Low-dose CT for assessing microstructural alterations in Renal Osteodystrophy among uremic patients: a retrospective...

Renal Osteodystrophy (ROD) is a severe complication among uremic patients. However, the gold standard for diagnosis, bone biopsy, is invasive and difficult to apply routinely. Although low-dose CT (LDCT) is widely used for other assessments in this population, its value in the non-invasive diagnosis of ROD remains to be elucidated.

To investigate the feasibility and diagnostic performance of quantitative assessment of lumbar vertebral microstructural changes using routine LDCT images for the identification and diagnosis of ROD and its subtypes.

This was a retrospective cohort study. A total of 155 uremic patients undergoing regular dialysis who concurrently underwent abdominal/pelvic LDCT and bone metabolic biochemical tests at the Affiliated Hospital of Jinggangshan University between June 2023 and June 2025 were consecutively enrolled. Based on the KDIGO guidelines combined with biochemical markers (iPTH, calcium, phosphorus, ALP, etc.), patients were categorized into a non-ROD group (n = 49), high-turnover ROD (n = 44), low-turnover ROD (n = 30), and mixed-type ROD (n = 32). Using 3D Slicer software, the following parameters were measured in the trabecular bone region of the L1/L2 vertebral bodies: volumetric bone mineral density (vBMD), bone volume fraction (BV/TV), trabecular thickness (Tb.Th), trabecular separation (Tb.Sp), trabecular number (Tb.N), and cortical thickness (Ct.Th). Inter-group differences were compared, correlations between CT parameters and biochemical markers were analyzed, and the diagnostic performance was evaluated using Receiver Operating Characteristic (ROC) curves.

The overall prevalence of ROD was 68.4% (106/155). Significant differences in biochemical markers and CT parameters were observed among the ROD subtypes (all P<0.001). High-turnover ROD exhibited the highest iPTH (552.60 ± 115.90 pg/mL), ALP (260.55 ± 50.29 U/L), and the worst bone microstructure (lowest BV/TV: 13.89 ± 3.09%, highest Tb.Sp: 1.03 ± 0.15 mm). Tb.Sp demonstrated the best performance in differentiating ROD from non-ROD, with an Area Under the Curve (AUC) of 0.902 (sensitivity 81.1%, specificity 89.8%, cutoff value 0.84 mm). CT parameters (e.g., BV/TV, vBMD, Tb.Sp) showed significant correlations with iPTH, ALP, and serum phosphorus (|r| = 0.429-0.579, P<0.001). When combining iPTH with BV/TV to construct a model, the diagnostic AUC improved to 0.888. Additionally, bone microstructure parameters showed significant correlations with coronary artery calcification (e.g., BV/TV: r = -0.383, P<0.001), supporting the bone-vascular axis concept.

Quantitative analysis of lumbar vertebral microstructure using existing LDCT images in uremic patients can effectively differentiate ROD subtypes and corroborate biochemical changes, with trabecular separation (Tb.Sp) showing outstanding diagnostic value. This method serves as a non-invasive and convenient auxiliary tool that may facilitate the clinical identification and precise management of ROD.