Machine learning proves a plus in cardiopulmonary care

Researchers have shown that machine learning of 3D patterns in cardiac MRI can improve accuracy when it comes to predicting survivability of—and potentially guiding care for—patients with high blood pressure in the lungs.

Timothy Dawes, FRCA, of University College London Hospitals in England, and colleagues had their findings published online Jan. 16 in Radiology.

The team enrolled 256 patients with newly diagnosed pulmonary hypertension. The patients underwent cardiac MRI, right-sided heart catheterization and six-minute walk testing with a median follow-up of four years.

The researchers used a complex 3D model of cardiac displacement and applied a machine-learning algorithm to identify recurring patterns indicative of various outcomes.

“From standard diagnostic imaging, a mathematical model of the relationship of cardiac function to survival can be generated,” the authors explain.

At the end of follow-up, 36 percent of the patients (n = 93) had died, and one underwent lung transplant.

The team found their supervised machine-learning survival model had enabled sharper prognosis than conventional imaging with hemodynamic, functional and clinical markers (area under the receiver operating characteristic curve, 0.73 vs 0.60, respectively [P < .001]; difference in median survival time between high- and low-risk groups, 13.8 vs 10.7 years, respectively [P < .001]).

They further found that, from conventional cardiac MR imaging, a disease-specific cardiac atlas can be used to create accurate and reproducible segmentations of the heart in pulmonary hypertension.

“Computational analysis of right ventricular motion in pulmonary hypertension can be used for risk stratification and demonstrates early prognostic signs of dysfunction,” Dawes et al. write. “Machine learning by using cardiac MR imaging should be evaluated as a tool to guide patient management.”

The authors acknowledge several limitations in the design of their study, which encompassed all noncongenital cases of, and all treatment regimens for, pulmonary hypertension. This pragmatic approach may limit applicability in selective groups, they note. Nevertheless, they add, the study surely does demonstrate the prognostic value of their machine learning model across an array of disease states and treatments.

Dave Pearson

Dave P. has worked in journalism, marketing and public relations for more than 30 years, frequently concentrating on hospitals, healthcare technology and Catholic communications. He has also specialized in fundraising communications, ghostwriting for CEOs of local, national and global charities, nonprofits and foundations.

Trimed Popup
Trimed Popup