ASCO 2015

Subgroup analysis of molecular markers in phase III study of high risk low-grade glioma

Dr Brigitta Baumert - Robert Janker Klinic, Bonn, Germany

You’ve been looking at the treatment of high risk low-grade glioma. What was the challenge that you were particularly wanting to illuminate?

We started this trial a long time ago and I’ve presented the data from these first result trials two years ago here at ASCO. Actually, we included patients with a high risk glioma, meaning patients in need of treatment having some predefined symptoms like age above 40, residual tumour, uncontrollable epileptic seizures. So this group was then randomised in an international trial, in a multi-institutional trial, conducted by the EORTC together with the UK and together with Australia, with the NCIC Canada. We randomised between radiotherapy alone and temozolomide alone and the results showed that progression free survival was equal, both treatments were equal. In addition, this trial which was new had a prospective question on molecular markers. So prospectively for all patients tumour tissue was collected and we did, in parallel, the molecular markers.

Right, tell me about these markers because there are three of them, aren’t there?

Yes. So the first one which we also had as the main endpoint was the 1p/19q deletion; this is one of the markers. Now, what we did during the last two years additionally we looked at the IDH1 and 2 mutations and also at the MGMT promoter methylation.

As a result of looking at molecular markers you’ve been able to report on slightly alternative outcomes in a subgroup, haven’t you? What did you find?

That was interesting, actually, because we found now a predictive marker. We found… actually we subgrouped patients according to their markers, we did not just look at one marker but we have found subgroups, so let me elude. The first one is the patients who have wild-type IDH1/2, that means they are not mutated. Those patients do worse; their median survival in this was only 19 months. That goes, if I can say so, to higher grade glioma, they have a similar median survival. Then we had two other subgroups which are grouped by the IDH1 mutated and these can be subdivided into the 1p/19q deleted and the 1p/19q non-codeleted. So if they’re both mutations, IDH1 plus 1p/19q, these are the best ones, they can live up to several years independent of the treatment.

And looking at whether they had radiation or temozolomide, what was the difference between the overall results?

For the other two subgroups there was no difference for the good ones. They had either radiotherapy and temozolomide for both mutations a good and long survival. For the temozolomide arm alone, for the non-codeleted ones, we have not enough events yet to have a conclusion on that one.

So what is coming out of this clinically then?

Clinically that would mean that those three subgroups would have to have different treatments and that is now under a big discussion - what should we propose now? For the worst group at least they need radiotherapy, that could be radiotherapy alone as my data showed that could be enough and not temozolomide but maybe we can also say we treat them like the glioblastomas with a different treatment scheme, that’s radiotherapy plus chemotherapy at the same time.

So what should cancer doctors read from these data at this point?

For the bad group, for the worse group, the IDH1 wild-type ones, they should learn that they should not give temozolomide in first [?? 4:01] but radiotherapy. And then as a second you can think whether you should add something on like temozolomide or not or whether you should give a little bit more aggressive treatment. For the other subgroups there may be one is enough, for the good prognosis one maybe it’s enough to give just simply alone temozolomide or radiotherapy. And the third arm, which is a mixture in-between, that will need, I would think, a combination of radiotherapy plus something and when we have data from another trial we could add in here.