Voxel-wise quantification of myocardial blood flow with cardiovascular magnetic resonance: effect of variations in methodology and validation with positron emission tomography

Table 3 Effect of method of deconvolution on myocardial blood flow quantification

A. Mean values
	Fermi	Tikhonov	TSVD	p value
Overall MBF (mL/min/g)	1.10 ± 0.64	1.15 ± 0.57	1.09 ± 0.56	0.023
Stress MBF (mL/min/g)	1.55 ± 0.63	1.55 ± 0.53	1.48 ± 0.54	0.105
Rest MBF (mL/min/g)	0.66 ± 0.20	0.76 ± 0.24	0.70 ± 0.20	<0.001
B. Mean differences and limits of agreement
	Mean difference	p-value	95% limits of agreement
Overall MBF (mL/min/g)
Fermi – Tikhonov	−0.05	0.160	−0.58 to 0.48
Fermi – TSVD	0.01	1.000	−0.41 to 0.44
Tikhonov – TSVD	0.06	0.025	−0.37 to 0.51
Stress MBF (mL/min/g)
Fermi – Tikhonov	0.00	1.00	−0.62 to 0.61
Fermi – TSVD	0.07	0.345	−0.44 to 0.59
Tikhonov – TSVD	0.08	0.148	−0.42 to 0.57
Rest MBF (mL/min/g)
Fermi – Tikhonov	−0.10	<0.001	−0.52 to 0.32
Fermi – TSVD	−0.04	0.021	−0.32 to 0.24
Tikhonov – TSVD	0.06	0.119	−0.32 to 0.45

95% limits of agreement represent mean difference ± 2 standard deviations. Other abbreviations as per Tables 1 and 2. p values in Part A refer to the significance of the mean difference between the three deconvolution methods; p values in part B are those obtained on post hoc analysis.

ISSN: 1532-429X