A report from the front line of cancer treatmentS

Want to know about cutting-edge cancer research that could impact you this year, instead of twenty years from now? Physician Michael Soble, a cancer specialist, recently attended the American Society of Oncology's annual conference. He wrote us this report on where cancer treatment is headed in the very near future.

Soble practices medicine at North Shore Oncology Hematology Associates in Illinois.

A Report from the ASCO Meeting

I am a clinician, not a researcher, so I am at the meeting to figure out how to best treat patients today, and catch up on anything I may have missed. I don't look at the posters or go to the oral sessions that trumpet a 72% response rate for compound XYZ-123 in a certain disease. Those drugs are years away from being beneficial, if they ever are, and there just isn't enough time to learn about them all. So while the geek in me finds the preclinical stuff interesting, the practical physician can't spend too much time on it.

The cynic in me also gets annoyed every time some news article suggests that frog slime will cure cancer, even though deep down I think it's pretty cool that drug development is so creative. With that in mind, here's a handful of new therapies that got me excited and why.

Targeted therapy is definitely the wave of the future.
That wave is already here and has been for about 10 years but (to continue the metaphor) it's sort of like when the water starts to recede from shore and everybody looks and says "Oh, isn't that interesting?" The tsunami is right behind it. ASCO's theme this year was "Patients, Pathways, Progress" in recognition of that.

The way this process works is that first, a pathway is identified that is working abnormally in the malignant cell. Most often these days, the pathways involve the epidermal growth factor receptors (EGFR). EGFR is a huge family of cell surface receptors that bind a ligand (a molecule in the blood that fits like a lock and key) which usually causes two receptors to join up (dimerize) and activate a signaling pathway. In a cancer, these processes can turn on without ligand binding or without dimerization; or once they turn on, they don't turn off, or there are too many of the receptors on the cell surface (overexpression). There are probably other ways the process can go wrong, too.

However, now that you have identified that this usually normal process as abnormal in the cancer cells, you can attack it. You can design a molecule (literally design the molecule with a computer's help just like in Neal Stephenson's The Diamond Age) that will either grab the ligand to keep it away from the receptor, block the receptor, prevent dimerization, or block the process downstream. Since some part of the malignant process is the result of mutation, it is theoretically possible to design a drug that will only affect the cancer cells, not normal ones. This is the fabled "magic bullet."

More and more of these processes are constantly being discovered as more cancers and normal genomes are being sequenced. In the next few years, the cost of sequencing a whole genome will fall under $1000 which will lead to even more information.

A pathway through the complexity
The problem comes in the complexity of the system. Some cancers are simple (Sledge called them "stupid"), like chronic myelogenous leukemia (CML). There is a single characteristic chromosomal translocation (t9:22) that every medical student knows and it causes CML. Imatinib blocks the pathway that makes the cells malignant, then the cells differentiate into normal cells, and it becomes very hard to find any residual evidence of leukemia. The patient is not cured, since if the drug is stopped the leukemia comes back, but people have been on the drug for over a decade now with no evidence that the drug will stop working for many of them.

Melanoma, lung cancers, and most others do not have a single abnormality like that. Sledge presented a statistic that a lung cancer develops one mutation for every three cigarettes a person smokes. Not three per day, but three in your life. A pack a day for 20 years is 146,000 cigarettes or nearly 50,000 mutations. Not all are important mutations, but more than one probably is. So when people ask if there will be a cure for cancer, they are asking the question tens of thousands of times in that one sentence. Each lung cancer is different. Some may be similar enough that we can develop a more general treatment, but this approach will still never lead to an imatinib for lung cancer. Since there are so many abnormal pathways in most cancers, it may take several drugs to control the disease.

How do you do a trial to determine safety and efficacy of a combination of pathway-blocking drugs? If you are dealing with two independent pathways that each have a 30% incidence, then the odds of a person having both is 9%. About 3% of cancer patients agree to clinical trials. 3% of 9% of the 200,000 new lung cancer patients a year is an unworkable number to determine anything statistically significant when you factor in everybody that is ineligible. Finding those people is an impossible task.

So one big challenge with pathway-based therapies is how do we establish safety and efficacy of drug combinations. It's not the drug companies' fault that this problem hasn't been solved. Drug companies want to create drugs, and many of these pathways exist across multiple tumor types, so new drugs will not be restricted to just lung cancer or breast cancer or whatever. It's just the complexity inherent in biological systems.

Why drugs fail
Complexity continues to work against you and explains why these drugs fail. The story of cetuximab (Erbitux) is illustrative. It was approved for metastatic colon cancer based on a low response rate. It is an antibody directed against EGFR so when it first came out it was used preferentially (but not exclusively) in patients whose cancers overexpressed EGFR. Still, it seemed to work just as well on other cancers, so use became more generalized. Eventually we learned that it only works in patients who didn't have a mutation of a gene in the EGFR pathway called K-RAS. In any event, we now don't bother to use it unless K-RAS isn't mutated. But we use it in head and neck cancers without bothering to check K-RAS, and it may have activity in lung cancer if there is an activating mutation in the EGFR gene – not K-RAS. Why? Again, I don't know. These systems are complex.

One reason these drugs fail is we have to apply them in the right situation. Tratsuzumab (Herceptin) is a wonderful drug for breast cancer, but works only for about 20% of women with breast cancer. If it had been tested in all women with breast cancer, the response rate would have been one fifth as high and the drug might have been abandoned. So you have to be right about your target.

A report from the front line of cancer treatmentS

Consider this image of the EFGR pathway (at left). These pathways are not linear; they are a spiderweb. When I give talks on this to the public, I'll ask a random person what route they took to get there. Then I will throw up a roadblock. I'll say, "What if there was an accident on Main Street and it was closed?" They can give me an alternate route. That's what cancer does. If there is a roadblock, it doesn't call in sick, it just takes a different route and gets where it wants to be a little later. That's anthropomorphic, but I like the metaphor.

So targeted treatment is full of promise, but there is a lot of work to do. It's not going to fall flat on its face like vaccines have, but it's probably not going to put me out of business either.

There were a handful of therapies I saw that also showed promise, for a variety of reasons. Here are a few.

*omics. Genomincs, epigenomics, proteomics and a dozen other "–omics."
We are finally learning what goes wrong in a cancer cell, why it goes down an abnormal pathway, and learning how to block those pathways with new drugs. Popular drugs like bevacizumab, vemurafinib, imatinib, ipilumimab and ruxolitinib are all products of that sort of research. A few chemotherapy drugs still come to market and most people are still treated with them, but the future is all about targeting specific metabolic pathways.

This is much harder than it sounds. For instance, crizotinib has been shown to be very effective in treating lung cancer in patients with a certain mutation. Unfortunately that mutation is only present in about 4 – 7% of lung cancer patients. There are challenges in bringing drugs to market that will treat so few people. While some cancers rely on a single pathway and can be controlled indefinitely with a single drug, most cancers have several critical pathways and we have not discovered them all. Even when we do, it may not be possible to block them all without harmful side-effects. It will be very hard (impossible) to study all possible combinations for efficacy and safety in large groups of people.

I find this fascinating not just because it offers new treatments, but because it is a wonderful example of how science works. We have made major breakthroughs, but they lead to even more questions. This was the topic of George Sledge's (now immediate past president of ASCO) presidential address at the conference.

Exemestane (Aromasin)for breast cancer prevention.
A placebo controlled study of high risk, post-menopausal, women found that the aromatase inhibitor (AI), exemestane, can reduce the risk of developing invasive breast cancer. It is interesting because the only other drugs approved for prevention are tamoxifen and raloxifen (Evista). AI's have a different side effect profile (less clots, more aches) so some women who were concerned about side effects but interested in prevention have another option. The control arm was placebo, not one of the other drugs, so we don't know which drug is more effective at prevention.

Bevacizumab (Avastin) for ovarian cancer.
Women with locally advanced ovarian cancer who received bevacizumab with chemotherapy after initial surgery lived longer than those who did not receive bevacizumab. Interesting for obvious reasons since we always want to improve on the standard of care, but this is only one study so we may need confirmation before this becomes a new standard of care. Especially since bevacizumab's cost makes it controversial. You may have heard about it in the news recently as the FDA revoked its breast cancer indication for lack of a survival advantage and the manufacturer is appealing that decision. This was one of the more bizarre sequence of events I have seen in drug development. The FDA panel voted 6 - 0 against the manufacturer's appeal but the head of the FDA still has to rule. Three separate panels have ruled against approving bevacizumab for breast cancer yet the FDA approved it once and The Center for Medicare and Medicaid services (CMS) says they will continue to pay for the drug for breast cancer patients. This is the sort of thing that makes people skeptical of the medico-industrial complex.

For the first time, there are drugs that have real activity in melanoma (skin cancer).
Up until last year, no drug had ever shown a survival advantage in metastatic melanoma. Now there are two. Ipilumimab (Yervoy) is an immunotherapy that prevents melanoma cells from avoiding the immune response. Typically, the human immune system is actually pretty good at detecting and eliminating early cancers. melanomas escape that mechanism by activating CTLA-4, a protein on lymphocytes that identifies cells as OK and not in need of being killed. Ipilumimab blocks the melanoma CTLA-4 receptor/CTLA-4 interaction so the Melanoma cell is unmasked and lymphocytes are free to destroy the cancerous cells. Vemurafinib (not yet approved so no trade name, but it probably will be approved later this year) blocks a critical pathway present in about half of melanomas. Cellular pathways are of critical and growing importance. Both help melanoma patients live longer. The last drug that I can recall being approved for melanoma was temozolomide (Temodar) which did not improve survival but was well tolerated. At the time I said that we had gone from toxic, ineffective drugs to less toxic, ineffective drugs and that constituted progress in melanoma. I can't say that anymore. These drugs actually help people.

Imatinib (Gleevec) in gastrointestinal stromal tumor (GIST).
After resection of high risk GIST, 3 years of imatinib is superior to one. This is interesting because it raises the question of whether even longer is better, a question that will probably never be answered. It may mean that we can take an incurable disease and prevent it from coming back with one pill a day. That's not a cure, since patients need to stay on the medicine indefinitely, but it's close and the drug is well tolerated.

Ruxolitinb for myelofibrosis (MF).
MF is a disease where the bone marrow gets scarred and fibrotic so no blood cells can be made. The role of blood production shifts to the spleen, but then the spleen gets enlarged and causes symptoms which affect quality of life like bloating, weight loss, decreased appetite, pain, sweating at night and fevers. We have identified a mutation intrinsic to this process (JAK-2) and for the first time we have a treatment that can ameliorate the symptoms. The drug is not yet FDA approved, but it is likely to be approved this year or next.

Negative trials.
These are just as important but get little press. Proving something untrue is always important; it just doesn't sell any drugs. The two I found most interesting: Giving chemotherapy for non-hodgkin's lymphoma every two weeks is no better than giving it every three; and adding oxaliplatin to treatment of rectal cancer before surgical resection did not increase the odds of cure. Admittedly, I probably found these interesting because I have accrued patients to these trials.

The point is that often the most interesting developments in cancer medicine, the therapies that save lives, aren't the flashy, futuristic curealls you see on television. They're the small, incremental breakthroughs in our understanding of systems that are so complex we're only just now discovering pathways through them.