I’m presenting the phase III comparison between two arms in the rEECur trial, which is a study in relapsed/refractory Ewing sarcoma, and it’s actually the first randomised comparison of chemotherapy in this setting. The trial initially had four arms and I’m presenting the phase III comparison between the two most successful arms.
The reason we set up this study is that although relapse refractory Ewing sarcoma is very rare, survival is very poor. So, less than 15% of patients with recurrent disease survive to five years. There were a lot of different regimens that have evolved the years, and within those regimens variations of dosing schedules, but there had been no previous randomised comparisons so we couldn’t really counsel patients about which was the most effective or the most toxic. For many of them we didn’t really have even good event free survival data based on single arm studies. So we agreed as an international community in Europe that we needed to do a randomised comparison so we could counsel the patients but also know which was the most effective arm or the most toxic arm, on which to base new trials with novel therapies.
Can you give me a bit of detail on the methodology, the number of patients?
We used a Bayesian approach and patients were initially randomised between any of the open arms, so we initially had four arms open. The phase II comparisons, we agreed that we would drop one arm after 50 patients were recruited to each arm, and the second arm after 75 patients. That was very much a pragmatic decision because we couldn’t have a four arm study open with such a rare cancer for many years because it would become out of date. There had been a lot of statistical modelling initially based on a one year event free survival of between 20-40% to see what sort of confidence we would have after 50 and 75 patients were recruited. Then for the phase III comparison we thought we needed at least 200 patients and if we had 200 patients, based on the number of events we would have, we thought if there was at least a 70% probability that one arm was better, we would detect it.
What did you find?
In the primary outcome, which was event free survival, there was a two month difference in median event free survival. So for the ifosfamide arm, median EFS was 5.7 months compared to 3.7 months for the topo-cyclo arm, and that equated to a 10% difference in six month event free survival. So 37% for the topo-cyclo arm and 47% for the ifosfamide arm. For the secondary outcome of overall survival, again the ifosfamide arm was better, so there was a five month difference in median overall survival between 10.5 months for topo-cyclo and 15.4 months for ifosfamide. Again that equated to a 10% difference in both one year and two year overall survival.
In terms of the secondary outcome of toxicity, the ifosfamide patients had a higher incidence of grade 3 toxicity overall. The rate of neutropenic fever was similar at about 25% in both arms, but ifosfamide patients were more likely to have grade 3 infections. We weren’t surprised to see that there was also more grade 3 renal toxicity and more encephalopathy in the ifosfamide arm, at 8% and 7% respectively.
Another important finding, although the study wasn’t powered for a formal subgroup analysis, when we examined by the stratifying factors which were site, type and timing of relapse, and also some other factors that had not been reported to be prognostic previously, so age, pubertal status, we found that there was an interaction between age and treatment arm. So the benefits of ifosfamide for both EFS and OS were more obvious in children under 14 than in adolescents and adults.
What could the implications of these results be?
At a very practical level, we now have randomised evidence for the first time, so when patients come into clinic we can tell them these are the options, this is the survival that we’ve found in the study, and these are the toxicities we’ve encountered. It’s important to point out that although we showed that ifosfamide was more effective than topo-cyclo, which again was more effective than irinotecan temozolomide, the differences are quite subtle. So actually it would be reasonable to give all three of those regimens.
Importantly, because we now have decent quality evidence for these three arms we can now move forward to develop randomised studies of targeted therapies with some confidence that we understand what the baseline survival is. Although both the clinical community and the patient advocate community really want us to test a lot of novel targeted therapies, actually the published survival of chemotherapy is still better than any of the targeted therapies. So the sad fact is that most patients with this condition still die and most of them still die within five years. What we actually need is more targeted agents, more novel combinations, but we now have a very strong basis to move forward with.
Is there anything on the horizon on that?
The multi-targeted tyrosine kinase inhibitors are definitely the most promising, so at single agent activity they are approaching the same efficacy as chemotherapy. So in the rEECur trial we will be bringing in a lenvatinib ifosfamide arm later this year. There are also studies ongoing or planned or soon to be started in first line therapy as well, so that’s very exciting. And there are several other drug classes that we’re keeping an eye on.