The impact of image resolution on computation of fractional flow reserve: coronary computed tomography angiography versus 3-dimensional quantitative coronary angiography

Lili Liu, Wenjie Yang, Yasuomi Nagahara, Yingguang Li, Saeb R. Lamooki, Takashi Muramatsu, Pieter Kitslaar, Masayoshi Sarai, Yukio Ozaki, Peter Barlis, Fuhua Yan, Johan H.C. Reiber, Shengxian Tu

Research output: Contribution to journalArticle

10 Citations (Scopus)

Abstract

Calculation of fractional flow reserve (FFR) based on computational fluid dynamics (CFD) requires reconstruction of patient-specific coronary geometry and estimation of hyperemic flow rate. Coronary computed tomography angiography (CCTA) and invasive coronary angiography (ICA) are two dominating imaging modalities used for the geometrical reconstruction. Our aim was to investigate the impact of image resolution as inherently associated with these two imaging modalities on geometrical reconstruction and subsequent FFR calculation. Patients with mild or intermediate coronary stenoses who underwent both CCTA and ICA were included. CCTA images were acquired either by 320-row area detector CT or by 128-slice dual-source CT. Two geometrical models were reconstructed separately from CCTA and ICA, from which FFRCTA and FFRQCA were subsequently calculated using CFD simulations, applying the same hyperemic flow rate derived from the ICA images at the inlet boundaries. A total of 57 vessels in 41 patients were analyzed. Average diameter stenosis was 43.4 ± 10.8 % by 3D QCA. Reasonably good correlation between FFRCTA and FFRQCA was observed (r = 0.71, p < 0.001). The difference between FFRCTA and FFRQCA was correlated with the deviation between minimal lumen areas by CCTA and by ICA (ρ = 0.34, p = 0.01), but not with plaque volume (ρ = −0.09, p = 0.51) or calcified plaque volume (ρ = 0.01, p = 0.95). Applying the cutoff value of ≤0.8 to both FFRCTA and FFRQCA, the agreement between FFRCTA and FFRQCA in discriminating functional significant stenoses was moderate (kappa 0.47, p < 0.001). Disagreement was found in 10 (17.5 %) vessels. Acceptable correlation between FFRCTA and FFRQCA was observed, while their agreement in distinguishing functional significant stenosis was moderate. Our results suggest that image resolution has a significant impact on FFR computation.

Original languageEnglish
Pages (from-to)513-523
Number of pages11
JournalInternational Journal of Cardiovascular Imaging
Volume32
Issue number3
DOIs
Publication statusPublished - 01-03-2016

Fingerprint

Coronary Angiography
Pathologic Constriction
Hydrodynamics
Coronary Stenosis
Computed Tomography Angiography

All Science Journal Classification (ASJC) codes

  • Radiology Nuclear Medicine and imaging
  • Cardiology and Cardiovascular Medicine

Cite this

Liu, Lili ; Yang, Wenjie ; Nagahara, Yasuomi ; Li, Yingguang ; Lamooki, Saeb R. ; Muramatsu, Takashi ; Kitslaar, Pieter ; Sarai, Masayoshi ; Ozaki, Yukio ; Barlis, Peter ; Yan, Fuhua ; Reiber, Johan H.C. ; Tu, Shengxian. / The impact of image resolution on computation of fractional flow reserve : coronary computed tomography angiography versus 3-dimensional quantitative coronary angiography. In: International Journal of Cardiovascular Imaging. 2016 ; Vol. 32, No. 3. pp. 513-523.
@article{e1573d1889374fa08da2337621188c3a,
title = "The impact of image resolution on computation of fractional flow reserve: coronary computed tomography angiography versus 3-dimensional quantitative coronary angiography",
abstract = "Calculation of fractional flow reserve (FFR) based on computational fluid dynamics (CFD) requires reconstruction of patient-specific coronary geometry and estimation of hyperemic flow rate. Coronary computed tomography angiography (CCTA) and invasive coronary angiography (ICA) are two dominating imaging modalities used for the geometrical reconstruction. Our aim was to investigate the impact of image resolution as inherently associated with these two imaging modalities on geometrical reconstruction and subsequent FFR calculation. Patients with mild or intermediate coronary stenoses who underwent both CCTA and ICA were included. CCTA images were acquired either by 320-row area detector CT or by 128-slice dual-source CT. Two geometrical models were reconstructed separately from CCTA and ICA, from which FFRCTA and FFRQCA were subsequently calculated using CFD simulations, applying the same hyperemic flow rate derived from the ICA images at the inlet boundaries. A total of 57 vessels in 41 patients were analyzed. Average diameter stenosis was 43.4 ± 10.8 {\%} by 3D QCA. Reasonably good correlation between FFRCTA and FFRQCA was observed (r = 0.71, p < 0.001). The difference between FFRCTA and FFRQCA was correlated with the deviation between minimal lumen areas by CCTA and by ICA (ρ = 0.34, p = 0.01), but not with plaque volume (ρ = −0.09, p = 0.51) or calcified plaque volume (ρ = 0.01, p = 0.95). Applying the cutoff value of ≤0.8 to both FFRCTA and FFRQCA, the agreement between FFRCTA and FFRQCA in discriminating functional significant stenoses was moderate (kappa 0.47, p < 0.001). Disagreement was found in 10 (17.5 {\%}) vessels. Acceptable correlation between FFRCTA and FFRQCA was observed, while their agreement in distinguishing functional significant stenosis was moderate. Our results suggest that image resolution has a significant impact on FFR computation.",
author = "Lili Liu and Wenjie Yang and Yasuomi Nagahara and Yingguang Li and Lamooki, {Saeb R.} and Takashi Muramatsu and Pieter Kitslaar and Masayoshi Sarai and Yukio Ozaki and Peter Barlis and Fuhua Yan and Reiber, {Johan H.C.} and Shengxian Tu",
year = "2016",
month = "3",
day = "1",
doi = "10.1007/s10554-015-0797-5",
language = "English",
volume = "32",
pages = "513--523",
journal = "International Journal of Cardiovascular Imaging",
issn = "1569-5794",
publisher = "Springer Netherlands",
number = "3",

}

The impact of image resolution on computation of fractional flow reserve : coronary computed tomography angiography versus 3-dimensional quantitative coronary angiography. / Liu, Lili; Yang, Wenjie; Nagahara, Yasuomi; Li, Yingguang; Lamooki, Saeb R.; Muramatsu, Takashi; Kitslaar, Pieter; Sarai, Masayoshi; Ozaki, Yukio; Barlis, Peter; Yan, Fuhua; Reiber, Johan H.C.; Tu, Shengxian.

In: International Journal of Cardiovascular Imaging, Vol. 32, No. 3, 01.03.2016, p. 513-523.

Research output: Contribution to journalArticle

TY - JOUR

T1 - The impact of image resolution on computation of fractional flow reserve

T2 - coronary computed tomography angiography versus 3-dimensional quantitative coronary angiography

AU - Liu, Lili

AU - Yang, Wenjie

AU - Nagahara, Yasuomi

AU - Li, Yingguang

AU - Lamooki, Saeb R.

AU - Muramatsu, Takashi

AU - Kitslaar, Pieter

AU - Sarai, Masayoshi

AU - Ozaki, Yukio

AU - Barlis, Peter

AU - Yan, Fuhua

AU - Reiber, Johan H.C.

AU - Tu, Shengxian

PY - 2016/3/1

Y1 - 2016/3/1

N2 - Calculation of fractional flow reserve (FFR) based on computational fluid dynamics (CFD) requires reconstruction of patient-specific coronary geometry and estimation of hyperemic flow rate. Coronary computed tomography angiography (CCTA) and invasive coronary angiography (ICA) are two dominating imaging modalities used for the geometrical reconstruction. Our aim was to investigate the impact of image resolution as inherently associated with these two imaging modalities on geometrical reconstruction and subsequent FFR calculation. Patients with mild or intermediate coronary stenoses who underwent both CCTA and ICA were included. CCTA images were acquired either by 320-row area detector CT or by 128-slice dual-source CT. Two geometrical models were reconstructed separately from CCTA and ICA, from which FFRCTA and FFRQCA were subsequently calculated using CFD simulations, applying the same hyperemic flow rate derived from the ICA images at the inlet boundaries. A total of 57 vessels in 41 patients were analyzed. Average diameter stenosis was 43.4 ± 10.8 % by 3D QCA. Reasonably good correlation between FFRCTA and FFRQCA was observed (r = 0.71, p < 0.001). The difference between FFRCTA and FFRQCA was correlated with the deviation between minimal lumen areas by CCTA and by ICA (ρ = 0.34, p = 0.01), but not with plaque volume (ρ = −0.09, p = 0.51) or calcified plaque volume (ρ = 0.01, p = 0.95). Applying the cutoff value of ≤0.8 to both FFRCTA and FFRQCA, the agreement between FFRCTA and FFRQCA in discriminating functional significant stenoses was moderate (kappa 0.47, p < 0.001). Disagreement was found in 10 (17.5 %) vessels. Acceptable correlation between FFRCTA and FFRQCA was observed, while their agreement in distinguishing functional significant stenosis was moderate. Our results suggest that image resolution has a significant impact on FFR computation.

AB - Calculation of fractional flow reserve (FFR) based on computational fluid dynamics (CFD) requires reconstruction of patient-specific coronary geometry and estimation of hyperemic flow rate. Coronary computed tomography angiography (CCTA) and invasive coronary angiography (ICA) are two dominating imaging modalities used for the geometrical reconstruction. Our aim was to investigate the impact of image resolution as inherently associated with these two imaging modalities on geometrical reconstruction and subsequent FFR calculation. Patients with mild or intermediate coronary stenoses who underwent both CCTA and ICA were included. CCTA images were acquired either by 320-row area detector CT or by 128-slice dual-source CT. Two geometrical models were reconstructed separately from CCTA and ICA, from which FFRCTA and FFRQCA were subsequently calculated using CFD simulations, applying the same hyperemic flow rate derived from the ICA images at the inlet boundaries. A total of 57 vessels in 41 patients were analyzed. Average diameter stenosis was 43.4 ± 10.8 % by 3D QCA. Reasonably good correlation between FFRCTA and FFRQCA was observed (r = 0.71, p < 0.001). The difference between FFRCTA and FFRQCA was correlated with the deviation between minimal lumen areas by CCTA and by ICA (ρ = 0.34, p = 0.01), but not with plaque volume (ρ = −0.09, p = 0.51) or calcified plaque volume (ρ = 0.01, p = 0.95). Applying the cutoff value of ≤0.8 to both FFRCTA and FFRQCA, the agreement between FFRCTA and FFRQCA in discriminating functional significant stenoses was moderate (kappa 0.47, p < 0.001). Disagreement was found in 10 (17.5 %) vessels. Acceptable correlation between FFRCTA and FFRQCA was observed, while their agreement in distinguishing functional significant stenosis was moderate. Our results suggest that image resolution has a significant impact on FFR computation.

UR - http://www.scopus.com/inward/record.url?scp=84957954093&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84957954093&partnerID=8YFLogxK

U2 - 10.1007/s10554-015-0797-5

DO - 10.1007/s10554-015-0797-5

M3 - Article

C2 - 26507326

AN - SCOPUS:84957954093

VL - 32

SP - 513

EP - 523

JO - International Journal of Cardiovascular Imaging

JF - International Journal of Cardiovascular Imaging

SN - 1569-5794

IS - 3

ER -