Is Platinum the Silver Lining for TNBC?
Dr. Rugo: In adaptive randomization, if patients are eligible, they are randomized. Then they sign another consent. They learn about the side effects of the agent and then they start their treatment. They have a baseline MRI and another after 3-4 weeks between the end of paclitaxel and beginning of the anthracycline, and then before surgery. Those MRI results are used and fed back into a predetermined algorithm, which randomly assigns patients to specific signatures on the basis of their response. So we are continuously learning from the previous patient and benefiting the next patient. At the end, we get the pCR data. That information is fed back into the MRI information as confirmation, so then if we see as we go along that a specific signature -- for example, triple-negative or hormone receptive-positive, HER2-negative, or HER2-positive -- derives more benefit than another, then more patients with that signature will be assigned to that arm.
Adaptive randomization is a novel idea that is based on success. A drug graduates from I-SPY if there is at least an 85% predicted probability of success with the new regimen compared with controls in the 300-patient, phase 3, 1:1 randomized trial. That's why the combination that I studied graduated. It's the combination of the oral PARP inhibitor veliparib and carboplatin along with paclitaxel and anthracycline.
Dr. Miller: What was the difference in pCR? This is a randomized phase 2 study. It's about learning and selecting, so it graduates, but still needs to be put to the test of the 300-patient trial. What sort of difference with this dual DNA-damaging hit did you get?
Dr. Rugo: That is a good question. In an adaptively randomized trial, you can't just calculate a pCR rate as N over the total. You are continuously changing the number of patients being randomized based on signature, so you have to weight it, and you get a distribution of pCR. If your curve looks like this, the top of the curve is your point estimate for pCR with a range. We got a point estimate, and those differences were quite striking. For the triple-negative group, it was 26%, as determined by pathologists who were trained in using the RCB approach (a very careful determination of pCR in breast and axilla) vs 52% in the patients who got veliparib and carboplatin. That results in a 99% predictive probability that the experimental arm is superior to the control arm, and a 92% chance that this will be superior in a 300-patient phase 3 trial.
Dr. Miller: With that probability of success, are you going to do the 300-patient trial or is that not needed anymore?
Dr. Rugo: It is needed because we have the trial with the carboplatin; we have some interesting data, in addition, with gemcitabine/carboplatin from the study that Melinda Telli presented. Many biomarkers are being looked at, and indeed I-SPY 2 is a biomarker-rich trial, so Funmi Olopade and Laura van 't Veer are looking at markers of DNA damage, including BRCA status, in terms of trying to figure out which subgroup within triple-negative will benefit more. We still need to show that adding veliparib to carboplatin is better than carboplatin alone. That is the design for a phase 3 trial that is planned by AbbVie, who makes and markets veliparib as an experimental agent with an international collaboration. I-SPY 2 is designed to lead into these trials, so Laura Esserman and the whole I-SPY 2 group have been working to design I-SPY 3, which will be a fluid transition from agents that graduate into a planned trial, in a similar mechanism, which will lead to the final results that we need to get these drugs to the patients who need them sooner.
Dr. Miller: You were an earlier believer in the power of platinums, and they have come a long way in the last year.
Dr. Rugo: Indeed.
Veliparib and Carboplatin Make the Grade
Dr. Rugo: In adaptive randomization, if patients are eligible, they are randomized. Then they sign another consent. They learn about the side effects of the agent and then they start their treatment. They have a baseline MRI and another after 3-4 weeks between the end of paclitaxel and beginning of the anthracycline, and then before surgery. Those MRI results are used and fed back into a predetermined algorithm, which randomly assigns patients to specific signatures on the basis of their response. So we are continuously learning from the previous patient and benefiting the next patient. At the end, we get the pCR data. That information is fed back into the MRI information as confirmation, so then if we see as we go along that a specific signature -- for example, triple-negative or hormone receptive-positive, HER2-negative, or HER2-positive -- derives more benefit than another, then more patients with that signature will be assigned to that arm.
Adaptive randomization is a novel idea that is based on success. A drug graduates from I-SPY if there is at least an 85% predicted probability of success with the new regimen compared with controls in the 300-patient, phase 3, 1:1 randomized trial. That's why the combination that I studied graduated. It's the combination of the oral PARP inhibitor veliparib and carboplatin along with paclitaxel and anthracycline.
Dr. Miller: What was the difference in pCR? This is a randomized phase 2 study. It's about learning and selecting, so it graduates, but still needs to be put to the test of the 300-patient trial. What sort of difference with this dual DNA-damaging hit did you get?
Dr. Rugo: That is a good question. In an adaptively randomized trial, you can't just calculate a pCR rate as N over the total. You are continuously changing the number of patients being randomized based on signature, so you have to weight it, and you get a distribution of pCR. If your curve looks like this, the top of the curve is your point estimate for pCR with a range. We got a point estimate, and those differences were quite striking. For the triple-negative group, it was 26%, as determined by pathologists who were trained in using the RCB approach (a very careful determination of pCR in breast and axilla) vs 52% in the patients who got veliparib and carboplatin. That results in a 99% predictive probability that the experimental arm is superior to the control arm, and a 92% chance that this will be superior in a 300-patient phase 3 trial.
Dr. Miller: With that probability of success, are you going to do the 300-patient trial or is that not needed anymore?
Dr. Rugo: It is needed because we have the trial with the carboplatin; we have some interesting data, in addition, with gemcitabine/carboplatin from the study that Melinda Telli presented. Many biomarkers are being looked at, and indeed I-SPY 2 is a biomarker-rich trial, so Funmi Olopade and Laura van 't Veer are looking at markers of DNA damage, including BRCA status, in terms of trying to figure out which subgroup within triple-negative will benefit more. We still need to show that adding veliparib to carboplatin is better than carboplatin alone. That is the design for a phase 3 trial that is planned by AbbVie, who makes and markets veliparib as an experimental agent with an international collaboration. I-SPY 2 is designed to lead into these trials, so Laura Esserman and the whole I-SPY 2 group have been working to design I-SPY 3, which will be a fluid transition from agents that graduate into a planned trial, in a similar mechanism, which will lead to the final results that we need to get these drugs to the patients who need them sooner.
Dr. Miller: You were an earlier believer in the power of platinums, and they have come a long way in the last year.
Dr. Rugo: Indeed.
SHARE
![Pediatric Supportive Care (PDQ®): Supportive care - Patient Information [NCI]-Overview](https://cdnus.oss-us-west-1.aliyuncs.com/libs/images/NewsJpg/cate_6/072_215x160.jpg)
