As a therapy to kill tumors that have formed it apparently works (as shown in this article) but that doesn't fix the underlying pathology that led to the emergence of the cancer (which in many cases can be purely genetic with a low environmental component) To cure cancer one would need to effect repair to these genetic causation factors. It makes no sense at all given what is *rapidly* being discovered about the pathological details that cause cancer as well as other diseases to stop at *treating* cancer when we will very shortly be able to cure it with methods just as non invasive. (in fact likely using denuded viral vectors instead of nanotech ones to deliver the genetic cure as done here as well as kill targeted cancer cells). There is also the potential danger of the constituents of artificially constructed nano-machines are they really broken up completely and flushed out harmlessly or can they serve as the seed for developing new pathology? A viral vector delivered gene therapy would not be subject to this fear as the default payload is removed.
Using nanobots to try and cure cancer is like choosing to hammer a nail by using a pair of vice grips clamped to the end of a hammer. It makes no sense at all given what is *rapidly* being discovered about the pathological details that cause cancer as well as other diseases. I give it 10 years tops to be cured using some form of genetic therapy and I am not restricting this prediction to just the easy cancers; brain cancer, pancreatic cancer, late stage melanoma and lung cancers...all of them are on the same chopping block. Just last year a team of researchers was able to sequence in detail the genetic changes that led to the emergence of lung and skin cancers, they could even tell when the genetic damage occurred and when the cancerous change was induced. I'll try and find the link but it easily stands as the most exciting paper from 2009.
The stem cell research of the last 6 years has opened up an amazing ability to parallel process the discovery phase of all the genes and proteins associated with cell development for particular types of tissue but more importantly allows the gene expression sequence to be spied as the cells go through their stages. The work now is simply performing the necessary reading of the expression patterns (and generated enzymes and proteins) and compare normal to pathological patterns to determine in which genes things go wrong most often and to devise ways to inject repairs using gene therapy in situ. 10 years ago when the human genome was sequenced the worry was that the next step of deciphering the proteome would be next to impossible (given the massive number of possible proteins that can be expressed with our tiny set of 25,000 genes) what was missing was knowledge of how genes might be compartmentalized such that their triggered activity is correlated with particular cellular processes.
Ironically (a huge irony) the investigations of stem cells to determine just what can be learned about their function was spurred in the US by the Bush administration cutting funding for embryonic stem cell development. So what did researchers do? They started playing with other types of stem cells and a few years ago were able to turn normal differentiated cells INTO stem cells in the process known as iPSC (for which the Japanese researcher who pioneered the work will SURELY win a Nobel in a few years time) so this is allowing the expression catalog of the entire human genome to be built by watching stem cells induced to be the required tissue and then "listening" (as in music) for the notes that play out the expressed genes to produce the particular behavior of that cell type. Imagine a floor full of baseball cards and then being tasked with categorizing them by team. One would start out by picking up the cards and looking at the colors on their faces, the player on the surface or even better the team. If they focus on colors they could get the cards cataloged over time but it would take far longer if they focused on team as the teams are relatively invariant over time compared to card color...also the set of teams is reasonably manageable compared to cataloging by publish year or player. The use of stem cells is performing the smart categorization for us of the genes, each tissue type is an analog to a team in which are expressed (from year to year) a particular set of players in positions to execute the game (in hopes of winning and becoming world series winners hopefully) ...If we know the genes that are expressed in the stem cells, it is a short leap to find the differences that make them unique from the differentiated cells of the same stem line which has associated genes which may or may not be consistent between lines (my guess they are) the point is the previously believed to be impossible task of processing through a huge set of possible protein combination blindly was illusory, the system is a record (through it's use of stem cells for each cell type) of the categorization process that is employed to produce all of us...and the smaller set of genes are easier targets to explore when studying pathologies like cancer or other diseases. We are on the verge of a sudden fall off in the lethal capability of tissue born disease like none ever seen before, my only worry is what society will do once these revolutions occur. The death rate due to pathological causes will plummet significantly over the next 10 - 15 years and thus people will live longer and reproduce more vigorously..the Earth is already strained from our misuses of her resources...will we be able to refine ourselves as a species to the point where we drop the selfish ways of the present and past in time to enjoy it?