by Gilbert Keith
Today in Mass Spectrometry class, Dr. Higgins talked about a 2002 study that was published in The Lancet (warning: UMN Access required. PubMed Link article access here.) The study used proteomics to generate the profiles of serum proteins from healthy individuals and from individuals suffering from ovarian cancer. The findings, from the abstract, state:
The algorithm identified a cluster pattern that, in the training set, completely segregated cancer from non-cancer. The discriminatory pattern correctly identified all 50 ovarian cancer cases in the masked set, including all 18 stage I cases. Of the 66 cases of non-malignant disease, 63 were recognised as not cancer. This result yielded a sensitivity of 100% (95% CI 93–100), specificity of 95% (87–99), and positive predictive value of 94% (84–99).
There was apparently a lot of hullabaloo raised over the publication of this article. Sorace and Zhan (2003) and Baggerly, Coombs, and Morris (2007) do a lot of good work going over the objections to the conclusions of the Lancet study.
Based on the results I have seen in the Mass Spectrometry course, I feel kind of convinced that in the short term, it will be very difficult for proteomics and mass spec profiling to make waves in the diagnostics world. The technologies still seem to be too expensive, and we still need to target the right variables for these tools (i.e. quantifying a protein might not be enough; we might need to look at ratios of different proteins, etc.). We have also been unable to come up with good solutions to address sample contamination by ubiquitous proteins such as albumin and keratins. I will admit, though, that I don’t know how big of an issue keratin and albumin present; I feel like any contamination would be detrimental, but experts might just be happy enough to acknowledge the problem and move on.
So, in one of my first “thinking about the future” posts, I will say this: I think that before 2020, we won’t have the right set of technologies that might help us come up with robust proteomics-based diagnostic techniques. I will also state that biologists who understand statistics and computational sciences will be worth twice their weight in gold. Biologists who understand biology will only be worth their weight in gold.