A Quantitative Approach to Distinguish Pneumonia From Atelectasis Using Computed Tomography Attenuation

J Comput Assist Tomogr. 2016 Sep-Oct;40(5):746-51. doi: 10.1097/RCT.0000000000000438.


Objective: It is known that atelectasis demonstrates greater contrast enhancement than pneumonia on computed tomography (CT). However, the effectiveness of using a Hounsfield unit (HU) threshold to distinguish pneumonia from atelectasis has never been shown. The objective of the study is to demonstrate that an HU threshold can be quantitatively used to effectively distinguish pneumonia from atelectasis.

Methods: Retrospectively identified CT pulmonary angiogram examinations that did not show pulmonary embolism but contained nonaerated lungs were classified as atelectasis or pneumonia based on established clinical criteria. The HU attenuation was measured in these nonaerated lungs. Receiver operating characteristic (ROC) analysis was performed to determine the area under the ROC curve, sensitivity, and specificity of using the attenuation to distinguish pneumonia from atelectasis.

Results: Sixty-eight nonaerated lungs were measured in 55 patients. The mean (SD) enhancement was 62 (18) HU in pneumonia and 119 (24) HU in atelectasis (P < 0.001). A threshold of 92 HU diagnosed pneumonia with 97% sensitivity (confidence interval [CI], 80%-99%) and 85% specificity (CI, 70-93). Accuracy, measured as area under the ROC curve, was 0.97 (CI, 0.89-0.99).

Conclusions: We have established that a threshold HU value can be used to confidently distinguish pneumonia from atelectasis with our standard CT pulmonary angiogram imaging protocol and patient population. This suggests that a similar threshold HU value may be determined for other scanning protocols, and application of this threshold may facilitate a more confident diagnosis of pneumonia and thus speed treatment.

MeSH terms

  • Adolescent
  • Adult
  • Aged
  • Aged, 80 and over
  • Algorithms
  • Computed Tomography Angiography / methods*
  • Diagnosis, Differential
  • Female
  • Humans
  • Lung / diagnostic imaging*
  • Male
  • Middle Aged
  • Observer Variation
  • Pattern Recognition, Automated / methods*
  • Pneumonia / diagnostic imaging*
  • Pulmonary Atelectasis / diagnostic imaging*
  • Radiographic Image Enhancement / methods
  • Radiographic Image Interpretation, Computer-Assisted / methods*
  • Reproducibility of Results
  • Sensitivity and Specificity
  • Young Adult