Automated Measurements of Foramina on CT

Purpose	Create automated measures of foraminal stenosis on cross-sectional imaging studies of the lumbar spine
Tag(s)
Panel	Musculoskeletal
Define-AI ID	20050034
Originator	Douglas Mintz
Lead	Douglas Mintz
Panel Chair	Jay Patti
Panel Reviewers	Musculoskeletal
License	Creative Commons 4.0
Status	Public Comment
RadElement Set	RDES163

                               

Value Proposition

Having an automated measure of stenosis would improve radiologists' efficiency and consistency in reading cross-sectional spine studies. It could also be the starting point for research, such as the creation of normative data.

Narrative(s)

A 60 year old patient with back pain comes for a pre-operative lumbar spine CT scan to assess the degrees of bony stenosis.

Workflow Description

The CT scan is analyzed by background AI software. The software verifies the numbering of the lumbar spine and provides automated measurement in square millimeters for the neural foramina at each imaged level left and right. A sample image would be provided in order to confirm the algorithm used the correct image. The algorithm should be able to provide a numeric value for the cross sectional area of the foramen in the sagittal plane defined by the location of the pedicle for the L5/S1 foramen and sequentially every visible foramen superior to this foramen.

The system auto-populates those parts of a report for the reading radiologist to confirm. The radiologist completes the CT report.

Inputs

 

DICOM study

Primary Outputs

Left Neural Foraminal Area Measurement

Right Neural Foraminal Area Measurement

• Expansion to other parts of the spine (cervical and thoracic) and to other modalities (MRI)

• Research potential includes the creation and confirmation of normative data for spinal canal and foraminal areas

• Include studies that have surgical hardware

No known related public datasets at this time,  please alert us if you know of any.