Cervicothoracic volumetric bone mineral density assessed by opportunistic QCT may be a reliable marker for osteoporosis in adults

Abstract

This study aimed to validate the correlation between volumetric bone mineral density in the cervicothoracic and lumbar spine using measurements from opportunistic CT scans. The bone density assessment proved feasible, allowing us to propose optimal cut-off values for diagnosing osteoporosis and predicting vertebral fractures in the cervical and thoracic spine.

Objectives: To investigate the performance of cervicothoracic volumetric bone mineral density (vBMD), obtained through opportunistic quantitative computed tomography (QCT), in discriminating patients with/without osteoporosis and with/without vertebral fractures (VFs), using lumbar vBMD as the reference.

Methods: Three hundred twenty-five patients (65.3 ± 19.2 years, 140 women) with routine non-contrast or contrast-enhanced multi-detector CT (MDCT) scans were included. Trabecular vBMD was automatically extracted from each vertebra using a convolutional neural network (CNN)-based framework (SpineQ software v1.0) with asynchronous calibration and contrast phase correction. The correlations of vBMD between each vertebra spanning C2-T12 and the averaged lumbar spine (L1-L3, or L4 and L5) vBMD values were analyzed, considering fracture status and degeneration. Vertebra-specific linear regression equations were used to approximate lumbar vBMD at the cervicothoracic spine.

Results: Cervicothoracic vBMD correlated well with lumbar vBMD (r = 0.79), with significant improvement after excluding degenerated vertebrae (p < 0.05; r = 0.89), except for C7-T3 and T9. Cervical (AUC = 0.94) and thoracic vBMD (AUC = 0.97) showed strong discriminatory ability for osteoporosis (vBMD < 80 mg/cm³). Excluding degenerated vertebrae at the cervical spine increased the AUC to 0.97. Cervical and thoracic vBMD (AUC = 0.74, AUC = 0.72) were comparable to lumbar vBMD (AUC = 0.72) in differentiating patients with and without prevalent VFs. Trabecular vBMD < 190 mg/cm³ for the cervical spine and < 100 mg/cm³ for the thoracic spine were potential indicators of osteoporosis, similar to < 80 mg/cm³ at the lumbar spine.

Conclusion: Cervicothoracic vBMD may allow for determination of osteoporosis and prediction of VFs.

Keywords: Cervicothoracic vBMD; Quantitative computed tomography (QCT); Trabecular vBMD.

Fig. 1

Flowchart for study inclusion. Exclusion criteria were as follows: incomplete visualization of larger parts of the spine (e.g., lumbar spine not depicted; n = 87), insufficient image quality (e.g., due to artifacts; n = 4), vertebral bodies with implanted hardware (n = 22), MDCT imaging not evaluable (e.g., degenerative disease with extensive sclerosis, spinal metastasis, or multiple myeloma, rendering representative vBMD measurements impossible; n = 109), and MDCT acquired for the purpose of myelography (n = 14). In total, 561 patients with routine CT-scans were collected for the study set. The final cohort consisted of 325 adults (140 women) with an age range from 18 to 99 years (mean age of 65.3 ± 19.2 years)

Fig. 2

A Non-contrast CT scan in sagittal view of a 78-year-old man with a VF at T12 (blue circle). B Spine segmentations, including subregion segmentation, were performed, and masks of the cortical bone (green) and masks of the trabecular compartment (light grey) were obtained. Posterior elements were also routinely segmented (various colors). Dental implants at C2 caused beam hardening artifacts (white arrows) that resulted in suboptimal segmentation of the vertebral compartment. Consequently, C2 was excluded from vBMD measurements. Linear regression equations, applied to cervical and thoracic vertebrae and averaged over the cervical and thoracic spine, yielded lumbar vBMD approximations. At the lumbar spine, vBMD values of L1–L3 were averaged. Lumbar approximations derived from the cervical spine vBMD_C3-5 = 42.9 mg/cm³ and from the thoracic spine vBMD_T1-11 = 34.2 mg/cm³. Lumbar vBMD_L1-3 = 55.4 mg/cm³. In this patient, all three measurements correctly identified the prevalent VF. CT, computed tomography; vBMD, volumetric bone mineral density; VF, vertebral fracture

Fig. 3

A, B Pearson’s correlation coefficients between vBMD of C2 through T12 with respect to the averaged lumbar vBMD of L1–L3. A The three distinct curves represent the correlation coefficients for all vertebrae (brown), after the exclusion of fractured vertebrae (turquoise), and after the exclusion of fractured and degenerated vertebrae (blue). The number of degenerated vertebrae that were subsequently excluded is shown in parentheses at the corresponding level. The grading and subsequent exclusion of vertebrae was simultaneously performed at the lumbar spine. B Two curves represent the correlation coefficients for patients aged younger than 40 years (turquoise) and patients aged older than 40 years (brown). *p < 0.05. vBMD, volumetric bone mineral density

Fig. 4

A, B Correlation between lumbar vBMD estimations versus lumbar vBMD (in mg/cm³). Lumbar vBMD estimations derived from the cervical spine (A, left) and the thoracic spine (B, right). Best lines of fit were y_cervical = 1.08*x − 3.48 and y_thoracal = 1.1*x − 5.69. Dotted lines indicate the 95% CI for the slope. Values for unenhanced scans are shown in blue, and values from contrast-enhanced scans are shown in red. The goodness of fit is r² = 0.82 with p < 0.0001 for the cervical spine; and r² = 0.87 with p < 0.0001 for the thoracic spine. C, D Bland–Altman analyses. Plots show the mean versus the difference for vBMD (in mg/cm³) measured at the cervical spine (C, left) and the thoracic spine (D, right). The solid black line represents the neutral line, while the dashed black lines indicate the 95% LOA. The dashed red line is the mean bias. The 95% LOA for C2–C7 ranged from − 39.6 to 47.1 mg/cm³, and for T1-12, they ranged from − 33.5 to 39.5 mg/cm³. CI, confidence interval; CT, computed tomography; LOA, limits of agreement; vBMD, volumetric bone mineral density

Fig. 5

A, B ROC analyses to discriminate patients with and without osteoporosis using averaged vBMD derived from cervical vertebrae (A, left), and averaged vBMD derived from thoracic vertebrae (B, right). C, D ROC analyses to discriminate patients with and without prevalent VFs using averaged vBMD derived from both the cervical, thoracic, and lumbar spine, including degenerated-vertebrae (C, left), as well as similarly from only non-degenerated vertebrae (D, right). For better readability, detailed results are provided in Table 2. AUC, area under the curve; ROC, receiver operating characteristic; vBMD, volumetric bone mineral density; VF, vertebral fracture

Fig. 6

ROC analyses to determine the optimal vBMD cut-off for osteoporosis. AUCs at various thresholds for the cervical spine (A, left) and the thoracic spine (B, right) in non-degenerated vertebrae, without applying any linear regression equations. The best AUC was achieved with a cut-off value of 190 mg/cm³ (AUC = 0.902, CI 0.87–0.93) for the cervical spine and 100 mg/cm³ (AUC = 0.896, CI 0.86–0.93) for the thoracic spine. AUC, area under the curve; CI, confidence interval; ROC, receiver operating characteristic; vBMD, volumetric bone mineral density