Related Experiment Video
Updated: Aug 6, 2026
Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
Published on: April 13, 2013
Assessment of an Automated CT Registration and Subtraction Tool to Improve Radiologic Evaluation of Patients with
Maram Alafif1, Lee J Hsieh2, Roland R Lee2
1From the Department of Radiology (M.A., R.R.L., J.D.R., R.T., N.F.), UC San Diego Health; UC San Diego School of Medicine (L.J.H.); Department of Radiology (R.R.L.), Medical Center, San Diego, VA and Scripps Clinic Medical Group (J.D.R.). maramalafif@gmail.com.
Background And Purpose:
CT of the head is the imaging modality of choice for assessing patients with intracranial hemorrhage. Relative differences in head positions can result in tilted and unaligned images, making it challenging to assess changes in intracranial hemorrhage. NeuroAlign CT is an FDA-approved software that performs automated alignment, co-registration, and subtraction of sequential CT images. In this study, we evaluated whether NeuroAlign CT could improve reader speed, confidence, and accuracy in detecting changes in intracranial hemorrhage.
Materials And Methods:
201 head CTs with intracranial hemorrhage and follow-up imaging were identified. Three neuroradiologists classified the evolution of intracranial hemorrhage with and without NeuroAlign CT with a washout period between. Each neuroradiologist was timed as they rated changes in intracranial hemorrhage as increased (+1), no change (0), or decreased (-1). They documented their confidence on a scale from low (0) to very high (3). Wilcoxon signed-rank tests were used to compare efficiency and confidence. McNemar's Chi-square tests were used to compare accuracy. A p-value of <0.05 was considered statistically significant.
Results:
Of 201 cases, intracranial hemorrhage increased in 55, was unchanged in 142, and decreased in 4 cases. Utilizing NeuroAlign CT decreased average read time (71 vs 98 seconds, p <0.0001), increased average confidence (2.81 vs 2.57, p <0.001), and improved accuracy for cases of increased ICrH (92.7% vs 80.0%, p =0.046). Of note, NeuroAlign CT increased the rate of "very high confidence" (81.1% vs 59.7%, <0.001). There was no significant difference in overall accuracy when considering all cases (p =0.25).
Conclusion:
Among experienced neuroradiologists, utilizing NeuroAlign CT improves speed and confidence when assessing intracranial hemorrhage on serial head CTs, and may improve accuracy in detecting worsening intracranial hemorrhage, which may ultimately lead to improved patient outcomes.
