While personalized medicine is seen as a way to target an individual cancer patient’s tumors, using targeted agents, immunotherapy, and chemotoxic drugs, researchers are increasingly finding that not only do the cancers of individuals vary widely, but that tumors within a patient can also be quite different at the molecular level. Today we are speaking with Charles Swanton, MD, PhD, of the London Research Institute, part of Cancer Research UK, and the University College London Cancer Institute. Dr. Swanton’s translational research focuses on understanding tumor heterogeneity, both how it arises on a molecular level and how it impacts prognosis and treatment. Last year, Dr. Swanton and colleagues published a paper in the New England Journal of Medicine that analyzed, on a genomic level, serial biopsies from primary and metastatic tumor samples from the same patients as they went through treatment. The analysis included several biopsies from the same tumor.
—Interviewed by Anna Azvolinsky, PhD
Cancer Network: Dr. Swanton, can you talk about the main findings of that particular study in terms of tumor evolution and heterogeneity?
Dr. Swanton: The main findings of the study were that, when we asked the simple question, how many mutations are present in every biopsy compared with those present in one biopsy or one or more biopsies but not all biopsies, we found that on the average, somewhere between 63% to 65% of mutations were not present in every biopsy sampled. So, approximately one-third of mutations were present in every region. Work over the last year or so has been based on this and looking in other tumor types besides renal cancer, and in renal cancer as well. What we find, and what we found reported in the New England Journal of Medicine paper, is that tumors evolve through a branched evolutionary trajectory. In contrast to the linear models of tumor evolution, what we are seeing is that the primary metastatic site evolves independently and that there are major differences between samples within the same metastatic site as well. What is interesting is what is driving heterogeneity at the metastatic site appears to be copy number changes or allelic imbalance differences between biopsies of the same metastatic chest wall tumor.
Cancer Network: You found this in renal cell cancer, and now you are also looking at this in other types of tumors?
Dr. Swanton: It is too early to say what we are finding in other tumors, but we will be submitting manuscripts on the basis of those findings over the course of the next year or so.
Cancer Network: There are still many important questions in terms of intratumor heterogeneity that remain unanswered. Do we know whether all solid tumor types display a range of heterogeneity or whether some tumors are more complex?
Dr. Swanton: There is a huge amount of literature on this already and a lot has been published in this area in the last 12 months in particular. What we are seeing is evidence of heterogeneity of individual tumors across many different tumor types, including breast cancer, colorectal cancer, ovarian cancer, esophageal cancer, and other tumor types. We are seeing a diversity of somatic copy numbers and mutational events that may be spatially separated or indeed heterogeneous within one biopsy; that is, one biopsy may contain one multitude of cancer subclones that harbor subtly different or majorly different copy numbers or somatic mutation events.
Cancer Network: What are some of the issues in dealing with tumor heterogeneity and analysis of a biopsy when patients are actually diagnosed? Is this less complicated in earlier stage cancer compared with late stage?
Dr. Swanton: We certainly don’t have a sufficient sample size to comment yet on the early stage to late stage issue regarding heterogeneity. Having said that, my hunch would be that heterogeneity within an individual tumor is likely to be less profound with less bulky tumors, ie, earlier stage tumors, but obviously that remains to be tested and remains to be proven. That is something that we are actively looking at.
Cancer Network: What does this mean as far as therapy approaches and identifying new therapeutic targets? Cancers are complex to begin with, and essentially, they are moving targets as they evolve and adapt to treatment.
Dr. Swanton: I think that the first impression that we have, and this is not just our impression, but those of colleagues, is that tumors evolve both spatially and temporally, so that they are changing over time and they can be, in some cases, diverse over space within one patient. And certainly, emerging genomic data in other tumor types when metastatic sites are compared, is beginning to show similar phenomena, that metastatic sites differ subtly or quite majorly in terms of their genomic profiles in some cases.
So, what can we take home from this? Well, what I think we can take home from this is that a single biopsy snapshot picture of the genomic basis of the disease of one point in time may obviously be useful but may also not give the full potential picture of the cancer’s potential evolutionary path for the following years in an individual patient. The question is, how do we manage this within a clinical setting, how do we think about structuring clinical trials optimally on the basis of this emerging complexity? And what we think is worth focusing on is this issue of clonal dominance. That is, focusing on what are the clonally dominant genetic events present in every subclone, the early founder events, and targeting those ruthlessly as it were to maximize cell kill and cell cycle arrest across the population.
But I think we are also keen to understand how heterogeneity changes over time. How a tumor’s phylogenetic tree adapts through therapy, how some of the branches of the phylogenetic tree may be selected over the course of the disease and what that tells us about the process of drug resistance. Because, ultimately, what we hope is, by focusing on those branches, we may be able to forestall drug resistance and perhaps stop it from happening, or at least limit its consequences in advanced solid tumors. That is the future we are aiming towards, and I think we are ultimately a long way away from that, but that is where we aim to head.
Cancer Network: It is still early, but has this view and these results changed your particular practice in any way yet?
Dr. Swanton: Well, it is hard to say whether it is changing my practice. I think that the evidence we provide in the New England Journal of Medicine paper confirms the suspicions we always had, that tumors are very heterogeneous. But, I must emphasize that this is still an anecdotal report of two patients, albeit in quite immense detail; in fact, we are still looking at these two patients. But we really need more evidence before we start changing practice. But that doesn’t stop us from thinking of ways in which we can think about focusing on heterogeneity as an endpoint in clinical trial design. So, what is the impact of heterogeneity on outcomes, what is the impact of heterogeneity on response to targeted therapies? That is something we are thinking very carefully about as secondary endpoints with trials, and how this will influence the way we are going to interpret clinical trial data. I think that is a first step.
Cancer Network: Thank you so much for joining us today, Dr. Swanton.
Dr. Swanton: A pleasure, thank you.