• AI
  • Molecular Imaging
  • CT
  • X-Ray
  • Ultrasound
  • MRI
  • Facility Management
  • Mammography

LDCT Pulls Double-Duty for Lung Cancer, CVD Screening


Using a deep learning algorithm enables providers to screen high-risk patients for both lung cancer and cardiovascular disease.

One the heels of research published last month, a new study reiterates that adding a deep learning algorithm to low-dose CT (LDCT) scans can enable both lung cancer and cardiovascular disease (CVD) screening.

Patients with cancer have a 10-fold greater risk of CVD mortality than the general population, so finding an easier, more efficient way to screen individuals. In a study published May 20 in Nature Communications, a team of investigators from Rensselaer Polytechnic Institute and clinicians from Massachusetts General Hospital show this scan can pull double-duty in patients at the greatest risk, potentially accelerating a diagnosis, facilitating treatment, and improving outcomes.

“Our work shows that, in high-risk patients, deep learning can convert LDCT for lung cancer screening into a dual-screening quantitative tool for CVD risk estimation,” said the team led by Pingkun Yan, assistant professor for biomedical engineering and member of the Center for Biotechnology and Interdisciplinary Studies (CBIS) at Rensselaer. “In this paper, we demonstrate very good performance of a deep learning algorithm in identifying patients with cardiovascular diseases and predicting their mortality risks.”

Related Content: Add Predicting Heart Disease Mortality Risk to LDCT Capabilities

Achieving dual screening is challenging, the team said, because LDCT produces lower-quality images, so picking out features can be difficult. To overcome this obstacle, the team trained their CVD risk predication model on 30,286 LDCT scans gleaned from the National Lung Cancer Screening Trail. From their work, they created an algorithm that can filter out unwanted artifacts and noise while pulling out the features that are fundamental for diagnosis.

Specifically, the team used a separate test set of 2,085 images from the screening trial and achieved an area under the curve (AUC) of 0.871. In addition, with an AUC of 0.768, the algorithm identified patients with high CVD mortality risk. They also validated the model with ECG-gated cardiac CT-based markers, including coronary artery calcification score, CAD-RADS score, and a MESA 10-year risk score.

The team tested their algorithm against the performance of radiologists from Massachusetts General. Not only did the algorithm effectively analyze the LDCTs for CVD risk in high-risk patients, but it also performed equally to the providers. Also, when tested against an independent dataset from 335 patients who had dedicated cardiac CT scans at Massachusetts General, the algorithm produced an equivalent performance.

These results, said Deepak Vashishth, CBIS director, point to the potential of integrated artificial intelligence.

“This innovative research is a prime example of the ways in which bio-imaging and artificial intelligence can be combined to improve and deliver patient care with greater precision and safety,” he said.

For more coverage based on industry expert insights and research, subscribe to the Diagnostic Imaging e-Newsletter here.

Related Videos
Can Fiber Optic RealShape (FORS) Technology Provide a Viable Alternative to X-Rays for Aortic Procedures?
Does Initial CCTA Provide the Best Assessment of Stable Chest Pain?
Nina Kottler, MD, MS
Practical Insights on CT and MRI Neuroimaging and Reporting for Stroke Patients
Related Content
© 2024 MJH Life Sciences

All rights reserved.