What we’re learning about cancer as we do genomics and really begin to understand cancers better than we did in the past is that the way we have designed clinical trials does not fit with the reality that the science is unveiling. So we need to redesign clinical trials to fit the scientific reality and that specific reality is that the genomic aberrations drive cancer, that’s number one, and number two is that each person’s tumour is different and each person’s tumour is complicated so we can’t put people in a box the way we used to.
When you’re looking at different stages of cancer you may have a completely different biologic environment in that cancer. So translating from one stage to another it might give you some information but you would be missing a lot of information. So we’re learning that not only are there a lot of differences between each patient’s tumour but even in one patient’s tumour, as the course progresses, there are very distinct alterations. We need to address what is there at the time. So it’s the right drug for the right person but it’s also at the right time.
How have new technologies helped?
The technology to make that assessment is very important. The technology that has really exploded is the genomic and next generation sequencing technology but now we can do it not just on tissue but we can do it on blood tests, something called the liquid biopsy. That opens up whole new doors to doing additional testing on patients because now we don’t have to biopsy them every time, we can also take a blood test.
Is artificial intelligence streamlining diagnosis?
The whole area of artificial intelligence is really interesting. I’m not quite sure that we’re there yet but we’re very close. This is going to be absolutely transformative. One of the things about having all this information is that the information is so complex that it’s beyond the ability of any human being to analyse it but this can be analysed by computer systems and computer systems that not just look at the information but can also learn from the information. We’re right at the cusp of being able to do that.
Can anything be learnt from the SHIVA trial?
The SHIVA trial was presented last year, actually in my session. It’s a very important trial because it was the first randomised trial. But the problem with the SHIVA trial is not the SHIVA trial, the problem is the way people over-interpreted the SHIVA trial. Even the first author would agree with that. So the SHIVA trial showed that it didn’t meet its endpoints but 80% of the patients were treated with single agent mTOR inhibitor everolimus or single agent hormone modulators. I think we can conclude from that trial that single agent hormone modulators or mTOR inhibitors do not work in patients with advanced, heavily pre-treated cancer. The SHIVA trial proved that definitively. To extrapolate from that to say that precision medicine doesn’t work is what people are doing and that makes no sense scientifically.
How does science affect the designing and reviewing trials?
It’s designing trials the right way but also making the right conclusions from those trials. We have to look at what are the conclusions. So the SHIVA trial had very important conclusions but we can’t jump from one conclusion and say all of precision medicine doesn’t work. I can give you hundreds of examples now of where precision medicine does work but I don’t want to extrapolate from where it does work and say it works everywhere. In the same way we can’t extrapolate from a limited couple of drugs that didn’t work out and say the whole field doesn’t work. We need good scientists doing good science and also people reporting that science in a way that is the right way to see the trials.