On 2014 Apr 05, Allison Stelling commented:

The problem lies in the tests themselves.

The pathologists can only be as right as the tests themselves. Immunohistochemistry and standard H&E staining both have a notorious and disturbing amount of inter-observer variability. The entire American medical enterprise is far, far too focused on "cures" rather than the underlying "causes", which results in high percentages of misdiagnoses- not because the pathologists are doing the tests wrong, but because the tests themselves are highly inadequate.

It's somewhat akin to having the world's deadliest arsenal but no map of the enemy territory. I'm working on this from my end with Raman and IR tissue mapping and analysis, should be far more accurate, reliable, and cheaper than anything that uses dyes or labels to sense disease phenotypes. (Esp. when combined with other non-destructive, optical methods.)

For a recent overview of the complexities of brain tumor diagnostics as an example, see Figure 1 in this lovely article on the field: "Refined brain tumor diagnostics and stratified therapies: the requirement for a multidisciplinary approach." Riemenschneider MJ, Louis DN, Weller M, Hau P.

(Note: I put this comment on the PLoS website when this article was first published [Posted by antistokes on 20 Nov 2013 at 13:40 GMT ], but thought I would place it in Commons as well.)

On 2014 Apr 05, Allison Stelling commented:

The problem lies in the tests themselves.

The pathologists can only be as right as the tests themselves. Immunohistochemistry and standard H&E staining both have a notorious and disturbing amount of inter-observer variability. The entire American medical enterprise is far, far too focused on "cures" rather than the underlying "causes", which results in high percentages of misdiagnoses- not because the pathologists are doing the tests wrong, but because the tests themselves are highly inadequate.

It's somewhat akin to having the world's deadliest arsenal but no map of the enemy territory. I'm working on this from my end with Raman and IR tissue mapping and analysis, should be far more accurate, reliable, and cheaper than anything that uses dyes or labels to sense disease phenotypes. (Esp. when combined with other non-destructive, optical methods.)

For a recent overview of the complexities of brain tumor diagnostics as an example, see Figure 1 in this lovely article on the field: "Refined brain tumor diagnostics and stratified therapies: the requirement for a multidisciplinary approach." Riemenschneider MJ, Louis DN, Weller M, Hau P.

(Note: I put this comment on the PLoS website when this article was first published [Posted by antistokes on 20 Nov 2013 at 13:40 GMT ], but thought I would place it in Commons as well.)