Garbage In, Gospel Out

There is an old saying called "Garbage In, Garbage Out" This saying was coined by George Fuechsel an IBM hack. Other such terms like FUBAR, SNAFU, and even KIBO reflect some of the issues we have going on with personalized medicine and even medicine as a whole.

This term is especially poignant today when the Wall Street Journal casts a shadow on the field of pathology and personalized medicine testing for breast cancers. The drug Herceptin is one of the Vanguards of what I called personalized molecular medicine therapy.

You see, Herceptin therapy is targeted towards a certain type of breast cancer. In this type of breast cancer your cells have a protein located on them that can encourage cells to grow. When this protein is blocked, cells die. Therefore, Cancer is beaten back. The catch, if you don't have this protein in your cancer, you don't respond to Herceptin.

This is the paradigm for personalized medicine. We test your cancer to see if it has this molecule. If it does great we give you Herceptin. If it doesn't, too bad, it's regular chemo for you. But what happens when GIGO takes over? What if the tumor sample is bad? What if the lab analysis is bad? What happens when the lab technicians are inexperienced? Garbage In...

So what do you get? Patients on Herceptin who may have not benefited from this therapy and perhaps would have benefited from the standard treatments. Simply because their test results were reported as positive. Even if this report was a mistake. You see, doctors got caught up in the worst acronym of all "Garbage In, Gospel Out" This term was coined to essential remind us how we often treat computer output as infallible. This is the very same case in medicine with pathology laboratories. I often go to the lab myself to double check results. Time consuming? Yes. Life saving? Absolutely!

The Sherpa Says:

What happens if the same thing occurs with whole genome analysis? Last time I checked, Copy Number Variation, Non-coding sequence, "Dark Matter" are all part of our genomes. Do we have a good handle on how well we can sequence this stuff? More importantly, we must not let genomes become Garbage In, Gospel Out"