SNRMeasurement

3D Slicer CLI to calculate signal-to-nose ratio of MR images using "difference image" method.

Contributors:

• Babak Matinfar babak@bwh.harvard.edu
• Junichi Tokuda tokuda@bwh.harvard.edu

Introduction to SNR measurement

There are several methods to determine the SNR of MR images suggested in the NEMA standard and the literature. The most commonly used method among the studies on MRI-compatible robotics is a single image method, where SNR is measured based on two regions of interest (ROIs) on a single image assuming the statistically and spatially uniform distribution of noise [Kaufman-1989, Henkelman-1985, NEMA-2008]. Usually, one ROI is defined in the tissue of interest, and the other is in the image background i.e. air. Although the method takes a practical approach requiring only a single image, it is know that the method does not yield accurate SNR measurement in a phased-array surface coil system and/or sum-of-square reconstruction due to the spatial and statistical variation of noise. Alternatively, one can measure SNR without assuming the statistically and spatially uniform distribution of noise with the following methods. The first method is to measure pixel-by-pixel SNR through the repeated acquisition [Reeder-2005, Kellman-2005, Constantinides-2004, Dietrich-2007]. Because the method relies less on the assumption of statistical and spatial noise distribution than any other methods, it is often used as the gold standard SNR measurement. A drawback of this method, however, is that it requires a large number of image acquisitions, typically a few hundreds, making it difficult to be used with a slow imaging sequence. The second method is the "noise only image" method [NEMA-2008, Constantinides-1997, Kellman-2005]. The method calculates SNR based on a single or multiple "noise only images", which are acquired without radiofrequency (RF) excitation. The "noise only image" method was further generalized by Kellman and McVeigh, where images are reconstructed in "SNR units" based on noise only images allowing direct pixel-by-pixel SNR measurement [Kellman-2005]. A disadvantage of the "noise only image" method is that most of clinical MRI system does not allow acquiring noise only images with standard pulse sequence. The third method is the "difference image" method, where noise is estimated by calculating pixel-by-pixel difference between two images acquired under exactly the same conditions [NEMA-2008, Murphy-1993, Firbank-1999]. Despite several drawbacks e.g. requirement for stationary tissue and limited precision due to small sample size, the method allows measuring local SNR with standard clinical sequence. The SNRMeasurement module calculates the SNR of MR images using the "difference image" method.

Module description

The SNR module is created to compute the SNR for MR sequences based on the difference image method. In this method, two MR images are acquired under identical conditions. These images are then added and subtracted from each other pixel by pixel. In the resulting images, a region of interest is defined and the mean and standard deviation of the ROI is computed. SNR is computed as:

Tutorial

1. Load the two MR images which SNR is measured for into Slicer.
2. Define region of interest (ROI) using Segment Editor. The ROI is defined by two segments, with the diameter of 80 mm, on two axial slices, separated 50 mm from each other. It does not matter which of the two MR images is used to define the ROI.

• Create a new segment, select Paint effect, click the % button to switch to setting the brush size in mm, set the diameter to 80 mm, select a slice on the axial window which is located at the edge of the volume and click in the slice view to paint. The slice at the edge of volume is based on the fact that SmartTemplate is placed on the edge of the phantom.
• Move the axial slice 50 mm from the selected slice (take a note of axial spacing). Create a new segment, and click in the slice view to paint a circular area on this slice.

3. Go to Data module, right-click on the Segmentation and select Export visible segments to binary labelmap
4. Open the SNRMeasurement module
5. Select the Input Volume 1 as one of the MR image, and Input Volume 2 as the other MR image
6. Select the Label Map as the exported labelmap
7. Click Apply and it will display the computed SNR in the Measured SNR field

If the MR images are the same size, there is no need to define the ROI for every image. Save the ROI as a .nrrd file and load it the next time SNRMeasurement module is needed. However if the sizes of the images are different, which is most of the time, the ROI should be defined again.

References

• [Kaufman-1989] Kaufman L, Kramer DM, Crooks LE, Ortendahl DA. Measuring signal-to-noise ratios in MR imaging. Radiology 1989;173:265-267.
• [Henkelman-1985] Henkelman RM. Measurement of signal intensities in the presence of noise in MR images. Med Phys 1985;12:232-233.
• [NEMA-2008] National Electrical Manufacturers Association (NEMA). Determination of signal-to-noise ratio (SNR) in diagnostic magnetic resonance imaging. NEMA Standards Publication MS 1-2008. National Electrical Manufacturers Association; 2008.
• [Reeder-2005] Reeder SB, Wintersperger BJ, Dietrich O, et al. Practical ap- proaches to the evaluation of signal-to-noise ratio performance with parallel imaging: application with cardiac imaging and a 32- channel cardiac coil. Magn Reson Med 2005;54:748-754.
• [Kellman-2005] Kellman P, McVeigh ER. Image reconstruction in SNR units: a general method for SNR measurement. Magn Reson Med 2005;54: 1439-1447.
• [Constantinides-2004] Constantinides CD, Atalar E, McVeigh ER. Signal-to-noise mea- surements in magnitude images from NMR phased arrays. Magn Reson Med 1997;38:852-857 (Erratum in: Magn Reson Med 2004; 52:219).
• [Dietrich-2007] Dietrich O, Raya JG, Reeder SB, Reiser MF, Schoenberg SO. Measurement of signal-to-noise ratios in MR images: influence of multichannel coils, parallel imaging, and reconstruction filters. J Magn Reson Imaging. 2007 Aug;26(2):375-85.
• [Murphy-1993] Murphy BW, Carson PL, Ellis JH, Zhang YT, Hyde RJ, Chenevert TL. Signal-to-noise measures for magnetic resonance imagers. Magn Reson Imaging 1993;11:425-428.
• [Firbank-1999] Firbank MJ, Coulthard A, Harrison RM, Williams ED. A comparison of two methods for measuring the signal to noise ratio on MR images. Phys Med Biol 1999;44:N261-N264.
• [Constantinides-1997] Constantinides CD, Atalar E, McVeigh ER. Signal-to-noise measure- ments in magnitude images from NMR phased arrays. [Published erra- tum in: Magn Reson Med. 2004 Jul;52:219]. Magn Reson Med 1997;38: 852-857.