Posted by Jeff Berk, BOLT International; It’s so cold and rainy here in the desert today we might as well be living in New Jersey. Anyway… One of our Pharma friends asked about diagnostics. Big topic, and we ask every one of BOLT’s Hematology / Oncology THOUGHT LEADER panels to touch on progress relevant to their tumor types. In this blog we’ve pulled together some comments from recent panels in non-small cell lung cancer and multiple myeloma. We chose these two because they exemplify two problems with diagnostics; and if you go back to our Amyvid™ post from March 19, 2011, you’ll see that the problems are not limited to cancer. First, knowing the predictive value of a biomarker often doesn’t translate into an effect on therapeutic choice. In the case of myeloma, knowing which patients are at high risk only goes as far as letting the physician inform the patient that their survival prognosis is unfavorable. Our myeloma example also highlights that some of these diagnostics, in this case PET imaging, while highly predictive, is difficult for the average oncologist (hematologist) to interpret. This is similar to the problem that Lilly has with their Amyvid™ Alzheimer’s diagnostic.
The use of diagnostics in NSCLC is the most advanced for any tumor we follow. Here there is a battery of markers that are driving therapy, and are actively used in screening patients for clinical trials.
So without further ado, some thoughts from BOLT’s NSCLC and MM THOUGHT LEADER panels:
Myeloma Prognostic Risk Signature (MyPRS) (Signal Genetics / Caris)
Ø Signal Genetics’ MyPRS is the first commercially available gene array that can risk stratify MM patients. As of TODAY, risk stratification can allay fear for standard risk patients, but cannot point to optimal therapeutic choice for high-risk patients. The hope is that once targeted therapy catches up with the diagnostic, it will make it easier for the community hematologist to choose appropriate therapy.
“I think what has changed in the last ten years is we are much better able to predict the high and the low risk groups and, again, to segregate them into different patient populations and clearly the low risk group do well. I mean you see those folks and you feel pretty comfortable telling them, listen that don’t fret, but what kind of therapy you are going to get you are probably going to live a long time no matter what we do. With the high risk people you secretly worry without telling the patient that this isn’t going to be good no matter what you do. I think that is where we are today. What is the value of having this prognostic risk signature? I think the main value is prognostic at this point. The main value is you are going to know through that test more than any other that we offer whether your patient is high or truly low risk and will therefore be more reassuring for those who are low risk. I think in academia it would change thinking about how to manage the disease. I don’t think the community is really mature enough now that they would take that information and necessarily change what they do based on it, but it is probably the single most effective test at telling whether your patient is high or low risk at the outset. I personally am supportive the test because I think it is good for the patient and the physician to know going forward whether they are high or low risk. But, I am sure there will be many folks you speak to that say we don’t see the value there yet in terms of whether we will change our treatment or not”.
“Signal Genetics producing the gene expression profiling as a commercial assay to risk stratify myeloma into high risk and standard risk now for the first time there will be, let’s say for the regular physician, the ability to risk stratify patients. Whether it will make a difference or not on what they do is another story. I think it is a watershed moment that there is that ability. So if you were going to advise the community hematologist on actions to take based on these risk stratifications and choices of first line therapy, second line therapy, etc, what is your take on…? I think for sure standard risk myeloma should get bortezomib-based induction and one transplant. Then if they are in complete remission the question of do they need maintenance is very appropriate. I would say this would provide strong argument that a patient with standard risk myeloma, low tumor mass, who had low level disease after transplant or had a CR after transplant, should probably be left alone and not be placed on maintenance. The converse to that is people with high tumor mass or high risk gene expression profiling (GEP) signature. They should probably be placed on maintenance post-transplant. With Revlimid? If they are not in CR Revlimid alone may not be enough and that they might be a combination maintenance with Revlimid and bortezomib. I think that to me was the practical aspect to that”.
Ø PET scanning can be diagnostic for risk stratification, but the biggest hurdle to universal adoption is that community physicians can’t interpret them as readily as they can a PET scan of a solid tumor.
“I don’t know where it stands, but I think the other thing that came out was the issue of PET scanning. So whether or not PET scans should be considered standard. Comment on that because others have as well. I think what came out is that PET scans could be effective because the trial that the Spaniards showed that it seemed to be helpful. I think what also came out loud and clear is that there is no standardization about the way PET scans are performed or read in North America. That in itself is probably the biggest barrier for universal use of PET scanning. It is still not considered standard. It is still not universally reimbursed. I think eventually it will be but the way I am looking at it is that probably somebody on the national level will have to create a task force to develop a guideline of how PET scans should be done and read in myeloma. And then after having done that they will have to look and see how it works. The Medicare Demonstration Project, I can’t say I have seen the definitive results, but I don’t think it was dramatically positive. And what happens, at least when I speak with community physicians, is many of them will tell you this really wasn’t really very helpful for me. It probably has more to do with the fact that the community radiologist doesn’t know how to read PET scans in myeloma patients because they read them like in regular solid tumors and there is a difference. Very good. Whether PET scanning for myeloma will be able to go out into the community I think is still a question.
Non-Small Cell Lung Cancer
Ø Diagnostics for mutational status are beginning to come into their own. Dana Farber, Mass General and MSKCC are running EGFR and FISH on every patient. The community centers are not yet equipped to routinely run these diagnostics but will come on board as the major centers validate the predictive value of the screens. Positive KRAS status is an important predictor for non-response to an EGFR inhibitor, while FISH positivity is an important predictor of response to these agents. Abbott signed a deal with Genentech/OSI to develop a FISH assay. Genzyme already has a package to use for mutation analysis of KRAS and EGFR. DSX is developing and validating an excellent KRAS BCR assay. Caris Diagnostics has a screen that detects KRAS mutations at codons (12, 13, and 61).
Ø Other mutations that are relevant to NSCLC are PI3 kinase, HER2, ALK-fusion and B-ras. This list is sure to grow.
“I think in lung a major tumor type right now we really are at a crossroads. We have a couple of things; cetuximab, Avastin, they all work just a little bit. I think it is going to be a time where we are going to see more personalization of medicine and more development of diagnostic tools and trying to match the right drug with the right patient. I am just not sure how we are going to do that yet. The key marker probably will be ras or EGFR-FISH”.
“EGFR mutation, EGFR-FISH, RAS we do them optionally, but I think they will become routine. How do you decide who is getting a RAS screen? I would look at someone who I was going to give an EGFR inhibitor to and if I couldn’t make a decision between the EGFR inhibitor and Alimta, I might look for RAS. If I know if they are RAS mutants their chance of benefiting from the EGFR is pretty poor”.
“I think there is going to be ever increasing teasing apart subsets of adenocarcinoma to develop more specific therapies. I think there will be a better understanding of who should get what EGFR therapy based on the biology of the tumor. For example, with Erbitux we may learn to exclude patients with KRAS mutations and we may know to exclude them from EGFR TKIs. So an adenocarcinoma with a KRAS mutation will get Avastin with chemotherapy. I think the testing will be more widespread”.
“I think basically we look at the NSCLC pie, particularly the adenocarcinoma pie, and what mutations do we know about. And there is EGFR and KRAS, and a small percentage of patients that have B-ras mutations, and a small percentage of PI3 kinase mutations, and a small percentage of HER2 mutations, and a small percentage have what is called a fusion protein, an ALK-fusion protein. And so basically the idea is when a patient comes in now, and we believe EGFR and KRAS are standard of care, they will sign a consent form that says we are going to be allowed to make DNA from your tumor and screen it for mutations and you may be contacted should an available therapy come up. We think that will be very appealing to industry because we will have a list of patients basically, a library of patients that we can contact if they want to come to us because they have a trial of an inhibitor for a certain mutation. But the other part is of course that we will have the DNA extracted so we can add that panel and just rerun someone’s specimen if we identify new things”.
“I think we are going to see the development of definite tests that will be sparked by the colon cancer. You are going to see at ASCO KRAS this year in colon cancer. Erbitux may not work very well in KRAS mutation patients. That might be 20% of lung cancer. It may not work very well in patients who have EGFR mutations, which are Asian patients, never smokers. Those you would want to give a TKI. Avastin has its group of patients based on largely toxicities and risk benefit assessment. And FSH now seems in some groups to be a substantial predictive factor for a positive response to Erbitux. I think what is going to happen 5-10 years down the road, and I think this is going to accelerate in both colon and lung cancer. I believe it is here and it is going to be diagnostics, companion diagnostics, or we are going to get a panel and we are going to look at EGFR mutation, ras mutation and FSH. I don’t know if you saw but Abbott just did a deal with Genentech OSI to develop a FSH assay. Genzyme already has package to use for mutation analysis of KRAS and EGFR. DSX is developing and validating a beautiful KRAS BCR assay. This is what is happening in the future”.
“We are doing EGFR and FSH in everyone who has tissue. We are the exception. You may think that is being done at most places. You will hear that from people at Dana Farber, Mass General and Memorial. We think we are sort of leading the way but time needs to validate that. There are a lot of the community centers that are not doing that right now. Yeah, they are just not equipped. But any of these things generally start at academic institutions and show better results and then you start to generalize and go broader setting and convince the community people or the scientific community at large that it is helpful and then you learn how to do it in the community or you figure out the mechanism, if they are send outs or whatever it might be”.
Ø The role of routine CT-screening in lung cancer is not by itself going to have an impact on outcomes. What is really needed is a way to molecularly characterize the tumors.
“I think the whole CT screening for lung cancer is in a lot of disarray right now because the question is whether CT screening will find lesions early enough that you can intervene and affect survival. Many are pretty skeptical about that. What is your personal opinion? There is the stuff that is coming out of New York where the NY-ELCAP where it is a single arm study and they do find cancers earlier. So in the short term there are fewer cancers seen. The survival for that group is great but they are looking short-term. There is a big randomized trial now which is what you need of CT scanning vs. chest x-ray with the national lung cancer screening trial of 50,000 patients. I guess there is a bit of bias in the trial. It is a single arm trial. There was also the whole thing last week because the tobacco company is funding the research. I don’t know if you saw that? Actually I don’t think that will have any effect. I think it was bias before that. I think that unless we know molecularly and prognostically about a tumor when you take it out which ones are going to be good and which ones are going to be bad I don’t think just screening is going to make that much difference. We have got to screen and then do something else. If you find an early cancer and it is the bad one it has probably already metastasized”.
Ø A 5-gene marker using DUSP6, MMD, STAT1, ERBB3, LCK is predictive for better survival on chemotherapy in high risk adjuvant NSCLC patients (Chen et al, NEJM 07). This marker can be used to decide who will and will not get chemo after surgery.
“The other significant story was looking at it, again not from a specific drug point of view, but there was a paper presented by the Canadian group that looked at a five gene marker in patients who were treated with adjuvant chemotherapy. That gene marker seemed to be quite predictive. These kinds of tests may be helpful in the future to decide who we are going to do chemotherapy after surgery”.