WIN 2016
Critique of the SHIVA trial
Dr Razelle Kurzrock - University of California, San Diego, USA
The main objective of many of the talks and of the conference itself is how to move forward with personalised or precision cancer medicine which many of us feel is transformative. The debate that we’re going to have today is about a recent publication that was quite high profile that was not able to show benefit from personalised medicine. My side, which has been labelled the ‘con’ side, not that I’m against personalised medicine because I absolutely believe in the power of this in the cancer field, but I don’t agree with the design of the study. I thought that the design of the study was such that it disadvantaged personalised medicine. So the ‘pro’ side is that the design of the study, I presume, is correct and perhaps the way we’re doing personalised medicine is not correct. So that’s what we’re debating.
What issues were there with the design of the SHIVA trial?
I will start first with the ‘pro’ side, what was good about the design. It was the first randomised trial of personalised medicine and randomised trials are considered the gold standard. That is why the SHIVA trial had such a large impact, because the claim was that all the other studies that have shown personalised medicine to benefit patients have not been randomised. But this study, that met the gold standard of randomisation, did not show a benefit. So that is the other side, the ‘pro’ side.
My view, however, is randomisation is not the only aspect of a trial that makes it well designed and it is true that this was a randomised trial but some other aspects of the trial, in my view, really disadvantaged it. I’ll give you an example that I think is really important, maybe I’ll give you two examples. So the essence of personalised medicine is matching patients to the right drugs but if you don’t make correct matches then it’s not really personalised medicine. I believe that the matches in many of the patients, maybe up to 80% of the patients, were sub-optimal. So you can randomise but if the underlying fundamental matching is suboptimal you’re not going to show a benefit.
Another aspect of the trial which I think is problematic, and is actually very commonly done in randomised trials, is that the personalised medicine arm followed an algorithm. In other words, the doctors were instructed if you see this genomic abnormality you’re going to give this drug. The other part arm, which was the control arm, was doctor’s choice. So patients got randomised to an algorithm versus doctor’s choice. In a real randomised trial both sides should be identical except for personalised or not personalised. So in my view both sides should have had an algorithm or both sides should have had doctors’ choice - doctor’s choice of regular therapy versus doctor’s choice of personalised therapy. By introducing this other variable where one side gets an algorithm and the other side gets doctor’s choice one is making the assumption, which I think is false, that doctor’s choice has no benefit to the patient. I believe that physicians, knowing a lot about medicine and a lot about patients, will make a choice that is beneficial more times than just an algorithm. Obviously the people who designed the trial don’t put a lot of credence in what physicians choose but I don’t think it’s debatable that in a randomised trial both arms should be the same except for the variable that you’re examining.
Do you think that the debates that have arisen in response to the SHIVA trial will impact the design of future trials?
It has to have impact on how to do trial design because if we’re going to do randomised trials we have to do them correctly. That’s number one and number two is if we’re going to move precision medicine forward the cornerstone is that we optimise matching. So we have to make a decision as to what optimised matching is and move forward with that.
What is the topic of your presentation?
The discussions will be again about clinical trial design and where we’re seeing benefits for personalised therapy and what the challenges are that remain. There are still a lot of challenges, obviously; cancer is a complex disease.
I’m hoping it will stimulate new studies and new ways of doing studies. One of the purposes of WIN is to stimulate and facilitate international studies in the personalised cancer field that are the cutting edge and that break out of the box. In many ways precision medicine is a disruptive field, it’s not just that we have new technology and new ways of measuring what’s in a tumour and new drugs but we actually need to change the design of our studies, individualisation of patient therapy, customisation for each patient, is not the way we’ve done things in the past. So doing things in a non-traditional way is something that I would like to see come out of this because until we change these models I don’t think we will realise the full potential of personalised medicine.
What would be your take home message?
WIN is a really special organisation, it brings together stakeholders from multiple different continents and multiple different organisations, from regulatory to patient advocacy to academics to the pharmaceutical industry. What’s amazing is to have all these people with different interests in the same room but then their interests converge in that we all want to make progress to eliminate the cancer problem.
