WIN 2016
Results of the MINDACT clinical trial
Dr Suzette Delaloge - Institut Gustave Roussy, Paris, France
We have been very active in the MINDACT and TRANSBIG consortium for ten years right now and active in the pre-MINDACT setting and then in the MINDACT trial. This is why I was so proud to present the results. There are no new results today, the paper will be released but it was the occasion to really think about it because this is a complicated trial and the application of the results are very important but still a bit discussed of how we do, how do we choose the prescription. So the primary endpoint was to verify that using the genomic test in patients who were classified as clinical high risk allowed to avoid chemotherapy safely, meaning that we expect in this population we would have 95% distant metastasis free survival at five years. We allowed ourselves a confidence interval with the lower bound being upper to 92%; we can’t set that lower than 92 wouldn’t be acceptable. This is a very conservative cut-off indeed but it was very important because if it’s lower than 92% or less there would be an important risk that chemotherapy benefit would be higher. So indeed MINDACT reached its primary endpoint which is the lower bound of the confidence interval is 92.7. But the DMFS at five years is 94.7, so very high, in this high risk population, almost half of them being node positive. So this is a very relevant population for which we have the question every day in which we still mostly prescribed chemo until today. So this is going to be very relevant, we just need the reimbursement in most countries. But provided those results we could change and provide the patients with genomic tests to try and avoid chemo among this category. So almost 50% of them avoid chemotherapy.
Was there anything surprising about the results?
It was exactly as expected regarding clinical load, genomic load, which is a very high DMFS rate at five years, almost 98%. So in this population there is no way anybody would prescribe chemo. For the clinical high genomic high, those patients have received chemo and we are at 90% at five years so this is very relevant, telling us that we need to do much better for this population. So it’s good to show that they need chemo. They probably benefit from chemo, we couldn’t show it in MINDACT but from previous results but they need additional treatments or replacing treatments instead of chemo because it’s still not perfect. 90% is not bad but it’s not at all perfect.
Very interestingly, the two groups, clinical high genomic low and clinical low genomic high did almost the same, around 95% at five years DMFS. This is especially questioning regarding the clinical low genomic high. We didn’t know really what to expect. This population is clinically very different, it’s mostly node negative, good tumours, and it’s interesting that it finally did almost the same. Some of them received chemo in that, indeed, but most of them did not.
It’s a small population, around 500 overall, and this population it does not appear that there is any benefit from chemo but the trial is not powered for that. I don’t think it’s useful to use genomic risk, the genomic assessment, for a clinical low population because at the end you don’t know what to do with it. So right now we would stick with clinical evaluation and that’s it for this population.
What results can we expect to see in the future?
For MINDACT there are some upcoming results but the main paper will be very detailed in terms of subgroups and everything. But the main results will be the secondary randomisations, so at San Antonio we’ll be presenting the chemo randomisation, it’s going to be published next year. For all patients who received chemo we proposed a randomisation between standard sequential anthracycline based treatment and taxanes for the node positive. The other arm was docetaxel plus capecitabine which is a very relevant question these days because the role of capecitabine has been questioned a lot recently. I don’t know anything about the results, I can’t tell you, and we will have all of this at the end of the year.
The secondary endpoint was a randomisation regarding endocrine therapy but it’s not for now, it’s going to be in almost five years. But the most interesting results, probably, are the next results regarding biological assessments and complete sequencing of the whole of tumours and assessing prognosis to sequencing, to other ways besides transcriptome. So this is going to be very interesting but this is exploratory; the main endpoints are here, the other ones are totally exploratory. It’s more science but it should provide us with a lot of information.
Do you have a take home message?
I think it’s always very important to mention that we won’t be able to do such trials any more again because it costs almost €50 million overall. It was a huge effort just to reach the level 1a biomarker so we just need another way to reach the level 1a. It’s almost impossible and I don’t think medicine and science can expect ten years before reimbursing genomic tests whereas finally the retrospective studies told the same thing. So we have to be a bit more modern regarding statistical tests and the way we assess biomarkers, if we have clear analytical studies etc. So this is very important.
