Improving Cancer Treatment Selection for At-Risk Patients

This transcript has been edited for clarity.
Hello. I’m David Kerr, professor of cancer medicine at the University of Oxford. I’d like to reflect on the concept of early tumor progression for patients who are being treated with chemotherapy or immunotherapy, many of whom are in clinical trials. This is something that I’ve had to think about over a long career as a cancer physician and researcher.
Recently, there’s been more focus on this again because there are starting to be small groups of patients who, on receiving immunotherapy treatment, seem to progress very rapidly. What could be the mechanisms underpinning this observation?
Stepping back from that particular subgroup, it is an interesting area. I’ve been looking recently in more detail at one of our clinical trials. It was a combination of our histone deacetylase (HDAC) inhibitor with an immune checkpoint inhibitor. The idea was that the HDAC inhibitor would reinstate immune reactivity in what were immunologically cold tumors. 
The hope was — actually, with some evidence, this turned out to be the case — that the HDAC inhibitor would reinstate tumor antigen presentation, turning a cold tumor immunologically hot, and therefore making it a target for combination therapies with immune checkpoint inhibitors. 
It was a phase 2 trial and we published it fairly recently. There were 55 patients. When I looked back recently, in real detail, at the cohort of patients that we treated, these were patients with advanced colorectal cancer who did not respond to all conventional chemotherapy and who were a type called microsatellite stable. These patients are resistant to immune checkpoint treatment, completely so; therefore, a very hard nut to crack in terms of trying to demonstrate returned immune reactivity. 
I discovered that about one third of the patients who entered the trial were late-stage patients, untreated perhaps, with a life expectancy of 3-5 months. It’s a fragile patient group. One third of patients received only one cycle of treatment before they progressed and died.
When we examined the results in the phase 2 trial, the average overall survival was about 8 months. In this group who received only one cycle of treatment, the average survival was 8 weeks, so quite a very marked difference. Of course, if in advance we knew there was a group of patients who were going to do particularly badly in response to this combination treatment, or indeed any treatment, then you would seek to avoid treating them.
The worst possible thing for a cancer patient is to have progressive disease and the side effects of an ineffective treatment. If we could identify these patients before we treated them, then with a degree of sensitivity, we would exclude them from the trial and from that treatment. There would be no point.
If they were going to progress to the point of death rapidly, then all we would be doing would be reducing the quality of their life with treatment, whether chemotherapy or immunotherapy. 
What is also interesting is that, although it’s a small number of patients, I looked at the characteristics of the patients who progressed rapidly, thinking about conventional factors. Is it poor performance status? Is it tumor burden? Is it multiple sites of disease? Is it age? I considered the sorts of things that one would imagine could enter a multivariate prognostic model. None of these played out. 
It was a very small sample size, and therefore, it was me just poking around the database for my own interest, but there was no obvious difference at all, looking at those conventional, staging, recognized prognostic factors. We did a bit of looking at RAS and RAF and some of the genetic tests, but there was no difference there.
It was a really small sample size. I’m going to keep saying that. It did make me think that, beyond the conventional things that we measure, we must be more aware of, or there must be better, more pharmacologically or biologically oriented tests that we could do to judge a patient’s likelihood of response. For example, are there wider means of testing the immune system, such as how potentially active it may be and so on? 
It’s a small thought to go beyond conventional prognostic factors in terms of patient selection for trials and thinking better, perhaps more pharmacodynamically, as to whether there are tests that we could do that may allow us to select patients that would work better.
As always, thanks for listening. For the time being, over and out.