Multimodal Modeling of the Knee Joint
- Conditions
- Knee Arthroplasty
- Registration Number
- NCT04179812
- Lead Sponsor
- University Hospital, Brest
- Brief Summary
The trial will contribute to the development of an innovative algorithm that aims to simplify and improve preoperative knee surgical decision. This will be made by automatically extracting anatomical informations from different images acquired on the patient (scanner, MRI, radiography).
- Detailed Description
The source population will consist of at least 50 patients whose following modalities of images of the leg are present in the database: scanner, MRI, radiography. No restrictions on age or gender are required.
A dedicated algorithm adapting statistical models of appearance to the specific morphology of the patient acquired through an MRI and / or scanner and radios will then be developed.
Recruitment & Eligibility
- Status
- UNKNOWN
- Sex
- All
- Target Recruitment
- 210
- Patients having the 3 modality of images : scanner, MRI, radiography
- Patients having osteoarthritis or knee trauma
- Images showing the articualtion of the knee
- having formulated his non opposition
- Presentation of a partial view of the knee or any view that does not allow a good segmentation of the total knee joint.
- Refusal to participate
Study & Design
- Study Type
- OBSERVATIONAL
- Study Design
- Not specified
- Primary Outcome Measures
Name Time Method Obtaining a similar knee segmentation for all patient regardless of the input image (MRI, CT or radiography) Inclusion - Day 0 Creation of an algorithm that will automatically segment and provide the same knee anatomical informations for all patients studied regardless of the input image.
This multimodal modeling will be carried out via the exploitation of an active model of appearance type algorithm.
- Secondary Outcome Measures
Name Time Method
Trial Locations
- Locations (1)
CHRU de Brest
🇫🇷Brest, France