I've only looked at the abstract (damn publishers), but superficially this seems to be a great study. I'd just like to underscore how amazing their samples are. The mean survival time of their patients, 2 years, is also surprising. (Survival bias?)
Getting at metastases is hard. In general it simply isn't medically necessary to biopsy them. Without medical necessity it's unjustifiable to put a patient through that risk and discomfort. Why are metastases important? Mostly because they're presumably cleaner than primary tumors.
In a primary tumor you start out some founder cell that has lost the ability to undergo programmed cell death. In general, DNA repair mechanisms are also awry. The cell divides producing daughter cells, the mutations add up and so on. The tumor becomes heterogeneous. The daughter cells diverge from one another. If you sample from the primary tumor you get plenty of mutations related to cancer, and plenty of mutations unrelated to cancer, simply because the genome is a big place. It is no fun to work with this data.
OK. Metastases. Some immortalized cell has lost basal lamina adhesion and migrated to another site in the body. This is a selective bottleneck -- all of those randomly mutated cells that didn't lose adhesion are stuck in the primary. The secondary should therefore be enriched for mutations in oncogenes and tumor suppressors. This is why this data is so important and could give some great insights into oncogenesis.
Heck, here's an experiment I would love to run:
* Collect biopsies from primary tumors in non-metastatic cases.
* Collect biopsies from metastases in another patient population.
* Look for mutations and changes in genome architecture that are unique/overrepresented in the metastatic cases.
Having lost my father to a pancreatic cancer, this is excellent news. His case doesn't seem to be an exception and it was utterly terrible (< 4 months between the first symptoms and death). Pancreatic cancers are awful, and don't get anywhere near enough attention.
Haven't had a chance to read the article yet, but I was under the impression that the lethality of pancreatic cancer (and a few others) was mainly due to the fact that it's so damned hard to identify until it's beyond the point of treatment. For example, during the period when the tumor is small, contained and treatable, it (again, AFAIK) doesn't tend to manifest any symptoms, or the symptoms which exist are identified as some other, more common issue.
> was mainly due to the fact that it's so damned hard to identify until it's beyond the point of treatment.
Yes indeed, which is why I talked about the first symptoms (though I could have added the precision of the first detected symptoms, in my father's case a strong pain in the lower back for no apparent reason). According to the article, pancreatic cancers are actually pretty slow and long-lived but they tend not to make themselves known until the very end when they've already started to metastasize and it's too late for treatment.
What I meant by "pancreatic cancers are awful" is mainly that the prognosis at the moment is one of the worst possible.
Great insight into tumor development. It reminded me of a scene in Dragon's Egg, a Sci-Fi story about life on a neutron star, where the aliens cure a female space travelers future breast cancer, by irradiating a tiny cluster of mutated cells.
Getting at metastases is hard. In general it simply isn't medically necessary to biopsy them. Without medical necessity it's unjustifiable to put a patient through that risk and discomfort. Why are metastases important? Mostly because they're presumably cleaner than primary tumors.
In a primary tumor you start out some founder cell that has lost the ability to undergo programmed cell death. In general, DNA repair mechanisms are also awry. The cell divides producing daughter cells, the mutations add up and so on. The tumor becomes heterogeneous. The daughter cells diverge from one another. If you sample from the primary tumor you get plenty of mutations related to cancer, and plenty of mutations unrelated to cancer, simply because the genome is a big place. It is no fun to work with this data.
OK. Metastases. Some immortalized cell has lost basal lamina adhesion and migrated to another site in the body. This is a selective bottleneck -- all of those randomly mutated cells that didn't lose adhesion are stuck in the primary. The secondary should therefore be enriched for mutations in oncogenes and tumor suppressors. This is why this data is so important and could give some great insights into oncogenesis.
Heck, here's an experiment I would love to run:
* Collect biopsies from primary tumors in non-metastatic cases.
* Collect biopsies from metastases in another patient population.
* Look for mutations and changes in genome architecture that are unique/overrepresented in the metastatic cases.