Image Quality and Dose • Image quality ‘Speckle and – Image noise sharpness’ – Spatial resolution – Contrast Image Quality and Dose Issues in MSCT – Artefacts • Radiation Dose S. Edyvean – Organ dose St. George’s Hospital London SW17 0QT – Effective dose ImPACT ImPACT 1 0 UKRC June 2007 UKRC June 2007 Image Quality and Dose Image Quality and Dose Issues in MSCT • Many issues are the same in ss and ms • Image quality – General comments – Image noise – Specific comments to msct – Spatial resolution • tend to relate to z-axis features – Contrast What we find is that they are all – Artefacts in a constant battle with each Whizzo CT Company x-ray tube other – each can only win at filtration focus to the expense of another axis • Radiation Dose – Organ dose focus to – Effective dose detectors detectors Z-axis 2 3 UKRC June 2007 UKRC June 2007 Scanner parameters affecting IQ and Dose Scanner parameters affecting IQ and Dose • Beam shaping filter • Beam shaping filter • mA • mA Whizzo CT Company Whizzo CT Company x-ray tube x-ray tube • Scan time • Scan time filtration filtration focus to focus to • kV • kV axis axis • Convolution kernel • Convolution kernel Dose • Detector size • Detector size focus to focus to detectors • No. of samples detectors • No of samples Noise • Image width • Image width • Beam width • Beam width detectors detectors • Pitch • Pitch 4 5 UKRC June 2007 UKRC June 2007
Scanner parameters affecting IQ and Dose Scanner parameters affecting IQ and Dose • Beam shaping filter • Beam shaping filter noise • mA • mA Whizzo CT Company z-axis resolution x-ray tube • Scan time • Scan time filtration focus to • kV axis • kV dose • Convolution kernel Noise • Convolution kernel Scan plane • Detector size • Detector size resolution focus to detectors • No. of samples • No. of samples • Image width • Image width • Beam width • Beam width detectors • Pitch • Pitch Z-axis 6 7 UKRC June 2007 UKRC June 2007 IQ and Dose in MSCT High contrast spatial resolution • Spatial resolution (z-axis) • How small can we go? • Pitch • Dose issues • Reconstruction algorithm • What image quality do we want? 8 9 Z-axis UKRC June 2007 UKRC June 2007 Spatial Resolution – 3D Z-axis spatial resolution • Scan plane (limited by pixel size) • Imaged slice width • Z-axis (image slice width) – Influences partial volume artefacts – Affects contrast and noise • In MSCT Picture – Flexibility of reconstructing different slice widths Element • In helical generally (SS and MS) (pixel) Slice Width – Optimised by reconstructing overlapping slices Volume Element 512 pixels (voxel) 10 13 UKRC June 2007 UKRC June 2007
Z-axis spatial resolution Z-axis spatial resolution • Thinner slice minimises partial volume artefacts Thin slice Thick slice Wider Narrower 14 15 UKRC June 2007 UKRC June 2007 Z-axis spatial resolution Thinner slice – improved contrast • Image width affects contrast and noise of object 10mm 5mm • Optimised slice width: imaged slice ≈ object size 4 mm 2mm low contrast more noise better contrast but more noise 16 17 UKRC June 2007 UKRC June 2007 Courtesy: Matthew Benbow, RBH Z-axis resolution in single-slice Thinner slice - higher noise • Object ~ 5 mm • Image width depended on beam width – And post patient collimation for thin slices 1mm 5mm 18 19 UKRC June 2007 Courtesy: Matthew Benbow, RBH UKRC June 2007
Z-axis resolution in multi-slice Z-axis resolution in multi-slice • Image width depends on detector acquisition width • Image width depends on detector acquisition width – eg 4 x 5mm, will not give a 2.5 mm slice! (Use 8 x 2.5) – eg 4 x 5mm, will not give a 2.5 mm slice! (Use 8 x 2.5) • May be optimised in helical • May be optimised in helical – with closer z-axis sampling – with closer z-axis sampling (eg z-sharp in Siemens, or certain overlapping pitches) (eg z-sharp in Siemens, or certain overlapping pitches) + + + + + + + + + ++ + + ++ + 20 21 UKRC June 2007 UKRC June 2007 Optimising z-axis spatial resolution Optimising z-axis spatial resolution • Visualisation optimised by overlapping reconstructions • Overlapping reconstructions recommended for (viewed by cine or 3-D) optimum contrast and z-axis resolution object transaxial images MPR • ½ to 2/3 rds overlap recommended 22 23 UKRC June 2007 UKRC June 2007 Effect of pitch Pitch – dose • SSCT vs MSCT • Overlapping pitch – average dose increases – Dose • Extended pitch – average dose lower – Noise – Image slice thickness • Artefacts Extended Overlapping Contiguous 24 25 UKRC June 2007 UKRC June 2007
Pitch – multislice (inc. pitch, mA const.) Pitch - single slice (increase pitch, mA const) • Dose decreases • Dose decreases • Same filter width • Noise constant with pitch – Image width remains the same – Two point interpolation regardless of spacing • Noise increases: • Image width increases – less projection data within filter width + + + + + + + + + + + + + ++ + + + ++ + + + + + + + + + + + + + ++ + + + ++ + + + + + ++ + + + ++ + + + + + + + + 26 27 UKRC June 2007 2-point interpolation (360LI shown) UKRC June 2007 Pitch – multislice (inc. pitch, inc. mA) Pitch – artefacts • Dose stays the same • Teflon (PTFE) rod in water • Same filter width – to simulate rib at an angle to scan plane – Image width remains the same • Spiral Artefacts in MPRs • Noise stays the same: – less projection data within filter width, – but more photons per projection Pitch 0.5 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + x-sectional MPR image + + + + + + + + 28 29 + + + + + + + + UKRC June 2007 UKRC June 2007 Pitch – artefacts Pitch – artefacts • Spiral Artefacts in MPRs of a Tilted Teflon Rod • Spiral Artefacts in MPRs of a Tilted Teflon Rod – image-width 3mm Volume Zoom Volume Zoom – acquired using 4*2.5mm (Siemens Volume Zoom) 4 slice 4 slice Pitch 0.5 Pitch 0.75 Pitch 1.0 image- width 3mm gradual decrease of image quality Pitch 0.75, Pitch 1.75, Pitch 1.25 Pitch 1.5 Pitch 1.75 coll. 4 x 2.5mm coll. 4 x 1mm • For a given image width: small detector acquisition width at higher pitch is better than wide acquisition width at lower pitch 30 courtesy Kalendar 31 images courtesy Kalendar UKRC June 2007 UKRC June 2007
Dose issues in MSCT Dose issues in MSCT - Beam width • Beam width (overbeaming) • Penumbra typically 3 mm for all beam widths – lower proportion of total dose with wider beam widths • Helical overscan (overranging) • Wider is generally better z-axis 32 33 4 slice 8 slice 16 slice UKRC June 2007 UKRC June 2007 Dose issues in MSCT - Overranging Effect of reconstruction filter • Except for short scan lengths and large pitches • Filter used in backprojection (convolution kernel) near sensitive organs – Smooth, standard, detail, bone – Use narrower beam widths, or axial scans – AH30, AH40, AB50 – FC41, FC43 etc, etc • Used to optimise spatial resolution against noise 34 35 Smooth Sharp UKRC June 2007 UKRC June 2007 Effect of reconstruction filter Tube current higher spatial frequency ⇒ more noise Lower mAs Smooth Sharp 200 to 100 mAs ⇒ noise x 1.4 Smooth → → Standard Sharp eg ~ 7 HU → → noise = 17 HU 70 HU 36 37 UKRC June 2007 UKRC June 2007
Low contrast detectability – recon filter Compromise depending on requirements • High spatial detail • Low contrast resolution Same mAs Smooth Bone 50 mAs 1600 mAs Similar noise 38 40 Noise x Noise x UKRC June 2007 UKRC June 2007 Image noise Image noise • What is an appropriate level of image noise ? • What is an appropriate level of image noise ? – too low – high dose 10 mGy 15 mGy 20 mGy – too high – no diagnosis / missed diagnosis • How do we find the optimum level? 25 mGy 30 mGy 35 mGy 41 42 Doses given are CTDI measured at surface of Catphan UKRC June 2007 UKRC June 2007 Image quality required for diagnosis Systematic addition of image noise • Systematic addition of noise to clinical images/raw data – Simulate mA • Studies for a variety of clinical conditions and scanners decreasing photons per projection → Ideal image 1,000,000 100,000 10,000 43 44 Scan Simulator: Courtesy of Toshiba UKRC June 2007 UKRC June 2007
Recommend
More recommend
