The International Journal of Cardiovascular Imaging

, Volume 30, Issue 7, pp 1373–1382

A novel method for non-invasive plaque morphology analysis by coronary computed tomography angiography

  • Shinichiro Fujimoto
  • Takeshi Kondo
  • Takahide Kodama
  • Yasuko Fujisawa
  • John Groarke
  • Kanako K. Kumamaru
  • Kazuhisa Takamura
  • Eriko Matsunaga
  • Katsumi Miyauchi
  • Hiroyuki Daida
  • Frank J. Rybicki
Original Paper

DOI: 10.1007/s10554-014-0461-5

Cite this article as:
Fujimoto, S., Kondo, T., Kodama, T. et al. Int J Cardiovasc Imaging (2014) 30: 1373. doi:10.1007/s10554-014-0461-5

Abstract

Coronary computed tomography angiography (CCTA) plaque morphology based on conventional Hounsfield units relies on absolute CT numbers is influenced by imaging and anatomical variables. The project describes and tests a novel alternative method, termed the “labeling method”, which uses relative CT numbers and 3-dimensional plaque structure. Using virtual histology intravascular ultrasound (VH-IVUS) as the reference standard, this study compares the labeling method to a conventional CT-number based method to determine coronary plaque morphology. Thirty-seven high-risk, non-calcified atherosclerotic coronary lesions were prospectively evaluated in 33 consecutive patients who underwent CCTA followed by VH-IVUS (mean interval 8.6 ± 13.3 days). CCTA-derived vessel and minimum lumen areas were compared to VH-IVUS measures. Fibrotic and necrotic core areas were calculated by both the labeling method to the CT-number based method; both were tested for agreement with reference standard VH-IVUS. Inter- and intra-observer correlations were assessed. CCTA significantly underestimated minimum lumen area when compared to VH-IVUS (mean difference −1.4 ± 0.9 mm2, p < 0.0001). Necrotic core and fibrous areas quantified using the labeling method demonstrated superior correlation with VH-IVUS compared to those quantified using the CT-number based method, Pearson’s r = 0.75 versus 0.42 and r = 0.80 and 0.59, respectively. Compared to VH-IVUS, limits of agreement for the labeling method-derived necrotic core (−2.0 to 2.5 mm2) and fibrous areas (0.6–8.0 mm2) were more narrow than those determined using the CT-number based method (−3.7 to 7.3 and −4.0 to 8.9 mm2, respectively). Inter- and intraobserver correlations were excellent for all CCTA derived measures (r = 0.85–0.98). A novel CCTA-based labeling method offers an alternative to conventional CT-number based analyses for plaque morphology. The labeling method demonstrates superior correlation to VH-IVUS for measures of fibrotic and necrotic core areas within non-calcified coronary atherosclerotic plaques.

Keywords

Vulnerable plaque Coronary artery disease Coronary computed tomography angiography Plaque analysis software Virtual histology intravascular ultrasound 

Supplementary material

10554_2014_461_MOESM1_ESM.doc (4.9 mb)
Supplementary material 1 (DOC 4967 kb)

Copyright information

© Springer Science+Business Media Dordrecht 2014

Authors and Affiliations

  • Shinichiro Fujimoto
    • 1
    • 2
  • Takeshi Kondo
    • 2
  • Takahide Kodama
    • 3
  • Yasuko Fujisawa
    • 4
  • John Groarke
    • 5
  • Kanako K. Kumamaru
    • 5
  • Kazuhisa Takamura
    • 2
  • Eriko Matsunaga
    • 1
  • Katsumi Miyauchi
    • 1
  • Hiroyuki Daida
    • 1
  • Frank J. Rybicki
    • 5
  1. 1.Department of CardiologyJuntendo University Graduate School of MedicineTokyoJapan
  2. 2.Department of CardiologyTakase ClinicTakasakiJapan
  3. 3.Cardiovascular CenterToranomon HospitalTokyoJapan
  4. 4.Application Research Group, Clinical Application Research Development CenterToshiba Medical Systems CorporationOtawaraJapan
  5. 5.Applied Imaging Science Laboratory, Department of Radiology, Brigham and Women’s HospitalHarvard Medical SchoolBostonUSA

Personalised recommendations