Most cancer therapies work by distinguishing between normal cells and cancerous ones. For example, radiation therapy distinguishes between tumor cells that can't repair themselves and healthy cells that can. Chemotherapy is more harmful to tumor cells than healthy cells because they replicate more frequently. And so on.

But perhaps there's a cleverer way to target tumors by thinking about them differently. Branislav Brutovsky at P. J. Safarik University in Slovakia and his pal Dragos Horvath outline just such a new way of thinking about cancer and say that it could have profound implications for cancer treatment. The key is to treat tumor growth as an evolutionary process and to think of therapy as a way of decreasing the evolutionary efficiency to an extent that the growth cannot survive.

Here's the background. According to Brutovsky and Horvath, the process of tumor growth is one of evolutionary optimization in which competing cells replicate with errors and in numbers that depend on their fitness qualities. Various research seems to indicate that certain cancerous tumors consist of cells with a huge variation in genomes. Some researchers have gone as far as to say there is no typical genotype for breast and renal cancers. It's this kind of evidence that suggests that cancer is somehow solving an evolutionary optimization problem.

That immediately points to a different way of tackling tumors. Instead of distinguishing between healthy and cancerous cells as conventional treatments do, the evolutionary view suggests that tumors can be inhibited by changing the evolutionary attributes of the entire system. Brutovsky and Horvath suggest strategies such as reducing mutation rates, reducing the effective population size, and increasing the generation time of the self-renewing cells. Another idea is to reduce the relative fitness of cancerous cells.

Those all sound like sensible suggestions, but the more difficult task is to translate them into real-world therapies. That's much more difficult, because it is hard to measure or model the evolutionary process at work. What exactly would the fitness landscape look like for cancerous cells, and how might it vary with time? Questions like these are likely to remain puzzles for some time.

But the new thinking does offer some interesting insight. Cancer specialists have long noticed that cancer cells that survive therapy are often more aggressive than those that were killed off, and that treated tumors regrow much faster than untreated ones.

Brutovsky and Horvath point out that when evolutionary algorithms are used to solve problems, engineers often remove a certain population of solutions to accelerate the convergence toward a specific goal. This may be exactly what is happening in therapies that do not remove a decisive population of cells. Instead evolution algorithms may just simplify the fitness landscape in a way that allows the surviving population to thrive.

As Brutovsky and Horvath put it: "A paradoxical consequence of the optimization facet of carcinogenesis is that applied anti-cancer therapy can be, in fact, the causative factor of its acceleration on longer time scales."

Sobering stuff.

Ref: arxiv.org/abs/0907.2004: Optimization Aspects of Carcinogenesis