (range), years Sex, male/female, n (%) WHO performance status, n (%) 0 1 2 Primary site of cancer, n (%) Bladder Other Metastatic site of cancer, n (%) Lymph nodes Lung Liver Bone Pleural/peritoneum Bladder Other Number of organs involved, n (%) 1 2 P3 Risk factors, n (%)b 0 1 2 3 Time since initial diagnosis of primary site median (range), years Time since most recent relapse/recurrence, median (range), months Prior antineoplastic therapy, n (%) Any Surgery Radiotherapy Chemotherapy settingc None Adjuvant Neoadjuvant Metastatic Prior platinum, n (%) Carboplatin Cisplatin
67 (53–77) 9 (75)/3 (25)
67 (46–81) 23 (74)/8 (26)
67 (46–81) 33 (75)/11 (25)
8 (67) 3 (25) 1 (8)
11 (36) 15 (48) 5 (16)
19 (43) 19 (43) 6 (14)
7 (58) 5 (42)
27 (87) 4 (13)
34 (77) 10 (23)
8 6 3 1 1 1 7
(67) (50) (25) (8) (8) (8) (58)
14 (45) 15 (48) 10 (32) 10 (32) 5 (16) 2 (6) 18 (58)
23 (52) 21 (48) 13 (30) 12 (27) 6 (14) 3 (7) 26 (59)
3 (25) 4 (33) 5 (42)
7 (23) 12 (39) 12 (39)
10 (23) 16 (36) 18 (41)
6 (50) 5 (42) 1 (8) 0 3.0 (1.4–11.9)
9 (29) 12 (39) 8 (26) 2 (6) 1.5 (0.5–5.0)
15 (34) 18 (41) 9 (20) 2 (5) 1.9 (0.5–11.9)
1.5 (0.3–4.2)
1.8 (0.1–14.1)
1.8 (0.1–14.1)
12 (100) 12 (100) 2 (17)
31(100) 30 (97) 12 (39)
44 (100) 43 (98) 15 (34)
0 1 (8) 2 (17) 10 (83)
1 (3) 16 (52) 4 (13) 19 (61)
1 (2) 17 (39) 6 (14) 30 (68)
5 (42) 10 (83)
10 (32) 25 (81)
16 (36) 36 (82)
FGFR3MUT, mutant fibroblast growth factor receptor 3; FGFR3WT, wild-type fibroblast growth factor receptor 3; WHO, World Health Organisation. a Includes one patient with unknown FGFR3 mutation status. b Risk factors are Eastern Cooperative Oncology Group performance status >0, haemoglobin level 10% of patients (all grades).a Adverse event
All patients (N = 44)
Preferred term, n (%)
Any grade
Grade 3
Grade 4
Any Diarrhoea Nausea Decreased appetite Vomiting Fatigue Asthenia Rash Anaemia Thrombocytopenia Alanine aminotransferase increased Hypertension Aspartate aminotransferase increased Constipation Dysgeusia
41 (93) 29 (66) 26 (59) 16 (36) 16 (36) 14 (32) 13 (30) 10 (23) 7 (16) 7 (16) 6 (14) 6 (14) 5 (11) 5 (11) 5 (11)
25 (57) 3 (7) 0 0 1 (2) 4 (9) 4 (9) 2 (5) 2 (5) 3 (7) 1 (2) 0 1 (2) 1 (2) 0
3 0 0 0 0 0 0 0 0 1 1 0 1 0 0
(7)
(2) (2) (2)
a Patients with multiple occurrences of an adverse event are counted only once at the highest grade.
in 25% of patients—most commonly diarrhoea (7%), vomiting (7%) and nausea (5%). AEs were similar between the FGFR3WT and FGFR3MUT groups (data not shown). The most common grade 3/4 shifts from baseline in laboratory abnormalities included lymphopenia (17%), increased alkaline phosphatase (14%) or triglyceride
(14%) levels, and anaemia (11%). Five patients (11%), all in the FGFR3WT group, exhibited a notable increase in QTcF/QTCB between 30 and 60 ms; however, there were no reports of intervals P500 ms during the study. Of the eight on-treatment deaths, which included those occurring up to 30 days after discontinuation of dovitinib, seven were in the FGFR3WT group. Seven
M.I. Milowsky et al. / European Journal of Cancer 50 (2014) 3145–3152
deaths were attributed to advanced UC (including one due to central nervous system metastases) and one to sepsis occurring 10 days after the last dovitinib dose, which was not suspected to be related to the study drug. 3.4. Pharmacokinetics and biomarker dynamics Serum dovitinib concentrations after the 500-mg dose (14–453 ng/mL) were in the expected range [13,14]. Nearly all statistically significant mean fold-changes from baseline in serum biomarker levels with treatment occurred in the FGFR3WT group, including increases in PlGF and VEGF and decreases in sVEGFR2 and CKIT (Table A.1; Fig. A.3). Although FGFR3MUT PlGF levels increased above baseline at all time points, most changes were not significant. 4. Discussion Dovitinib inhibited tumour growth and proliferation in preclinical UC models with FGFR3 fusions, FGFR3activating mutations or FGFR3 overexpression [8,9]. In addition to FGFR, dovitinib targets the angiogenesis-associated receptors VEGFR and PDGFR [6] and was thus hypothesised to provide antitumour activity in patients with UC, irrespective of FGFR3 mutation status. However, in this two-stage study, dovitinib showed very limited clinical efficacy in patients with advanced UC, regardless of FGFR3 classification. In stage 1, only one patient in the FGFR3WT group showed a PR by investigator review, and most patients in the FGFR3MUT group received
Available at www.sciencedirect.com
ScienceDirect journal homepage: www.ejcancer.com
Clinical Trial
Phase 2 trial of dovitinib in patients with progressive FGFR3-mutated or FGFR3 wild-type advanced urothelial carcinoma Matthew I. Milowsky a,⇑, Christian Dittrich b, Ignacio Dura´n c, Satinder Jagdev d, Frederick E. Millard e, Christopher J. Sweeney f, Dean Bajorin g, Linda Cerbone h, David I. Quinn i, Walter M. Stadler j, Jonathan E. Rosenberg g, Melissa Lochheed k, Paramita Sen k, Matthew Squires l, Michael Shi k, Cora N. Sternberg h a Department of Medicine, Division of Hematology/Oncology, Lineberger Comprehensive Cancer Center, University of North Carolina, Chapel Hill, NC, USA b LBI-ACR VIEnna & ACR-ITR VIEnna, Center for Oncology and Haematology, Kaiser Franz Josef-Spital, Vienna, Austria c Centro Integral Oncologico Clara Campal, Universidad CEU San Pablo, Madrid, Spain d St. James’s Institute of Oncology, Leeds, UK e Department of Medicine, Division of Hematology–Oncology, University of California, San Diego, San Diego, CA, USA f Center for Genitourinary Oncology, Dana-Farber Cancer Institute, Boston, MA, USA g Department of Medicine, Memorial Sloan-Kettering Cancer Center, New York, NY, USA h Department of Medical Oncology, San Camillo and Forlanini Hospitals, Rome, Italy i Division of Medical Oncology, University of Southern California, Norris Comprehensive Cancer Center, Los Angeles, CA, USA j Department of Medicine, University of Chicago, Chicago, IL, USA k Novartis Pharmaceuticals Corporation, East Hanover, NJ, USA l Novartis Pharma AG, Basel, Switzerland
Received 25 June 2014; received in revised form 25 September 2014; accepted 10 October 2014 Available online 30 October 2014
KEYWORDS Urothelial carcinoma Dovitinib Receptor tyrosine kinase Fibroblast growth factor FGFR3 VEGFR
Abstract Background: Second-line treatment options for patients with advanced urothelial carcinoma (UC) are limited. Fibroblast growth factor receptor 3 (FGFR3) is dysregulated in UC by activating mutations or protein overexpression in non-mutant tumours. In this study, the efficacy, pharmacodynamics and safety of dovitinib—a broad-targeted inhibitor of tyrosine kinases, including FGFR3—were evaluated in patients with previously treated advanced UC with and without FGFR3 mutations. Methods: Forty-four adults with advanced UC who had progressed after one to three platinum-based and/or combination chemotherapy regimens were classified as having mutant
⇑ Corresponding author at: University of North Carolina at Chapel Hill, 170 Manning Drive, 3rd Floor, Physician’s Office Building, Chapel Hill, NC 27599-7305, USA. Tel.: +1 919 843 7942; fax: +1 919 966 6735. E-mail address: [email protected] (M.I. Milowsky).
http://dx.doi.org/10.1016/j.ejca.2014.10.013 0959-8049/Ó 2014 Elsevier Ltd. All rights reserved.
3146
M.I. Milowsky et al. / European Journal of Cancer 50 (2014) 3145–3152
(FGFR3MUT; n = 12), wild-type (FGFR3WT; n = 31), or unknown (n = 1) FGFR3 status. Patients received 500 mg dovitinib once daily on a 5-days-on/2-days-off schedule. The primary end-point of this two-stage study was the investigator-assessed overall response rate (ORR). Results: Most of the patients were men (75%) and over half of the patients were aged P65 years (61%). All patients had received P1 prior antineoplastic therapy for UC. The study was terminated at the end of stage 1, when it was determined by investigator review that the ORR of both the FGFR3MUT (0%; 95% confidence interval [CI], 0.0–26.5) and FGFR3WT (3.2%; 95% CI, 0.1–16.7) groups did not meet the criteria to continue to stage 2. The most common grade 3/4 adverse events, suspected to be study-drug related, included thrombocytopenia (9%), fatigue (9%), and asthenia (9%). Conclusion: Although generally well tolerated, dovitinib has very limited single-agent activity in patients with previously treated advanced UC, regardless of FGFR3 mutation status. clinicaltrials.gov NCT00790426. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Urothelial carcinoma (UC) accounts for P90% of cases of urinary bladder cancer [1]. Despite initial sensitivity to standard first-line combination platinum-based chemotherapy in patients with advanced disease, the overall prognosis is poor; the median survival is approximately 15 months in chemotherapy-treated patients [1,2]. Although taxanes are widely used in cisplatinrefractory patients, efficacy is modest and toxicities are limiting [2]. Vinflunine is registered for relapsed/refractory UC in Europe but is not approved in the United States. Thus, there is a significant need for effective and well-tolerated agents for patients with advanced, previously treated UC. A promising target in UC is fibroblast growth factor receptor 3 (FGFR3)—one of four highly conserved FGF receptor tyrosine kinases (RTKs) with known regulatory roles in tumour growth and survival [1]. Bladder cancer is associated with FGFR3 protein overexpression in FGFR3-mutant (85%) and -non-mutant (42%) tumours, and approximately 70% of low-grade non-invasive and 15% of high-grade UC tumours are associated with 11 different FGFR3-activating missense mutations [1,3,4]. Furthermore, transcriptome sequencing of UC tumours revealed recurrent fusion of FGFR3 with sister chromatid cohesion and segregation component TACC3 [5]. FGFRs also regulate angiogenesis (along with vascular endothelial growth factor receptor [VEGFR] and platelet-derived growth factor receptor [PDGFR]) [6]. The modest phase 2 clinical activity of sunitinib, a multitargeted inhibitor of RTKs (including VEGFR and PDGFR, but not FGFR), in previously treated metastatic UC suggests that the VEGFR axis may represent a viable target in advanced disease [7]. Therefore, broader inhibition of angiogenesisassociated RTKs, including FGFR, may provide more potent antitumour effects in patients with advanced UC, regardless of FGFR3 mutation status. Dovitinib (TKI258; Novartis Pharmaceuticals), an inhibitor of FGFR, VEGFR, PDGFRb, CSF-1R,
CKIT, RET, TrkA, and FLT3, has preclinical activity in FGFR3-mutant (FGFR3MUT), FGFR3-fused and FGFR3-overexpressing bladder cancer cell lines and mouse xenografts [8,9]. This phase 2 study evaluated dovitinib in patients with previously treated advanced FGFR3MUT or FGFR3 wild-type (FGFR3WT) UC. 2. Patients and methods 2.1. Patients Patients aged P18 years with histologically confirmed bladder, urethra, ureter or renal pelvis UC who had progressed after one to three platinum-based and/ or combination chemotherapy regimens were eligible for this study. All patients had P1 non-irradiated measurable lesion, archival tumour tissue available for central FGFR3 mutation analysis, adequate blood chemistry, acceptable cardiac and liver function and a World Health Organisation (WHO) performance status of 62. Patients with known or suspected brain metastases, history of another malignancy within 3 years (except adequately-treated cervical, prostate or non-melanomatous skin cancer), uncontrolled diarrhoea, or those receiving anticoagulant therapy were not eligible. All patients provided written informed consent. 2.2. Study design and treatment This phase 2, open-label, two-arm, Simon’s two-stage design, multicenter study evaluated the safety and efficacy of dovitinib in patients with previously treated advanced FGFR3MUT or FGFR3WT UC. The protocol and all amendments were reviewed by the Independent Ethics Committee or Institutional Review Board for each study site, and the study was conducted in accordance with the ethical principles of the Declaration of Helsinki. Patients received 500 mg dovitinib orally once daily on a 5-days-on/2-days-off schedule in 28-day cycles until
M.I. Milowsky et al. / European Journal of Cancer 50 (2014) 3145–3152
disease progression, start of new cancer therapy or unacceptable toxicity. For patients unable to tolerate 500 mg dovitinib, dose adjustments to 400 or 300 mg or interruptions of 621 days were permitted until toxicity resolution. The primary end-point was overall response rate (ORR) in FGFR3MUT and FGFRWT groups by investigator review. Secondary end-points were ORR by central radiology review, PFS and overall survival (OS) by investigator review, disease control rate (DCR), safety and tolerability in both groups. Exploratory objectives included pharmacokinetics and pharmacodynamic effects.
2.3. Patient assessments Tumour assessments were conducted every 8 weeks until disease progression. ORR was defined as the proportion of patients with complete response (CR) or partial response (PR) by investigator or central review using Response Evaluation Criteria in Solid Tumors 1.0. DCR was defined as the proportion of patients with CR or PR, or stable disease (SD) lasting P16 weeks after first dose, by investigator review. PFS was defined as the time from treatment initiation to first-documented progression or death due to any cause. For patients who did not progress or die at the analysis cutoff date, PFS was censored at the last adequate tumour assessment date before data cutoff or at the start date of a new antineoplastic therapy, whichever occurred first. OS, defined as the time from first dose to date of death due to any cause, was censored at the last contact date prior to data cutoff for patients alive at the analysis cutoff date. Adverse events (AEs) were recorded throughout the study. Complete blood counts, blood chemistry, urinalysis, vital signs, weight, WHO performance status, physical condition and cardiovascular function were regularly monitored. AEs, coded using Medical Dictionary for Regulatory Activities v15.0 terminology, were summarised by primary system organ class, preferred term and maximum grade according to the National Cancer Institute’s Common Terminology Criteria for Adverse Events v4.0.
2.4. FGFR3 mutation status SNaPshot analysis (Life Technologies) [10] was initially used to evaluate patients’ archival tissues for nine commonly occurring FGFR3 mutations (R248C, S249C, G372C, Y375C, A393E, K652E, K652M, K652T and K652Q), but was later replaced by Sanger sequencing when two atypical (not found in the literature) and unconfirmable SNaPshot-identified mutations were discovered during screening. SNaPshot-evaluated patient samples were reassessed by Sanger sequencing and FGFR3 mutation status was reclassified, if necessary.
3147
Patients with tissue not analysable by Sanger sequencing were classified as having ‘unknown’ FGFR3 status. 2.5. Pharmacokinetics and biomarker assessments Blood samples were collected on days 1 and 26 of cycle 1, pre- and postdose on day 26 of cycle 2, and on day 26 of cycle 3 for limited-sampling pharmacokinetic central analysis of plasma exposure to dovitinib. For plasma pharmacodynamic analysis, blood samples were collected at baseline, days 1 and 26 of cycle 1, day 26 of cycle 2, day 26 of every other subsequent cycle and at the end of treatment. Circulating growth factors (VEGF, placental growth factor [PlGF] and FGF23) and soluble receptors (sVEGFR1, sVEGFR2 and CKIT) were evaluated as core pharmacodynamic biomarkers and measured by enzyme-linked immunosorbent or multiplex assays (Meso Scale Discovery, Rockville, MD, United States of America (USA)). 2.6. Statistical methods Simon’s two-stage design [11] was used for each group to test the literature-supported null hypothesis ORR 6 0.10 [12] using a one-sided test with 10% significance and 90% power at the alternative ORR = 0.25. Twenty patients were planned for stage 1 and an additional 20 patients for stage 2. Safety and study analysis sets included patients receiving P1 dovitinib dose. Observation of P2 responders from stage 1 was required for stage 2 continuation. Descriptive statistics and frequency counts were used to summarise patient demographics, baseline characteristics and AEs. The Kaplan– Meier product-limit method was used to describe PFS and OS. An analysis of variance model with an estimated-parameter covariance matrix was used to calculate model-adjusted mean biomarker fold-changes from baseline (day 1, cycle 1) at defined time points. 3. Results 3.1. Patient characteristics and disposition Of 259 patients screened (Fig. 1), 231 had FGFR3 mutation data reported. Of these, 206 were FGFR3WT and 25 were FGFR3MUT (14 had S249C, seven had R248C, two had G372C, one had Y375C, and one had A393E). A total of 44 patients from 20 centres in eight countries were enrolled in the study (Table 1) and stratified by FGFR3 mutation status. Eight patients with SNaPshot-identified FGFR3 mutations were reclassified to FGFR3WT status by Sanger sequencing, and one patient with a tissue sample not analysable by Sanger sequencing was reclassified as having unknown FGFR3 status. This reclassification led to an overrecruitment of patients with FGFR3WT (n = 31) relative to
3148
M.I. Milowsky et al. / European Journal of Cancer 50 (2014) 3145–3152
Fig. 1. Flow diagram of progress through enrolment, stratification, follow-up and analysis in the trial. aScreen failure includes closure of the wildtype fibroblast growth factor receptor 3 (FGFR3WT) group.
FGFR3MUT (n = 12) status. The majority of patients were men (75%), aged P65 years (61%), and had UC of the bladder as the primary site (77%). All patients received prior antineoplastic therapy, most commonly surgery (98%) and/or chemotherapy (98%). In general, baseline characteristics were well-balanced between the FGFR3WT and FGFR3MUT groups; however, a WHO performance status P1 (64% versus 33%) and stage IV disease at diagnosis (55% versus 8%) were observed more often in patients with FGFR3WT than FGFR3MUT UC. All patients discontinued study drug (Fig. 1), most commonly due to disease progression (67% and 42%) or AEs (17% and 32%) in the FGFR3MUT and FGFR3WT groups, respectively. Dose reductions or delays were reported for 41% and 48% of patients, respectively, mostly due to AEs, including diarrhoea and asthenia.
3.2. Efficacy Primary efficacy during stage 1, by investigator assessment, showed no responses in the FGFR3MUT group and one PR in the FGFR3WT group (Table 2; Fig. A.1). SD was observed in 42% and 32% of patients in the FGFR3MUT and FGFR3WT groups, respectively; DCRs were similar between groups (25% versus 26%, respectively). Secondary-response analysis by central review showed similar overall results. The number of responders in each group failed to meet the criteria for study continuation to stage 2. Because most patients in the FGFR3MUT group did not receive >6 months of treatment, and meeting the response threshold to proceed to stage 2 was highly unlikely, the study was terminated. The median PFS by investigator review was 3 months (95% confidence interval [CI], 1.6–3.6) for
M.I. Milowsky et al. / European Journal of Cancer 50 (2014) 3145–3152
3149
Table 1 Patient characteristics by FGFR3 mutation status. Demographic variable
FGFR3MUT (n = 12)
FGFR3WT (n = 31)
All patients (N = 44)a
Age, median (range), years Sex, male/female, n (%) WHO performance status, n (%) 0 1 2 Primary site of cancer, n (%) Bladder Other Metastatic site of cancer, n (%) Lymph nodes Lung Liver Bone Pleural/peritoneum Bladder Other Number of organs involved, n (%) 1 2 P3 Risk factors, n (%)b 0 1 2 3 Time since initial diagnosis of primary site median (range), years Time since most recent relapse/recurrence, median (range), months Prior antineoplastic therapy, n (%) Any Surgery Radiotherapy Chemotherapy settingc None Adjuvant Neoadjuvant Metastatic Prior platinum, n (%) Carboplatin Cisplatin
67 (53–77) 9 (75)/3 (25)
67 (46–81) 23 (74)/8 (26)
67 (46–81) 33 (75)/11 (25)
8 (67) 3 (25) 1 (8)
11 (36) 15 (48) 5 (16)
19 (43) 19 (43) 6 (14)
7 (58) 5 (42)
27 (87) 4 (13)
34 (77) 10 (23)
8 6 3 1 1 1 7
(67) (50) (25) (8) (8) (8) (58)
14 (45) 15 (48) 10 (32) 10 (32) 5 (16) 2 (6) 18 (58)
23 (52) 21 (48) 13 (30) 12 (27) 6 (14) 3 (7) 26 (59)
3 (25) 4 (33) 5 (42)
7 (23) 12 (39) 12 (39)
10 (23) 16 (36) 18 (41)
6 (50) 5 (42) 1 (8) 0 3.0 (1.4–11.9)
9 (29) 12 (39) 8 (26) 2 (6) 1.5 (0.5–5.0)
15 (34) 18 (41) 9 (20) 2 (5) 1.9 (0.5–11.9)
1.5 (0.3–4.2)
1.8 (0.1–14.1)
1.8 (0.1–14.1)
12 (100) 12 (100) 2 (17)
31(100) 30 (97) 12 (39)
44 (100) 43 (98) 15 (34)
0 1 (8) 2 (17) 10 (83)
1 (3) 16 (52) 4 (13) 19 (61)
1 (2) 17 (39) 6 (14) 30 (68)
5 (42) 10 (83)
10 (32) 25 (81)
16 (36) 36 (82)
FGFR3MUT, mutant fibroblast growth factor receptor 3; FGFR3WT, wild-type fibroblast growth factor receptor 3; WHO, World Health Organisation. a Includes one patient with unknown FGFR3 mutation status. b Risk factors are Eastern Cooperative Oncology Group performance status >0, haemoglobin level 10% of patients (all grades).a Adverse event
All patients (N = 44)
Preferred term, n (%)
Any grade
Grade 3
Grade 4
Any Diarrhoea Nausea Decreased appetite Vomiting Fatigue Asthenia Rash Anaemia Thrombocytopenia Alanine aminotransferase increased Hypertension Aspartate aminotransferase increased Constipation Dysgeusia
41 (93) 29 (66) 26 (59) 16 (36) 16 (36) 14 (32) 13 (30) 10 (23) 7 (16) 7 (16) 6 (14) 6 (14) 5 (11) 5 (11) 5 (11)
25 (57) 3 (7) 0 0 1 (2) 4 (9) 4 (9) 2 (5) 2 (5) 3 (7) 1 (2) 0 1 (2) 1 (2) 0
3 0 0 0 0 0 0 0 0 1 1 0 1 0 0
(7)
(2) (2) (2)
a Patients with multiple occurrences of an adverse event are counted only once at the highest grade.
in 25% of patients—most commonly diarrhoea (7%), vomiting (7%) and nausea (5%). AEs were similar between the FGFR3WT and FGFR3MUT groups (data not shown). The most common grade 3/4 shifts from baseline in laboratory abnormalities included lymphopenia (17%), increased alkaline phosphatase (14%) or triglyceride
(14%) levels, and anaemia (11%). Five patients (11%), all in the FGFR3WT group, exhibited a notable increase in QTcF/QTCB between 30 and 60 ms; however, there were no reports of intervals P500 ms during the study. Of the eight on-treatment deaths, which included those occurring up to 30 days after discontinuation of dovitinib, seven were in the FGFR3WT group. Seven
M.I. Milowsky et al. / European Journal of Cancer 50 (2014) 3145–3152
deaths were attributed to advanced UC (including one due to central nervous system metastases) and one to sepsis occurring 10 days after the last dovitinib dose, which was not suspected to be related to the study drug. 3.4. Pharmacokinetics and biomarker dynamics Serum dovitinib concentrations after the 500-mg dose (14–453 ng/mL) were in the expected range [13,14]. Nearly all statistically significant mean fold-changes from baseline in serum biomarker levels with treatment occurred in the FGFR3WT group, including increases in PlGF and VEGF and decreases in sVEGFR2 and CKIT (Table A.1; Fig. A.3). Although FGFR3MUT PlGF levels increased above baseline at all time points, most changes were not significant. 4. Discussion Dovitinib inhibited tumour growth and proliferation in preclinical UC models with FGFR3 fusions, FGFR3activating mutations or FGFR3 overexpression [8,9]. In addition to FGFR, dovitinib targets the angiogenesis-associated receptors VEGFR and PDGFR [6] and was thus hypothesised to provide antitumour activity in patients with UC, irrespective of FGFR3 mutation status. However, in this two-stage study, dovitinib showed very limited clinical efficacy in patients with advanced UC, regardless of FGFR3 classification. In stage 1, only one patient in the FGFR3WT group showed a PR by investigator review, and most patients in the FGFR3MUT group received
