N.rteas.: 85 Period : 195 dags

Phantom Statistics Result

Hean Phantom BMC Mean Phantom iiHT Mean Phantom AREA HfIC Variation BHD Uariation AREfl Variation

3.526 grans

0.390 g/CB1

9.04 cm1

N.rteas.: 85 Period : 195 dags

Phantom Statistics Result

Hean Phantom BMC Mean Phantom iiHT Mean Phantom AREA HfIC Variation BHD Uariation AREfl Variation

3.526 grans

0.390 g/CB1

9.04 cm1

Fl-Help F3-E4it F4-St*t F6-0ptions

F8-Print F9-Pi> i nt Later

Fig. 4-5. The DTX-200 quality control screen. This is a Shewhart chart with control limits of ±1.5%.

a single phantom BMD value is more than three SDs from the average. This is a violation of the Shewhart three SD rule. "WARNING 2" is a violation of either the Shewhart two SD twice or range of four SD rule and "WARNING 3" is a violation of the Shewhart four ±1 SD rule.

The lower graph reflects the accuracy of the system (16). The solid horizontal line represents the phantom BMD value that was entered into the computer during the setup of the system. The dashed horizontal lines indicate a range of ±1.5% about this value. The values plotted on this graph are the average BMD values for the last 16 phantom scans. If the average value for the 16 most recent phantom scans falls within ±1.5% of the true phantom value, "OK" will be seen next to the word "ACCURACY" at the bottom of the graph. An "OUT OF RANGE" message will appear if the value falls outside those limits. If eight consecutive values fall on the same side of the true phantom value, a "TREND WARNING" message will appear.

The quality control graphs and calculations for the Norland pDEXA® are very similar to those of the XR-Series. The control limits for the accuracy of the pDEXA® system are ±2.5% instead of 1.5% (17).

Hologic scanners also provide automated quality control graphing procedures (18). The BMD of a phantom is established during the initial calibration procedures for the scanner. The control limits of ±1.5% of the phantom BMD value are defined on a graph onto which subsequent spine phantom BMD data is plotted. Underneath the graph, two tables are displayed. The table titled "Reference Values" lists the average or mean value and SD for the spine phantom established during machine calibration. The table titled "Plot Statistics" lists the number of phantom scans plotted (n), the mean, SD, and %CV for those scans. There are no sensitizing rules built into the quality control program in the computer. With this automated plot, however, Shewhart rules are easy to apply.

Other manufacturers have automated charting of phantom values. Figure 4-5 is such a chart from the Osteometer DTX-200 DexaCare®, a dedicated DXA forearm scanner.

Table 4-6

Recommendations from the Canadian Panel of the International Society for Clinical Densitometry for Documentation of a Quality Control Program

1. Current operating manual from equipment manufacturer

2. Appropriate positioning devices

3. Appropriate calibration standard

4. Calibration history for specific densitometer

5. Precision data and estimates of site-specific precision errors

6. Maintenance and upgrade records

7. Software version/upgrade records

8. Cross-calibration records in the event of an equipment change

9. Data and database archiving procedures

10. Local, provincial, and federal licensures of equipment as required

11. Medical physicist inspection reports as required

Reproduced from ref. 3 with permission of the publisher.

The dashed horizontal lines on the graph represent control limits of ±1.5%. None of the 85 phantom values has fallen outside the control limits and the values appear to be randomly scattered about the average value. If such charts are not available, they are easily created using the information in this chapter.

All densitometry centers should implement quality control procedures that minimally consist of control tables or charts with defined control limits of ±1.5% for the average of 10 phantom scans performed on 1 day or 25 scans performed on consecutive days. Shewhart rules with a filter can then be implemented, using rules defined on the basis of percent or SD, to further strengthen the quality control program. The application of CUSUM charts and calculations as performed at professional quality control centers is more labor intensive and not necessarily of greater benefit to the clinical densitometry center. The recommendations (3) from the Canadian Panel of the International Society for Densitometry for a complete quality control program include not only the creation of a control chart with limits of 1.5% but the maintenance of logs and manuals for each densitometer that include the items listed in Table 4-6. These recommendations are certainly appropriate for densitometry facilities in the United States as well as Canada.

Was this article helpful?

0 0

Post a comment