Medical imaging monitors specification guidelines

Transcription

1 Medical imaging monitors specification guidelines

2 Document details Contact for enquiries and proposed changes If you have any questions regarding this document or if you have a suggestion for improvements, please contact: Contact officer: David Thiele Title: Manager Statewide Medical Physics Business area Biomedical Technology Services, Health Services Support Agency Phone: David_Thiele@health.qld.gov.au Version history Version Date Changed by Description 1.0 July 2010 Lawrie Sim, Ben Keir, David Thiele First issue 1.1 November September 2013 David Thiele David Thiele New template New template 1

3 Contents 1. Purpose Scope Definitions Monitor specification guidelines Non-Mammography Monitors Mammography Monitors Notes Records Associated documents References Monitors for radiological diagnosis minimum specification

4 1. Purpose These guidelines recommend monitor specifications for the viewing of medical images. 2. Scope These guidelines apply to monitors used for the viewing of medical images from all diagnostic modalities used in Queensland Health facilities. 3. Definitions GSDF MP Small matrix images SOE Grayscale Standard Display Function optimised relationship between pixel value and monitor luminance Megapixel: 1 MP = 1,000,000 pixels Images of matrix size 1024x1024 e.g. CT, MRI, ultrasound Standard Operating Environment 3

5 4. Monitor specification guidelines 4.1 Non-mammography monitors Type Tier 1 Primary diagnostic Tier 2 Clinical review Tier 3 Remote diagnostic Purpose Used for the interpretation of non-mammography medical images i.e. radiologist workstation or instances where the primary treatment decision is made in the absence of a interpretative report e.g. Emergency Department, ICU, Orthopaedic Clinic Used for the viewing of nonmammography medical images for non-interpretative purposes where a higher standard of image quality is required than can be provided by a standard (e.g. SOE) PC. Typically used by medical staff and consultants (non-radiologists) when an interpretative report is available. Medical imaging monitors specification guidelines Not ideal but used for the interpretation of small matrix medical images (e.g. CT, MR, ultrasound) when a primary diagnostic monitor is not available Size (diagonal) 21 inch (54 cm) 21 inch (54 cm) 17 inch (43 cm) Matrix Size 3 MP: 1536x MP: 1600x MP: 1280x1024 Max Luminance 500 cd/m 2 (a) 300 cd/m cd/m 2 (c) Contrast Ratio 600:1 500:1 500:1 (d) Luminance Uniformity 15% variation from centre 15% variation from centre 15% variation from centre Calibration Auto-GSDF Auto-GSDF Ability to accept 3 rd party GSDF calibration Bit Depth 10 bits 10 bits Brightness / Contrast Controls Colour/Monochrom e Locked out Locked out Manual Monochrome (b) Either Either Backlight Saving Yes Yes Tier 4 Basic image viewing Used for basic viewing of medical images e.g. patient consultation, chart review. Existing SOE monitors can be used for the basic viewing of medical images. However caution should be exercised as the image quality will be inferior to that of Tier 1 and Tier 2 monitors. a) This minimum standard should be easily achieved as primary diagnostic monitors are generally manufactured with higher luminance (up to 1000 cd/m 2 ). b) Monochrome recommended as backlight life double that of colour. c) Rationale for this minimum standard is discussed in Sim et al (2007). d) To achieve this contrast ratio, the monitor must be able to render very low luminance blacks. 4

6 4.2 Mammography monitors Type Purpose Tier 1M primary diagnostic Used for the interpretation of mammography images by authorised readers Tier 2M Clinical review Size (diagonal) 21 inch (54 cm) 21 inch (54 cm) Matrix Size 5 MP: 2048x MP: 1600x1200 Max Luminance 700 cd/m cd/m 2 Contrast Ratio 800:1 600:1 Luminance Uniformity 15% variation from centre 15% variation from centre Calibration Auto-GSDF Auto-GSDF Bit Depth 10 bits 10 bits Brightness / Contrast Controls Locked out Medical imaging monitors specification guidelines Used for the viewing of mammography images for non-interpretative purposes where a higher standard of image quality is needed than can be provided by a standard (e.g. SOE) PC. Locked out Colour/Monochrome Monochrome* Either Backlight Saving Yes Yes *Monochrome recommended as backlight life double that of colour. 4.3 Notes 1. The manufacturer s dead pixel policy should be reviewed as part of the purchase. 2. Use a graphics card recommended by the monitor manufacturer. 3. For image viewing in operating theatres use a Clinical Review monitor taking into account screen size, viewing distance, ingress, and sterility issues. Highest spatial resolution monitors may not represent best value for money. Contact Biomedical Technology Services for advice. 4. Performance of the monitor will depend on environmental conditions e.g. ambient lighting. 5. To maintain optimum performance, quality control testing is recommended. Contact Biomedical Technology Services for advice 5

7 5. Records Nil 6. Associated documents Nil 7. References AAPM On-Line Report No. 03, Assessment of Display Performance for Medical Imaging Systems, PS3.14 DICOM Part 14: Grayscale Standard Display Function Sim L.H., Manthey K., Stuckey S. (2007), Comparison of performance of computer display monitors for radiological diagnosis; diagnostic high brightness monochrome LCD, 3MP vs clinical review colour LCD, 2MP, Australas. Phys. Eng. Sci. Med. 30(2), Monitors for radiological diagnosis minimum specification. When considering viewing monitors for radiology applications it is necessary to also consider the modality. These can be defined as: 1. Small Matrix (The image matrix is less than or equal to 1K x 1K. e.g. CT, MRI, Ultrasound) 2. Large Matrix (The image pixel array is greater than 1K x 1K e.g. Plain Images (CR, DR). The performance of a monitor in a radiological viewing application can be characterised by a small number of quantitative metrics that are often used to describe Image quality. These metrics generally fall into the categories of spatial resolution and contrast resolution. Spatial resolution The spatial resolution characteristic involves more than simply specifying specific minimum monitor matrix sizes (e.g. 2MP or 3MP) for each modality. The physical size of the monitor is also an important consideration. (e.g. there are notebook screens in excess of 1600 x 1200 that have a small pixel pitch that will display a chest x-ray but at an inappropriate image size). It is more appropriate to define the screen size (e.g. 21 diagonal) as well as the required matrix size. This will determine the pixel pitch which determines the smallest sided object that can be resolved when the image is displayed in a 1:1 spatial map from image pixel data to image display data. Note though that magnification features of the display application software on the workstation will determine achievable displayed spatial resolution. The monitor should be capable of displaying the complete image at close to life size (not necessarily at full spatial resolution) and should be capable of displaying the entirety of a region of interest (e.g. pathology) and its relationship with its surroundings in a 1:1 spatial map from image pixel data to image display data (i.e. maximum achievable spatial resolution). It is not necessarily a requirement to display the complete image in the same 1:1 relationship. 6

8 Contrast resolution Monitors applied to radiographic interpretation and/or clinical review of medical images should be capable of the following: 1. Contrast ratio of greater than or equal to 500:1 2. Maximum luminance of not less than 500 cd.m -2 for interpretation and 300 cd.m -2 for clinical review. The recommendation of ACR 1 for this parameter is171 cd.m -2. In practice this value is readily achieved. 3. Luminance uniformity of less than +/- 15% deviation from the central measured luminance value across the area of the screen. (High quality AMLCD panels should be capable of better. This is an important conformance issue) An ability to have its luminance transfer characteristic conformed to the DICOM Part 14 Grayscale display function (GSDF) Minimum 10 bit greyscale output from Look Up Table (LUT) 6. The monitor(s) should be supplied complete with automatic luminance calibration capability and automatic DICOM GSDF conformance software. The maximum luminance output (white level or L max ), minimum luminance output (black level or Lmin) and luminance uniformity. The ratio L max /L min is the dynamic range of the monitor. The interval L max L min on the GSDF will define the maximum achievable number of Just Noticeable (contrast) Difference indices (JNDs) for the display (ignoring any contrast reduction effects of ambient light reflected from the screen). ACR 1 recommend a value of 50 foot lamberts (171 cd.m -2 ) as the minimum value of L max for a monitor for application in radiological diagnosis. ACR does not provide a recommendation for L min. Inspection of the GSDF demonstrates that a luminance value of 171 cd.m -2 will provide 550 JNDs. Most LCD monitors have a minimum luminance output of at least 0.3 cd.m -2 to 0.4 cd.m -2, corresponding to 40 JND indices above zero luminance. This leaves about 510 JNDs achievable in the displayed image. A majority of current display applications use 8 bit presentation values (or 255 luminance intervals) in any presented images. Thus a monitor with luminance characteristics of L max = 171 cd.m -2 and L min = 0.4 cdm -2 will allow a transfer of presentation values to luminance values at the monitor output with approximately two (2) JND indices per luminance interval. The parameter L min is important as a higher value of L min will have a similar effect to a higher level of ambient light; i.e. both circumstances will reduce the effective dynamic range of the display and reduce achievable display contrast. Higher luminance capable monitors can achieve higher numbers of JND indices per luminance interval and deliver increased image contrast. Currently available monitors can readily achieve the luminance values recommended above. A 10 bit greyscale output (remembering that the presentation values on the input side are generally only 8 bit) can be mapped to wider separations on the output greyscale thereby achieving a smoother approximation to the GSDF curve and improved use of available JND indices. There is a body of opinion that suggests the use of higher performance commercial colour monitors for radiographic interpretation (i.e. diagnosis) may be appropriate. There is some recent scientific evidence in support of this view 6, 7 and there is also ongoing work that suggests caution 8. The specification of colour or monochrome may be quantifiable in terms of noise characteristics but this may be less important than the performance specification of the monitor in terms of luminance characteristics and greyscale rendering ability. Consideration should be given to standardisation of monitors across a practice to enable reporting of all (or most) modalities from any workstation. Another advantage is that all monitors will display the same image similarly so consistent presentation of images is facilitated. 7

9 (NOTE: Mammography may be an exception as it is likely that cost would be an impediment to fitting a practice out completely with mammography capable monitors). 1. ACR, (2001), ACR Standard for Digital Image Data Management, American College of Radiology 2. Kimpe T., Sneyders Y., (2006), Effect of non-uniformity on DICOM GSDF compliance, Int. J Cars 1:p NEMA, (2003), Digital Imaging Communications in Medicine (DICOM) Part 14: Grayscale Standard Display Function, PS Leachtenauer J.C., (2004), Electronic Image Display: Equipment Selection and Operation, SPIE Press, Chapter Sim L.H, Manthey K.L., Keir B., (2006), Soft Copy Radiology How do I know which monitor to use? (Proc.) Computer Assisted Surgery & Radiology, Osaka, Japan. 6. Hirschorn D.S., Dreyer K.J., Smith G., The Evaluation of Consumer Displays for the Primary Interpretation of Radiography. RSNA2006; Abstract SSQ Sim L.H., Manthey K., Stuckey S., (2007); Comparison of Performance of Computer Display Monitors for Radiological Diagnosis; Diagnostic High Brightness Monochrome LCD 3MP vs Clinical Review Colour LCD, 2MP; Australasian Physical and Engineering Sciences in Medicine, In Press Accepted for Publication March Roehrig H., Krupinski E., Fan J., Yoneda T. (2007), Are color LCD displays ready for radiology?, Hot Topic presented at CARS2007, Berlin. 8