The topic of new drugs for cancer has come up repeatedly around here - and naturally enough, considering how big a focus it is for the industry. Most forms of cancer are the very definition of "unmet medical need", and the field has plenty of possible drug targets to address.
But we've been addressing many of them in recent years, with incremental (but only rarely dramatic) progress. It's quite possible that this is what we're going to see - small improvements that gradually add up, with no big leaps. If the alternative is no improvement at all, I'll gladly take that. But some other therapeutic areas have perhaps made us expect more. Infectious disease, for example: the early antibiotics looked like magic, as patients that everyone fully expected to die started asking when dinner was and when they could go home. That's what everyone wants to see, in every disease, and having seen it (even fleetingly), we all want to have it happen again.
And it has happened for a few tumor types, most notably childhood leukemia. But we definitely need to add more to the list, and it's been a frustrating business. Believe me, it's not like we in the business aiming for incremental improvements, a few weeks or months here and there. Every time we go after a new target in oncology, we hope that this one is going to be - for some sort of cancer - the thing that completely knocks it down.
We may be thinking about this the wrong way, though. For many years now, there have been people looking at genetic instability in tumor cells. (See this post from 2002 - yes, this blog has been around that long!) If this is a major component of the cancerous phenotype, it means that we could well have trouble with a target-by-target approach. (See this post by Robert Langreth at Forbes for a more recent take). And here's a PubMed search - as you can see, there's a lot of literature in this field, and a fair amount of controversy, too.
That would, in fact, mean that cancer shares something with infectious disease, and not, unfortunately, the era of the 1940s when the bacteria hadn't figured out what we could do to them yet. No, what it might mean is that many tumors might be made of such heterogeneous, constantly mutating cells that no one targeted approach will have a good chance of knocking them down sufficiently. Since that's exactly what we see, this is a hypothesis worth taking seriously.
There are other implications for drug discovery. Anyone who's worked in oncology knows that the animal tumor models we tend to use - xenografts of human cell lines - are not particularly predictive of success. "Necessary but nowhere near sufficient" is about as far as I'd be willing to go. Could that be because these cells, however vigorously they grow, have lost (or never had) that rogue instability that makes the wild-type tumors so hard to fight? I haven't seen a study of genetic instability in these tumor lines, but it would be worth checking.
What we might need, then, are better animal models to start with - here's a review on some efforts to find them. From a drug discovery perspective, we might want to spend more time on oncology targets that work outside the cancer cells themselves. And clinically, we might want to spend more time studying combinations of agents right from the start, and less on single-drug-versus-standard-of-care studies. The disadvantage there is that it can be hard to know where to start - but we need to weigh that against the chances of a single agent actually working