aBMD measured by DXA as a reference standard in osteoporosis diagnosis
DXA measures the aBMD by obtaining the X-ray attenuation of patients exposed to low levels of X-ray radiation. The attenuation of each pixel in the predefined region of interest is summarized and converted to aBMD in units of g/cm
2 using the previously established tissue quality attenuation coefficient [
39,
40]. aBMD at the spine and hip is the reference standard for osteoporosis diagnosis. However, as our comprehension of bone microstructure advances, its significance in fracture risk prediction has encountered challenges [
41,
42]. The prevalence of vertebral fractures varies across the vertebral level, with peaks in the middle of the spine (T7-T8) and the thoracolumbar junction (T12-L1). Vertebral fractures at these specific sites are more strongly associated with the risk of new fractures in the upper than in the lower spine [
43]. It's important to note that DXA has certain limitations: it lacks accuracy in measuring BMD in the thoracic spine due to the influence of soft tissues and chest coverage [
44]. Furthermore, DXA cannot provide detailed information about bone microstructure [
45], consequently failing to adequately explain the correlation between aBMD and bone structure parameters [
46]. The vertebral fracture assessment technique is an extension of DXA that attempts to evaluate vertebral fractures from a more comprehensive point of view, but certain limitations still exist, including its inability to identify vertebrae between T7-L4 and to detect common osteoporosis-related sclerosis or lytic changes [
47].
vBMD measured by QCT in osteoporosis
vBMD measured by QCT requires calibrating the computed tomography values in Hounsfield units (HU) according to known density standards to calculate the equivalent density [
48]. QCT can selectively measure trabecular or cortical BMD based on the region of interest, allowing further assessment of the parameters related to bone microstructure [
49]. Moreover, it can accurately distinguish soft tissue compartments, minimizing the impact of extraosseous components [
50]. A cross-sectional study found that QCT was more sensitive compared to DXA in detecting osteoporosis in postmenopausal women [
51]. In a cohort study, it was found that QCT performed better in the risk assessment of imminent vertebral fractures than DXA [
52].
Considering radiation exposure, it is advisable to minimize the number of vertebrae measured when utilizing QCT, with L1-L2 being the most commonly recommended site for evaluation [
53]. According to the American Society of Radiology, the average BMD of L1-L2 is used for radiological diagnosis of osteopenia and osteoporosis [
54]. A previous study has demonstrated high concordance between thoracic and lumbar BMD measured by QCT [
55]. Another study also showed a strong correlation between the average BMD of L1–L2 and the other lumbosacral vertebrae (L3–S1) [
56].
Due to disparities in equipment, imaging software, and the expertise of analysts, vBMD estimated by QCT is still individualized. Nevertheless, promising strides toward standardization have been highlighted through several multi-center clinical trials [
48]. We look forward to an official recommendation to provide more details such as patient variables and calibration methods to guide clinical use of QCT in measuring vBMD.
In this study, we analyzed the ability of vBMD detected by QCT and aBMD detected by DXA in distinguishing between osteoporosis patients with vertebral fractures and those without fractures. Our findings indicated that both vBMD assessed by QCT and aBMD assessed by DXA can effectively differentiate osteoporotic fractures regardless of the scanning site, and vBMD (WMD = − 27.08; 95% CI − 31.24 to − 22.92) performed better than aBMD (WMD = − 0.05; 95% CI − 0.08 to − 0.03). Among the six articles, the scanned site of two of QCT [
27,
28] and one of DXA were at the femur, and the others scanned more than two vertebrae between L1 and L4, which was consistent with the recommendation [
13,
54]. Regrettably, the precise location of vertebral fractures was not described in detail in the six articles, thereby limiting us to deeply analyzing the correlation between BMD and vertebral fractured site. Interestingly, we found that even the BMD obtained from femur scans exhibited an association with vertebral fractures, suggesting that osteoporosis was an overall degeneration of bone health.
There are limitations needed to be improved in our research. In the analysis of evaluating the ability of QCT and DXA to distinguish fractures, it would be inappropriate to conclude that QCT is superior to DXA in identifying individuals with fractures based solely on vBMD and aBMD because we didn’t clear whether different units will lead to different results related to fracture risk. Also, the analysis of fracture risk based on BMD was still insufficient, and the quality of bone microstructure should be another important factor in fracture risk. However, we can deduce that osteoporosis associated with vertebral fracture risk may exhibit, on average, deficits of − 27.08 in vBMD and − 0.05 in aBMD, in comparison to osteoporosis without fracture. In the future, we may focus on patients with fractures and further study the application of QCT in osteoporotic fractures.
This paper demonstrated the strong correlation between BMD and osteoporotic fracture risk obtained by QCT and DXA, and identifies the superiority of QCT in predicting osteoporotic fracture risk.