Multidisciplinary approach to oncology improves accuracy of lung cancer screening

Share :
Published: 10 Nov 2014
Views: 2499
Prof Peter Kuhn - University of Southern California, Los Angeles, USA

Prof Kuhn talks to ecancertv at NCRI 2014 about a multidisciplinary approach to oncology involving maths, physics, biology, medicine and engineering, with a particular focus on lung cancer screening.

The presentation yesterday really focussed on both the conceptual framework of our work as well as on three specific results that we have achieved over the past couple of years now. The conceptual framework, what we call quantified oncology, is this aspect of convergent science where we are bringing together mathematics, physics, biology and medicine and engineering to really provide maybe a new perspective, an additional perspective, on how we can understand cancer and how it progresses in the human body.

So if you take that as a conceptual framework then, of course, the next question arises where everybody says, “OK, that’s an interesting framework but what do you do with it?” The three very specific examples I showed at the plenary were really triggered by very specific problems that are observed in the clinical setting. The first problem that we were looking at was really during the time of diagnosis in patients with the suspicion of lung cancer. So that is currently, of course, a significant debate because of two issues and that is that, first of all, the screening procedures that we have in place with chest X-rays or low dose CT scans produce a very large false positive rate on the one hand. On the other hand, most patients by the time they are being diagnosed are actually being diagnosed with stage 4 disease which is the lethal form of the disease. So you have this paradox, this paradox situation, where screening is not really performing to the level that you would want it to perform at and then diagnosis is overall happening too late. So with lung cancer the challenge is of course, as we have heard in other talks and will continue hearing throughout this conference, is that many patients who are at high risk of lung cancer have poor lung function, they have overall lung disease, a burden of lung disease overall, so many of them have shadows on their lungs etc. so how do you differentiate between that and actual lung cancer? So we teamed up with a group at Stanford University under Dr Sam Gambhir who is really an expert in radiation and imaging and FDG-PET imaging in particular. The discussion that we had together with Sam was one should be able to combine the imaging of the primary lesion together with an assessment of what’s happening in the circulatory system, in the blood itself. Those two together should give us a more precise representation of what’s going on with these patients and really improve the accuracy of lung cancer diagnostics. And that is indeed what turned out to be the case in a study that we conducted over the past two years that we just concluded quite recently where patients who have circulating tumour microemboli, so these are clusters of tumour cells that float around the blood, patients who have these clusters and have a particular level of glucose metabolism as detected by PET imaging, there all of a sudden your accuracy after diagnostic test is substantially improved as opposed to using either one of those two by itself.

So if you now think about how does cancer progress you’ve now answered a question of is it cancer. So in this combination of diagnostic approaches, if the answer is yes then the next question is what do I do? What do I do about that? In order to understand that in particular in lung cancer we were trying to figure out what are the most likely progression patterns because maybe from the progression patterns we can actually learn more about how to go about treatment plans overall. That again developed as a very interesting story for patients, for a subset of patients with lung cancer and in particular those who have early metastases in the adrenal glands. That is something that occurs very early on and, as our mathematical model has shown, based on autopsy data, as the mathematical model has shown, this was work done by Paul Newton in our centre at USC who really showed that the adrenal gland metastasis happens early, very consistent with clinical observation, but he also showed that the adrenal gland is a spreader. A spreader is a particular metastatic site that is quite likely involved in spreading the disease further in this overall traffic pattern. That is then a result that our thoracic oncologists and pulmonologists use to go back into literature and come back with this observation that patients who do have adrenal gland mets and undergo adrenalectomies seem to benefit from that as a survival benefit. Now this is something that we would have to show in a prospective clinical trial but this would be very exciting because it’s a small percentage of those lung cancer patients but there’s a lot of lung cancer patients so overall it will affect a fairly large number of patients.

So that was the second story of the plenary and then the third story was really focussed on how can we monitor what happens in a response to treatment in patients so that we can actually understand the fundamental biology that’s taking place in the particular patient undergoing targeted or chemotherapy. We are doing this by really using what we call high content analysis at the single cell level. So we identify single cells in the blood circulation of cancer patients and then we sample that over time. So we take multiple blood samples over time. We also have tissue samples from the primary tumour and then we perform this high content analysis and that’s a combination of really looking at the morphology of these cells; looking at it in prostate cancer in particular we tried this out first, looking at the androgen receptor, which is a key driver in prostate cancer, and then together with Jim Hicks and his team, who just recently moved to USC as well, who is really a world expert in doing single cell genomics, so together with Jim we are taking these individual cells, we are doing whole genome analysis on that and then with that we realised that we can actually truly understand what’s going on in a patient. That is, of course, what we believe will make all the difference from the current state of affairs where oftentimes we have to say, “It worked!” or “Oh, it didn’t work,” without the understanding on the basic sciences of what has actually just happened to the patient. And that is really where we’re going now with this overall approach.

Are there trials planned for this adrenal mets research?

Trials planned, yes, in our heads. Making those trials a reality is a little bit more complicated in part and for all the wrong reasons, as that sometimes happens. It is our expectation that this kind of approach will most dramatically benefit a fraction of the lung cancer population, probably maybe 3-5% and that means that that next trial has to involve a number of different cancer centres around the country or the countries, whoever gets involved with that. It’s not very controversial at all, most interventional oncologists are just like, “Yes, that makes sense. We can do that. We do it quite often for other reasons.” So it’s not very controversial at all but organising a trial like that and financing a trial like that is actually the challenge because there will not be, in this particular case, there is not a drug that gets paid for afterwards so there is no direct financial benefit to any given stakeholder. This will just save a bunch of lives.

What we need to figure out, we need to identify those responsible for overall healthcare economics because the research hypothesis we have proven so now the actual impact is going to be at the human level for the individual patient and the financial impact will be on the overall system so it’s a healthcare economics question. But there we just haven’t quite figured out who we need to talk to.

How could it impact the clinic?

That last one is where the fundamental concept that is really needed in order to make this next level of personalised cancer care a reality is to really bring the basic biology of the disease within the patient, together with what’s happening in the clinic. So it’s really bringing the bench and the bedside together. It’s not translating back and forth but it’s really bringing them truly together. So with the analytical accuracy that we have, and the sensitivity and specificity that we have, and our ability to measure, can we with that truly assess the patient more or less in real time? In this particular case what that means is that we can identify resistance early on and not only identify the resistance but also have suggestions around what might come next. So given that we have more and more therapeutics available to us to really pick and choose from, it really means that the biggest challenge now is, as we go forward, to figure out which patient benefits from which treatment at which time point and at which dose. That is what we think we can do only if we can actually access the disease in real time, and that’s where we believe it really is through the blood sampling, to understand what has happened and what is likely to happen.

Are we a long way from having this ‘real time genomics’?

That’s a difficult question to answer because there are lots of other parameters that play into this question, that is really the time to impact in clinical care. We have been driving this programme forward as part of a signature initiative of the National Cancer Institute in the United States over the past five years and the NCI very early on encouraged us to really think carefully about if this all works out how quickly can you actually make this a reality, Peter. So we worked early on with commercial partners to have everything ready to roll as we go forward but of course we do need the outside partners to really develop these kinds of research concepts into regulatory compliant processes, procedures and products. We are doing that in parallel. It gets a little tricky to put an actual timeline out; what we find absolutely fascinating, especially because we are so much on the basic science side, but we live day by day with our oncologists going side by side with us. Our oncologists are saying more and more often that they really want to use this data and they are the guys who really know the good, the bad and the ugly because we always say, “We are just research people. We know a little bit but it’s a little bit.” But the oncologists are now really at a point where they say, “No, no, Peter, this is actually hanging together very, very nicely. I can understand my patient better if I add the information to it.” So part of the answer to your question is really for some amount of time fairly shortly it will be complementary. At a slightly longer time frame it will be definitive and we have to very carefully distinguish between those two.