That was the quaint phrase used by one of the presenters at the recent AACR-EORTC-NCI Molecular Targets meeting in San Francisco.

Apparently, some drug or two was considered, too toxic (fair enough) or lacking in efficacy, hence the requisite binning of a multi-million dollar program to the scrapheap.

Yesterday’s post, however, reminded me that maybe sometimes, it’s not that the efficacy was lacking but the clinical trial design or tumor type or even line of therapy was the best one.  Let’s consider a couple of recent ideas here:

  1. The aurora kinase inhibitor PHA-739358 didn’t show any efficacy in adenoncarcinoma of the prostate, but the target, aurora kinase A may be a key one in some neuroendocrine tumours of the prostate.  These are very different subsets requiring a different approach to patient selection criteria and screening, which might potentially lead to a higher response rate in a small subset.
  2. At the above AACR meeting, I was discussing mTOR inhibitors in breast cancer with a few people.  Everyone noted how interesting it was that Wyeth’s temsirolimus failed to show any efficacy in a large phase III trial in women with ER/PR+ newly diagnosed breast cancer when given an aromatase inhibitor and the mTOR.  In contrast, Novartis took a different approach and used the AI and mTOR combination in second line therapy using everolimus and exemestane and saw dramatic responses. Why the difference?  Well, mTOR is known to cause resistance to AI over time, so it would make more sense to add it in later, rather than upfront.

There are many many other examples like this.  Sometimes, the key is in better understanding of the underlying processes from basic research.

For me then, dog drug heaven might not always be due to a poor molecule, but a failure to figure out where and how the drug might have worked effectively.  Dr Len Saltz (MSKCC) summed this up nicely at the NY Chemotherapy Symposium earlier this month:

Screenshot 2011 11 29 Gone to dog drug heaven...

Now, while Dr Saltz was specifically discussing the potential role (or lack of) for PI3K inhibitors in colorectal cancer, I do think his maxims hold very true for any targeted agent being evaluated in the clinic and something that cannot be emphasized enough.

The first point is obvious, but many sadly seem to miss it!  More preclinical and translational research is key to determining what the targets are and which ones matter in which tumor types.  Without that rational approach, you might as well throw mud at a wall and see what happens.  The second point speaks to the therapeutic index of the drug and whether we are shutting down the pathway enough to stop aberrant activity.  The final point is absolutely crucial – is the target a driver or a passenger?  If it’s the latter, the first two will not matter a jot no matter what we throw at it, in fact all that happens there is more toxicities are introduced and that’s not a good thing for the patient on the receiving end.

These issues become even more pertinent when we consider how regimens and increasingly, clinical trials, are moving more towards double and perhaps even triple combination therapies in an effort to shut down a pathway more completely.

In the meantime, the dog drug heaven days will likely continue.