3.5 DXA Scanner Changes & Cross-Calibration
Key Takeaways
A same-technology replacement or hardware change uses ten repositioned phantom scans before and after. A mean difference greater than 1% calls for manufacturer correction.
One ISCD approach for a different-technology replacement uses 30 representative patients scanned once on the original and twice on the replacement within 60 days.
Adding a scanner has a different protocol. Match same-technology scanners using a spine phantom on 20 days, or use the two-scans-per-device in vivo approach for different technology.
Without valid site-specific cross-calibration, establish a new baseline and an intra-system LSC on the new scanner.
Select the protocol for the equipment change
When a DXA scanner changes, the apparent difference in bone mineral density (BMD) may reflect the equipment. Cross-calibration establishes whether and how measurements made before and after the change can be compared. First determine whether hardware changed on one scanner, a scanner was replaced, or a second scanner will be used while the original stays in service. Then determine whether the devices use the same technology.
| Situation | Measurement design in the ISCD Adult Positions | What the design establishes |
|---|---|---|
| Hardware change on an existing system or replacement with the same manufacturer and model | One technologist performs 10 repositioned phantom scans before and 10 after the change | Compare mean phantom BMD. A difference greater than 1% calls for manufacturer service or correction. |
| Replacement with a different technology, including a different manufacturer | One approach measures 30 representative patients once on the original system and twice on the replacement within 60 days | Establish the inter-system BMD relationship and least significant change (LSC) for each monitored skeletal site. |
| Addition of a second scanner with the same manufacturer and model to support scanning across devices | Scan the same spine phantom on the index scanner and each additional scanner on 20 different days | A mean difference greater than 0.5% calls for manufacturer correction and repeat verification with the same process. |
| Addition of a different-technology scanner while retaining the original | Preferred approach designates an index system and measures at least 30 representative patients twice on each system in a 60-day study window, no more than 30 days apart for any patient | Determine the LSC across devices for each commonly monitored site and monitor scanner stability after cross-calibration. |
Compare measurements within the right calibration
For a different-technology replacement or addition, measure the sites used for monitoring in that practice, such as the lumbar spine and proximal femur. The ISCD DXA Machine Cross Calibration Tool estimates the relationship and LSC between instruments. Quantitative comparison is supported only for the sites actually cross-calibrated. A result for the lumbar spine does not certify comparability at the hip.
Inter-system LSC applies when the two patient scans come from the two systems included in the applicable cross-calibration. Intra-system LSC applies when the scans come from the same system. After replacement, establish local precision and an intra-system LSC for later comparisons confined to the new scanner. If different-technology systems stay in service together, maintain stability with frequent checks using a suitable external phantom. ISCD specifies that the running mean on each system should stay within 0.5% of its value at cross-calibration. Establish a separate same-device LSC from duplicate scans on the added system for patients compared within that system.
When a facility adds an equivalent scanner on the same campus with the same technologists, the original scanner's LSC may be used for inter-scanner comparisons under the ISCD position. If locations, technologists or patient populations differ, assess precision at each site. ISCD allows an average LSC using the relevant technologists' precision assessments under its multi-technologist rules. Compliance with local radiologic safety rules remains part of the acquisition protocol.
Apply the results to patient care
Consider a hardware replacement in an existing scanner. The mean of ten repositioned phantom scans before the change is 1.004 g/cm² and the post-change mean is 1.017 g/cm². The difference is 0.013 g/cm², or about 1.29% of the pre-change mean. This exceeds the 1% service threshold for this protocol. Contact the manufacturer for correction. The result is a phantom calibration check and does not by itself yield a correction to apply to each patient's BMD.
Now consider a patient returning to a scanner that replaced the original with different technology. If the site was cross-calibrated and the studies are technically comparable, use the applicable inter-system procedure and LSC. If no suitable cross-calibration was done, report that quantitative comparison with the old instrument is unavailable. Establish the patient's current BMD as a new baseline and use the new scanner's intra-system LSC for subsequent comparable studies. Matching a make, model, or software name alone does not establish the necessary measurement relationship.
For outside studies, identify the scanner and check whether valid cross-calibration exists. When feasible, the patient should return to the original device if its clinical precision and LSC are known and acceptable. Compare the same anatomy and assess technical quality even when the scanner is the same. A valid inter-system threshold addresses differences between systems and cannot correct a mismatched region of interest.
A hospital replaces its DXA scanner with a system using different technology from another manufacturer. Which approach is described in the ISCD Adult Positions for preserving quantitative comparison?
Use ten phantom scans on each machine and apply a fixed 5% offset to patient results.
Use T-score changes without a device comparison.
Use lateral spine radiographs in place of a device comparison.
Scan 30 representative patients once on the original and twice on the replacement within 60 days, and calculate the site-specific inter-system BMD relationship and LSC.
A clinic will keep its current scanner and add a second one of the same manufacturer and model. Which phantom protocol supports comparison across the two scanners?
Scan each instrument once and consider equal T-scores sufficient.
Scan one spine phantom on both systems on 20 different days, and seek correction if mean BMD differs by more than 0.5%.
Scan the phantom on only the original scanner for 20 days.
Scan the two scanners with different phantoms once each and subtract the difference from all patient measurements.
The old scanner was removed before any cross-calibration could be performed. A patient returns for follow-up on the new scanner. How should the interpreter handle the quantitative comparison?
Compare the absolute BMDs using the older scanner’s intra-system LSC.
Compare the T-scores because the devices measured the same anatomy.
Document that quantitative comparison is unavailable and use the new study as the baseline for subsequent comparable scans.
Estimate an inter-system correction from the patient’s two scan results.
Sections you finish are checked off in the contents.