• DocumentCode
    3747066
  • Title

    Three-dimensional segmentation and quantification of the anatomic regurgitant orifice in mitral regurgitation using 3D ultrasound images

  • Author

    Miguel Sotaquira;Mauro Pepi;Gloria Tamborini;Enrico Caiani

  • Author_Institution
    Universidad de San Buenaventura, Bogota, Colombia
  • fYear
    2015
  • Firstpage
    13
  • Lastpage
    16
  • Abstract
    Currently, clinical assessment of mitral regurgitation (MR) makes the use of 2D color Doppler imaging for the estimation of the vena contracta (VC) and of the effective regurgitant orifice area (EROA)..Since the anatomic regurgitant orifice (ARO) has a 3D shape and it´s not circular, it cannot be accurately represented by these 2D parameters. We developed a novel semiautomated method for 3D ARO segmentation and quantification using 3D transesophageal echocardiographic (TEE) datasets, and validated it vs manual planimetry on a set of 25 patients with mild to severe MR. ARO 2D projected area and circularity index (CI) correlated well with planimetry results (r2=0.77 and 0.90 and bias of -0.02 and 0.02, respectively). In 19/25 patients ARO planarity index was less than 0.9, confirming its 3D morphology. Only 1/25 patients exhibited an almost circular ARO (CI > 0.92), while 24/25 patients had more elongated orifices (CI<;0.8). In conclusion, volumetric quantification of 3D ARO is feasible, and the results confirm the potential of this technique for the estimation of real 3D ARO shape. The proposed method could be a useful alternative for the assessment of patients with MR, given the intrinsic shape of the ARO and the limitations present in current approaches.
  • Keywords
    "Image segmentation","Magnetic resonance","Biomedical imaging","Sociology","Statistics","Cardiology","Three-dimensional displays"
  • Publisher
    ieee
  • Conference_Titel
    Computing in Cardiology Conference (CinC), 2015
  • ISSN
    2325-8861
  • Print_ISBN
    978-1-5090-0685-4
  • Electronic_ISBN
    2325-887X
  • Type

    conf

  • DOI
    10.1109/CIC.2015.7408574
  • Filename
    7408574