STEREOTACTIC BREAST BIOPSY EQUIPMENT SURVEYS JAMES A. TOMLINSON, M.S. Diagnostic Radiological Physicist American Board of Radiology Certified Medical Physics Consultants, Inc.
Bio 28 yrs experience 100% Dx Rad >100 mammo systems/yr 15 are dedicated stereotactic units
SBB-Testing Requirements Currently exempt from MQSA certification and ACR accreditation (voluntary). Some states require testing, all tests that are applicable. This focus is on dedicated units with digital IR.
Testing Procedures The ACR-SBB manual contains testing procedures for technologists and physicists. There are options for filmscreen units and those with digital receptors.
SBB Units Dedicated prone table units are made by Fischer and Lorad. Other re-salers have put their label on the Lorad unit. Add-on stereotactic devices are also used. We will only discuss the prone table units.
FISCHER MAMMOTEST
Multi-Angle Biopsy
Image Receptor 9-10 lp/mm 1024 2 Min-R screen 5x5cm Fiber-optic taper CCD 1 x1 1024x1024
MTF
LORAD STEREOGUIDE
Breast support Detector assembly
Image Receptor ~5 lp/mm 512 2 7-10 lp/mm 1024 2 Lens Mirror CCD 1 x1 1024x1024 Lanex screen
EQUIPMENT TESTING 1. Unit assembly evaluation 2. Collimation assessment 3. Focal spot performance/system resolution 4. kvp performance 5. Beam quality (HVL) 6. AEC or manual exposure performance
EQUIPMENT TESTING 7. Digital image uniformity 8. Breast radiation dosimetry 9. Image quality 10. Artifact evaluation 11. Localization accuracy
1. Unit Assembly Evaluation Free-standing dedicated unit is mechanically stable All moving parts move smoothly All locks and detents work properly Image receptor holder assembly is free from vibrations Image receptor is held securely by assembly Compressed breast thickness scale accurate to 5 mm Patient or operator not exposed to sharp or rough edges or other hazards
2. Collimation Assessment Radiation to digital image alignment Radiation beam confinement to image receptor
Coin Method
Monitor Film
Collimator Tool Method
Monitor Film
Collimation Tolerance The radiation field should extend beyond the image receptor But not more than 5 mm on any edge
Radiation Beam Confinement Often, the rad field will exceed the image receptor at the chest wall breast unnecessary radiation radiation field image field metal compression paddle
3. Focal Spot Performance The focal spot performance is tested with the res pattern and film (no screen)
Focal Spot Performance 11 lp/mm in the direction perpendicular to the anode cathode axis 13 lp/mm in the direction parallel to the anode cathode axis 13 parallel 11 perp
System Resolution (Digital) The resolution test pattern is used with the digital receptor to find limiting resolution of the 50x50 mm image size. For 512 image, pixel size is = 0.098 mm = 5.1 lp/mm For 1024 image, pixel size is 0.049 mm = 10.2 lp/mm Typically see 7-10 lp/mm
Q1- The maximum limiting system resolution for a dedicated SBB device in 1024 mode would be: 20% 20% 20% 20% 20% 1. 1 lp/mm 2. 3 lp/mm 3. 5 lp/mm 4. 7 lp/mm 5. 10 lp/mm 10
5. 10 lp/mm Bushberg, Jerrold T, Ph.D., et al; The Essential Physics of Medical Imaging, Second Edition; Lippincott, Williams & Wilkins, 2002, p. 284
4. KVp Performance
kvp Accuracy and Reproducibility Mammography units should have clinically used kvp accuracy to within 5% of the indicated kvp From four measurements at most commonly used kvp, compute mean, s.d., and coefficient of variation (sd/mean) C.V. should be within 0.02
Q2- Tube potential accuracy should be within: 20% 20% 20% 20% 20% 1. 1 % 2. 1 kvp 3. 2 % 4. 2 kvp 5. 5 % 10
5. 5 % American College of Radiology; Stereotactic Breast Biopsy Quality Control Manual; 1999, p. 65
5. Beam Quality (HVL) The HVL requirements are slightly different from filmscreen mammo, because the image port is open (no paddle attenuation) The HVL should be above HVL kvp/100 (mm Al) The ACR suggests an upper limit of HVL < kvp/100 + 0.12 (mm Al) for Mo/Mo At 28 kvp, 0.40 HVL 0.28
Half-Value Layer High-purity aluminum
Half-Value Layer Logarithmic interpolation can be are used to calculate HVL Using the logarithmic method, where E 0 =0 mm, E a = first added filtration, E b =second total added filtration HVL = t b ln (2E a / E o ) - t a ln (2E b / E o ) ln ( E a / E b )
Q3- The half-value layer (HVL) at 28 kvp must be greater than or equal to: 20% 20% 20% 20% 20% 1. 0.26 mm Al 2. 0.28 mm Al 3. 0.30 mm Al 4. 0.32 mm Al 5. 0.35 mm Al 10
2. 0.28 mm Al American College of Radiology; Stereotactic Breast Biopsy Quality Control Manual; 1999, p. 68
6. AEC Performance (if used) For the performance test, 4, 6, and 8 cm of lucite is evaluated with clinical technique factors
AEC System Performance For digital systems: read the signal value in the center of the image signal from 6 and 8 cm should be within ± 20% of the 4 cm reading if not, a technique chart should be developed and posted
Technique Chart Posted techniques should be evaluated and adjusted to keep image quality high, yet exposure times < 2 seconds
7. Digital Image Uniformity 4 cm lucite image Signal analysis in center and quadrants Signal Standard deviation (sd) SNR Calculate quadrant deviance from the center values Systems without ROI capability, use wire mesh pattern
Lorad Tolerances SNR in each corner ± 15% of center (early manuals show 128x128 box). I ve never found a Lorad unit that will pass with this protocol. Later test protocols use a 32x32 box and specifies the corner areas to sample. The specified tolerance was that the SNR in each corner not exceed ± 20% of the central SNR. Most units will pass with these parameters.
6% of total area 0.4% of total area 128x128 512 image 32x32 512 image
Newer Lorad Units On the latest units, the operator manual indicates to take signal in the center of 1024 image with 32x32 box (0.1% of total area). This also requires significantly more radiation to keep signal at 4k, when quadrupling image matrix. Signal in each quadrant should be within 15% of the center. All of them appear to be able to pass this protocol.
Fischer Tolerances SNR in each corner ± 15% of center (ACR manual shows 128x128 box) Most units will pass this tolerance
Q4- The ACR digital receptor uniformity tolerance for signal value is: 20% 20% 20% 20% 20% 1. 5 % 2. 10 % 3. 15 % 4. 20 % 5. 25 % 10
3. 15% American College of Radiology; Stereotactic Breast Biopsy Quality Control Manual; 1999, p. 75
8. Breast Radiation Dosimetry Find technique factors that give manufacturer recommended signal values for the ACR phantom or the mas given from an AEC technique. Position ion chamber in port at 4.5 cm from breast support. Measure manual exposure 4 times. For AEC techniques, use closest mas. Calculate average, s.d., and c.v. < 0.05
Breast Dose
Breast Dosimetry For AEC technique, calculate ESE from: ESE aec = ESE man (mas aec /mas man ) Calculate the average glandular dose from look-up table in ACR manual (kvp vs HVL) for Mo/Mo. Typical conversion factor is 0.18 at 28 kvp. Dose should be less than 3.0 mgy (300 mrad). (Michigan requires 2.0 mgy for all mammography)
9. Image Quality Make an exposure of the ACR phantom using the kvp and density normally used clinically. For the standard ACR phantom, image in each quadrant to determine IQ score. For the ACR mini-phantom, only one image is needed.
ACR Phantom Screen-film: Digital: 4.0 fibers, 3.0 specks, 3.0 masses 5.0 fibers, 4.0 specks, 3.5 masses
Mini-Phantom
Mini-Phantom Screen-film: Digital: 2.0 fibers, 2.0 specks, 2.0 masses 3.0 fibers, 3.0 specks, 2.5 masses
ACR Phantom Scoring Fibers are scored down to the smallest full or half fiber visible. A full point is given for full fiber, 0.5 for half. Specks are scored down to the smallest visible group. A full point is given if specks 4, 5, & 6 are visible, 0.5 if only 2 & 3 are visible, and 0 if only 1 speck is visible. Masses are scored down to the smallest visible. A full point is given if round and in the right location, a 0.5 point if in the right location but not round.
Score Subtraction for Artifacts If a fiber-like artifact is as apparent as an actual fiber, it is subtracted from the last half or full fiber. If speck-like artifacts are present, they are deducted on a one for one basis from the last group counted. If a mass-like artifact is as apparent as an actual mass, it is subtracted from the last real mass.
Mag Tool
Q5- The digital image quality required for the dedicated mini-phantom: 20% 20% 20% 20% 20% 1. 4.0 fibers, 4.0 specks, 4.0 masses 2. 4.0 fibers, 3.5 specks, 3.5 masses 3. 3.5 fibers, 3.5 specks, 3.5 masses 4. 3.0 fibers, 3.0 specks, 3.0 masses 5. 3.0 fibers, 3.0 specks, 2.5 masses 10
5. 3.0 fibers, 3.0 specks, 2.5 masses American College of Radiology; Stereotactic Breast Biopsy Quality Control Manual; 1999, p. 91
10. Artifact Evaluation For digital IR image lucite sheet large enough to cover IR Adjust the window width and level to look for defects in the image
Very narrow window
11. Localization Accuracy
Mammotome
Scout Image (0 0 )
Stereo Scouts (+15 0, -15 0 )
Oh Baloney!
Thin metal pieces
Stereo Scouts
Pre-Fire
Post-Fire
Technologist s QC Zero Alignment - before each patient Localization Accuracy (XYZ) - each day of use Phantom Imaging - each week of use Visual Checklist - monthly Hardcopy QC - monthly (if applicable) Compression - semiannual Repeat Analysis - semiannual
Coordinates Calibration (XYZ)
Hardcopy QC (monthly) Print SMPTE pattern from console Print SMPTE from printer or Radiograph a step-wedge and print
Q6- The localization accuracy (in air) and phantom image quality should be conducted by the technologist: 20% 20% 20% 20% 20% 1. Daily, daily 2. Daily, weekly 3. Each day of use, weekly 4. Each day of use, each week of use 5. Weekly, weekly 10
4. Each day of use, each week of use American College of Radiology; Stereotactic Breast Biopsy Quality Control Manual; 1999, pp. 17&21
SUMMARY SBB-ACR accreditation is voluntary, for now Even though exempt from MQSA testing regimen, dedicated units should be tested routinely by the Medical Physicist, and may be required by state agencies.