Z-axis geometric efficiency in CT Characterises the extent of the - PDF document

Comparison of Definitions of Geometric Efficiency in Computed Tomography Scanners Sue Edyvean, Nicholas Keat ImPACT* (Imaging Performance Assessment of CT Scanners) London UK *An MHRA Evaluation centre for the UK Department of Health (Medicines and Healthcare products Regulatory Agency) www.impactscan.org WC 2003 August 2003 Z-axis geometric efficiency in CT • Characterises the extent of the radiation beam that is used for image creation (along the z-axis) z-axis WC 2003 August 2003

Geometric efficiency (g.eff.) • Single slice – Usually little wasted dose, as all of beam is used for imaging, except for narrow slices where sometimes post-patient collimation is used WC 2003 August 2003 Geometric efficiency (g.eff.) • Multi slice – Usually more unused dose, as an even irradiation of slices is required – Penumbral region of the beam therefore can’t be used for imaging WC 2003 August 2003

Definitions for z-axis geometric efficiency (g.eff.) • IEC 60601- 2- 44 Ed. 1 1999: Original definition – “the full width at half maximum of the sensitivity profile expressed as percentage of the full width at half maximum of the dose profile.” • IEC 60601- 2- 44 Ed. 2 Am.1 2003: New definition – “the integral of the dose profile along the z-direction, integrated over the range subtended by the detector elements used during acquisition, expressed as a percentage of the total integral of the dose profile in the z-direction” • g.eff. to be displayed on scan console when less than 70% WC 2003 August 2003 Definitions – paraphrased • Old definition: – Ratio of the measured imaged width to the measured dose profile width (add imaged widths for multi-slice) • New definition: – Integral of dose profile over range of nominal slice width divided by integral of whole dose profile integral of dose profile over dose range of nominal profile slice z-axis width integral of whole dose profile add -> total imaged width measured imaged width total nominal acquisition width nominal slice width (nominal total collimation) WC 2003 August 2003

Purpose of this study • Identify problems associated with old g.eff. • Compare two calculation approaches for new g.eff. • Compare values between old and new WC 2003 August 2003 Methods of measurement – dose profiles • Kodak X-Omat V Radiotherapy film • Scanned at iso-centre • Read out with scanning micro-densitometer • Optical density profile converted to dose profile dose profile 8 7 6 5 dose 4 3 fwhm 2 1 0 0 5 10 15 20 25 30 35 40 distance along z-axis WC 2003 August 2003

Methods of measurement – imaged slice width • Angled metal plates scanned in axial mode • Width projected into image → Z-sensitivity profile WC 2003 August 2003 Problems with old g.eff. • Different measurement conditions (air vs phantom) • For multi slice scanners – Many image width combinations for one beam width • Many measurements needed – e.g. Philips Mx8000 four slice scanner Beam width Image width combinations Beam width (mm) No. x slice (mm) (mm) 1 1 2 x 0.5 1 x 1 4 4 4 x 1 2 x 2 1 x 4 10 4 x 2.5 2 x 5 1 x 10 10 16 2 x 8 16 20 4 x 5 2 x 10 20 WC 2003 August 2003

Problems with old g.eff. (continued) • Different focal spots – Potentially doubles the number of values • Sixteen slice scanners – narrow slices – Need accurate measurement for imaged widths ~ < 1 mm – Not many existing test objects will do that ImPACT’s ‘thin slice’ tool 50 µm Titanium plates 8° angle to scan plane WC 2003 August 2003 Potential problems for calculation of new g.eff. • Dose profile – ..‘integrated over the range subtended by detector elements used’ ( � nominal total acquisition width) – Position of dose profile relative to detectors is not defined – Shape asymmetric � assume a reference position dose profile 8 7 6 5 dose 4 3 2 1 0 0 5 10 15 20 25 30 35 40 distance along z-axis WC 2003 August 2003

Two approaches for new g.eff. calculation • Could either – Centre integration range on centre of dose profile • Define as halfway between fwhm – Position of range which gives maximum geometric efficiency A. B. dose profile dose profile 8 8 7 7 6 6 5 5 dose dose 4 4 3 3 2 2 1 1 0 0 0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40 distance along z-axis distance along z-axis WC 2003 August 2003 Two approaches for new g.eff. calculation • Comparison of two techniques – Maximum difference = 0.04 % (ss), 0.6 % (ms) – Mean difference = 0.1% (ms) A. B. dose profile dose profile 8 8 7 7 6 6 5 5 dose dose 4 4 3 3 2 2 1 1 0 0 0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40 distance along z-axis distance along z-axis 1: Centred on dose profile 2. Maximum g.eff. WC 2003 August 2003

Results • Analysed data from 13 scanners (all manufacturers) – 4 single slice – 5 four slice – 1 eight slice – 3 sixteen slice • Total of 123 different slice width combinations – Old g.eff. • Compared new and old g.eff. WC 2003 August 2003 Single slice - old g.eff. (imaged width / dose width) – Most scanners have values of approximately 100% – Values higher than 100% probably due to scatter – Low for narrow slices using post patient collimation 120 Geometric Efficiency (%) 100 80 70% 60 Scanner A 40 Scanner B Scanner C 20 Scanner D 0 0 1 2 3 4 5 6 7 8 9 10 Slice width (mm) WC 2003 August 2003

Single slice - new g.eff. (dose in nominal width/whole profile) – g.eff. values lower than 100% – Radiation falling outside the nominal width – Non rectangular dose profiles as widths gets thinner 100 Geometric Efficiency (%) 80 70% 60 Scanner A 40 Scanner B 20 Scanner C Scanner D 0 0 1 2 3 4 5 6 7 8 9 10 Slice width (mm) WC 2003 August 2003 Multi slice - old g.eff. (imaged width / dose width) • Unused penumbra – values less than 100% • Many values for each scanner – e.g. Siemens Sensation 16 (small focus only) Slice width Imaged width 1 slice 2 slice 3 slice 4 slice 6 slice 12 slice (mm) (mm) 0.6 71% 1 106% 0.75 77% 1.5 77% 88% 3 88% 76% 4.5 77% 87% 5 95% 6 87% 9 77% 88% 10 96% WC 2003 August 2003

Multi slice - old g.eff. (continued) • Same beam width – Different values Slice width Imaged width 1 slice 2 slice 3 slice 4 slice 6 slice 12 slice (mm) (mm) 9 mm 0.6 71% beam 1 106% 0.75 77% 1.5 77% 88% 3 88% 76% 18 mm 4.5 77% 87% beam 5 95% 6 87% 9 77% 88% 10 96% WC 2003 August 2003 Multi slice - new g.eff. (dose in nominal width/whole profile) • Values for each beam width only • Unused penumbra – values less than 100% 100 dose within nominal Geometric Efficiency (%) width equals total 80 dose 70% 60 40 Scanner A dose within nominal Scanner B width less than total 20 Scanner C dose Scanner D 0 0 2 4 6 8 101214161820222426283032 Collimation (mm) WC 2003 August 2003 Four slice scanners only

Multi slice - new g.eff. versus old g.eff. • Mean ratio = 0.99, but standard deviation = 10% • Biggest differences at narrow collimations – total measured imaged width matches dose width, but both greater than nominal 120 • Some values 70% old Geometric Efficiency (new) (%) now below 100 70% 2 x 1 mm 80 70% new 2 x 0.63 mm 60 2 x 0.63 mm 40 20 0 0 20 40 60 80 100 120 Geometric efficiency (old) (%) WC 2003 August 2003 Conclusions • Old g.eff. definition – Worked well enough for single slice systems – Drawback is many slice width combinations for multi slice • New g.eff. does not use imaged width; only the dose profile and nominal width – Two calculation techniques are comparable – Results for single slice are all generally lower • spread of dose outside of nominal width not always identified before – Multi-slice new g.eff. values similar to old values • except for some thin slices where both imaged slice width and dose width are > nominal � Some values now below 70% threshold WC 2003 August 2003

Comparison of Definitions of Geometric Efficiency in Computed Tomography Scanners Sue Edyvean, Nicholas Keat ImPACT* (Imaging Performance assessment of CT Scanners) London UK *An MHRA Evaluation centre for the UK Department of Health (Medicines and Healthcare products Regulatory Agency) www.impactscan.org WC 2003 August 2003