Scientists developing machine-based algorithm for heart failure and emphysema Dx

As the med tech industry casts its eye toward quick, point-of-care diagnostics, scientists at MIT are teaming up with physicians from Harvard Medical School to create new technology that could help paramedics more quickly distinguish between acute emphysema and heart failure by measuring an individual's breath.

Researchers from MIT's Research Laboratory of Electronics and physicians from Harvard Medical School and Philadelphia's Einstein Medical Center are developing an algorithm that can determine whether a patient is suffering from emphysema or heart failure by using readings from a capnograph, a machine that measures the carbon dioxide concentrations in exhalations.

The findings, published in the December issue of IEEE Transactions on Biomedical Engineering, could lead to an alternative to a traditional blood test and provide a more accurate diagnosis, MIT News reports.

George Verghese

In the past, physicians have used capnograph machines to make sure breathing tubes are inserted correctly into sedate patients. If a tube is inserted properly, the capnograph displays a regular wave pattern; if it hasn't been, the capnograph shows a flatline reading. The machine is found in every emergency room and operating room, George Verghese, one of the study's co-authors and professor of biomedical engineering at MIT, told MIT News.

MIT scientists and their colleagues compared capnogram readings from healthy patients with those of sick patients. Over time, the researchers observed subtle but consistent difference between the readouts. The scientists used the information to create an algorithm that looks at patterns in capnogram readings and can determine a patient's ultimate diagnoses.

In particular, researchers found that the algorithm yielded a 0.98 true positive rate when distinguishing healthy subjects from those with emphysema, and a 0.89 true positive rate when separating patients with congestive heart failure from those with emphysema.

Scientists are testing their findings in the field, analyzing capnograms from paramedics to see whether the algorithm can measure the severity of asthma attacks and sedation in patients undergoing medical procedures. In the future, the technology could be used to warn physicians in advance about potential complications or help paramedics more accurately attribute shortness of breath in patients, Steven Green, a professor of emergency medicine at Loma Linda University, told MIT News.

"I am convinced that in 10 years, the fruits of this work will be commonplace in medical practice," Green said. "There are two real, pressing needs for which this work has great promise, and likely more will be identified as the technology progresses."

- read the MIT News story
- here's the IEEE Transactions on Biomedical Engineering abstract