Search Results

Search Constraints

Start Over
Filtering by: Creator Cevidanes, Lucia Remove constraint Creator: Cevidanes, Lucia Keyword Machine learning Remove constraint Keyword: Machine learning

Search Results

  • A Comprehensive Patient-Specific Prediction Model for Temporomandibular Joint Osteoarthritis Progression

    work
    Creator:
    Cevidanes, Lucia
    Description:
    Image Pre-Processing To allow reliable detection and comparison of changes between several individuals or within the same individual at different time points, before extracting the quantitative bone texture/morphometry features, all hr-CBCT scans were pre-processed using validated protocols. Extraction of Trabecular Bone Texture-based and Morphometry Imaging Features Using the “crop-volume” tool in 3D Slicer, a rectangular shaped volume of interest (VOI) was cropped from the trabecular bone in the mandibular condyles and the articular fossa. Then, using the average minimum and maximum intensity values of all VOIs, we standardized the grey level intensities of the VOIs to eliminate inaccuracies of textural features calculation and possible dependency on the global characteristics of the images. Lastly, imaging markers were extracted from the standardized VOIs using “BoneTexture” module in 3D-slicer. Measurement of the 3D Articular Joint Space To assess the progression/improvement of osteoarthritic changes in the affected individuals, we measured the 3D superior joint space. We pre-labelled two landmarks in the sagittal view of the oriented CBCT scans: on the most superior point of the condyle and on the opposing surface of the articular fossa. To avoid biasing the landmarks’ placements, pre-labelling was performed simultaneously on T1 and T2 scans, using two independent windows in ITK-SNAP. After the volumetric reconstruction of the identified landmarks, linear measurements were obtained in millimeters using the Q3DC tool in 3D Slicer. Three-dimensional Shape Analyses and Quantification of Remodeling in the Condyles SPHARM-PDM software was used to compute the correspondence across 4002 surface points among all condyles. The output point-based models displayed color-coded maps that enabled visual evaluation of consistent parametrization of all condyles. An average condyle shape for the TMJ OA and control groups was calculated through propagation of original surface point correspondences across all stages of deformations and averaging the condyle surface meshes. For visualization of the 3D qualitative changes of the average models within the same group at different time points or among different groups, semi-transparent overlays were created using 3D Slicer software. The vector differences were presented on the condyle surfaces, scaled according to the magnitude of difference, and pointing towards the direction of bone change. For quantification of remodeling in the condyles, calculation of signed distances across condyles surface meshes reflected the quantitative bone changes in the TMJ OA and control samples. To quantify regional bone changes across the lateral and anterior surfaces of the condyles, we used the Pick ‘n Paint tool in 3D Slicer to propagate regional surface points to the corresponding regions of shapes across all subjects and time points.
    Keyword:
    Degenerative joint disease, Temporomandibular joint osteoarthritis, TMJ OA, Machine learning, and Prognosis
    Citation to related publication:
    Al Turkestani N, Li T, Bianchi J, Gurgel M, Prieto J, Shah H, Benavides E, Soki F, Mishina Y, Fontana M, Rao A, Zhu H, Cevidanes L. A comprehensive patient-specific prediction model for temporomandibular joint osteoarthritis progression. Proc Natl Acad Sci U S A. 2024 Feb 20;121(8):e2306132121. doi: 10.1073/pnas.2306132121. Epub 2024 Feb 12. PMID: 38346188; PMCID: PMC10895339.
    Discipline:
    Health Sciences