Detection of pneumothorax on ultrasound using artificial intelligence

Sean Montgomery*, Forrest Li, Christopher Funk, Erica Peethumangsin, Michael Morris, Jess T. Anderson, Andrew M. Hersh, Stephen Aylward

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

5 Scopus citations

Abstract

BACKGROUND Ultrasound (US) for the detection of pneumothorax shows excellent sensitivity in the hands of skilled providers. Artificial intelligence may facilitate the movement of US for pneumothorax into the prehospital setting. The large amount of training data required for conventional neural network methodologies has limited their use in US so far. METHODS A limited training database was supplied by Defense Advanced Research Projects Agency of 30 patients, 15 cases with pneumothorax and 15 cases without. There were two US videos per patient, of which we were allowed to choose one to train on, so that a limited set of 30 videos were used. Images were annotated for ribs and pleural interface. The software performed anatomic reconstruction to identify the region of interest bounding the pleura. Three neural networks were created to analyze images on a pixel-by-pixel fashion with direct voting determining the outcome. Independent verification and validation was performed on a data set gathered by the Department of Defense. RESULTS Anatomic reconstruction with the identification of ribs and pleura was able to be accomplished on all images. On independent verification and validation against the Department of Defense testing data, our program concurred with the SME 80% of the time and achieved a 86% sensitivity (18/21) for pneumothorax and a 75% specificity for the absence of pneumothorax (18/24). Some of the mistakes by our artificial intelligence can be explained by chest wall motion, hepatization of the underlying lung, or being equivocal cases. CONCLUSION Using learning with limited labeling techniques, pneumothorax was identified on US with an accuracy of 80%. Several potential improvements are controlling for chest wall motion and the use of longer videos. LEVEL OF EVIDENCE Diagnostic Tests; Level III.

Original languageEnglish
Pages (from-to)379-384
Number of pages6
JournalJournal of Trauma and Acute Care Surgery
Volume94
Issue number3
DOIs
StatePublished - 1 Mar 2023
Externally publishedYes

Keywords

  • Ultrasound
  • artificial intelligence
  • pneumothorax

Fingerprint

Dive into the research topics of 'Detection of pneumothorax on ultrasound using artificial intelligence'. Together they form a unique fingerprint.

Cite this