imgflow - tpcclib 0.6.6 © 2016 by Turku PET Centre
Estimation of rate constants K1, k2 and Va from dynamic PET image in
ECAT 6.3, ECAT 7.x, NIfTI-1, or Analyze 7.5 file format using
linearized two-compartment model [1].
Lawson-Hanson non-negative least squares (NNLS) method [2] is used to solve
general linear least squares functions.
When applied to dynamic [O-15]H2O studies, the resulting K1 image
equals perfusion (blood flow) image. K1 image can be divided by tissue
density (g/mL) (option -density) and multiplied by 100 (option -dL)
to achieve the blood flow image in units (mL blood)/((100 g tissue) * min).
When applied to dynamic [O-15]O2 brain studies, the resulting K1 image
can be converted to oxygen consumption image by multiplying it by
arterial oxygen concentration [3] (ml O2 / dL blood) to get the
parametric image in units mL O2 / ((100 ml tissue) * min).
The model assumptions hold only when oxygen consumption is 1-6.7
mL O2/(100g * min) and fit time is set to 300 s or less [3].
Arterial blood TAC must be corrected for decay and delay, with sample times
in seconds. Dynamic PET image must be corrected for decay. Fit time must
be given in seconds.
Usage: imgflow [Options] btacfile imgfile fittime flowfile
Options:
-k2=<filename>
Parametric k2 image is saved; in some situations perfusion calculation
from k2 can be more accurate than the default assumption of f=K1.
Perfusion can be calculated from k2 using equation f=k2*pH2O, where
pH2O is the physiological partiton coefficient of water in tissue.
-Va=<filename>
Parametric Va image is saved.
Set -Va=0, if Va=0 is assumed; otherwise Va is always fitted.
-Vd=<filename>
Parametric K1/k2 image (apparent pH2O) is saved.
-thr=<threshold%>
Pixels with AUC less than (threshold/100 x max AUC) are set to zero.
Default is 5%.
-max=<Max value>
Upper limit for flow values in final units.
-filter
Remove parametric pixel values that are over 4x higher than
their closest neighbours.
-noneg
Pixels where K1 estimates are negative are fitted again with
constraint Va=0.
-mL or -dL
Units in flow and Va images will be given per mL or per dL,
respectively. By default, units are per mL.
-density[=<value>]
With option -density the flow is calculated per gram or 100g tissue.
Tissue density can be changed from the default 1.04 g/ml.
-h, --help
Display usage information on standard output and exit.
-v, --version
Display version and compile information on standard output and exit.
-d[n], --debug[=n], --verbose[=n]
Set the level (n) of debugging messages and listings.
-q, --quiet
Suppress displaying normal results on standard output.
-s, --silent
Suppress displaying anything except errors.
By default, the units of pixel values in the flow (K1) image is
(mL blood)/((mL tissue) * min), in DV image (mL blood)/(mL tissue),
in k2 image 1/min, and in Va image (mL blood/mL tissue),
but the flow and Va units can be changed with above listed options.
Example 1. Calculation of perfusion and arterial blood volume image,
stopping fit at 180 s:
imgflow -Va=s2345va.v s2345abfit.kbq s2345dy1.v 180 s2345flow.v
Example 2. Dynamic image is precorrected for vascular activity,
and all available data is used:
imgflow -Va=0 s2345abfit.kbq s2345dy1_vacorr.v 9999 s2345flow.v
References:
1. Blomqvist G. On the construction of functional maps in positron
emission tomography. J Cereb Blood Flow Metab. 1984;4:629-632.
2. Lawson CL & Hanson RJ. Solving least squares problems.
Prentice-Hall, 1974.
3. Ohta S, Meyer E, Thompson CJ, Gjedde A. Oxygen consumption of the
living human brain measured after a single inhalation of positron
emitting oxygen. J Cereb Blood Flow Metab. 1992;12:179-192.
See also: imgbfh2o, imgcbv, fit_h2o, b2t_h2o, imglhdv, fitdelay, eabaort
Keywords: image, modelling, perfusion, blood flow, radiowater, NNLS