Evolving Biologic Diversity Generates New Challenges

Ben Leach;

Evolving Biologic Diversity Generates New Challenges

October 16, 2012

Oncology Live®, September 2012, Volume 13, Issue 9

The heterogeneity of lung tumors generates challenges in terms of the costs and logistics of developing targeted therapies, given that relatively few patients may harbor a particular mutation.

D. Ross Camidge, MD, PhD

Director, Thoracic Oncology Clinical Program, University of Colorado, Denver, CO

In less than a decade, lung cancer has been transformed from a disease broadly characterized by tumor histology to an intricate molecular mosaic in which at least 10 genetic driver mutations or abnormalities have been identified in adenocarcinomas alone.

D. Ross Camidge, MD, PhD, director of the Thoracic Oncology Clinical Program at the University of Colorado Denver, has been at the forefront of efforts to identify driver mutations and develop therapies that target those abnormalities.

In 2009, Camidge helped define the impact of rearrangements in the anaplastic lymphoma kinase (ALK) gene in non-small cell lung cancer (NSCLC) (J Thorac Oncol. 2009;4[12]:1450-1454). He then played a leading role in the clinical trial that led to the FDA’s approval last year of crizotinib (Xalkori), the first ALK inhibitor for advanced NSCLC.

This year, the Bonnie J. Addario Lung Cancer Foundation honored Camidge with its fifth annual lectureship award during the 13th International Lung Cancer Congress. In announcing the award, the foundation leadership said Camidge is “a relentless advocate of molecular testing, redefining the standard of care for lung cancer, and aligning lung cancer patients with targeted treatment regimens.”

In his presentation at the congress, Camidge said the heterogeneity of lung tumors generates challenges in terms of the costs and logistics of developing targeted therapies, given that relatively few patients may harbor a particular mutation.

Camidge said multiplex testing is needed, but there may be regulatory requirements to validate each element of an assay before a panel is approved. In addition, he said participation in clinical trials could be enhanced by “leveraging patient communities” through Internet tools and by making the delivery of care more accessible through telemedicine and a hubstyle research network.

In this interview with OncologyLive, Camidge further discussed emerging issues in molecular testing and drug discovery.

OncologyLive: Please give us a little history about the groundbreaking molecular testing program that you and your colleagues at the University of Colorado Denver established. Camidge: We started testing every patient with nonsmall cell lung cancer regardless of histology starting in early 2008. The menu was pretty short to begin with. We did direct sequencing for EGFR and for KRAS. We did p53 as well at that time, although that sort of dropped off the menu a little bit.

By 2009, we started doing ALK-FISH [fluorescence in situ hybridization] testing because we were one of only two sites in the United States that could really get that assay up and running, and we helped to optimize that assay. And then very quickly, with the Lung Cancer Mutation Consortium coming on board, we took on board the snapshot platform for multiplex testing of mutations, and it’s still a constantly expanding menu.

We now have FISH testing for ALK, for ROS1, for MET gene amplification, and we have RET gene rearrangement FISH testing coming on board any day now.

Which patients do you test for molecular mutations?

I tend to test everybody with advanced-stage disease, mostly stage IV, but people with stage III because I think they’re very likely to relapse. I test all histologies, partly because these assays are multiplex, so you don’t have to think about whether you should do a PI3 kinase assay on a squamous, and an EGFR on an adenocarcinoma. It’s just one big assay, and I do that when they walk through the door.

I think we’re shying away from testing stage I patients. We think it’s probably a waste of money because most of those are likely to be cured. It’s slightly controversial as to whether you should test nonadenocarcinomas. I think you should test them, partly because deciding on whom to do a definitive molecular assay on the basis of a subjective determination of histology makes less sense to me.

There are a lot of data showing how often pathologists, particularly community oncologists, will disagree about the histology. So I don’t think we should decide on whom to do a definitive test on the basis of a nondefinitive test.

Would you test for a panel of mutations?

I would test for a panel of mutations for two reasons. One is you actually use up more tissue if you do sequential assays, so you get more “bang for your buck,” and you don’t waste as much time. And, secondly, the cost effectiveness gets much better as your hit rate, albeit for a series of different abnormalities, goes up dramatically for only a minor increase in the cost.

I think as these assays become more sophisticated and actually cheaper, we’re very much going to be moving toward multiplex platforms, maybe what’s called next-generation sequencing. We won’t have to think about whether I should send this test off or that test. We’re just going to check the lung cancer box, or maybe in the future, just the cancer box.

Is FISH the best way to test for a patient’s ALK status?

There are clearly different ways you can test for ALK. The assay that was developed for clinical trials and became the licensed companion diagnostic was the Vysis Break Apart FISH assay developed by Abbott. There are other assays that are out there—immuno-histochemistry for the protein and RT-PCR [reverse transcription-polymerase chain reaction] for the abnormal fusion transcript—and there’s also a new PCR assay looking to quantify expression of the ALK kinase domain, which has recently been released and is now commercially available.

What I suspect is that if you’re positive on these things, there’s a very good chance that you might respond to crizotinib. However, if you’re negative, sometimes these assays may miss it. In addition, occasionally there may be cases where you have a false-positive and you are inappropriately directing somebody to crizotinib. But, unless there’s a head-tohead comparison, we can’t tell. The only way to tell whether that’s a false-positive or a false-negative or a true positive or a true negative is to actually treat them with crizotinib. There’s a planned study from SWOG looking at these atypical negatives or where there’s disagreement between assays, and asking whether they respond to crizotinib or not.

The illustration shows the geographic reach of the crizotinib clinical trials conducted at the University of Colorado (red circle), which functioned as a research hub. Participants came from Colorado (62%), other US states (36%), and international locations (2%).

Source: West HJ, Camidge DR. Have mutation, will travel: utilizing online patient communities and new trial strategies to optimize clinical research in the era of molecularly diverse oncology. J Thorac Oncol. 2012;7(3):483-484.

How has your institution been able to enroll so many patients in clinical trials?

We’ve been running at about 30% to 40% of patients going on clinical trials for pretty much the last five or six years. We tried to choose the best trials, the ones that had the greatest chance of benefiting our patients, be they cooperative group, be they pharma-sponsored, be they investigator-initiated. And the goal was to have a mosaic that overlapped with almost every possible clinical indication so that you had a clinical trial to offer every patient.

When you get offered a clinical trial, the investigator brochure describes the preclinical data associated with it, and maybe some preliminary clinical data. If you’re experienced, you get to say, “Well, this is going to go nowhere” or “This is looking very promising.” Fortunately, we got relatively good at picking the winners in advance so that some people would travel to us just because we had the most exciting trials. Nothing succeeds like success. All of the physicians in the program are also very charming, and that helps a little, too!

You’re identifying potential targets in small percentages of the population who are affected by these mutations. How do you work with other hospitals or academic centers to get patients onto clinical trials?

The world of clinical research is changing. As the tremendous molecular heterogeneity in lung cancer becomes clear, you are not going to find enough patients with a given molecular abnormality within the traditional catchment area of any of the major cancer centers. So we have to address how we’re going to do clinical research in an era of geographically dispersed patients.

The first challenge is figuring out how to get the message out to patients to even come to you in the first place and have this molecular testing done. There is ClinicalTrials.gov, which is a preexisting government- run website, listing most of the studies going on in the United States. It does have a search engine, but it’s not really designed for a patient to work in. You have to know the code number of your drug. You can’t enter your zip code and say these are the closest centers. [The site was redesigned in mid-September with more patient-friendly features, including some map functions and contact information for study centers.]

The other factor is that, even before you get to that point, the patients have to know what they’re looking for, and that’s where there’s some really well-informed, patient-based websites. A very good example is Dr Jack West’s Global Resource for Advancing Cancer Education, GRACE, website [www.cancergrace. org]. He calls it an expert-moderated forum. He has a number of physicians who will go in there, post topics, and field questions from patients.

For example, with our crizotinib trials, when the story was just breaking, when the first glimmer of the phase I results were coming out in 2009, Jack asked me to do a podcast, and I didn’t actually know what a podcast was. It involved my doing my clinic at work, and then Jack phoned me from Seattle to Denver, and I gave a talk on the telephone while he showed my slides, captured it all digitally, and then posted it on the Internet. That’s since been downloaded thousands of times, and I would say at least a third of the patients who came to see us heard about us from that podcast.

There’s still a challenge that even if the patient knows what their abnormality is and knows where they need to go, we still have 20th century trial designs that say you have to travel hundreds of miles to come to Denver to sign a piece of paper before I can send your tumor off to testing. Why don’t we have remote consenting built in as a line item in the budget for clinical trials? Most of these patients are paying on their own dollar for travel and accommodations. Given that the pharmaceutical industry wants to put people on the clinical study, and these are trivial costs to them, that should be reimbursed. Why should a patient travel for a toxicity check? Why can’t we just use local laboratories and use telemedicine? We just speak to them on the phone or even a video conference through Skype to say, “How are you doing?” It’s totally feasible and could really be built into trial designs in the future and make it useful for patients.

Figure. Pies Within Pies

These pie charts illustrate driver mutations in histologic types of lung cancer. The picture is changing rapidly as new studies add more slices to each pie. "Lung cancer has become a series of orphan diseases nestling within the newly discovered orphanage of non-small cell lung cancer," said D. Ross Camidge, MD, PhD, at the 13th International Lung Cancer Congress. "We have heterogeneity in lung cancer and each slice of the pie produces more heterogeneity."

Classification of Lung Cancers by Histology¹

The pie chart at left depicts categories of lung cancer: non-small cell lung cancer (NSCLC), which includes adenocarcinoma, squamous cell, and carcinoid/large cell/large cell neuroendocrine, and small cell lung cancer. The pie charts for driver mutations within each histologic type represent recent research findings. New studies will add more slices to each pie.

*Percentages are approximate. Mt indicates mutation

Driver Mutations*

References

Bunn PA. Current status of advanced lung cancer therapy. Presented at: 13th International Lung Cancer Congress; July 19-22, 2012; Huntington Beach, CA.
Pao W, Hutchinson KE. Chipping away at the lung cancer genome. Nat Med. 2012;18:349-351.
Molecular profiling of lung cancer. My Cancer Genome website. http://www.mycancergenome.org/content/disease/lung-cancer. Accessed September 19, 2012.

Sos ML, Thomas RK. Genetic insight and therapeutic targets in squamous-cell lung cancer [published online ahead of print January 23, 2012]. Oncogene. 2012:1-4. doi:10.1038/onc.2011.640.

Research Resources Small Cell¹

The Lung Cancer Mutation Consortium, a network of 14 leading cancer centers in the United States, is leading efforts to discover driver mutations in NSCLC. Clinical trial information and updates are available at www.golcmc.com.
The Cancer Genome Atlas Resource Network has made characterizing lung squamous cell carcinoma one of its priorities. New research was published online ahead of print September 9, 2012, in Nature (doi:10.1038/nature11404). www.cancergenome.nih.gov
Three research papers published in September have shed fresh light on driver mutations and proteins in small cell lung cancer. Two articles were published online ahead of print on September 2 in Nature Genetics (doi:10.1038/ng.2396; doi:10.1038/ng.2405), and a third article was published online ahead of print September 6 in Cancer Discovery (doi: 10.1158/2159-8290.CD-12-0112).

If such telemedicine tools were to become more prevalent, how would community oncologists be affected in the way they interact with their patients?

I think a community oncologist wants to do the best for their patient, and I think it’s going to get easier and easier for them to do the molecular testing. They’re going to get a piece of paper that comes back and says, “My patient has got this abnormality.” Then what they need is a useful website, and there are several fledgling companies and the Jack West GRACE website, and various others where you’ll be able to go and find out what trials work for this molecular abnormality and where they are located.

If we can make the contact details better, they should be contacting their sites. I now routinely get multiple e-mails every day from people saying, “I’ve just done this molecular abnormality, what does it mean?” or “I’ve got this patient and I want to put them on a trial. How do I get them on?” Nobody likes sending their patients away. A good colleague of mine at the Cleveland Clinic sent me an e-mail saying, “I really like this patient. I know you’re a good doctor, but I don’t want to send him to you, but I have to.” And I said, “Don’t worry, because at some point you’re going to have a trial that I don’t have, and I’m going to have to send my beloved patients to you, too.”

It’s a team game. You have to throw the ball to people, and they run with it, and then they throw it back to you.

Do you think targeted agents should be evaluated in clinical trials earlier in the disease process?

One of the biggest controversies at the moment is whether if somebody, for example, has an EGFR mutation and they have early-stage disease and they’re resected, should you give this in the adjuvant setting. The same argument with ALK-positive disease is whether patients with early-stage resected disease should get crizotinib.

The number one answer is we don’t know. There are a little bit of data. There was some retrospective analysis of some adjuvant studies with gefitinib in the small proportion with an EGFR mutation, which didn’t seem to suggest an improvement, but that was kind of a retrospective subgroup analysis.

At this year’s ASCO conference, there was a very interesting small study of about 36 patients called the SELECT study, which was stage I through III, resected EGFR-mutant patients who received two years of erlotinib [J Clin Oncol. 2012;30(suppl; abstr 7010)]. At two years, they had a 94% disease-free survival rate, which sounds fantastic. However, they stopped the drug at two years, and at three years it had dropped down to about 60%, and by four years it had dropped down to about 40%. The strong suggestion is that the adjuvant use of these agents is not killing off these cells in the same way that chemotherapy in the adjuvant setting does. What it’s doing is what we know it can do in the advanced disease setting, which is control the cancer.

Now, that means one of two things. Either you’re going to have to put these people on these drugs for their rest of their life, which is unbelievably expensive, and then you have to think about what the side effects would be, or we’re kind of missing the picture here, which is that there are sensitive cells that are surviving this drug. We’re good at controlling it, but somehow the sensitive cells survive, and unless we find targets other than just the things that drive growth, we’re not going to be able to cure it. And some of the assays that we do to look for mechanisms of resistance will not find those.

We’re going to have to do different assays and probably some synthetic lethal experiments where you knock out genes and see what completely deletes a colony in the presence of the drug to find the things that allow you to maximize the kill rate. Until we do that, I think we shouldn’t be giving these things in the adjuvant setting because we don’t yet know what we’re doing, and they’re not increasing the cure rate.

Selected Later-Stage Clinical Trials of Targeted Agents in Lung Cancer

Target

Agent

Trial Description

Stage (Identifier)

Evolving Biologic Diversity Generates New Challenges

Which patients do you test for molecular mutations?

Would you test for a panel of mutations?

Is FISH the best way to test for a patient’s ALK status?

How has your institution been able to enroll so many patients in clinical trials?

You’re identifying potential targets in small percentages of the population who are affected by these mutations. How do you work with other hospitals or academic centers to get patients onto clinical trials?

Figure. Pies Within Pies

Classification of Lung Cancers by Histology1

Driver Mutations*

References

Research Resources Small Cell1

If such telemedicine tools were to become more prevalent, how would community oncologists be affected in the way they interact with their patients?

Do you think targeted agents should be evaluated in clinical trials earlier in the disease process?

Selected Later-Stage Clinical Trials of Targeted Agents in Lung Cancer

Classification of Lung Cancers by Histology¹

Research Resources Small Cell¹