Monitoring bone mineral density (BMD) over time on treatment is valuable to determine if the treatment is having the desired effect. Similarly, monitoring BMD over time during a planned interruption in bisphosphonate therapy is valuable in order to determine when to resume drug treatment. However, in order for changes in BMD to be valid, they must exceed the least significant change (LSC) at the 95% confidence limit; otherwise, the probability of error in the difference between measurements is unacceptably high. The International Society for Clinical Densitometry (ISCD) has long advocated that each densitometry unit perform its own precision study to determine local LSC which can vary based on patient characteristics, technologist skill, and densitometer model1.

ISCD has also advocated that when reporting spine BMD, vertebrae with artifacts should be excluded. There is ample data to suggest though that as vertebral bodies are excluded, the precision of the spine BMD worsens and the LSC rises2, 3, 4, 5. The question arises: if a center performs a precision assessment and determines the LSC for L1-4, how should the LSC for < 4 vertebrae be determined? This is the basis of the two Key Questions for our task force:

1) How should LSC be modified when vertebral bodies are removed from analysis due to technical limitations, and 2) Can a spine BMD based on a solitary vertebral body be reported for monitoring or diagnosis?