Analyzing large quantities of data is just not what humans were meant to do. We are, quite simply, not good at it. Anyone that has spent any time with a hefty spreadsheet knows what I mean. Our brains can never be trained to parse through thousands of pieces of information and put it to good use. We are, after all, creative beings.
Data science, on the contrary, excels at this task. Combing through data, analyzing the information and creating models — we are just getting started. In fact, data science and artificial intelligence (AI) work better with more information, not less. Data science, you see, is an analytical machine.
So what if we combined the two? What if we augment a human being’s experience, expertise and creativity with the parsing and pattern detecting power of data science and artificial intelligence? That, my friends, is where the magic happens.
Here in San Diego, a medical research company is applying this augmented reality approach to saving lives. They are challenging the idea of a standard mammogram to transform how doctors detect cancer. This local company, CureMetrix, has developed a technology called cmTriage, which is the first FDA-cleared triage software tool. The tool works with a human, not in place of a human, to identify critical cases and move them up the list for further analysis.
Normally, radiologists read mammograms using the “first in, first out” method; meaning they are read in the order they are performed in the clinic. But with cmTriage, AI identifies any suspicious cases and alerts the radiologist. By flagging the mammogram, it can be read before the patient leaves the clinic, which can end up saving her life — the magic of combining efforts.
Maybe you are thinking — yeah, yeah, but does it work? Here is what they did:
Radiologist without AI:
To test their technology, they worked with a clinic to analyze 2,129 mammograms screenings that were taken in 2013. That year, without the help of data science and AI, 285 patients were asked back for further screening. Out of these, 61 patients underwent a biopsy; and of those, 15 patients had malignant breast cancer.
Radiologist with AI, the augmented approach:
Examining those same 2,129 mammograms, their AI technology correctly identified 100% of the cancers on the spot. That means the patients would have been flagged immediately and not have to return for the screening. Additionally, the AI technology identified an additional three cancers that were initially missed by the radiologists in 2013 (but diagnosed later). That’s the power of being an analytical machine — improved accuracy.
As we continue down the path of advancing technology, data science, and AI, it’s important to remember that the future doesn’t have to be one or the other. Human or Machine. Creative or Analytical. Experienced vs. New. By combining the two, we draw from the strengths of each, and the results can save lives.