Johns Hopkins unveils new computational method for precision oncology
One of the ongoing dilemmas faced by provider organizations with precision medicine is that for all the advances made in genomic research, sometimes it can still be hard to translate into routine clinical practice: Physicians don't always know how best to turn genetic-based data into appropriate treatments.
A key challenge for clinicians is that each primary form of cancer, such as breast or prostate, may have multiple subtypes, each of which responds differently to a given treatment.
Healthcare IT News is reporting this week from the HIMSS Precision Medicine Summit in Washington, D.C. Also this week, researchers at nearby Johns Hopkins announced what they say is a new computational strategy that can help translate complex precision medicine data into a more simplified format that keeps the focus on patient-to-patient variation in the molecular signatures of cancer cells.
"One of the things that people in this field have noticed over the past 10 years – and, in fact, it has been startling – is how much heterogeneity there is even between two patients with the same subtype of cancer," said Donald Geman, professor in Johns Hopkins' Department of Applied Mathematics and Statistics, in a statement announcing the findings.
"By that, I mean that in two patients who were both diagnosed with melanoma, the skin lesions may look quite similar to the naked eye but the cancerous cells may be very different at the molecular level. They may have different forms of dysregulation, including different genetic variants and different gene expression profiles."
But helping physicians to know as much as possible about the genetic makeup and biological pathways specific patients can help them make more informed decisions about their prognoses and treatments, helping to adjust them to the particular molecular profile.
So Geman and his team have developed an mathematics-based approach, which is detailed in the Proceedings of the National Academy of Sciences, for helping bring some simplicity and order to that complex data.
The result is something that's not unlike the bloodwork summaries commonly produced for docs when a patient has an annual physical exam, according to Johns Hopkins – the lab tests that show in a simple format when blood sugar, cholesterol and other results are outside of the appropriate range.
The researchers say they've developed a way to greatly simplify the data – on tens of thousands of molecular states – by converting them to binary labels that basically show whether a patient's measurement falls within or beyond healthy levels.
Geman sees a big future for other similar research using advanced math to help simplify complex medical data for integration in routine clinical care.
“The goal is taking classification problems of genuine clinical interest and producing an algorithm that is accurate, interpretable and makes sense biologically," he said.