Heart disease and lung cancer are serious and potentially deadly health conditions that share many of the same risk factors and are often found in the same people.

If doctors could screen for both conditions at the same time, using the same imaging technology, diagnosis could happen sooner and lives could be saved. Researchers at Rensselaer Polytechnic Institute are working on a project, supported by the National Heart, Lung, and Blood Institute (NHLBI) under the National Institutes of Health (NIH), to combine screening for both cardiovascular disease and lung cancer into one low-dose CT scan.

“The long-term goal, I would say, is that hopefully one day we can make this into the clinical guidelines,” said Pingkun Yan, assistant professor of biomedical engineering at Rensselaer, and lead investigator on this project.

Low-dose CT scans have been an effective technique for screening for lung cancer, Yan said, reducing lung cancer-related deaths by 20 percent. What that technology currently isn’t catching, however, are the patients who also have cardiovascular disease — which can be a larger risk.

“Overall cancer deaths compared with cardiovascular is just a small portion,” Yan said. “More patients in that population died from cardiovascular diseases instead of lung cancer.”

Traditionally, a specialized cardiovascular CT scan has been needed to diagnose heart disease. Now, Yan and his team will apply deep learning methods to low-dose CT technology in order to produce the clearer images that are needed to identify heart disease.

In the first part of this project, Yan and his team will develop, train, and evaluate an algorithm that will be used to produce those clearer images, using a large data set released by the NIH National Cancer Institute that was gathered during a national lung screening trial. 

In the second part of the project, the team will work with Dr. Mannudeep K. Kalra at Massachusetts General Hospital (MGH). Dr. Kalra and his team will share images from patients who have undergone low-dose CT and specialized cardiac CT so that the Rensselaer team can apply their newly developed algorithm to the low-dose CT images and then compare those to the ones produced by expert formulas on the specialized cardiac CT.

From there, Yan said, the researchers hope to produce a road map for doctors so they can read the low-dose CT images and identify cardiovascular disease when it’s there.

“Cardiologists are so used to this specialized cardiac CT they actually don’t have these clinical guidelines of how to read the low-dose CT images and the features that are correlated with cardiovascular images,” Yan said. “The main thing here is we are going to identify those imaging biomarkers and convert those biomarkers into a risk score that can be easily evaluated by doctors.”