Quantitative Texture Analysis of Trabecular Bone Alterations Surrounding Vertical Root Fractures on Panoramic Radiographs

Objectives Vertical root fractures (VRFs) are difficult to diagnose on panoramic radiographs because fracture lines are often not clearly visible. This study aimed to quantitatively evaluate trabecular bone alterations surrounding VRF roots using texture analysis of panoramic images and to identify indirect imaging features that may assist in VRF assessment. Materials and Methods This retrospective single-center study included 38 patients with surgically confirmed 40 VRF roots and 40 matched control roots without clinical or radiographic signs of VRF. Seven anatomically standardized Region of Interests (ROIs) were manually defined around each root. Six texture features, including gray-level co-occurrence matrix (GLCM) parameters (contrast, ASM, entropy, IDM), fractal dimension (FD), and lacunarity, were extracted. Inter-observer reliability was assessed using intraclass correlation coefficients (ICC) and Bland–Altman analysis. Group differences were evaluated using ROI-level comparisons, relative percentage change (Δ%), and inter-feature correlation analysis. Results Overall ROI-averaged analysis of texture features demonstrated substantial overlap between groups. However, ROI-based analysis revealed consistent directional differences. GLCM contrast decreased by approximately 11–29% across all ROIs in the VRF group, whereas IDM increased by 14–19% (p < 0.001). Lacunarity showed region-specific differences. Although FD did not differ significantly in absolute terms, moderate shifts in inter-feature correlations involving FD (Δr = 0.20–0.30) were observed. Inter-observer reliability demonstrated good-to-excellent agreement. Conclusions ROI-based quantitative texture analysis of panoramic radiographs identified spatially organized trabecular bone alterations associated with VRF, even in the absence of a clearly visible fracture line. These findings suggest that texture-based features derived from routinely acquired panoramic images may provide complementary information in VRF evaluation.