A novel computed tomography-based algorithm for the quantitative identification of lumbar spine-localized trabecular...

Conventional bone health assessment based on average bone mineral density (BMD) cannot capture localized trabecular abnormalities, and BMD is not always decreased in patients with type 2 diabetes mellitus (T2DM). To overcome these limitations of using BMD alone in diagnosis, this study aimed to apply a novel computed tomography (CT)-based algorithm to quantitatively evaluate lumbar spine-localized trabecular bone loss (LTBL) in individuals with and without T2DM.

This study retrospectively enrolled individuals with and without T2DM, and their basic clinical information was collected. Lumbar spine BMD was measured via quantitative CT (QCT). LTBL was quantified via a novel algorithm, with LTBL defined as trabecular regions with BMD <40 mg/cm³ and volume ≥16.5 mm³. Parameters of LTBL, including the number and volume of LTBL areas, were recorded. Mann-Whitney tests were used to compare BMD and LTBL parameters between the T2DM and non-T2DM groups. Spearman correlation analysis was conducted to assess the relationships between age, BMD, and LTBL parameters. Regression analysis was used to evaluate factors influencing LTBL. All statistics are expressed in terms of median and interquartile range.

A total of 166 participants were included, comprising 94 with T2DM and 72 without T2DM. The T2DM group had a slightly higher BMD than did the non-T2DM group (P>0.05). The T2DM group, as compared to the non-T2DM group, had a significantly higher total number of LTBL areas [12.00 (8.00, 24.00) vs. 10.00 (5.00, 16.50), P<0.05], number of vertebrae with LTBL [5.00 (4.00, 5.00) vs. 4.00 (3.00, 5.00), P<0.05], and total volume of LTBL areas [576.00 (303.00, 1,276.00) vs. 448.50 (199.00, 853.50) mm³, P<0.05]. There was a slight positive correlation between age and total number (r=0.35; P<0.05) and volume (r=0.33; P<0.05) of LTBL areas in the T2DM group. BMD was moderately negatively correlated with both the total number (r=-0.47; P<0.05) and volume of LTBL areas in the T2DM group (r=-0.47; P<0.05). Regression analysis showed that age was positively associated with the total number of LTBL areas (β=0.336; P<0.05), total volume of LTBL areas (β=34.778, P<0.05), and number of vertebrae with LTBL (β=0.019; P<0.05); meanwhile, BMD was negatively associated with these parameters (β=-0.252, β=-14.778, and β=-0.014, respectively; P<0.05), while T2DM was positively associated (β=6.222, β=497.558, and β=0.533, respectively; P<0.05). In contrast, gender and BMI showed no significant associations with any of these measures (all P values >0.05). Finally, only age was significantly correlated with the mean volume of LTBL (β=0.732; P<0.05).

Individuals with T2DM have more severe lumbar trabecular network injury despite a similar BMD to those without T2DM, and the effect of T2DM on LTBL is independent of age and BMD. LTBL may represent a promising imaging marker for detecting early microstructural deterioration in the bone of individuals with diabetes.