Big Data to Predict Coronary Artery Disease

One common complaint and issue with our healthcare system is the prevalence of unnecessary testing on patients for what turns out to be both relatively harmless and serious diseases. All of this extra testing that is currently done costs patients and insurance companies untold amounts of money every year. One example of this is patients who visit the doctor complaining of chest pain being recommended for further testing. A common current test for patients who are suspected to have obstructive coronary artery disease (CAD) is called stress myocardial perfusion imaging or MPI. In 2010 alone, more than 10 million MPI’s were performed in the United States. Following an MPI, patients whose results show probable signs of CAD are subjected to further invasive testing, which can be dangerous to the patient’s health and expensive, but also completely necessary if the patient is thought to have such a serious disease. However, when that sort of invasive testing is done on patients who actually exhibit little to no signs of CAD upon anatomical examination the testing is just dangerous to the patient’s health and expensive, with little to no benefits. A reason this occurs is because MPI can be unreliable in accurately predicting whether a patient truly needs further invasive testing or not. More food for thought is that if the MPI test is generating false positive results, chances are it is also generating false negative results, which means people are not getting the further testing they need to help diagnose potential heart issues.

 

 Source: doctorstock.photoshelter.com

Fortunately for us, Big Data has been set upon the this issue and researchers at a company called CardioDx, Inc. used big data to research a new way to better predict those who need additional invasive testing for CAD. The company used the open-source R statistics language to analyze 10 gigabytes of human genome data in order to find genes linked with CAD. After completing research, they were able to identify 23 genes which were able to predict the presence of CAD in patients. From this, the team of researchers was able to create a simple blood test to find those patients who needed more invasive CAD testing, which they termed the Corus CAD®. The creation of this test was announced in January of 2011, but clinical testing of the test was only recently completed, with the results of these clinical studies released about a month ago.

 

 Source: http://www.nature.com/nrc/journal/v3/n3/box/nrc1010_BX1.html

The results of the clinical study were clear: the new blood test was more accurate than the current CPI in predicting the occurrence of coronary artery disease. With a high negative predictive value and high sensitivity, the test has the potential to be much more efficient and accurate than the current CPI test. By having a negative predictive value, this means the Corus test will accurately give an accurate negative result to the patient. By using this test, patients and insurance companies will be saved from millions of unnecessary dollars being spent on perfectly healthy patients forced to undergo invasive testing.  With nearly 10,000 patients coming into American medical centers every day complaining of chest pain, it is more important than ever to be able to quickly distinguish those who have serious heart problems from those who are healthy, or have some other cause of their pain. This is just another example of how Big Data can transform our lives for the better, with all of this progress coming from a company using the same open-source software that all of us have access to if we wish to use it. It makes you think of the possibilities a single dedicated individual with a good set of data could create to change the world.  

 

Sources:

http://www.revolutionanalytics.com/news-events/news-room/2011/CardioDX-uses-Revolution-Analytics-to-Predict-Coronary-Artery-Disease.php

http://www.marketwatch.com/story/cardiodx-announces-publication-of-compass-study-demonstrating-the-corus-cad-test-outperforms-myocardial-perfusion-imaging-in-overall-diagnostic-accuracy-for-obstructive-coronary-artery-disease-2013-02-19