I85PREP
⌘ K
ISCD CCD study materials
ISCD CCD study resources

The Repeat Scan That Looks Like a Big Loss

ISCD CCD · Put your knowledge to work in context.

All case studies

The Repeat Scan That Looks Like a Big Loss

A 66-year-old man had a DXA 3 years ago: lumbar spine L1–L4 T-score −1.9, total hip −1.7, femoral neck −1.8. He was started on alendronate. Today he returns for follow-up. The new scan shows lumbar spine T-score −2.6, total hip −2.1, femoral neck −2.2. The report software flags 'significant decrease at all sites.' On review of the images, the L1–L4 region of interest on the new scan includes a vertebra that was not included previously; the technologist notes the patient's positioning was slightly different and the spine software version was updated between visits. His weight is unchanged, and he reports good adherence.

1. Is this a real loss of BMD? Walk through what must be checked before concluding anything.

Read the worked answer

Most likely not a real loss - this is a textbook 'apparent' change driven by technical factors. Before interpreting a longitudinal change, verify:

  1. Same scanner, same software version. A software update between scans can shift BMD values and reference data; the comparison may be invalid. If software changed, the difference cannot be attributed to the patient.
  2. Same region of interest. If L1–L4 now includes a vertebra that was excluded before (or vice versa), the mean is not comparable. Re-analyze both scans using an identical, matched ROI.
  3. Positioning and rotation. Slight rotation or altered lordosis changes spine BMD substantially, especially at the spine.
  4. Adherence and confounders. He reports good adherence; confirm with pharmacy refill data if possible, and check for new medications, weight change (none), or new fractures.
  5. Least significant change (LSC). The change must exceed the facility's LSC at that site to be considered real. A 0.7 T-score drop at the spine may or may not exceed LSC depending on the precision of the center - but it is moot if the ROI and software are not matched.

Only after all of these are addressed can a real loss be concluded. The correct action is to re-analyze with matched ROI and software, and if still discrepant, repeat the scan.

2. What is the appropriate response to the software's 'significant decrease at all sites' flag, and what do you recommend?

Read the worked answer

The automated flag is a comparison of numbers, not a clinical judgment. It does not know that the ROI changed or that the software was updated. The interpreter should:

  • Not report 'significant bone loss' or change therapy based on this comparison.
  • Re-analyze the prior and current scans with identical ROI and, if possible, the same software version. If the software cannot be reverted, state in the report that the comparison is limited by a software change and that the difference cannot be reliably interpreted.
  • If, after matching, a real decrease remains and exceeds the facility's LSC, then consider secondary causes (vitamin D, calcium intake, malabsorption, myeloma, glucocorticoids, hypogonadism) and reassess adherence before escalating therapy.
  • Schedule the next DXA per ISCD interval guidance (typically every 1–2 years on treatment, and not sooner than the LSC allows meaningful comparison).

Bottom line: the flag is a prompt to check, not a finding to report.