Automated Leg Length Discrepancy Assessment on Radiographs

Purpose	Quantify the lower extremity limb lengths on standing radiographs
Tag(s)	Limb length, leg length, discrepancy, lower extremity
Panel	Musculoskeletal Panel
Define-AI ID	19050029
Originator	David Sandman
Lead	David Sandman
Panel Chair	Jay Patti
Panel Reviewers	Musculoskeletal Panel
License	Creative Commons 4.0
Status	Public Commenting
RadElement Set	RDES102

                               

Value Proposition

Leg or lower extremity limb length discrepancy is a correctable problem affecting the pediatric and adult populations. Several methods of measuring the lower extremities have been described, each of which is prone to both intra-reader and inter-reader variability and can be a tedious task to complete. Automated measurements of the lower extremities can provide an accurate and reproducible measuring technique and result.

Narrative(s)

65 year old male with severe medial compartment osteoarthritis of the left knee. Total knee arthroplasty is planned, goals of which include attaining limb length equality and physiologic alignment. Weightbearing long-leg/lower extremity radiographs is ordered as part of the preoperative work-up. An algorithm measures the lower extremity lengths, auto-populating the data when a radiologist or other clinician reviews the study.

Workflow Description

Images obtained from the radiography unit are sent to PACS and the AI engine. The images are analyzed by AI. Coordinates of areas of interest are assessed using a measuring tool. Annotated measurements are saved as an image along with a summary of the measurement data set.

Procedures(s)	{ lower extremity limb length radiographs, long-leg radiographs}
View(s)	{ Upright/standing-radiographs}
Sex at Birth	{Male,Female}
Age	[0,90]

Inputs

 

DICOM Study

 

Primary Outputs

 

Bilateral Lower Extremity Length

 

 

Bilateral Lower Extremity Length Coordinates

 

 

Bilateral Femur Length

 

Bilateral Tibia Length

 

Bilateral Bone Length

 

 

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