I think that depends on the point of view. If you take the view of the cancer cells it might seem that they are more fit to survive inside the whole organism. However, cancer cells don't procreate, thereby their mutated genotype is unable to pass on to next generation anyway, and therefore they do not have a fitness in their own right. If they in effect kill their organism, the metazoan organism will be less fit to pass on its genotype to the next generation.

Yes, Michael makes the key distinction here, between fitness of the organism and the fitness of a particular cancer cell. If a cell survives better and has more progeny than other cells in the tumor, then it has increased fitness, in terms of the cancer. In traditional evolutionary biology terms, though, if a cancer susceptibility allele does not affect the organism's reproductive potential, then it has no affect on fitness.

BRCA2 mutations predispose people to cancer, but that cancer doesn't typically kick in until later in life, often post-menopause. So even though it shortens the lifespan of the individual, it doesn't have much effect on reproductive fitness, which is why it persists in the population.