The Long Tail of Cancer Research

By MaverickNY on July 14, 2010

"What's really amazing about the Long Tail is the sheer size of it. Combine enough nonhits on the Long Tail and you've got a market bigger than the hits."

Source: Chris Anderson, Wired

The other other day, while barely lucid in the early hours of the morning, I was thinking about the herd instinct, the tendency to follow the masses and the counterpoint to that, i.e. the long tail.

If you're wondering what the long tail is, think of those statistical distribution plots where everything is bunched around the median, creating 50% on one side and 50% on the other. At either end the plots tail off into infinity. That's the long tail.

The long tail looks small at first, but creative marketers realise that added up (as Chris Anderson's quote above shows), you can in effect still have quite a large niche market while every one else focuses on minute shares in the middle.

I first realised the long tail effect as a sales rep in the tough anti-cholesterol market. You can spend fruitless hours chasing the majority of the primary care doctors who see reps but don't remember you when the relevant patient appears, or you can carefully invest time in the ones who don't see reps but are busy with appropriate patients and eventually crack if you try enough times. Once convinced, they often switch wholesale and your job is done with a smaller group of physicians who others can't see :-).

If we apply the principle of the long tail to cancer research, we can see things more clearly. In the old days, the majority of patients got treated pretty much the same with various chemo doublets, irrespective of whether they might work or not. In fact, in many cases, there wasn't even biomarkers such as ERCC1 to determine which patient should get platinum or not, but things are changing and now it's much easier to make these decisions and start segmenting patients according to their biochemical profile and make decisions based on the profile.

Today, we can take this niche idea to treatment a lot further.

In colon cancer, for example, we now know that patients with wild type KRAS are more likely to respond to an EGFR therapy than those with mutant KRAS. However, in lung cancer, the old adage was that erlotinib (Tarceva) was best suited for patients who were female, asian, non-smokers with adenocarcinoma is giving way to a more precise definition, i.e. do they have the EGFR mutation or not? If they do, they're more likely to respond, irrespective of smoking status. Offering these patients maintenance treatment may also be an effective treatment strategy that impacts outcomes.

We can see this effect in other cancer types too. Trastuzumab (Herceptin) is approved for women with Her-2 breast cancer, imatinib (Gleevec) and other TKIs such as dasatinib (Sprycel) and nilotinib (Tasigna) for Philadelphia-chromosome positive chromic myeloid leukemia (CML).

Histology is a very crude way to select patients, but looking at the aberrant mutations essentially creates niche long tail opportunities to treatment for pharma and biotech companies. Patients who are more likely to respond to a given therapeutic get appropriate treatment, without having to expose others who would not to unnecessary systemic effects. This is a win-win solution all around.

Why?

Well, it's good news for patients as you increase the chances of successful outcomes rather than relying on hope alone. It's also good news for manufacturers because smaller patient populations ultimately involve fewer patients in clinical trials, thereby making clinical development more cost effective and as cancer therapy moves from an acute to chronic disease, so longer term revenues are generated despite a small patient base.

As more oncology companies start looking at their pipelines, we can see many asking the critical questions – which patients are more likely to respond to a given treatment and why? As we learn more about the underlying biology of the disease, so companion diagnostics are also evolving in sophistication and sensitivity.

The other marketing advantage of developing niche targeted therapies and diagnostics is that often, you see less competition for smaller subsets of disease because it creates a high barrier to entry. Diagnostics also create barriers to entry because of the extra costs involved in their development. For smaller biotechs this can be prohibitive, and many are more actively seeking Pharma and Biotech partners to fund late stage research and clinical trials. Overall, the long tail opportunities offer big and small hurdles, depending on the circumstances.

Roche's VEGF monoclonal antibody bevacizumab (Avastin) has a very high barrier to entry in colon cancer, for example, as other VEGF inhibitors have fallen by the wayside, unable to beat the results already obtained by the first to market drug.

At the other end of the scale, the barrier to entry is much lower in renal cell cancer with numerous targeted therapies now approved for a relatively small niche indication, including sorafenib (Nexavar), sunitinib (Sutent), temsirolimus (Torisel), everolimus (Afinitor), pazopanib (Votrient) and bevacizumab (Avastin), probably reflecting the improvement over IL2, without completely reducing unwanted side effects or dramatically improving efficacy.

Some of the marketed therapies mentioned in this post are now billion dollar blockbusters despite cancer being a relatively niche market opportunity compared to the much bigger primary care markets such as metabolic or cardiovascular disease, proving that there are valuable nuggets to be found, even in the long tail.

The future in cancer research is not in broad acting systemic chemotherapies that target normal cells as well as cancer cells, but in the niche development of better and less toxic targeted therapies based on the underlying biological abnormalities with easy to use diagnostic technology based on fluid-based biomarkers. To achieve this though, will take a lot of bright smart people with expertise in oncology who dare to think differently and boldly, whether they be scientists, marketers or clinical research professionals.

Watch this space!

The new paradigm of requiring a companion diagnostic as a condition for approval of new targeted therapies has placed pressure so great that the companion diagnostics they’ve approved often have been mostly or totally ineffective at identifying clinical responders (durable and otherwise) to the various therapies.
It should be in the FDA’s interest in saving the healthcare system perhaps billions of dollars a year (and thereby the healthcare system itself) by ensuring that expensive treatments are used appropriately. It should serve their interest not only in discovering new cancer treatments, but also using currently-available cell culture technologies to improve the effectiveness of existing drugs and save lives today by administering the right drug to the right patient at the right time.
The headlong rush to develop companion diagnostics to identify molecular predisposing mechanisms does not guarantee that a cancer drug will be effective for an “individual” patient. Nor can they, for any patient or even large group of patients, discriminate the potential for clinical activity among different cancer agents of the same class.
The drug discovery model over the last number of years has been limited to one gene/protein, one target, one drug. The “cell” is a system, an integrated, interacting network of genes, proteins and other cellular constituents that produce functions. You need to analyse the systems’ response to drug treatments, not just one target or pathway.
The decoding of the human genome in 2000, sparked hopes that a new era of tailored medicine was just around the corner. However, uncovering the genetic differences that determine how a person responds to a drug, and developing tests, or biomarkers, for those differences, is proving more challenging than ever. As a result, patients with cancer are still being prescribed medicines on a trial-and-error basis or one-size-fits-all.
The key to understanding the genome is understanding how cells work. The ultimate driver is “functional profiling” (is the cell being killed regardless of the mechanism) as opposed to “molecular profiling” (does the cell express a particular target that the drug is supposed to be attacking). While a “molecular profiling” tells you whether or not to give “one” drug, “functional profiling” can find other compounds and combinations and can recommend them from the one test.
The core of “functional profiling” is the cell, composed of hundreds of complex molecules that regulate the pathways necessary for vital cellular functions. If a “targeted” drug could perturb any one of these pathways, it is important to examine the effects of the drug within the context of the cell. Both genomics and proteomics can identify potential new thereapeutic targets, but these targets require the determination of cellular endpoints.
Cell-based “functional profiling” is being used for screening compounds for efficacy and biosafety. The ability to track the behavior of cancer cells permits data gathering on functional behavior not available in any other kind of testing.
Molecular profiling, important in order to identify new therapeutic targets and thereby to develop useful drugs, are years away from working successfully in predicting treatment response for “individual” patients. Perhaps this is because they are performed on dead, preserved cells that are never actually exposed to the drugs whose activity they are trying to assess.
It will never be as effective as the cell “function” methodology, which has existed for the last twenty years and is not hampered by the problems associated with gene expression tests. That is because they measure the net effect of all processes within the cancer, acting with and against each other in real-time, and it tests “living” cells actually exposed to drugs and drug combinations of interest.
It would be more advantageous to sort out what’s the best “profile” in terms of which patients benefit from this drug or that drug. Can they be combined? What’s the proper way to work with all the new drugs? If a drug works extremely well for a certain percentage of cancer patients, identify which ones and “personalize” their treatment. If one drug or another is working for some patients then obviously there are others who would also benefit. But, what’s good for the group (population studies) may not be good for the individual.
Patients would certainly have a better chance of success had their cancer been “chemo-sensitive” rather than “chemo-resistant,” where it is more apparent that chemotherapy improves the survival of patients, and where identifying the most effective chemotherapy would be more likely to improve survival above that achieved with “best guess” empiric chemotherapy through clinical trials.
It may be very important to zero in on different genes and proteins. However, when actually taking the “targeted” drugs, do the drugs even enter the cancer cell? Once entered, does it immediately get metabolized or pumped out, or does it accumulate? In other words, will it work for every patient?
All the validations of this gene or that protein provides us with a variety of sophisticated techniques to provide new insights into the tumorigenic process, but if the “targeted” drug either won’t ‘get in’ in the first place or if it gets pumped out/extruded or if it gets immediately metabolized inside the cell, it just isn’t going to work.
To overcome the problems of heterogeneity in cancer and prevent rapid cellular adaptation, oncologists are able to tailor chemotherapy in individual patients. This can be done by testing “live” tumor cells to see if they are susceptible to particular drugs, before giving them to the patient. DNA microarray work will prove to be highly complementary to the parellel breakthrough efforts in targeted therapy through cell function analysis.
As we enter the era of personalized medicine, it is time to take a fresh look at how we evaluate new medicines and treatments for cancer. More emphasis should be put on matching treatment to the patient, through the use of individualized pre-testing.
Upgrading clinical therapy by using drug sensitivity assays measuring cell-death (apoptosis) of three dimensional microclusters of ‘live’ fresh tumor cell, can improve the situation by allowing more drugs to be considered. The more drug types there are in the selective arsenal, the more likely the system is to prove beneficial.

One Response to “The Long Tail of Cancer Research”

Gregory Pawelski July 16, 2010

The new paradigm of requiring a companion diagnostic as a condition for approval of new targeted therapies has placed pressure so great that the companion diagnostics they’ve approved often have been mostly or totally ineffective at identifying clinical responders (durable and otherwise) to the various therapies.
It should be in the FDA’s interest in saving the healthcare system perhaps billions of dollars a year (and thereby the healthcare system itself) by ensuring that expensive treatments are used appropriately. It should serve their interest not only in discovering new cancer treatments, but also using currently-available cell culture technologies to improve the effectiveness of existing drugs and save lives today by administering the right drug to the right patient at the right time.
The headlong rush to develop companion diagnostics to identify molecular predisposing mechanisms does not guarantee that a cancer drug will be effective for an “individual” patient. Nor can they, for any patient or even large group of patients, discriminate the potential for clinical activity among different cancer agents of the same class.
The drug discovery model over the last number of years has been limited to one gene/protein, one target, one drug. The “cell” is a system, an integrated, interacting network of genes, proteins and other cellular constituents that produce functions. You need to analyse the systems’ response to drug treatments, not just one target or pathway.
The decoding of the human genome in 2000, sparked hopes that a new era of tailored medicine was just around the corner. However, uncovering the genetic differences that determine how a person responds to a drug, and developing tests, or biomarkers, for those differences, is proving more challenging than ever. As a result, patients with cancer are still being prescribed medicines on a trial-and-error basis or one-size-fits-all.
The key to understanding the genome is understanding how cells work. The ultimate driver is “functional profiling” (is the cell being killed regardless of the mechanism) as opposed to “molecular profiling” (does the cell express a particular target that the drug is supposed to be attacking). While a “molecular profiling” tells you whether or not to give “one” drug, “functional profiling” can find other compounds and combinations and can recommend them from the one test.
The core of “functional profiling” is the cell, composed of hundreds of complex molecules that regulate the pathways necessary for vital cellular functions. If a “targeted” drug could perturb any one of these pathways, it is important to examine the effects of the drug within the context of the cell. Both genomics and proteomics can identify potential new thereapeutic targets, but these targets require the determination of cellular endpoints.
Cell-based “functional profiling” is being used for screening compounds for efficacy and biosafety. The ability to track the behavior of cancer cells permits data gathering on functional behavior not available in any other kind of testing.
Molecular profiling, important in order to identify new therapeutic targets and thereby to develop useful drugs, are years away from working successfully in predicting treatment response for “individual” patients. Perhaps this is because they are performed on dead, preserved cells that are never actually exposed to the drugs whose activity they are trying to assess.
It will never be as effective as the cell “function” methodology, which has existed for the last twenty years and is not hampered by the problems associated with gene expression tests. That is because they measure the net effect of all processes within the cancer, acting with and against each other in real-time, and it tests “living” cells actually exposed to drugs and drug combinations of interest.
It would be more advantageous to sort out what’s the best “profile” in terms of which patients benefit from this drug or that drug. Can they be combined? What’s the proper way to work with all the new drugs? If a drug works extremely well for a certain percentage of cancer patients, identify which ones and “personalize” their treatment. If one drug or another is working for some patients then obviously there are others who would also benefit. But, what’s good for the group (population studies) may not be good for the individual.
Patients would certainly have a better chance of success had their cancer been “chemo-sensitive” rather than “chemo-resistant,” where it is more apparent that chemotherapy improves the survival of patients, and where identifying the most effective chemotherapy would be more likely to improve survival above that achieved with “best guess” empiric chemotherapy through clinical trials.
It may be very important to zero in on different genes and proteins. However, when actually taking the “targeted” drugs, do the drugs even enter the cancer cell? Once entered, does it immediately get metabolized or pumped out, or does it accumulate? In other words, will it work for every patient?
All the validations of this gene or that protein provides us with a variety of sophisticated techniques to provide new insights into the tumorigenic process, but if the “targeted” drug either won’t ‘get in’ in the first place or if it gets pumped out/extruded or if it gets immediately metabolized inside the cell, it just isn’t going to work.
To overcome the problems of heterogeneity in cancer and prevent rapid cellular adaptation, oncologists are able to tailor chemotherapy in individual patients. This can be done by testing “live” tumor cells to see if they are susceptible to particular drugs, before giving them to the patient. DNA microarray work will prove to be highly complementary to the parellel breakthrough efforts in targeted therapy through cell function analysis.
As we enter the era of personalized medicine, it is time to take a fresh look at how we evaluate new medicines and treatments for cancer. More emphasis should be put on matching treatment to the patient, through the use of individualized pre-testing.
Upgrading clinical therapy by using drug sensitivity assays measuring cell-death (apoptosis) of three dimensional microclusters of ‘live’ fresh tumor cell, can improve the situation by allowing more drugs to be considered. The more drug types there are in the selective arsenal, the more likely the system is to prove beneficial.

Comments are closed.