The thoracic spine is usually scanned during chest CT examinations, which could supply the opportunity for spine BMD measurement. However, the ACR recommended criteria are based on LBMD. The thresholds of TBMD for osteoporosis diagnosis have not been clarified. In the present study, we explored two approaches to define low bone mass or osteoporosis based on TBMD. Our data showed that the diagnosis of osteopenia or osteoporosis using TBMD had good agreement with LBMD. Two methods could be used. One is to calculate the threshold of TBMD in identifying osteopenia or osteoporosis. The other is to normalize the data of TBMD to LBMD, and then use the ACR criteria to define osteopenia or osteoporosis. The two methods showed good agreement with that based on LBMD.
TBMD is usually higher than LBMD [3, 6, 9]. In addition, TBMD is highly correlated with LBMD (r > 0.80) [6, 10]. Our data also showed higher correlation coefficients (r = 0.93). The highly positive correlation makes it feasible to normalize TBMD to LBMD for osteoporosis definition. Our data showed that the performance of osteoporosis identification was good using the normalized LBMD. A previous study also showed a comparable T-score results between measured BMD and translated LBMD [6].
Few studies have shown the thresholds of lower thoracic vertebral BMD in defining osteopenia or osteoporosis. Therefore, we also calculated the threshold of TBMD in identifying osteopenia and osteoporosis using the LBMD ACR definition as the reference. The cutoff values were 127.7 mg/cm3 for osteopenia and 90.9 mg/cm3 for osteoporosis. The two thresholds showed good agreement with LBMD in defining osteopenia or osteoporosis (kappa = 0.80). Interestingly, Rühling et al. reported BMD thresholds at T11 (127.8 mg/cm3 for osteopenia and 91.4 mg/cm3 for osteoporosis) and T12 (121.0 mg/cm3 for osteopenia and 83.5 mg/cm3 for osteoporosis) for the identification of osteopenia and osteoporosis using linear regression analysis [7]. Unfortunately, the diagnostic performance was not shown in their study, and they did not show the thresholds of combined T11 and T12. In Rühling et al.’s study [7], a linear regression equation was obtained between T11/T12 and L1-L3. Then, the thresholds were calculated based on these equations and the ACR criteria (120 mg/cm3 and 80 mg/cm3). We calculated the thresholds using the ROC curve which has been widely used to choose the optimal cutoff value. We also defined the thresholds by substituting the values of 120 mg/cm3 and 80 mg/cm3 into a linear regression equation in our population and evaluated their performance in identifying osteoporosis or osteopenia. The agreement was lower than that of our other two methods. The value of this method must be confirmed. On the whole, a previous study and our data indicated that ACR criteria may be not fit for TBMD in identifying osteoporosis. Because of the relative small sample size, further studies with a larger sample sizes are needed to obtain the thresholds of TBMD in identifying osteopenia and osteoporosis. Our study is an exploration.
Osteoporosis is associated with bone fractures. We also used the thresholds of TBMD (91 mg/cm3) and TTBMD to predict the outcomes of incident fracture in an independent population in this study. Osteopenia and osteoporosis defined by the thresholds or the TTBMD were associated with the risk of VCF, indicating that our methods are credible. Our data showed that approximately 53.6–60.7% of subjects with VCF had osteoporosis. VCF is also associated with muscle quality [11, 12] or early menopause and current smoking [13]. Previous data showed that osteoporosis accounts for approximately half of all hip fractures [14]. Our data showed a slightly better performance than the previous study. Those data considering clinical endpoints, such as fracture, further indicated that the ACR criterion of 80 mg/cm3 was not fit for TBMD. A new threshold should be set for TBMD in defining osteoporosis. Although our study obtained two thresholds that were close to a recent report (127.8 mg/cm3 for osteopenia and 91.4 mg/cm3 for osteoporosis) [7], many efforts should be undertaken to define a generalizable TBMD threshold for osteoporosis.
Lung cancer screening with low-dose CT has been gradually increased over the past decade. The U.S. Preventive Services Task Force (USPSTF) recommends that adults aged 50 to 80 years old who have a 30 pack-year smoking history and currently smoke or have quit within the past 15 years should undergo annual screening for lung cancer [15]. Tobacco use in China is a critical public health issue. There are more than 300 million smokers and 740 million nonsmokers exposed to second-hand smoke in China [16]. Moreover, lung caner screening is recommended for subjects with chronic obstructive pulmoriary disease China [17]. Screening is also high in women because the incidence of lung cancer in women and nonsmokers is high in China [18]. Lung cancer is the leading cause of cancer-related deaths in China [19]. Therefore, a large population undergoes lung cancer screening every year in China. Old age is a common risk factor for osteoporosis and lung cancer. Osteoporosis evaluation during CT scans for lung cancer screening does not cause additional radiation exposure or increase scanning time. Some studies have shown that cardiac CT is useful to identify individuals with low BMD and individuals with a high risk of fracture [2, 20]. Our study first reported that TBMD based on CT for lung cancer screening is also useful to identify individuals with low BMD and individuals with a high risk of fracture.
The ACR criteria for osteoporosis diagnosis do not consider the sex differences. In this study, we also did not separately calculate the TBMD threshold in women and men. A recent study showed that there were no significant differences in TBMD between men and women in old age (62.2 ± 12.1 years), which was very close to our population study [20]. Budoff et al. also reported similar results between men and women aged 50–75 years old [6]. Our data showed that the TBMD of women was lower than that of men. However, no significant differences were found in our populations. These results indicated that the thresholds of TBMD obtained in our study were fit for both men and women.
Our study had several advantages. First, we used two methods to define osteoporosis based on lower thoracic vertebral BMD. Second, we also validated our results in a population with VCF. Our study also had limitations. First, the sample size was not large in calculating the threshold of TBMD or translating TBMD into LBMD. Our results might need to be validated in studies with large sample size or in an independent population. Second, only lower thoracic spine (T11-T12) were included in osteoporosis analysis because the two thoracic spine locations were close to the lumbar spine, and the mean BMD of the two thoracic spine location was the lowest in thoracic spine [20]. The thresholds in our study only apply to the mean BMD of T11-T12 because volumetric BMD in vertebrae increases at increasingly higher levels. Third, all our subjects were older than 50 years old. The sample size was small for young adults because BMD measurement was usually performed on old adults. Whether our results can be generalized to younger adults will also require further exploration. However, a recent study showed that the BMD of T11 and T12 were higher than 150 mg/cm3 in subjects < 50 years old [7]. The addition of young adults may not change our results.
In conclusion, our study showed that the ACR criteria might not be suitable for TBMD in defining osteoporosis. However, BMD based on the lower thoracic spine can be used to identify osteopenia and osteoporosis if a slightly higher threshold or TTBMD is used. Two approaches, based on TBMD threshold and data regarding translated LBMD, both showed good diagnostic performance in identifying osteoporosis. TBMD can also predict the incident new vertebral fractures. More studies are needed to obtain generalizable TBMD thresholds for osteoporosis.