Technical quality assurance in digital mammography screening in the

Technical quality assurance in digital mammography screening in the Netherlands R.E. van Engen section Physics and quality control LRCB, Nijmegen National Expert and Training Centre for Breast Cancer Screening Outline • Dutch Breast Cancer Screening programme • Technical Quality Control • Digital mammography - Screening trials - Introduction of digital mammography • Contrast threshold visibility measurement • Limiting values National Expert and Training Centre for Breast Cancer Screening Dutch Breast Cancer Screening programme • Breast cancer screening was introduced in 1989 (age group: 50-70) • Fully implemented in 1997 • 1997: age group 50-75 • Nine regions in the Netherlands • Reference centre: Nijmegen National Expert and Training Centre for Breast Cancer Screening Dutch Breast Cancer Screening programme • Reference Centre: National Expert and Training Centre for Breast Cancer Screening (LRCB) (Nijmegen, the Netherlands) - Training radiographers/radiologists - Research projects - Quality Control (medical and technical) National Expert and Training Centre for Breast Cancer Screening Technical Quality Control • Quality Control tests in digital mammography: - Type tests - Acceptance tests - Six monthly test - Homogeneity test - Monitor check National Expert and Training Centre for Breast Cancer Screening Technical Quality Control Type test • The applicability of equipment is checked • Test outside the screening • All equipment in screening must be approved • Dutch type test protocol: - Physico-technical measurements (based on acceptance test protocol) - Clinical test (stability of equipment, evaluation of clinical images) National Expert and Training Centre for Breast Cancer Screening Technical Quality Control Type test • Euref type testing - Physico-technical measurements Based on European Guidelines, 4th ed. Some additions and adaptations - Clinical test Based on clinical tests of EU countries • If a system is approved by Euref: -> allowed in Dutch screening National Expert and Training Centre for Breast Cancer Screening Technical Quality Control Acceptance test • Equipment which is installed on screening units is subject to an acceptance test • Performed by physicists of reference centre • Equipment checked against limiting values • Baseline for subsequent tests • Dutch protocol for acceptance tests (almost identical to European Guidelines) National Expert and Training Centre for Breast Cancer Screening Technical Quality Control Six monthly test • On every screening unit all equipment is subject to physico-technical tests every six months • Dutch protocol for acceptance tests (not all measurements are performed) • Equipment checked against limiting values and baseline National Expert and Training Centre for Breast Cancer Screening Technical Quality Control Homogeneity test • In this test the stability of X-ray system and image receptor is checked • Dutch protocol for acceptance tests Based on European Guidelines • Frequency: Weekly and before and after calibration of the system (if performed by the user) • Important test National Expert and Training Centre for Breast Cancer Screening Homogeneity test (Weekly) In ROI: - Average pixel value - Variation in pixel value - Signal-to-noise ratio - Deviating pixel values Between ROI: - Average pixel value - Signal-to-noise ratio National Expert and Training Centre for Breast Cancer Screening Homogeneity test National Expert and Training Centre for Breast Cancer Screening Homogeneity test CR systems • Two monthly (?) homogeneity images of all screens to detect screen artefacts National Expert and Training Centre for Breast Cancer Screening Technical Quality Control Monitor check • In this quick test monitor calibration and ambient conditions are checked • Dutch protocol (Based on EG) • AAPM Test image (www.euref.org) National Expert and Training Centre for Breast Cancer Screening Introduction of Digital Mammography (Screening) trials • Radboud University Nijmegen Medical Center (1999) • Preventicon static screening unit (2002) • BBNN mobile screening unit (2004) • SBBZWN mobile screening unit (2004) • Sanadome trial (2006) National Expert and Training Centre for Breast Cancer Screening (Screening) trials Radboud University Nijmegen Medical Center • Started in 1999 • The applicability of digital mammography was tested in a clinical environment • General Electric Senographe 2000D • At start: many problems • Image quality sufficient National Expert and Training Centre for Breast Cancer Screening (Screening) trials Utrecht Peventicon • Started in 2002 • The applicability of digital mammography was tested in a screening environment (static unit) • Lorad Selenia • At start: many problems • IQ and workflow sufficient National Expert and Training Centre for Breast Cancer Screening (Screening) trials Mobile units • Started in 2004 • Digital mammography was tested in a mobile screening environment • BBNN: Agfa DM 1000 • SBBZWN: Fuji Profect • Digital mammography equipment can be used in mobile units National Expert and Training Centre for Breast Cancer Screening (Screening) trials Nijmegen Sanadome • Started 2006 • Connectivity/functionality issues and typetesting • Lorad Selenia • Planmed Nuance • IMS Giotto • GE Senographe Essential • Kodak Directview CR 950 • Fuji Profect National Expert and Training Centre for Breast Cancer Screening Introduction of Digital mammography • Outcome of trials was positive • Decided in 2006: -> the Dutch screening will go digital between 2008 and 2010 • European tender in 2007 National Expert and Training Centre for Breast Cancer Screening Introduction of Digital mammography In European tender, demands on (among other): • Quality of systems: - must comply to Dutch type test protocol - must comply to European Guidelines, 4th ed. • Functionality/connectivity: - must comply to Integrating the Healthcare Enterprise (IHE) Mammography Image Profile National Expert and Training Centre for Breast Cancer Screening Why IHE? National Expert and Training Centre for Breast Cancer Screening National Expert and Training Centre for Breast Cancer Screening National Expert and Training Centre for Breast Cancer Screening National Expert and Training Centre for Breast Cancer Screening Introduction of Digital mammography • Six manufacturers chosen for mammography unit, one manufacturer for Image management system • In proof of concept functionality and connectivity is checked (November 2007) • Proof must be shown that systems comply to Dutch type test protocol and European Guidelines, 4th ed. National Expert and Training Centre for Breast Cancer Screening Threshold contrast visibility (CDMAM phantom) National Expert and Training Centre for Breast Cancer Screening Threshold contrast visibility CDMAM phantom version 3.4 • Cylindrical objects on Aluminium plate • Diameter: 0.06 – 2.00 mm • Thickness: 0.03 – 2.00 µm • Wedged between 4.0 cm PMMA (2 cm below, 2 cm above) National Expert and Training Centre for Breast Cancer Screening CDMAM scoring of images Acceptance test: • 16 CDMAM images, the phantom is repositioned after each exposure • The exposure factors as chosen by the AEC for 5 cm PMMA are used • If manual mode is used to mimic fullautomatic mode: Take pre-exposure into account National Expert and Training Centre for Breast Cancer Screening CDMAM scoring of images • Scored by human and computer readout • Human readout is used to check the image quality against the limiting values • Computer readout is baseline for future tests (Subsequent tests only computer readout) National Expert and Training Centre for Breast Cancer Screening CDMAM scoring of images Human scoring • 8 CDMAM images • 4 readers, each reader scores 2 different images • For each square % of detection is determined • For each diameter on the phantom threshold contrast visibility is determined as 62.5% detection point on a psychometric curve National Expert and Training Centre for Breast Cancer Screening Psychometric curve for 0.40 mm diameter objects Correctly identified objects (%) 100 80 60 40 20 0 0.01 p( t ) = 0.75 + 0.25 1 + e − f (C ( t )−CT ) 0.1 1 10 gold thickness (µm ) CDMAM scoring of images Human scoring • Resulting threshold contrast visiblity points for each diameter are fitted with a polynomal function • The fitted curve is checked against limiting values National Expert and Training Centre for Breast Cancer Screening Contrast-detail curve Human re adout 10 Limiting value GE Senographe 2000D; 27 Mo/Rh, 2.1 mGy Gold thickness [um] 1 0.1 0.01 0.1 Diam eter [mm ] 1 CDMAM scoring of images computer readout • 16 images are used • For each diameter on the phantom threshold contrast visibility is determined as 62.5% detection point on a psychometric curve • Resulting threshold contrast visiblity points for each diameter are fitted with a polynomal function to obtain the contrastdetail curve National Expert and Training Centre for Breast Cancer Screening Limiting values European Guidelines: ‘Limiting values have been derived as much as possible from practice using screen-film mammography: it is assumed to be a requirement that digital mammography should perform at least as well as screenfilm mammography.’ National Expert and Training Centre for Breast Cancer Screening Limiting values • The threshold contrast visibility limiting values for digital mammography have been chosen such that 97.5% of the screen-film systems would comply • If the image quality of a digital system is on the image quality limit -> the overwhelming majority of screen-film system does have higher image quality (K. Young et al, Proceedings IWDM 2004, 149-155) National Expert and Training Centre for Breast Cancer Screening Summary • Dutch Breast Cancer Screening programme • Technical Quality Control • Digital mammography - Screening trials - Introduction of digital mammography • CDMAM • Limiting values National Expert and Training Centre for Breast Cancer Screening Thank you National Expert and Training Centre for Breast Cancer Screening