Quality outcomes of reinterpretation of brain CT studies by subspecialty experts in stroke imaging

AJR Am J Roentgenol. 2012 Dec;199(6):1365-70. doi: 10.2214/AJR.11.8358.


Objective: The purposes of this study were to determine the clinical importance and relative value of reinterpretation of brain CT studies by subspecialty experts by assessing the accuracy of interpretation by general radiologists at primary stroke centers and to assess interpretive quality outcomes as a function of change in the treatment of patients with stroke diagnoses or acute presentations of suspected stroke.

Materials and methods: Computerized medical records for the years 2009-2010 at four major community hospitals were queried for primary interpretation of brain CT studies of stroke patients with an acute presentation of either stroke or suspected stroke as diagnosed by board-certified general radiologists (nonneuroradiologists). A central database was queried that allowed one to query by clinical history or symptoms. Secondary interpretation of images of the identified patient sample was then performed by three experienced neuroradiologists. Each case was initially interpreted as an emergency or urgent study by a general radiologist. The reinterpretations performed by a neuroradiologist were scored as concordant or discordant. The discordant studies were categorized as a major discordance if there was a change in clinical management or as a minor discordance if there was no effect on or change in clinical management. The assessment was limited to brain CT studies without contrast administration. CT angiography and perfusion CT studies were not included in the analysis. Patients with hemorrhagic stroke, brain tumors, abscesses, and AIDS or HIV infection were excluded to limit the assessment to ischemic nonhemorrhagic disease.

Results: Of the 560 studies reviewed, 14 studies (2.5%) were identified as discordant. Of those discordant studies, four (0.7% of the original 560) were categorized as major discrepancies necessitating a change in clinical management. Ten (1.78%) were categorized as minor discrepancies, for which there was no change in management. There were no permanent adverse outcomes with respect to morbidity and mortality as a result of a discrepant interpretation, as determined by chart review or communication with the attending or referring physician by the secondary reader.

Conclusion: Most of the interpreted head CT cases read by board-certified general radiologists for patients presenting with stroke or stroke symptoms did not result in discordant interpretations as verified by subspecialty experts. Discordant interpretations did not result in changes in clinical management in most cases. Double reading of head CT scans for these patients by subspecialty experts appears to be an inefficient method of substantially improving imaging health quality outcomes in stroke.

Publication types

  • Multicenter Study

MeSH terms

  • Adult
  • Aged
  • Aged, 80 and over
  • Algorithms
  • Clinical Competence
  • Diagnosis, Differential
  • Diagnostic Errors / statistics & numerical data
  • Female
  • Hospitals, Community
  • Humans
  • Male
  • Medical Records Systems, Computerized
  • Middle Aged
  • Neuroradiography / standards*
  • Radiology / standards*
  • Reproducibility of Results
  • Retrospective Studies
  • Stroke / diagnostic imaging*
  • Tomography, X-Ray Computed*