Rumination 13 - This is Science?

Rumination 13. This is Science?

by

Thomas P. Vogl

May 9, 2008

This week I had a repeat PET scan as well as a CT with contrast. My record for correctly predicting the results remains intact, although my level of certainty was lower this time around. (I base my predictions on the scientifically tenuous association between the activity of my immune system and itchy skin eruptions.) I remain stable with marginally significant progression. At this rate, there is plenty of time for me to get run over by a street car rather than a rogue cell. Dr. Shapiro and I discussed alternatives, and the option of trying an Hsp90 inhibitor is worth considering (see future Ruminations). We agreed to stay with AZD-1152 for another two months, changing the monitoring criterion from PET to CT with contrast, to better follow the liver lesions. After the repeat scan in two months we'll reconsider the options.  Hopefully by then there will have been an advance in the immunologic treatment options or the Hsp90 inhibitor studies will have progressed to the point where maximum dose levels will have been established. It is comforting that the progression is slow enough to allow the luxury of waiting for alternatives to develop.

As some of you know, I have been puzzling over the mechanism of action of AZD-1152. The underlying, oversimplified, theory is that cancer cells over-express Aurora-B and that the drug suppresses Aurora-B expression. All well and good, but all cells express Aurora-B; they need it to divide successfully. Details in references in prior Ruminations.

By the law of mass action [http://www.biochem.northwestern.edu/holmgren/Glossary/Definitions/Def-L/law_of_mass_action.html or, in detail, http://en.wikipedia.org/wiki/Law_of_mass_action],

it would follow that a given concentration of Aurora-B inhibitor will have greater effect on normal cells than on those cells that over express it. However, two other considerations override this effect. One is the observed phenomenon called "oncogene addiction" which suggests that the cells that over express a protein do so because they need it to compensate for some other deficiency. They are addicted to it – they need the excess to survive and hence reducing the amount available to the cell will cause it to die (apoptosis). [Combined Depletion of Cell Cycle and Transcriptional Cyclin-dependent Kinase Activities Induces Apoptosis in Cancer Cells, D. Cai et al., Cancer Research, 66: 9270-9280 (Sept. 15, 2006), Also see Oncogene Addiction: Setting the Stage for Molecularly Targeted Cancer Therapy, S.V. Sharma & J. Settleman, Genes and Development, 21: 3214 – 3231 (2007). Addiction to Oncogenes – The Achilles Heal of Cancer, B. Weinstein, Science, 297: 63-64 (5 July 2002) and Mechanism of Disease: Oncogene Addiction – a Rationale for Molecular Targeting in Cancer Therapy, B. Weinstein and A. K. Joe, Nature Clinical Practice 3: 446-457 (August 2006)].

A second factor overriding the law of mass action are the mechanisms for monitoring and repairing errors in cell division or deleting cells that cannot be repaired. This group of evolutionary mechanisms is essential for maintaining viability. One of the cascades in this monitoring system is known as p53, and is known to be defective in many cancer cell lines. [See http://en.wikipedia.org/wiki/P53. An Organometallic Protein Kinase Inhibitor Pharmacologically Activates p53 and Induces Apoptosis in Human Melanoma Cells, K. S. M. Smalley et al., Cancer Research 67: 209 – 217 (Jan. 1 2007)]. This is evolutionary belt and suspenders. When AZD-1152 suppresses Protein Kinase-B in cells it causes the chromosomes to fail to line up properly for cell division. (I'm somewhat oversimplifying.) If the p53 cascade is functioning properly, it will cause the cells with the misalignment to die. If the p53 cascade is not working properly, the cells will divide, but contain so many duplicate chromosomes (polyploidy) that they will be killed by other mechanisms on the next cell cycle.

All very neat. Except, if this is the whole story, why does AZD-1152 (and its analogues) not cure cancer?  Why am I stable rather than in regression? Why is it not obviously the blockbuster everyone has been looking for? Because evolution provides so many alternate pathways for processes that are important for survival and when they go wrong they can go wrong in so many different ways; things are never simple, nor have simple answers. Probably most important from the clinical perspective, one solution will never fit all; each solution, it may turn out, will fit only a small fraction of the patient population.

An additional, well recognized but rarely discussed, problem is that even if a treatment kills 99.999% of all the cancer cells, 0.001% will survive. Those that survive will, of course, be the ones that most effectively resisted the treatment. The cells that are not destroyed by the treatment survive to reproduce – evolutionary pressure in action – and the disease will come back and the treatment that worked before will now fail.

From this, I would conclude that detailed observation of each patient's response, with an emphasis on the mechanisms that cause the treatment to provide only partial success or to fail entirely, is far more important than observing that the expected result ensues. On top of their other problems, current clinical trial design actively avoids these issues.

Both the medical literature and the popular press have been publishing articles about the high failure rate of Phase III clinical trials and the consequent exorbitant cost of these drugs as the drug companies try to recoup the costs of all the failed starts. These astronomical costs have vast side effects on the behavior of drug companies (recall that all side effects, by definition, are undesirable). Three are particularly pernicious.

First, the craving for 'blockbuster' drugs that, ideally, everyone needs to take. Aspirin and the statins come close to this ideal, from the perspective of the drug companies. (The actual need for all this cholesterol lowering therapy is a different question.) The driving force being that unless a new drug has billions of dollars in sales, the drug companies will lose money because of all the drugs that do not make it through the pipeline.

Second, the irresistible temptation to market drugs in order to increase sales, by advertising to doctors and to the public. These marketing efforts employ, lets call a spade a spade, falsified results of the trials to make the drug appear more effective than it really is by fudging the statistics using questionable end-points and/or misrepresenting results by creative writing of scientific papers. The case of Merk's Rofecoxib (Vioxx) (See several articles in the April 16, 2008 issue of JAMA) is just the tip of the iceberg

A review by D Murray et al, Design and Analysis of Group-Randomized Trials in

Cancer: A Review of Current Practices, J Natl Cancer Inst 100: 483 – 491 (2998) found that more than one-third of the trials contained statistical analyses that they considered inappropriate to assess the effects of an intervention being studied. Of those studies, 88% reported statistically significant intervention effects that, because of analysis flaws, could be misleading to scientists and policymakers.

"We cannot say any specific studies are wrong. We can say that the analysis used in many of the papers suggests that some of them probably were overstating the significance of their findings," Murray said. "If researchers use the wrong methods and claim an approach was effective, other people will start using that approach. If it really wasn't effective, then they're wasting time, money, and resources and going down a path that they shouldn't be going down." The review identified 75 articles published in 41 journals that reported intervention results based on group-randomized trials related to cancer or cancer risk factors from 2002 to 2006. Thirty-four of the articles (45%) reported the use of appropriate methods used to analyze the results. Twenty-six articles (35%) reported only inappropriate methods were used in the statistical analysis. Eight percent of the articles used a combination of appropriate and inappropriate methods, and nine articles had insufficient information to even judge whether the analytic methods were appropriate or not.

Third, the abrupt stopping of trials when it appears that blockbuster status will not be met is particularly hard on patients who have volunteered to participate in the trial and are left high and dry even though the drug is showing success in subpopulations of patients. This problem is discussed in a paper by Trotta et al [Stopping a Trial Early in Oncology: For Patients or for Industry? Annals of Oncology, Advanced on-line publication, April 5, 2008 doi:10.1093/annonc/mdn042]. Their conclusion: "Though criticism of the poor quality of oncological trials seems out of place [sic!], unfortunately early termination raises new concerns. The relation between sparing patients and saving time and trial costs indicates that there is a market driven intent. We believe that only untruncated trials can provide a full level of evidence which can be translated into clinical practice without further confirmative trials."

It seems to me that all of these problems would be greatly ameliorated if a way were found to increase the success rate of drug development, especially if the failure rate of Phase III trials initiated because of the apparent success in Phase I and II trials could be diminished. I believe they can, because Trotta's explicit avoidance of commenting on the scientific quality of the trials is disingenuous.

One of the first lessons I learned as a young scientist is that a well designed experiment yields useful information irrespective of the outcome of the experiment. Clinical trials as they are currently conducted violate this cardinal rule. They are designed to address the question 'on what fraction of a population that has disease X does this drug work?' (and we'll discard the drug if the fraction is not blockbuster enough) and that only by the time (with great expense and effort) they get to Phase II/III. Looking at the literature and talking to the clinical trial staff I am struck by the fact that the emphasis is entirely on 'does it work' and at a higher level 'what makes it work' and never on 'why does it not work' or, more generally, 'what modulates how well it works'.

Some examples from my personal experience.

As I reported in previous Ruminations, last Spring and Summer I had several months during which my disease, as measured by repeated PET scans, had not only stabilized but regressed. During that time I was not taking any cancer specific drugs. The metastases started to proliferate in the Fall and I started my participation in the Phase I trial of AZD-1152 in January. {A word of explanation- A Phase I study is designed to determine the maximum tolerable dose of a drug. In practice that means that a few patients get a low dose, and succeeding small groups of patients get increasing doses until the side effects become unacceptable.}The March PET scans demonstrated stable disease as reported in Rumination 12. When I asked one of the senior members of the study staff whether it might not be appropriate, given this history, to check my immune response to see whether it was in overdrive and therefore may account for the stability rather than the drug, I was told  that there is no point in doing that since it would not make any difference in my treatment. This is an astounding response in a Phase I trial which should be occasion for the most wide ranging investigation of the effects, and the side effects, of the drug. Yet the emphasis is entirely on side effects. I get the feeling that the underlying assumption is that we've put so much effort into getting the drug this far it has to work.

I recently asked whether it was not time to schedule another MRI of my brain to check for metastases. I was informed that this would not be a good idea, because if a metastasis were found, I would be kicked out of the trial of AZD-1152. So, as far as clinical trials are concerned, what you know may hurt you. As long as the drug company does not know about the brain metastases, all is well; if they find out, you're out of luck both in terms of getting eliminated from the trial and of the possibility of having appropriate radiation therapy while the lesion is small,. Amazingly, this attitude is the norm.  My friend Sarah was told that because a long standing, slow growing carcenoid tumor was incidentally discovered, she would no longer be eligible for treatment with the Ipi that had already been demonstrated to work successfully on her melanoma.

This is science? I am outraged, discouraged, and disgusted. Who is responsible for scientific curiosity being banned from clinical trials? What a waste of time, money, and opportunity!