Name: Christof von Kalle
Title: Director, translational oncology, National Center for Tumor Diseases and German Cancer Research Center, Heidelberg, since 2005
Chair, NCT board of directors, since 2005
Experience and Education:
Group leader, internal medicine, University of Freiburg, 1996-2005
Postdoc, University of Cologne, 1994-95
Research Associate, Fred Hutchinson Cancer Research Center, 1992-94
Researcher, University of Cologne and DKFZ, 1988-1992
MD, University of Cologne, 1987
As director of translational oncology at the National Center for Tumor Diseases and the German Cancer Researcher Center, or DKFZ, in Heidelberg, Christof von Kalle is at the forefront of taking new research results and technologies to the clinic.
The NCT, a collaboration between the DKFZ, Heidelberg University Hospital and medical school, and German Cancer Aid, a charity, is a comprehensive cancer center that is committed to translating new research findings quickly into "innovative methods for diagnosis, treatment and prevention of cancer," and to offer these in clinical trials, according to its website.
On a recent visit to Heidelberg, Clinical Sequencing News met with von Kalle and asked him about his plans and expectations for high-throughput sequencing in cancer care. Below is an edited version of the conversation.
How are you using, or planning to use, the new sequencing technologies here at the National Center for Tumor Diseases?
The DKFZ has formulated a strategy that we call the 'Heidelberg Personalized Oncology Center,' which will be the scientific and technological platform to provide high-throughput analyses for all avenues of cancer research.
We think that eventually, high-throughput molecular diagnostics — in a sense, molecular imaging of the genome, and perhaps later of the proteome — should be available for every patient at every major milestone of the treatment process. First, to guide stratification of patients with particular molecular lesions into particular therapeutic trials; later on, to guide the development of combination therapies that try to hit more than one target at once; and eventually, for personalized oncology, meaning that we would like to adjust the treatments specifically to the molecular make-up of each individual cancer.
We would like to try and make available high-throughput molecular analyses for every patient at the center by 2015. We started in 2010/11 with the International Cancer Genome Consortium projects, and now we get to analyses like targeted resequencing, all-exome studies, and, later on, whole-genome studies of patients who are actually treated at the center.
Can you provide a few more details about projects that involve patients treated here?
They are currently research projects, but they are prospective, so they look at samples from patients at the beginning of therapy, and not as much retrospectively. We are focusing on diseases where there is a particular interest here, like pancreatic cancer, chronic lymphocytic leukemia, breast cancer, colon cancer, and myeloma, where there are research groups with ongoing genetic studies, and we are trying to deepen the information available to these investigators by adding these sequencing studies.
One of the first areas of interest that we have identified and where people have already started sequencing are some collections where we have samples from the pre-neoplastic lesion, the early manifestation of disease, and full-blown cancer from the same individual, so we can compare these with normal tissue and can start identifying lesions that have existed at the precancerous stage, and see how that has transitioned into the stages of disease. There are very good tumor bank collections of existing patients in some areas here, and some of the investigators are really interested in this.
How are you analyzing these tumors at their various stages?
There are a number of exome studies going on, some of them locally, some of them outsourced to commercial companies. There is also high-throughput sequencing of whole genomes, again, both internally and externally. The technological platforms that are being used are largely the same as at other centers, in particular the Illumina HiSeq, and for exome sequencing, we're currently using the Agilent SureSelect platform.
With regard to the investments of the center, the focus is on bioinformatics. Our philosophy is that the technologies will change, but the genome is going to stay the same. So for us as a cancer center and a cancer research center, the true art in mastering this lies more in our ability to analyze and understand and interpret the biology of the observations and their ramifications for pathways and actionable strategies in the clinic. We are trying to be conservative about our estimates on how many of the additional analyses will be actual game-changers for individual patients.
How much do you think sequencing tumors will actually impact patient treatment?
In the beginning, I would think a single-digit percentage of patients, depending on the disease. But the success of what we are trying to do does not hinge on that as much. I'm trying to compare it with other imaging technologies. We have also introduced magnetic resonance imaging, and we went from 0.1 Teslas to 1 Tesla to 1.5 Teslas to 3 Teslas to 7 Teslas, with significant expenditure, not because the next level of resolution would cure all cancers but with the philosophy that the more you see, the more you know, and eventually, this is going to improve the quality of treatment.
I think in molecular biology, it's a little bit of the same perspective. To me, it's an imaging modality, just like an X-ray or magnetic resonance imaging. And just like you analyze creatinine levels in kidney failure, you analyze gene damage in cancer. Some of these measurements are very valuable for treating that individual patient, and others are just a waypoint along the path of scientific discovery, and we don't know which is going to be what at each time. But it becomes clear that these genetic mutations are root causes of cancer development, so we need to know.
Some centers in the US seem to have opted for more targeted sequencing approaches. You seem to be going a different route with whole-exome and even whole-genome sequencing.
I think if you talk to these centers, they are also thinking about whole exome for next year or so. It's just that we are entering a little bit later than some of the big centers, and probably also with less money, so the question is, how many transitional phases can we have?
I think the strategy that these centers have chosen is to first establish the entire process of identifying the sample, ordering the test, making sure the right analytes get to the lab, sorted from the right section of the tumor by the pathologist, to have the quality of that process in place, and then also the bioinformatics after the wet lab phase. Once those processes are established, it is a lot easier to then switch the technology you actually apply. Exactly which technology you use in the wet lab phase is almost secondary, though of course it's not. But if you have spent time with clinical processes, you know that the biggest complexity is not in the wet lab phase.
So I think these colleagues who have started earlier and are looking at limited resequencing panels are acting very smartly, because they provide interpretable information and make sure they have all their processes lined up before the delve into then more complex data generation and interpretation, which still hinges on the exact same processes.
In the ongoing projects you mentioned, are you reporting back any information that could be useful for the treatment of the patients?
No, that is legally not possible, and I don't think it makes sense, also. But we have created an umbrella protocol for consenting patients that states that if we discover, say, an actionable gene lesion at the research level, and there is a targeted drug available, or we know of a trial protocol that this patient might fit in, we can crack the pseudonym of the scientific data and recontact the patient. But to use this information for the decision-making of putting somebody onto the trial, we will always run a second validation study, say, targeted resequencing by a PCR- or Sanger-based method. We are not taking the research-grade study at face value until proven by a second step.
In the mid-term, of course, we expect that industrial development will provide validated sequencing technologies. Like any other lab test, there will probably be [sequencing] devices that have a degree of reproducibility that would allow that. But that is clearly something that we are not developing as an academic center.
Some labs in the US have started to use next-gen sequencers in a CLIA lab. What is the equivalent to a CLIA lab here in Germany?
The way this works here in Germany is that pathology departments run molecular studies that are connected to histomorphological diagnoses. So confirmation sequencing or confirmation PCR studies that have to run at validated levels, at least at our center, and probably in many places in Germany, are focused within the institutes of pathology. We have a very good collaboration with our pathologists here, who are very important both at the beginning of the process, to identify the area of the tumor for the research-grade study, and then in the second half of the process, if we have identified a suspicious lesion that we might want to use clinically, they will provide confirmatory studies.
Do you believe high-throughput sequencing — be it exome or whole-genome sequencing — will be used routinely in cancer diagnostics the future? If so, when do you think this will be the case?
Yes, we believe so, and I think it will probably still be very much an academic fact-finding or cartography mission for the next two or three years, with the exception of proof-of-concept studies.
We think that those basic discoveries, with the availability of more and more targeted drugs, will in the future very rapidly lead to clinical testing of hypotheses. So if we see a mutation, we will see a lot of trials in the future that test the hypothesis that the presence of a mutation would indicate that a targeted drug may work in a particular clinical setting.
Some trials will be succeeding, others will be failing, but the more of these kinds of diagnoses are becoming available, and the more reliable and the less expensive sequencing tests will be, it will be just a question of economy to sequence a new tumor or a new metastasis upfront, rather than trying to run several molecular studies. Probably, these will lead to more informed, and then, we hope, better therapeutic strategies as we go forward.
Anything else you would like to mention with regard to the use of high-throughput sequencing here at the NCT?
In general, there is a lot of hype around the question of personalized medicine, and I am pretty sure there are going to be some sobering moments in what we can and cannot do a couple of years down the road. I would like to try and make sure that we do not oversell the idea.
On the other hand, I think there is no alternative to trying to get the best possible molecular information for everything we do because it's in the best interest of the patient. Will it solve the cancer problem within the next two or three years? No, it will not. I think it's very important to keep that in mind. But again, like the next level of resolution in MRI or any other imaging technology, that has never prevented us from trying to go to the next step. So I think we should not overburden the notion that we can now do a more precise molecular analysis with the idea that we can find everything, solve everything, and save money at the same time in the next couple of years.
What will be the bottleneck to make a change in cancer care?
The bottleneck will be the huge variety in the biology of cancer. There are a couple of red lines, where you can follow certain lesions in more than one cancer, and they are potentially actionable, but there are also types of mutations that we know are very frequent, like KRAS or p53, that we do not really know how to counteract. And then there is a third category, where there is no clear picture and a lot of mutations in many cancers. That, of course, is a problem that will be difficult to address. So the true bottleneck, I think, will continue to be our comprehension of the biological process.