Sledge Expects "Big Data" to Generate Big Changes

Oncology Live®, April 2013, Volume 14, Issue 4

The technological advances in analyzing the human genome have spawned a new era in breast cancer as well as other types of malignancies that will affect oncology practice and will necessitate dramatic changes in the clinical trials system.

George W. Sledge Jr, MD

The 30th Annual Miami Breast Cancer Conference® provided a forum for exploring the latest research, techniques, and analyses of all aspects of caring for patients with breast cancer, including updates in surgical, medical, and radiation oncology. Physicians’ Education Resource (PER) hosted the conference, which took place March 7-10, 2013, in Miami Beach, Florida. The technological advances in analyzing the human genome have spawned a new era in breast cancer as well as other types of malignancies that will affect oncology practice and will necessitate dramatic changes in the clinical trials system, according to George W. Sledge Jr, MD.

Sledge told attendees at the 30th Annual Miami Breast Cancer Conference that 2012 marked a leap forward in understanding breast cancer as the results of many genomic analyses became available. The range of mutations uncovered in individual tumors will require moving beyond battling cancer by identifying a particular molecular process, as has been the case in the targeted therapy era.

“We’re clearly entering a new age, and that age is what I consider to be the genomic era,” said Sledge, who is chief of the Oncology Division at Stanford University School of Medicine in California and a past president of the American Society of Clinical Oncology. “This is an era of great promise. We’re at the point where we’ll be able to tell an individual what’s driving their cancer, but it’s going to require a whole lot more of us.

In developing new therapeutics, researchers will have to focus not only on qualitative mutations but also quantitative aberrations, Sledge said. “We don’t need a magic bullet, we need a magic shotgun,” he said. “We need something that can shoot pellets at a lot of different targets, and do so more or less simultaneously.”

He said the current clinical trials system is poorly equipped to take advantage of advances in knowledge about cancer genomics, and that many changes are needed. His ideas for overhauling the system include trials designed around multitargeting, greater collaboration among research entities, an information network for clinical trials, a redesigned informed consent process, and a “fundamentally different regulatory apparatus.”

“We have a next-generation sequencing. We need a next-generation clinical trials system, “ Sledge said.

Sledge noted that technological advances have delivered an explosion of information at an ever-decreasing cost. He said the sequencing of the first human genome took 13 years and cost approximately $3 billion; in the next several years, researchers likely will be able to sequence a genome in less than two weeks at a cost of about $1000. He said the price would drop further and that the challenge would be using the information generated.

“The evaluation of that gene chip that you order will be incredibly complicated and will require a significant amount of playing out over the next decade in terms of how we use it,” he said.

In an interview with OncologyLive, Sledge elaborated on his remarks.

Please expand upon your description of the new genomic era in breast cancer, as well as other types of cancer.

We’ve moved through a number of eras in the history of breast cancer. We started off with local-regional therapies, such as surgery and radiation. We then went to systemic therapies with drugs such as chemotherapy. The first targeted therapies were actually hormonal therapies dating back now several decades. In the last decade, the pace of research has picked up as we’ve learned more about the biology of the disease. This has led, of course, to an explosion of new, targeted therapies, not just for breast cancer but for all cancers.

I would argue that we’re moving into a new era. And this genomic era is going to be an era in which big data will come to the fore, where we will do deep sequencing or something like it on every patient’s tumor, and for that matter, on the host genome as well. We will have exquisite details and understanding of what is driving the biology of an individual’s cancer, and, hopefully we’ll have the tools available that will allow us to interfere with that biology on the very individual level.

Now, that’s got real pluses and minuses at the same time. The plus, of course, is that when you truly individualize therapy you won’t be treating the patient with drugs that might make them sick. You will be treating them with drugs that might keep them well. That therapeutic individualization with targeted therapies is certainly what we hope for and what we aspire to.

At the same time, because this is going to represent such narrow casting, such focus on the individual, the clinical trials of the past will no longer be possible. In the past we’ve done, for instance, 7000-patient trials to compare two different aromatase inhibitors for an allcomers estrogen receptor-positive population. We won’t be able to do that in the genomic era. In fact, it may be very hard to do anything other than pilot trials for some rare mutations that exist in certain types of breast cancers. So, the clinical trial piece of this will get much more difficult even as it gets much more exciting.

How do you think oncology practice will change in the genomic era? Will patients be visiting their oncologist with their tumor analysis on a computer chip?

This is probably every doctor’s fear: Mrs Jones walks in and says, “Here’s the memory stick. It’s got all 3.2 billion base pairs for my cancer. Here’s my host genome on another memory stick. Tell me how you’re going to treat me, doctor.”

Part of the change that we’re going to need to make is not just in how research approaches this, but how physicians approach it. Physicians are not trained to analyze deep sequencing, and probably never will be. We’re going to need bioinformatic support that’s more or less real time, where once one has this data available, that bioinformatics support is going to be able to sort through these complicated genomes and give us answers that are applicable and useful in a real-time way in the clinic. We don’t have that yet, but that’s certainly something that can be developed over the next few years as we learn more.

The "Genomic Chaos" of Breast Cancer

Graphic courtesy of The Genome Institute at Washington University in St. Louis/Newswise

Researchers are able to characterize the genomic complexities of breast cancer in greater detail than ever, noted George W. Sledge Jr, MD.

For example, researchers at the Siteman Cancer Center and The Genome Institute at Washington University in St. Louis found that women who did not respond to estrogen-lowering therapy had tumors with higher mutation rates, as shown for one patient in plot at right (Nature. 2012;486[7403]:353-360).

"Smart tumors" with higher mutational loads result in "genomic chaos," Sledge said in his presentation. Below, Sledge listed key aspects of breast cancer mutations (Nature. 2012;486[7403]:400-404).

Mutational Landscape

  • 100 breast cancer genomes analyzed
  • Driver mutations found in at least 40 different cancer genes
  • 73 different combinations of driver-mutated cancer genes
  • 28 cancers had a single-driver mutation, but some had as many as 6 driver mutations
  • WE HAVE NEVER TARGETED 6 DRIVERS!

Please explain the concept of gene mutational load and how this expanded understanding might change therapeutic approaches.

This has been an area of great fascination to me. If you look at human cancers, and if you measure the number of mutations that there are in the cancer and come up with what are known as mutations per mega base⎯the mutation rate for a million base pairs in someone’s genome⎯that is sort of a new currency of being able to measure how disordered the patient’s genome is. What one finds is that there’s a thousand-fold difference between the least mutated and the most mutated of human cancers. Now, you’ve got to imagine that having a thousand more mutations is going to affect how you’re going to respond to a particular therapy, and whether or not you’re going to respond to a particular therapy.

People like myself have spent pretty much their entire career looking at qualitative aspects of oncology: Is HER2 amplified? Is the patient estrogen receptor- positive? This is somewhat of a change because we now need to think in terms of quantitative as well as qualitative analysis when we’re looking at these tumors.

A good example of this is triple-negative breast cancer where the cancers on average are highly disordered and have lots of mutations. Those cancers, we’re finding out, are the cancers that are really good at finding ways around the roadblocks that we throw up. They’re really good at resisting the drug therapies that we have available for patients.

During one of your presentations, you noted that analyses of tumors by mutational load has led to the concept of “stupid cancers” versus “smart cancers.” Please elaborate.

If we look at human cancers, they are now in broad terms dividing themselves out into cancers that are relatively “stupid” or “smart.” A great example of a stupid cancer is chronic myelogenous leukemia, where you have a single genetic alteration with bcr-abl translocation that is responsible for the cancer and which represents the way that you treat the cancer. You come up with a small-molecule receptor tyrosine kinase inhibitor of that fusion product, and the very first drug that came on the market gave you an 89% five-year survival for widespread disease. And, if that cancer no longer responds to that drug, most of the mutations that cause resistance are in exactly the same ATP [adenosine triphosphate] pocket of that particular kinase. So, you come up with a different kinase inhibitor and the patient goes back into a remission.

That’s not a very smart cancer, is it? That is a cancer that flunked out of grade school, in a way.

Key Aspects of "Stupid" and "Smart" Cancers

Stupid Cancers

Smart Cancers

✓ Single dominant mutation

✓ Multiple mutational drivers

✓ Small mutational load

✓ Large mutational load

✓ Monotherapy effective

✓ Multitargeted therapy required

✓ Resistance rare, late,

and in same pathway

✓ Resistance common

and early

In contrast, we have these really smart cancers that have all these loads of mutations. They have a bunch of different drivers that are driving the cancer simultaneously.

If you think of this as when you treat a patient, it is as though the criminal is trying to escape town and you throw up a roadblock. The stupid criminal gets caught at the roadblock and never goes any farther. The smart criminal is just going to find another route out of town. I think that’s what we’re faced with.

Very interestingly, many of the hematologic malignancies and childhood cancers tend to fall at the end of the spectrum that has relatively few mutations. In contrast, as perhaps we might expect, cancers such as colorectal cancer, which is caused by a lifetime of exposure of the gut to Big Macs, or melanomas caused by too much time in a tanning salon or too much time out on the beach with, again, that lifetime of mutagenic exposures, or head and neck cancer or lung cancer caused by a lifetime of cigarette smoking⎯these cancers have a large mutational load.

What changes in the clinical trials system might be necessary to take advantage of the expanded knowledge available in this genomic era?

We may have to change the standard of evidence. The FDA already has the capacity when we’re dealing with an orphan disease where there are relatively few cases to accept drugs based on what we might otherwise consider a lower standard. In cases where you have a true orphan disease, for instance, response rate might be the criteria for approval rather than overall survival or progression-free survival. If you don’t have the patients to allow you to do a phase III trial, then you’re probably going to have to accept different endpoints if we’re going to be able to move this forward.

How might approaches to therapy change?

We’ve been doing largely a kinase-based approach, such as targeting HER2 and the epidermal growth factor. Whether or not you’re talking about interfering at the cell membrane or downstream, there is still a kinase-based approach. But, if a patient has multiple genetic drivers of their cancer, it may be very difficult to do these trials. We may be talking about such thin slices of the pie that mounting anything other than a pilot trial will be really difficult.

If we’re thinking in terms of having multiple kinases involved, we have to think about using either multiple drugs or alternatively going to drugs that are perhaps a little bit less targeted. Many of our drugs, many of our kinase inhibitors, are not particularly clean drugs; in fact, they are targeting several different kinases simultaneously. It may be that we need to be thinking in terms of, if I can use a phrase, “selectively promiscuous drugs,” which are capable of targeting several kinases simultaneously.

Doing that is going to require a different way of thinking about drug development, both from a pharmaceutical company’s standpoint in terms of how you develop the drugs, but also from a regulatory standpoint in terms of how you approve the drugs. We’re not quite there yet.

The changes you are discussing seem to point to a dramatic overhaul of the current system.

The reality is that if we’re going to continue down this path of interfering with kinases, of interfering with the mutational drivers of the cancer, we’re going to need to think about a significant overhaul of our system.

Of course, the other alternative is not to continue down that path. We certainly have other paths that we can look at. For instance, immune-based approaches⎯new drugs such as anti-PD-1, anti-PD-L1 [programmed death and its ligand], that sort of nonspecifically invoke the body’s immune system to attack the cancer⎯ might be highly active for some cancers. There are certainly many other approaches that are being looked at right now that wouldn’t involve a kinase-based approach.

What role has federal funding played in advancing research into the human genome, and are you fearful about the prospects for future advancements in light of the current debates about government support of oncology research?

It’s safe to say that we would be nowhere near where we are in genomics without federal funding, not just in the United States but in other countries as well. Particularly in the United States, the first human genome from the Human Genome Project came about as a result of a largescale federal project. The Cancer Genome Atlas project that has looked at 20 different cancers⎯including perhaps somewhere close to 20,000 human genomes at this point⎯ came about as a result of federal funding in recent years. It has been crucial in terms of being able to understand the biology of human cancer.

I think all of us who are in this area, as we look not just at the sequestration issues but also at the broader issues in terms of federal funding for cancer research, are concerned that we might lose an entire generation of investigators. Young investigators looking at this field may just simply say, “This is too tough. If I have a one in 20 chance of getting a research grant, why on earth should I do that? I have better odds at the racetrack.”

And, a nation that treats its young investigators as if they were people who should be going to the racetrack probably will deserve what it gets in the long run. We need to think more seriously about how we fund scientific research in general, and cancer research in particular, if we’re going to be successful going forward.

Do we need the level of support that was provided in the “War on Cancer” of the 1970s?

It’s interesting that if you go back to Richard Nixon’s original idea for the “War on Cancer,” we were flamingly ignorant about cancer at that point. We knew nothing about oncogenes, for instance. Can you imagine declaring a war on cancer without knowing what’s causing cancer? I think the great analogy is actually the space program. We went to the moon right around the same time as we declared war on cancer, but we went to the moon based upon Newtonian physics, something that was discovered in the 1680s. Basically what you needed to get to the moon was stuff that was discovered several hundred years ago. But we had to start anew when we started the war on cancer, and we’re still learning it at a huge pace.

I’m confident that we’re almost there. I’m confident that we have learned so much about cancer biology that we’re going to see a lot of these diseases falling over the next few years.

Despite what I’ve already mentioned about how tough some of these cancers are going to be, I do believe that knowledge is power, and I do believe that the huge amount we’re learning is going to be applicable to individual patients in the not-too-distant future. I’m a cautious optimist when it comes to this. But to do that, we’re going to need the infrastructure, we’re going to need the support for smart young researchers because they’re dealing with very smart tumors. We’re going to need to think about how we move this into the clinic in an efficient and effective way from a clinical research standpoint, and ultimately from a patient care standpoint. None of that’s going to be easy, or simple, or cheap.