CT AUTO QALite
Supports the following analysis
- Pixel size (geometric distortion)
- CT linearity or constancy
- Slice thickness
- Localizer and laser alignment accuracy
- Spatial resolution (MTF)
- Noise and Mean CT #
- Slice sensitivity profile
- Low Contrast – Contrast-to-noise ratio (CNR) (Catphan 700)
Output available as;
- PDF format, exportable to Excel format
- Data is stored in database folders for trend analysis
- Trend analysis supported
- Phantoms: Catphan CTP700, CTP605, CTP604, CTP600, CTP500, CTP504(Varian), CTP503 (Elekta) and Gammex 464 ACR , GE, Siemens and Philips system QA phantoms
- Supports images folders with images that contain different matrices.
- Includes fixes for trend analysis linearity report, Catphan604 manual MTF, floating point calculation error slice width measurement (Windows7) when processing large FOV image
Standard Processing Features
- Pixel Size – test verifies the expected pixel size (information stored in the DICOM header) from the measured set of four calibration pins positioned at a known distance location as specified in the Catphan® section CTP401 and CTP404. Not supported by Catphan 700.
- CT Linearity (Sensitometry) – Mean CT values are calculated for each reference material in the phantom test section. The measured CT numbers are fit to the linear attenuation coefficients using a linear least squares for energies from 40 to 100 keV. The variance representing the lowest variance was designated as the scanner’s effective energy.
- Slice Thickness – is determined from a processing algorithm that calculates the full-width at half-maximum (FWHM) of the CT number profile for each wire ramp. A trigonometric conversion is calculated based on the known ramp angle to yield the slice width. The individual ramp and average ramps calculated slice widths are reported along with the expected slice width. This test also provides information on the position of the phantom and the vertical and horizontal tilt values. A rotation of the phantom about an axis perpendicular to the ramps is also computed
- Localizer and Table Incrementation Accuracy – is assessed from the slice width section when the ramps are paired at opposing angles. The slice width test outputs a parameter called Table Position Offset, which is an offset in the z-axis relative to the center of the opposing wire (or test section). Verification of the scanner table incrementation accuracy can be checked by incrementing the weighted table by 30 mm and then returning the table to the starting position and scanning the slice width module.
- Spatial Resolution (MTF) – The modulation transfer function (MTF) is calculated from the discrete Fourier transform of the average vertical and horizontal LSF’s of the point spread function from the bead or wire test section. The program reports the 50%, 10% and 2% MTF cutoff values.
- Noise and Mean CT Number – is calculated from several regions of interest (ROI) positioned over a water/uniformity phantom section. The number, size and location of these ROI’s are variable, but typically five are defined: one at the phantom center and the other four along the axes at the same radius.
- Uniformity – Vertical and horizontal profiles 10 pixels wide are generated and averaged through the phantom’s center. The fractional uniformity of the profile is calculated as the percentage of the pixels within an acceptable range determined by + 2 times the central noise or + 10H, whichever is smaller. The new IrisQA Report tool will provide the uniformity as the maximum deviation between the 4 peripheral and central ROI measurements.
- Contrast Detail (Low Contrast Resolution) – Theoretical Contrast-Detail data is calculated based on the measured noise of the low contrast modules (CTP263 and CTP515).
- Low Contrast – CNR measurements for low contrast targets in the CTP515 module
- Slice sensitivity profile data processing from an input source (bead) in a series of axial images
- Manual processing mode for the MTF analysis – with variable target (wire/bead) locations and zoom factor
- Printed or saved reports are now available with data grouped (sorted) by test type as an additional option to the standard display format based on image number
ACR Phantom Processing
- CT Linearity – similar to the method described for the standard processing using 5 materials
- Slice thickness – is determined from the CT number profile of the 2 pairs of angled series of wires visible in 0.5 mm z-axis increments.
- Laser Alignment – is verified using the central wire in the slice width ramps and determining the location relative to the other wires on a slice less than 2 mm in thickness.
- Low Contrast – CNR is measured for the 25mm target, plus the system receives user input for the smallest visible target.
- Noise/Uniformity – The mean CT# and standard deviation (SD) is calculated for one central ROI and 4 peripheral ROI’s at 3, 6, 9, 12 o’clock locations and displayed as the mean/SD. Uniformity is calculated for each peripheral locations based on the difference between the central mean and each peripheral mean values and reported as the ‘Diff (HU)’.
- Pixel Size – Is reported as the measured pixel size based on a know distance and the number of pixels between the 2 BB’s in module three.
- High Contrast Spatial Resolution – is reported by the measured noise levels of each of the 8 bar resolution patterns plus the user inputs the smallest pattern resolved.
- Table incrementation – assessment is based on the detection of the 4 steel BB’s at 3, 6, 9, 12 o’clock location in the slice (less than 2 mm).
Trend Analysis Reports
- Auto populates the hospital names and CT serial numbers
- Provides a thumbnail list of all the unique data sets based on the scan parameters used for easy selection
- User defined baseline values by either selecting the option for averaging the first 5, 10 or manually select data records or entering an arbitrary baseline
- User selects tolerances from a list of operators; + %, + , >, <, =
- Trend graphs provided of key fields
- Exportable reports to Microsoft’s Excel including trend graphs
QALite Release Notes
v188.8.131.52 release (September 2016)
- CNR (#24) Background value used for CNR is the average of three background samples, each one is adjacent to the 15 mm target for each contrast level
- CNR (#24) Reduced the background sample diameter from 14 mm to 10 mm for the Catpha700
- Fixed AutoSelect function for the Catphan 700 phantom
- Fixed problem with the saved PDF report file, when the CNR test results are displayed last with the last line of data is missing
- Catphan 604 now supported, AutoSelect function selects single slice per module
- Catphan 605 now supported
- Angiography CBCT images (XA) now supported (Options/Allow XA uncompressed images )
- Added CNR test (#24) to Catphan 504 phantom configuration
- Corrected the scan mode displayed on the image file listing and in the QALite results page.
- Scan mode field added to the scan parameters string (axial, Nornal-CT or helical) which now allows scan date to be sorted by scan mode for Trend Analysi
- AutoFind first module center function enabled when series exceed 20 images
v3.1.5 release (February 2015)
- Images with non-standard matrices (e.g. 384 x 384) when processed where multiple files are located in the selected image folder now supported.
- ‘Bead’ or ‘Wire’ was added to the scan parameters in the slicethk.dbf and resolut.dbf files when bead was used for the MTF test #4 and Slice Thickness test #16 and wire was used for MTF test# 21 and Slice Thickness Test # 2.
- Updated the AutoSelect function for Catphan 600; the ‘Autoselect’ was not picking the correct slice from CTP528; the solitary bead 10mm offset from the gauge and 80mm offset from the CTP404 module center, not picking the uniformity slice from CTP486 160mm from CTP404 module center, Catphan 500; not picking the solitary bead 10mm offset from the gauge and 40mm from CTP404 module center, also picking an extra slice from uniformity module CTP486 40mm offset from center.
- Device serial number DICOM data element (0018,1000) if not present, instead of adding ‘unknown’ the Station Name tag (0008,1010) is used as the device serial number for the SN locations in reports and the database SN field.
- A modification has been implemented for the Slice Width measurements using the wire ramps under specific conditions – that is noise and expected slice width dependent. Noise (sampled in center CTP404) dependent profile smoothing has now been included in our processing algorithm. The impact of this change is seen in typically slice thickness 1.0 – 1.25 mm where the range of change was +3.8% to +16.8% for 1.0 mm slice and +8.8% to +16.8% for 1.25 mm slice. In all cases the image noise (SD) is > 10HU. The following conditions for the profile smoothing are defined below;
- Noise < 6.0 = no smoothing
- Noise > 6.0 and < 10.0 smoothing is Expected SW dependent
- > 1.0 mm and < 2.0 mm = 1 smoothing cycles
- > 2.0 mm and < 4.0 mm = 2 smoothing cycles
- > 4.0 mm = 4 smoothing cycles
- Noise > 10.0 smoothing is Expected SW dependent
- < 1.0mm = no smoothing
- > 1.0 mm and <2.0 mm = 1 smoothing cycle
- > 2.0 and < 4.0 mm = 4 smoothing cycles
- > 4.0 mm = 8 smoothing cycles
- Expanded phantom/patient listing now includes a breakdown by series available by clicking the check box ‘Sort Images by Series’ found below DICOM Directory listing.
- Support DICOM files names up to 116 characters in length which can also include brackets [ ].
- Expanded the energy range for CT contrast scale/effective energy calculations from 15 (40-100 keV) to 20 energies (40-140keV)
- A new test was added that supports the IEC noise and uniformity measurements. The current noise test (#5) remains implemented for Catphan 410, 411 and 412 models using the default values for a square 15×15 pixel ROI at center and 4 peripheral locations on axis 50mm from center. This is the setup that has always been used for these phantoms. For all other phantoms Test#5 has been turned off and Test#14 is now implemented. The database file associated with the output for Test#14 when doing Trend Analysis will be ACRnoise.dbf. If the user prefers to not use test #14 and turn on Test#5 the phantom configuration files for your phantom can be modified. See our Technical notes page for details on this modification. The new test (#14) Image Uniformity generates 6 circular ROI’s; one at center 40% of phantom diameter which reports the noise, the other 5 are 10% of phantom diameter positioned at the center and approximately 1 cm from phantom edge on axis. These 5 ROI’s report the mean/noise(SD) and calculate the image uniformity, the peripheral ROI and center ROI mean difference. The sizes and locations of the ROI’s used can be modified if necessary. See the Phantom File Description File included with the new 3.1.4 Operators manual.
- Images with a 196.5 mm FOV (< phantom 200mm diameter) were considered targeted images would only process the images from CTP591 or CTP528 for the MTF measurements. This issue has been seen with Toshiba’s images when the DFOV is set to 200 mm. These images will now process all phantom configured tests.
- Exclude Toshiba Aquilion localizer images (image type OriginalPrimaryLocalizier) from the image selection list.
- No longer search for the water sample used in the Catphan 600. The sample size is a problem when the vial is not completely full. If the user wants to add this sample to their phantom configuration please contact QABenchmark
V3.1.01 release 1d (3.19.2013)
- Modified Noise test and Noise/Uniformity test to support IEC requirements which supports different size ROI’s for the center sample and the peripheral samples, with either square or circular ROI’s
- Increased the scale for the CT number axis on the CT contrast scale graph to include air and Teflon
- Updated attenuation coefficients for all Catphan phantoms (except CTP424) due to change in density values recommended by the Phantom Laboratory. Changed attenuation coefficients; Bone50%, Bone20%, Teflon, Polystyrene, Delrin, Lung materials
Added water to the measured CT sensitometry materials in the Catphan 600 and 700 models
V3.1.01 release 1c (2.7.2013)
- Corrected floating point error for Philips Brilliance 64 (version 4.0)
V3.1.01 release 1b (1.17.2013)
- Supports Catphan 700 phantom analysis
- Modified the slice width analysis (used in version prior to v3.1.01) for bead ramps for the CPT591 module in the Catphan 600 phantom by adjusting the threshold used for bead detection
- IrisQA Reporting update includes option to save unique baseline and tolerances values for each CT scanner in the trend analysis database
- Output vertical and horizontal uniformity profiles as text files
- Automated processing for the Catphan 700 low contrast (CNR target measurements) and high contrast resolution modules
- Added CNR – Low Contrast trend analysis to the IrisQA Reporting application for Catphan 700 users
- Manual processing mode for MTF calculation
- Slice sensitivity profile (SSP) manual processing mode
- Tests results sorted by test name in addition to sort by slice order
- Added window level display controls for thumbnail images
- Introduced new trend analysis report function – ‘IrisQA Reporting’
- Trend analysis allows user to define baseline and tolerances values
- Improved ROI centering routine for CT sensitometry samples
- Created data folder ‘AutoQA’ in ProgramData folder to store sample phantom images (Images),PDF report files (Test Reports) database files, text and screen output files created by QAlite
- Replaced the USB software protection device with a software encrypted license
- Supports DICOM image file names >100 characters
- IrisQA Reporting updates includes Boolean operator correction for negative data values, ‘average all’ function fix and removed sort function from column data header