Evaluation of 3-year changes in panoramic radiomorphometric indices and fractal dimension values in primary hyperparathyroidism patients

Longitudinal radiographic data on how primary hyperparathyroidism (PHPT) affects the jawbone are scarce. This study aimed to quantify 3-year mandibular changes in PHPT using radiomorphometric indices and fractal dimension (FD) analysis on panoramic radiographs. A retrospective, age- and gender-matched, 3-year longitudinal comparison was performed using 120 panoramic radiographs from 60 individuals (30 PHPT; 30 controls). Each participant contributed baseline and 3-year images (mean interval: 3.26 ± 0.13 years). Radiomorphometric indices (SI, MI, AI, PI, PMI) and FD from three standardized regions of interest (ROI1: condyle; ROI2: angle; ROI3: mental) were measured. Intra-observer reliability was excellent (ICCs 0.80–0.96). Between-group and within-group comparisons used Mann–Whitney U and Wilcoxon signed-rank tests (two-sided, p < 0.05). The antegonial index (AI) was lower in PHPT than controls at baseline (3.26 ± 0.43 vs. 3.62 ± 0.81 mm; p = 0.034) and remained lower at 3 years (3.24 ± 0.35 vs. 3.55 ± 0.43 mm; p = 0.050). Other indices (SI, MI, PI, PMI) showed no between-group differences at either timepoint and no significant within-group change (all p > 0.05). FD analysis showed a lower ROI2 (mandibular angle) value in PHPT at baseline (p = 0.050) and a significant difference at 3 years (p < 0.01); ROI1 and ROI3 did not differ between groups at either timepoint and exhibited no temporal change (all p > 0.05). In a 3-year age- and gender-matched comparison, PHPT was associated with consistently lower AI and reduced FD in the mandibular angle, while other indices and regional FDs were largely unchanged. These findings identify panoramic AI and ROI2-FD as adjunctive, non-invasive markers that could prompt biochemical evaluation when clinical suspicion exists; confirmation and clinical integration require larger multicenter prospective studies with concurrent biochemical profiling.