A semi-automated approach for measuring splenic injury using computed tomography

Biomed Sci Instrum. 2007:43:13-7.

Abstract

The purpose of this study was to establish a semi-automated method for determining the extent of damage to the spleen from computed tomography (CT) scans. This method uses a software program to evaluate the differences in splenic tissue based on Hounsfield units. In this study, five patients that suffered blunt abdominal injury with splenic injury from a motor vehicle crash were evaluated. First, a trained radiologist identified the damaged areas of the spleens of each patient. From this assessment a semi-automated method of assessment was created using a wide range of Hounsfield units to create a mask of the entire spleen. The contrast range of the injured tissue, consisting of laceration, hematoma, contusion, or other injury was also identified. A percent injured tissue was created by dividing the injured tissue mask volume by the total spleen mask volume. The results demonstrated as the AIS grade increased, the percent injury volume increased. In conclusion, it is possible to calculate injury volumes from abdominal CT scans with minimal user inputs.

MeSH terms

  • Accidents, Traffic
  • Algorithms*
  • Artificial Intelligence*
  • Humans
  • Pattern Recognition, Automated / methods*
  • Radiographic Image Interpretation, Computer-Assisted / methods*
  • Spleen / diagnostic imaging*
  • Spleen / injuries*
  • Wounds, Nonpenetrating / diagnostic imaging*