Back to all papers

Deep learning model for osteoporosis screening on chest CT with low tube voltage.

November 10, 2025pubmed logopapers

Authors

Zhang Y,Shao R,Zhang K,Wang L

Affiliations (3)

  • Department of Radiology, Nantong First People's Hospital, Nantong, Jiangsu, China.
  • School of Electrical Engineering, Nantong University, Nantong, Jiangsu, China.
  • Department of Radiology, Nantong First People's Hospital, Nantong, Jiangsu, China. [email protected].

Abstract

The purpose of this study is to develop a deep learning (DL) model for osteoporosis screening based on images of thoracic vertebrae obtained from chest CT with a low tube-voltage of 100 kV, with quantitative computed tomography (QCT) as a reference standard. From May 2022 to January 2024, a total of 649 patients who underwent low tube-voltage chest CT and lumbar QCT were retrospectively included in this study, with 518 cases in the training set and 131 in the testing set. Patients were classified as normal bone mineral density (BMD), osteopenia and osteoporosis according to QCT. The CT images of T5-T8 were input and DL models were established through a two-level network of Bone-PSPNet and Ost-ClassNet. Receiver operating characteristic curve analysis was used to evaluate the classification performance of the DL models. For distinguishing patients with osteoporosis from non-osteoporotic ones (normal BMD and osteopenia), the basic DL model achieved a high sensitivity of 1 and an AUC value of 0.978. For distinguishing low BMD (osteopenia and osteoporosis) from normal BMD, the basic DL model achieved a sensitivity of 0.969 and an AUC value of 0.945. This study developed a DL model based on low tube-voltage chest CT to automatically identify osteoporosis, which has shown good classification performance.

Topics

Journal Article

Ready to Sharpen Your Edge?

Join hundreds of your peers who rely on RadAI Slice. Get the essential weekly briefing that empowers you to navigate the future of radiology.

We respect your privacy. Unsubscribe at any time.