Results of the phase IIa RADICAL trial of the FGFR inhibitor AZD4547 in endocrine resistant breast cancer

Abstract

We conducted a phase IIa, multi-centre, open label, single arm study (RADICAL; NCT01791985) of AZD4547 (a potent and selective inhibitor of Fibroblast Growth Factor Receptor (FGFR)-1, 2 and 3 receptor tyrosine kinases) administered with anastrozole or letrozole in estrogen receptor positive metastatic breast cancer patients who had become resistant to aromatase inhibitors. After a safety run-in study to assess safety and tolerability, we recruited 52 patients. The primary endpoint was change in tumour size at 12 weeks, and secondary endpoints were to assess response at 6 weeks, 20 weeks and every 8 weeks thereafter and tolerability of the combined treatment. Two partial responses (PR) and 19 stable disease (SD) patients were observed at the 12-week time point. At 28 weeks, according to centrally reviewed Response Evaluation Criteria in Solid Tumours (RECIST) criteria, five PR and 8 SD patients were observed in 50 assessable cases. Overall, objective response rate (5 PR) was of 10%, meeting the pre-specified endpoint. Fourteen patients discontinued due to adverse events. Eleven patients had retinal pigment epithelial detachments which was asymptomatic and reversible in all but one patient. Exploratory ribonucleic acid sequencing (RNA-Seq) analysis was done on patients' samples: 6 differentially-expressed-genes could distinguish those who benefited from the addition of AZD4547.

Conflict of interest statement

N.R. is co-founder and officer of Inivata, a company that commercialises technologies for ctDNA analysis. D.L. consults to Astrazeneca (AZ) and was an employee in the past. His research team has also received grant funding from AZ. He also has a Director role and is an employee of his own consultancy company DeLondra Oncology Ltd. RCC currently has a grant from AZ. I.M. consults to AZ. R.B. consults to AZ and has received grant funding, travel, accommodation, expenses from AZ. A.A. has received grant funding to her Institution from AZ. P.B., I.Z., X.L., J.P.L.K., H.N., M.S., N.C., E.P., L.M., A.L., F.M., O.P., R.A., J.G.C. and V.F. have indicated no competing interests.

© 2022. The Author(s).

Figures

Fig. 1. Boxplots of phosphate changes, showing the proportion of phosphate change at different time points using cycle 1 day 1 phosphate as baseline.

Two patients without follow-up data are not included in the graph. The centre lines are defined by the median. The ‘outsider’ dots are the data points that fall outside the line called “whiskers” in the boxplot. The whiskers include all data points within 1.5 IQR (Interquartile range) of the nearer quartile (e.g. 3rd Quartile or 75th percentile). Source data are provided as a Source Data file.

Fig. 2. AZD4547 in metastatic breast cancer: time on treatment and best overall response.

a Time on Treatment. Each bar represents one subject in the study and their length of treatment in months. *Subject with the longest follow-up has 43 months of treatment duration. § Two subjects had known partial response at week 12 follow-up. Eight subjects (black squares) had known stable disease at week 28 (or 7th month). Three subjects (black diamonds) had known partial response at week 28, while two subjects had known partial response at an earlier follow-up time. During the whole study period, 5 subjects had partial response at any time point. A durable responder (red square) is a subject who has confirmed response for at least 6 months. The two subjects marked with continued response (black arrows) had partial response as their last known response prior to their withdrawal from the study. It is not known when the partial response in these subjects ended. Source data are provided as a Source Data file. b Treatment Response – Waterfall Plot. Best overall response and maximum percentage of tumour reduction based on central review of subjects with at least two tumour measurements (n = 45); Each bar represents one subject in the study. Subjects with only baseline tumour measurement or who have withdrawn before follow-up are not included in the graph. Source data are provided as a Source Data file. Two subjects (black squares) had partial response at week 12. *Three subjects had partial response at week 28 among the five subjects who experienced partial response at any time point during the study. Two of the subjects with partial response (green bars without *) have stopped treatment prior to week 28. § Eight subjects had stable disease at week 28.

Fig. 3. Modifications in the ctDNA tumour fraction measured with the t-MAD score from the baseline and early post-treatment samples (~week 4 and week 6).

p values for the Wilcoxon tests are indicated above the boxplots. Samples are separated based on the best overall response on RECIST (blue = PR, yellow = SD, grey = PD). There were n = 44 biologically different samples at baseline, 35 at week 4 and 34 at week 6. All samples are indicated by dots. Boxplot centre is the median, bounds of box represent 25 and 75 percentile, lines are 10 and 90 percentile. Source data are provided as a Source Data file.