Nature Communications (Feb 2024)

A deep-learning-based framework for identifying and localizing multiple abnormalities and assessing cardiomegaly in chest X-ray

  • Weijie Fan,
  • Yi Yang,
  • Jing Qi,
  • Qichuan Zhang,
  • Cuiwei Liao,
  • Li Wen,
  • Shuang Wang,
  • Guangxian Wang,
  • Yu Xia,
  • Qihua Wu,
  • Xiaotao Fan,
  • Xingcai Chen,
  • Mi He,
  • JingJing Xiao,
  • Liu Yang,
  • Yun Liu,
  • Jia Chen,
  • Bing Wang,
  • Lei Zhang,
  • Liuqing Yang,
  • Hui Gan,
  • Shushu Zhang,
  • Guofang Liu,
  • Xiaodong Ge,
  • Yuanqing Cai,
  • Gang Zhao,
  • Xi Zhang,
  • Mingxun Xie,
  • Huilin Xu,
  • Yi Zhang,
  • Jiao Chen,
  • Jun Li,
  • Shuang Han,
  • Ke Mu,
  • Shilin Xiao,
  • Tingwei Xiong,
  • Yongjian Nian,
  • Dong Zhang

DOI
https://doi.org/10.1038/s41467-024-45599-z
Journal volume & issue
Vol. 15, no. 1
pp. 1 – 14

Abstract

Read online

Abstract Accurate identification and localization of multiple abnormalities are crucial steps in the interpretation of chest X-rays (CXRs); however, the lack of a large CXR dataset with bounding boxes severely constrains accurate localization research based on deep learning. We created a large CXR dataset named CXR-AL14, containing 165,988 CXRs and 253,844 bounding boxes. On the basis of this dataset, a deep-learning-based framework was developed to identify and localize 14 common abnormalities and calculate the cardiothoracic ratio (CTR) simultaneously. The mean average precision values obtained by the model for 14 abnormalities reached 0.572-0.631 with an intersection-over-union threshold of 0.5, and the intraclass correlation coefficient of the CTR algorithm exceeded 0.95 on the held-out, multicentre and prospective test datasets. This framework shows an excellent performance, good generalization ability and strong clinical applicability, which is superior to senior radiologists and suitable for routine clinical settings.