MBF using [15O]H2O and dynamic PET


PET enables noninvasive quantification of myocardial blood flow (MBF). PET radiopharmaceuticals that are used for estimation of blood flow (perfusion) include diffusible radiotracer [15O]H2O and partially extracted radiotracers 82Rb and [13N]ammonia, and in case of myocardium also [11C]acetate.

Measurement of myocardial perfusion with [15O]H2O is based on the general model for radiowater. The differential equation for tissue radioactivity concentration curve (CT),

can be used in heart studies as such, with arterial blood concentration curve (CA) as input function. Perfusion (f) and tissue-to-blood partition coefficient (p) are the model parameters. CT cannot be directly measured by PET, because the radioactivity concentration curve from region-of-interest (ROI) drawn on myocardial muscle is affected by spillover and partial volume effects. These can be accounted for by a model for the left ventricular (LV) muscle ROI curve,

, where α is the water-perfusable tissue fraction (PTF), and VA is the vascular volume fraction, including the spillover from the blood in LV cavity (Iida et al., 1988, 1991, and 1992). The resolution of modern PET may allow using the blood curve derived from a tiny ROI placed in the LV cavity directly as the arterial input function. If the curve from LV cavity ROI is also affected by spillover and partial volume effects, these can be accounted for by a model for the LV cavity curve,

, where β is the recovery coefficient of LV ROI, which is dependent on the image resolution, but can be measured with [15O]CO PET. The model is first fitted to the data from a large ROI covering most of the left ventricular myocardium (possibly excluding septum and areas with marked spillover from the liver), enabling calculation of CA, which is then used as input function for smaller myocardial ROIs (Iida et al., 1988, 1991, and 1992).

The methods that are specifically developed for myocardium differ in the correction methods for spillover and partial volume effects, and determine the tissue perfusion from the washout rate (k2, f/p) of the radiotracer (Iida et al., 1988, 1991, and 1992). Myocardial perfusion estimation based on k2 could even be applied to PET image that is not attenuation corrected (Lubberink et al., 2010; Tuffier et al., 2016).

An alternative but less used method is to use factor analysis to correct for spillover effects (Hermansen et al., 1998b; Ahn et al., 2001; Lee et al., 2005).

ROI drawn on RV cavity can be used to correct septum for the additional spillover from RV cavity (Hermansen et al., 1998a; Harms et al., 2011). In that case, Eq 2 could be replaced by

The radiowater model for LV muscle in Carimas accounts for spillover from LV cavity by default, and optionally both LV and RV cavity.

Perfusion in RV muscle can be estimated using the corrected input function from LV analysis and RV cavity for the spillover and partial volume effect correction (Kudomi et al., 2019). Carimas does not yet implement this model for calculation of perfusion in RV muscle.

[15O]H2O model provides not only the perfusion, but also the water perfusable tissue fraction (PTF). PTF can be used as an alternative to FDG PET or resting MBF to assess myocardial viability (Iida et al., 1991 and 2012; Grönman et al., 2019). In nonviable myocardial region, resting state PTF×MBF is markedly reduced, but in hibernating region it is only slightly lower than in normal myocardium (Nowak et al., 2003; Schaefer et al., 2003). PTF is, however, more sensitive to scatter, movement, and misalignment between PET and attenuation map than MBF (Koshino et al., 2012; Hirano et al., 2012).

[15O]CO study is often combined to myocardial radiowater PET to get a blood volume image. Anatomic tissue fraction (ATF) represents the mass (density) of extravascular tissue in a region of interest, and it can be calculated by subtraction of blood volume image from transmission image (Iida et al., 1991). Perfusable tissue index (PTI) can be calculated as the ratio of PTF and ATF (Iida et al., 1991). PTI represents the fraction of extravascular tissue being perfused by water, and is a marker of viable myocardium (Knaapen et al., 2003; Grönman et al., 2021). PTI is reduced also in advanced state of dilated cardiomyopathy because of fibrosis (Knaapen et al., 2004). Parametric PTI image can be calculated from a single [15O]H2O study (Harms et al., 2011).

Coronary flow reserve

2013 ESC guidelines state that detection of ischemia is important in the management of patients with suspected coronary artery disease (CAD). Radiowater PET has higher diagnostic accuracy for this purpose than SPET or coronary CT angiography (Danad et al., 2017).

The short halflife of 15O enables repeating the MBF study without problems of background radioactivity from the previous study. Coronary flow reserve (CFR), or myocardial flow reserve (MFR), can therefore be easily measured in a rest-stress study setting:

PET CFR imaging can act a noninvasive gatekeeper for fractional flow reserve (FFR) measurement using invasive coronary angiography; PET imaging cannot allocate mendable culprit lesions like FFR and it does not distinguish focal epicardial from diffuse and small-vessel disease, and therefore the additional FFR measurements are required to drive revascularization strategy (Driessen et al., 2018). Coronary revascularization improves CFR, correlating with improvement in angiographic CAD curden (Aikawa et al., 2019).

Stress perfusion

Absolute stress perfusion (MBFstress, sMBF) alone has been shown to be superior to perfusion reserve in detection of coronary artery disease (Joutsiniemi et al, 2014; Cho et al., 2018). In symptomatic patients with suspected obstructive CAD, reduced global sMBF identifies those at the highest risk of adverse cardiac events, whereas reduced segmental sMBF with preserved global sMBF is associated with intermediate event risk (Harjulahti et al., 2021). Among patients evaluated for chronic chest pain, combined coronary CTA and sMBF predict outcomes equally in women and men (Kujala et al., 2023) and in non-diabetic and diabetic patients (Mäenpää et al., 2023). Hyperaemic MBF alone also has better diagnostic value than longitudinal flow gradient, the abnormal decrease in hyperaemic MBF from the base to the apex of the left ventricle (Bom et al., 2018). Global and regional hyperaemic MBF is prognostic factor for death and myocardial infarction (Bom et al., 2019). Classification of ischaemia is feasible from sMBF polar maps with convolutional neural network (Teuho et al., 2022).

[15O]H2O stress perfusion measurements have low low sensitivity to image reconstruction method, scanner type, and PET/CT misalignment, and therefore diagnostic sMBF cutoff values can be used consistently used (Nordström et al., 2023).

LV volumes and ejection fraction

Gated [15O]H2O PET can provide good estimates of left ventricular end-systolic volume (ESV), end-diastolic volume (EDV), stroke volume (SV), and ejection fraction (EF) (Nordström et al., 2017).

Motion corrections methods could also greatly improve the accuracy of MBF measurements (Armstrong & Memmott, 2018).

Recommended analysis method


Our currently recommended analysis method (MET5817) is to use Iida's MBF model (Iida et al., 1988, 1991, and 1992), as implemented in Carimas™ (Nesterov et al., 2009; Harms et al., 2014; Tarkia et al., 2015). Measurements of MFR are highly reproducible within and between two observers with different experience. Full hybrid model combining MFR PET and CTA further improves discrimination between significant and non-significant CAD at vessel level (Thomassen et al., 2018).

CLI program

Alternatively, Iida's method can be applied outside Carimas using command-line interface and program fitmbf:

If you have previously used older versions of fitmbf (2.0 or less), please note that in the current version:

  1. Only one study can be analyzed with one command. If α needs to be constrained to a basal study, it can be done with option -f
  2. Existing result file with the same name is overwritten
  3. Result files can be processed further, e.g. mean and s.d. from different subjects or differences between studies can be calculated before importing results to spreadsheet programs
  4. fitmbf can optionally provide confidence limits and/or s.d. for the MBF, α and Va

Partition coefficient of water, p, is needed in the calculation, and by default it is set to value p=0.9464 mL/mL, which is based on the p value 0.91 mL/g used by (Iida et al (1988), multiplied by myocardial tissue density 1.04 g/mL. Bergmann et al (1989) used value p=0.92 mL/g, equalling p=0.96 mL/mL, which was also used by Lammertsma et al (1992).

Older fitmbf version (2.0)

To use this, enter the command with version number: fitmbf_2_0_0. The results between this version and versions 2.1.* or later are similar inside three digits. The differences are in the usage:

  1. Two or three studies may be calculated at one run, constraining α to the α estimate from the first (basal) study.
  2. Result file format is different. If more than one study is analyzed at one run, all results are written in the same file.
  3. If result file exists previously, it is not overwritten, but new results are appended to the end.

Even older software version(s)

Even older versions exist, but only on Solaris/SUN. Usage of those is not recommended.

Blood curves from RV and LV cavities.

Blood time-activity curve (BTAC) from LV cavity represents the arterial input function for all organs, including the myocardial muscle itself, but excluding the lungs, where BTAC from the RV cavity must be used as the input function. BTAC from RV cavity is also be needed for spillover correction of TACs derived from ROIs placed on the RV muscle and on septum.

After a successful [15O]H2O bolus infusion the BTAC from RV cavity is well-formed and can be easily fitted with functions, including the exponential function that incorporates the injection schedule (Wong & Feng, 2005; Wong et al., 2006),

, where

, and the initial delay (radiotracer appearance time, Tap) and the duration of radiotracer infusion (Tin) define the three parts of the function. Program fit_sinf with option -n=2 could be used for the fitting. Alternatively, the BTACs from RV and LV cavities could be fitted simultaneously, assuming that the LV cavity BTAC can be derived from RV cavity BTAC, assuming that the passage of radiowater through the lungs and left atrium can be modelled with just two additional parameters, accounting for the delay and dispersion. This fitting to the RV and LV cavity BTACs can be performed using program fit_wrlv. The delay and dispersion between RV and LV cavity BTACs could be fitted using program fit_disp.

Example of RV and LV cavity BTACs fitted concomitantly
Figure 1. Example of RV and LV cavity BTACs from a [15O]H2O PET study, fitted concomitantly using program fit_wrlv.
Fitted functions can be used to produce input functions for simulations.

See also:


Bergmann SR, Fox KA, Rand AL, McElvany KD, Welch MJ, Markham J, Sobel BE. Quantification of regional myocardial blood flow in vivo with H215O. Circulation 1984; 70(4): 724-733. doi: 10.1161/01.CIR.70.4.724.

Boellaard R, Knaapen P, Rijbroek A, Luurtsema GJ, Lammertsma AA. Evaluation of basis function and linear least squares methods for generating parametric blood flow images using 15O-water and positron emission tomography. Mol Imaging Biol. 2005; 7(4): 273-285. doi: 10.1007/s11307-005-0007-2.

Danad I, Uusitalo V, Kero T, Saraste A, Raijmakers PG, Lammertsma AA, Heymans MW, Kajander SA, Pietilä M, James S, Sörensen J, Knaapen P, Knuuti J. Quantitative assessment of myocardial perfusion in the detection of significant coronary artery disease: cutoff values and diagnostic accuracy of quantitative [15O]H2O PET Imaging. J Am Coll Cardiol. 2014; 64(14): 1464-1475. doi: 10.1016/j.jacc.2014.05.069.

Harms HJ, Nesterov SV, Han C, Danad I, Leonora R, Raijmakers PG, Lammertsma AA, Knuuti J, Knaapen P. Comparison of clinical non-commercial tools for automated quantification of myocardial blood flow using oxygen-15-labelled water PET/CT. Eur Heart J Cardiovasc Imaging 2014; 15: 431-41. doi: 10.1093/ehjci/jet177.

Herrero P, Markham J, Bergmann SR. Quantitation of myocardial blood flow with H215O and positron emission tomography: assessment and error analysis of a mathematical approach. J Comput Assist Tomogr. 1989; 13(5): 862-873. PMID: 2789240.

Iida H, Kanno I, Takahashi A, Miura S, Murakami M, Takahashi K, Ono Y, Shishido F, Inugami A, Tomura N, Higano S, Fujita H, Sasaki H, Nakamichi H, Mizusawa S, Kondo Y, Uemura K. Measurement of absolute myocardial blood flow with H215O and dynamic positron emission tomography. Strategy for quantification in relation to the partial-volume effect. Circulation 1988; 78: 104-115. doi: 10.1161/01.CIR.78.1.104.

Iida H, Rhodes CG, de Silva R, Yamamoto Y, Araujo LI, Maseri A, Jones T. Myocardial tissue fraction - correction for partial volume effects and measure of tissue viability. J Nucl Med 1991; 32:2169-2175. PMID: 1941156.

Iida H, Rhodes CG, de Silva R, Araujo LI, Bloomfield P, Lammertsma AA, Jones T. Use of the left ventricular time-activity curve as a noninvasive input function in dynamic oxygen-15-water positron emission tomography. J Nucl Med 1992; 33:1669-1677. PMID: 1517842.

Joutsiniemi E, Saraste A, Pietilä M, Mäki M, Kajander S, Ukkonen H, Airaksinen J, Knuuti J. Absolute flow or myocardial flow reserve for the detection of significant coronary artery disease? Eur Heart J Cardiovasc Imaging 2014; 15: 659-665. doi: 10.1093/ehjci/jet274.

Kajander S, Joutsiniemi E, Saraste M, Pietilä M, Ukkonen H, Saraste A, Sipilä HT, Teräs M, Mäki M, Airaksinen J, Hartiala J, Knuuti J. Cardiac positron emission tomography/computed tomography imaging accurately detects anatomically and functionally significant coronary artery disease. Circulation 2010; 122(6): 603-613. doi: 10.1161/CIRCULATIONAHA.109.915009.

Kajander SA, Joutsiniemi E, Saraste M, Pietilä M, Ukkonen H, Saraste A, Sipilä HT, Teräs M, Mäki M, Airaksinen J, Hartiala J, Knuuti J. Clinical value of absolute quantification of myocardial perfusion with 15O-water in coronary artery disease. Circ Cardiovasc Imaging 2011; 4(6): 678-684. doi: 10.1161/CIRCIMAGING.110.960732.

Nesterov SV, Han C, Mäki M, Kajander S, Naum AG, Helenius H, Lisinen I, Ukkonen H, Pietilä M, Joutsiniemi E, Knuuti J. Myocardial perfusion quantitation with 15O-labelled water PET: high reproducibility of the new cardiac analysis software (Carimas™). Eur J Nucl Med Biol Mol Imaging 2009; 36(10): 1594-1602. doi: 10.1007/s00259-009-1373-9.

Oikonen V. Model equations for myocardial perfusion studies with [15O]H2O PET. TPCMOD0005. doi: 10.13140/2.1.1738.2402.

Watabe H, Jino H, Kawachi N, Teramoto N, Hayashi T, Ohta Y, Iida H. Parametric imaging of myocardial blood flow with 15O-water and PET using the basis function method. J Nucl Med. 2005; 46: 1219-1224. PMID: 16000292.

Tags: , , , , ,

Updated at: 2023-09-05
Created at: 2007-05-09
Written by: Vesa Oikonen, Chunlei Han