Akademik Veri Yönetim Sistemi
Oral Surgery, Oral Medicine, Oral Pathology and Oral Radiology, 2026 (SCI-Expanded, Scopus)
Aim This study evaluated whether fractal analysis of periapical radiographs can detect periradicular disease. Materials and methods Sixty periapical radiographs of mandibular molars were divided into four groups based on the Periapical Index (PAI) ( n = 15 per group; scores 1-4). Fractal analysis of periapical and furcation regions was performed using the box-counting method in ImageJ. Multiple region of interest (ROI) sizes were tested (20 × 20, 25 × 25, 30 × 30, 40 × 40, 50 × 50, 60 × 60, 70 × 70, 20 × 50, and 20 × 40 pixels) to determine the best correlation with PAI scores. Statistical analyses included one-way analysis of variance with Tukey’s post hoc test, Student t-test, and ROC analysis (p < 0.05). External validation was conducted using 33 additional radiographs. Results Fractal dimension (FD) values from periapical ROIs (20 × 20, 25 × 25, and 30 × 30) and bifurcation ROIs (20 × 20 and 25 × 25) significantly differed between healthy (PAI 1-2) and diseased (PAI 3-4) teeth ( P < .05). ROC analysis supported the diagnostic value of these ROIs, with the highest AUC for the bifurcation 20 × 20 ROI. In the internal dataset, periapical and bifurcation 20 × 20 ROIs achieved the highest accuracies (0.82 and 0.80, respectively), and the bifurcation 20 × 20 ROI showed the greatest sensitivity (0.93). External validation confirmed the robustness of the bifurcation 20 × 20 ROI (accuracy 0.76, sensitivity 0.65, and specificity 0.88). Conclusion Fractal analysis can identify periradicular disease on periapical radiographs. Diagnostic accuracy depends on the analyzed region and ROI size, with the 20 × 20-pixel bifurcation ROI yielding the best performance.