Data Access Among Keys to Personalized Cancer Therapies

| Comments (0)

By Dean Lee, M.D., Ph.D., assistant professor, Pediatrics, Cell Therapy Section

dean anthony.jpgPediatric brain tumor research has made some important advances recently. Articles on pediatric brain cancers regarding genomic sequencing, hedgehog pathways, p53 degradation and EGFR signaling have shown up this year in a variety of respected scientific publications.

One editorial on this progress proclaimed "The Time Has Come for Many Things in Medulloblastoma." Translating those advances to clinically proven therapies is, as always, the hard part.

As you may have heard or seen, IBM's supercomputer, Watson, made an impressive showing recently on the "Jeopardy" game show, beating two previously undefeated champions. Watson was able to answer trivia questions by applying pattern recognition to analyze the question and sort it into key data components. Then, the computer quickly interrogated multiple large databases to select the most likely answer to that question.

Last week, the folks at IBM presented Watson to the public at a special seminar at Rice University with the explicit purpose of discussing the potential real-life applications of such computers.

Calling on computers
While we may not really care whether IBM is able to create a computer that can beat our best "Jeopardy" geniuses, we would care about a computer that can help physicians interrogate large sets of diagnostic or outcomes data as they pertain to the questions we have about a given patient.

This futuristic concept goes far beyond data mining. Instead, it applies pattern-recognition technology to very large data sets to determine the diagnosis that best fits the patient data. Based on its analysis, it could potentially determine a treatment plan that best fits unique disease characteristics.

Here's what it could look like:

  • Watson could take my question, "What is wrong with me?" and the symptoms I input (for instance, fever, anemia, bruising, bone pain and a high white blood count), and all of that goes into its key data components.
  • A query of the world of data is conducted to determine what disease I most likely have. Do I have cancer or a viral infection?
  • Once a diagnosis is selected, Watson could then tell me which treatment is most likely to cure me.

This is, perhaps, the critical leap needed to achieve truly personalized medicine. Computerworld magazine recently published an article illustrating more of Watson's potential applications in medicine.

What about the data?
The actual success of such a system, however, may lie not in what technology we can create, but on what data we are willing to release. The more data we have to compare, the more sure we can be of the answer.

At the moment, privacy laws including HIPAA (Health Insurance Privacy and Portability Act) go to great lengths to keep our health information private and secure. If we want Watson to come up with an answer he is sure about, and do it in time to hit the buzzer, he must have access to the data.

While my mind may not be as iron-clad for information as Watson's is, I have the distinct opportunity as a physician to base my decisions on years of patient data -- the data I have in my mind about my own patients. That is data and experience that, at the moment, is off limits to Watson.

MD Anderson is leading the way in developing personalized therapies for cancer. At some point, we as a society will have to decide if having our personal health data available to such systems is worth the improved outcome promised.

Perhaps we will be more trusting of Watson, perhaps not.

Leave a comment


Connect on social media

Sign In