Thank you very much for the invitation to share EMBRACA. Talazoparib is a highly potent dual-mechanism PARP inhibitor. It inhibits the PARP enzyme and traps PARP on single-stranded DNA breaks. It prevents the repair of DNA damage and results in cell death. The phase I trial established a tolerable dose of 1mg orally once daily for continuous dosing. The phase II ABRAZO trials showed encouraging efficacy in safety in patients with germline BRCA1 and 2 mutations with either a prior platinum therapy or at least three prior cytotoxic regimens.
Here is the study design. EMBRACA is a phase III international open label study that randomised 431 patients in 16 countries and 145 sites. Patients were eligible with locally advanced or metastatic HER2 positive metastatic breast cancer and germline BRCA1 or 2 mutations. Patients were stratified on three factors: number of prior chemotherapy regimens, tissue receptor subtype, and history of central nervous system metastases. This was a two-to-one randomisation to talazoparib, which is a once daily, single dosing daily of 1mg, to physician’s choice of therapy of either capecitabine, eribulin, gemcitabine or vinorelbine. The primary endpoint was progression free survival by RECIST by blinded central review with key secondary efficacy endpoints of overall survival, overall response rate by the investigator, and safety. We’re also going to talk about two of the exploratory endpoints today which is duration of response and quality of life.
Here are the baseline characteristics in the intention to treat population. As you can see, patients were well-balanced between the two arms but I will point out a slight increase in the percentage of patients who were randomised to talazoparib who were either under the age of 50, as well as a higher percentage with a shorter disease free interval of less than 12 months. I also want to point out that patients with treated CNS metastases were eligible, with 15% of patients in the talazoparib arm and almost 14% in the physician’s choice. Here is the patient disposition. I do want to point out that this was an open label trial because there were both oral and IV chemotherapies in the physician’s choice arm. Because of that, eighteen patients who were randomised to the physician’s choice arm did not receive study drug, compared to one in the talazoparib arm. I would also point out that patients still have a higher percentage who are ongoing still receiving talazoparib at the time of this review, as well as a further increase in patients still ongoing in the long-term survival follow up. For study drug exposure the protocol specific physician’s choice was determined prior to randomisation for each patient, you can see the majority of patients received capecitabine and eribulin, with a smaller percentage with gemcitabine and vinorelbine. The median relative dose intensity was 87.2%.
Here are the results. The primary endpoint was progression free survival by blinded central review. As you can see, the curves do separate early and the separation continues, with a hazard ratio of 0.54 which was statistically significant with a median of 8.6 months versus 5.6 months. This is with a median follow-up of 11.2 months. In key subgroup analysis it remains consistent, with all favouring talazoparib, with only the prior platinum with a confidence interval that passes one. This was a pre-planned interim overall survival analysis as a secondary endpoint if the primary endpoint was positive. Here we present the interim overall survival analysis with 51% of projected events. The hazard ratio does favour talazoparib at 0.76, it does not reach statistical significance, but we can notice that the curves do separate at the end and so it will be important to follow this as the data matures. A landmark analysis shows that at 24 months of survival probability is 45% at talazoparib versus 37% at overall physician’s choice.
Here are the secondary and exploratory endpoints. There were confirmed complete responses in the talazoparib arm and a higher percentage with partial response. For the objective response by the investigator with measurable disease, the odds ratio was 4.99, which was statistically significant, and a clinical benefit rate at 24 weeks in the intention to treat population with an odds ratio of 4.28, also statistically significant. In the duration of response by the investigator assessment, we do see the hazard ratio favouring talazoparib at 0.43 that was statistically significant with a median of 5.4 months versus 3.1 months. The one-year probability of sustained response was 23% for patients on talazoparib versus zero for patients on physician’s choice of therapy.
I do want to point out the toxicity related to talazoparib; the most common toxicity was anaemia. This was managed clinically, and only two patients had to be removed from study due to toxicity from anaemia. Neutropenia was higher in the physician’s choice, we can see that the rate of febrile neutropenia was low in both arms. There were no diagnoses of either MDS or AML in patients randomised to talazoparib, and one patient diagnosed with AML who received capecitabine while on the therapy. For the non-hematologic toxicities the most common, fatigue, nausea and headache, I did also want to also point out alopecia which was equal between the two arms, but the vast majority on talazoparib was grade 1, with only 2.4% being grade 2. There was higher hand foot, there was higher diarrhoea, and I think that that really relayed the higher percentage of using capecitabine in the physician’s choice arm.
In summary of adverse events, these are similar between the two arms with only a very small percentage of adverse events resulting in permanent drug discontinuation. I also wanted to highlight the patient reported global health status. Patients who were randomised to talazoparib had an improvement in their global health status versus patients who had a deterioration when randomised to physician’s choice of therapy. There was a statistically significant improvement in the estimated overall mean change from baseline for those patients on talazoparib versus physician’s choice. The time to deterioration showed a statistically significant delay in the time to clinically meaningful deterioration – that’s a greater than or equal to ten-point decrease that persisted – with a hazard ratio of 0.38 and a median of 24.3 months versus 6.3 months.
In conclusion, EMBRACA is the largest randomised trial evaluating a PARP inhibitor in patients with advanced breast cancer and a germline BRCA1 or 2 mutation. It resulted in prolonged progression-free survival versus physician’s choice of therapy with a hazard ratio of 0.54. All key secondary efficacy endpoints demonstrated benefit with talazoparib. The overall survival is immature, with only 51% of projected events, and it will be important to follow this as this data matures. The global health safety and quality of life showed overall improvement from baseline and a delay in the time to clinically meaningful deterioration in patients receiving talazoparib, with a hazard ratio of 0.38. Talazoparib was generally well-tolerated with minimal non-hematologic toxicity and few adverse events resulting in treatment discontinuation.