Midline Shift
| Purpose | Detect and measure midline shift |
| Tag(s) |
|
| Panel | Neuroradiology |
| Define-AI ID | 18030001 |
| Originator | Sumit Niogi |
| Panel Chair | Alex Norbash |
| Panel Reviewers | Neuroradiology Panel |
| License | |
| Status | Published |
| Related RadElement Sets | RDES39 |
Clinical Implementation
Value Proposition
Midline shift is closely associated with intracranial pressure (ICP), leading to brain stem compression and eventual death if untreated. It is critical to manage ICP as quickly as possible. Identifying midline shift, especially early or subtle onset, can significantly improve patient outcomes. In emergency situations, it is not always a neuro-specialist taking the first look at the study. Automated midline shift measurement would provide value to nonspecialists, especially those who read images in emergency situations.Narrative(s)
A 27-year-old male pedestrian is struck by a motor vehicle. The patient is reported to have five minutes of loss of consciousness after the injury. At presentation to the ED, the patient does not remember the accident or how he arrived there. The patient has obvious left-frontal-scalp hematoma and swelling. Given physical exam findings and history of present illness, the ED physician references the ACR Appropriateness Guidelines for Traumatic Brain Injury and orders a noncontrast CT.Workflow Description
The patient receives CT for head trauma. An algorithm receives the entire CT data set. If the algorithm can determine a result (the presence or absence of midline shift, distance of midline shift, and direction of shift), it is returned. A reliability metric, assessing the accuracy of the measurement, would also be helpful. If the midline shift is significant (larger than 5 mm), a modest alert should notify the user.
Additional considerations are as follows: The algorithm executes after the exam is verified on PACS. The algorithm optimally integrates on PACS and dictation or reporting software. The user is then able to automatically populate the report or manually input the results. An indicator image may save to PACS as part of the medical record.
Considerations for Dataset Development
Procedures(s): CT, Head
Sex at Birth: {Male, Female}
Age: [16,90]
Comorbidities: {Hematoma, tumor, abscess}
Technical Specifications
Inputs
DICOM Study
| Procedure | CT, Head |
| Data Type | DICOM |
| Modality | CT |
| Body Region | Head |
| Anatomic Focus | Brain |
Primary Outputs
Midline Shift
| RadElement ID | RDE237 |
| Definition | Measurement of midline shift at maximum (septum pellucidum) and at level of foramen of Monro. |
| Data Type | Numeric |
| Value Set | [0,10] |
| Units | mm |
Secondary Outputs
Midline Shift Direction
| RadElement ID | RDE238 |
| Definition | The direction of midline shift, if present |
| Data Type | Categorical |
| Value Set | 0–Unknown 1–No midline shift 2–Left |
| Units | N/A |