For quantitative MRI assessing image quality we developed quality assurance (QA) phantoms to control the following parameters and characteristics: accuracy of the linear velocity and volumetric flow of liquids measured by phase-contrast MRA (PC MRA),
the spatial resolution and signal to noise ratio for various pulse sequences used in MRA.
The rotating disk phantom for MRA (Fig.
1) is filled by agarose gel and rotated by belt transmission.
It provides a range of velocities from two speed ranges: +/- 20 cm/sec and +/- 72 cm/sec (provided the change of pulleys) - at a radius of disc equal to 55 mm that gives an opportunity to assess the accuracy of measurement of linear and volumetric flow rates of liquids.
During the scan,
the phantom is centered on OY axis and the central axis of the shaft in order to avoid the consideration of the non-parallelism of all the linear velocities in the voxel (Fig.
2).
Each pixel in the resulting MR-image corresponds to the average value in the OX axis velocity projections, and with slice thickness of 1 mm the ratio of the velocity projections on the OX is less than a 1%.
If we consider the flow compensation gradients,
the resulting value of the linear velocity will contribute to the variables that depend on the velocity and acceleration of moving spins.
At the disk rotation frequency v = 2,3 Hz,
distance between the center of the disk and voxel r = 50 mm,
on-time of flow compensation gradients = 3.5 ms the velocity = 0.72 m/s,
acceleration a = 10,37 m/s2.
When the slice thickness equals 1 mm the projection on OX axis is less than 0.2%.
Such solution provides the entire set of linear velocities in a single scan and can be used for the construction of the calibration curve for PC MRA (Fig.
3).
The calibration curve for linear velocity is built as follows: marked area of interest evenly spaced perpendicular to the axis of rotation of the disk phantom.
The linear flow rate obtained for each pixel with the corresponding spatial coordinates is compared with the given diagrams of linear velocities.
After that,
statistical indicators are calculated to assess the accuracy of measurement of the linear velocity and the correction factors are determined.
Described above phantom also allows to evaluate image quality characteristics of 2D TOF and 3D TOF MRA sequences.
The modern methods of 3D TOF angiography use different technologies suppression signal from fat tissue,
and other stationary tissues (brain,
muscle).
For example,
some manufacturers use WET (Water excitation technique) technology,
applied a binomial pulses,
multi chunk is used to increase the relative signal from moving spins (blood flow).
There are many other technologies which allow to enhance inflow effect and reduce the signal attenuation in the distal vessels of the distal arterial bed.
To control the image quality of these modes we propose to measure the following parameters: coefficient of in-flow effect,
coefficient of saturation,
coefficient of fat sat effectiveness,
spatial resolution,
contrast to noise ratio (CNR) (Tabl.
1).
Parameter
|
Coefficient of in-flow effect
|
Coefficient of saturation
|
Coefficient of fat sat effectiveness
|
Contrast to noise ratio
|
Spatial resolution,%
|
Formula
|
|
|
|
|
|
Table 1.
Quantification of MR angiographic sequences (2D TOF and 3D TOF) according to a phantom study.
– mean intensity for three tubes with inflowing MR-contrast fluid (slice № i);
– mean intensity for three tubes with outflowing MR-contrast fluid (slice № i);
– mean intensity for two vials with stationary water (slice № i);
- mean intensity for two vials with fat (slice № i);
- mean intensity of the background (slice № i);
N –number of slices;
i – account number of slice.
For these purposes,
the rotating disc is placed to the isocenter of the magnet,
is completed with three parallel silicone tubes of equal diameter (from 3 to 10 mm) filled with MR-contrast fluid and connected to the ring using an adapter.
To avoid saturation of the spins the additional pulley structure involved in the movement of the tubes and situated outside of the RF coil.
For assessing the spatial resolution,
the ratio of the difference between the mean value of maximum intensities (Imax) for three tubes and the mean value of minimum intensities (Imin) of the spaces between them to Imax is measured (Fig.
4).