Cerebral Hemorrhage
Conditions
Keywords
Artificial Intelligence, Computed Tomography
Brief summary
The goal of this observational retrospective study is to evaluate artificial intelligences (AI)'s proficiency in identifying and annotating brain bleeds in computed tomography (CT) images. The main questions it aims to answer are: * Whether AIs at present are capable of analyzing and recognizing cerebral traumas in CT images? * If they are, how competent are they and how can humans take advantages of that? CT images were selected during the normal diagnosis and treatment process of patients, and no one needed to undergo any additional procedures connected to the study.
Interventions
None listed
Sponsors
RenJi Hospital
Study design
Observational model
OTHER
Time perspective
RETROSPECTIVE
Eligibility
Sex/Gender
ALL
Age
16 Years to 92 Years
Healthy volunteers
No
Inclusion criteria
* People who had cerebral traumas with brain bleedings shown in CT images.
Exclusion criteria
* Brain bleedings were not clear or representative in CT images.
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| The identification completeness of the annotated images. | 1 months | Use Photoshop to calculate the area of hemorrhage and the marked area, and then computing the ratio, import into Graphpad Prism for further analysis of the mean percentage and the standard deviation. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| The evaluations from professionals for outcomes produced by AIs | 1 months | The outputs will be evaluated by professional radiologists on a 4-point scale questionnaire from the completeness, accuracy and success of the annotation. Then the results of the questionnaire will be further analyzed in the Graphpad Prism to get the average score and the standard deviation. |
Countries
China
Outcome results
None listed