L1 vertebral density on CT is too variable with different scanning protocols to be a useful screening tool for osteoporosis in everyday practice

Abstract

Objective: To determine whether a single L1 density threshold can be used to screen all patients undergoing CT at a busy tertiary referral centre for those at risk of osteoporosis.

Methods: 200 patients, who had been randomly selected for an audit of CT reporting of incidental vertebral fractures, had their L1 density measured on 864 routine CT examinations. These had been performed with a variety of kVp and intravenous (i.v.) contrast protocols, reflecting the wide range of imaging indications.

Results: L1 density measured on thick axial, thin axial or sagittal images had an excellent intraclass correlation coefficient (0.996). For the same patients imaged twice within 6 months, there was mean intraexamination L1 density difference of 27.5 HU. Variability due to i.v. contrast medium administration resulted in a mean difference of 24.5 HU. Mean difference due to acquisition kVp was 24.1 HU. Once matched for sex, age, kVp and i.v. contrast, there was a significant difference between the L1 density in patients with vertebral fractures compared to those without fractures (mean 30.19 HU).

Conclusion: There is significant variability in the L1 vertebral body CT density due to differences in acquisition variables such as kVp and timing post-i.v. contrast medium. Advances in knowledge: Previous studies suggested that L1 CT density could be used to screen for osteoporosis. The current study cautions that it is not possible to define a single L1 density threshold for screening, due to the number of variables within a wide range of scanning protocols for different imaging indications in everyday practice.

Figure 1.

Region of interest in L1 in (a) thick axial, (b) thin axial and (c) sagittal reformatted images.

Figure 2.

Increase in L1 density post-i.v. contrast medium. HU, Hounsfield Unit; i.v., intravenously.

Figure 3.

L1 density decreases with age but is not significantly different in males compared to females. All L1 densities measured on images acquired at 120 kVp post-i.v. contrast medium. Error bars show 95% confidence limits for the mean L1 density. HU, Hounsfield Unit; i.v., intravenously.

Figure 4.

L1 density compared to DXA or QCT BMD for 37 patients including those with ( grey diamonds) and without ( black squares) vertebral fractures. The WHO DXA thresholds are indicated for osteoporosis (T < −2.5) and osteopaenia (−1 < T > −2.5). BMD, bone mineral density; DXA, dual energy X-ray absorptiometry; HU, Hounsfield Unit; i.v., intravenously; QCT, quantitative CT.