Work

Assessing Condylar hyperplasia management outcomes using semi-automatic segmentation methods: A reliability study

Year: 2026
Type: article
Abstract: OBJECTIVES: To assess the reliability of two condylar segmentation methods in patients with Unilateral Condylar Hyperplasia (UCH) treated with proportional condylectomy. METHODS: A sample of 10 patients with UCH who underwent proportional condylectomy was randomly selected from a private practice. Landmarks were defined and volumetric measurements were performed from cone-beam computed tomography (CBCT) from affected (n = 10) and non-affected sides (n = 10), as well as before and after proportional condylectomy. Condyles were segmented using two methods: method SN (sigmoid notch method) and method ML (mandibular lingula method). A single blind operator performed the procedure twice, with a four-week interval between the evaluations. Random error was assessed using the Dahlberg formula, and intra-rater reliability was evaluated using the Intraclass correlation coefficient (ICC) in a two-way mixed-effects model. RESULTS: An excellent reliability was observed for both methods (SN: ICC = 0.96, CI 95% 0.92-0.98; ML: ICC = 0.91, CI 95%, 0.79-0.96). The method ML exhibited a slightly higher variability in measurement agreement, indicating higher random measurement error than method SN. On average, measurements with method ML were 9.9% higher on the second evaluation, while method SN measurements were 4.7% higher. CONCLUSIONS: Condylar segmentation is a reliable approach for evaluating volumetric changes in complex mandibular condyles. While, both SN and ML segmentation methods demonstrated strong intra-rater reliability, the method SN exhibited slightly lower variability. Nonetheless, this study highlights the potential of method ML as a clinically relevant alternative to condylar assessment.
Source: PLoS ONE
Language: English
Subfield: Orthodontics
Field: Dentistry
Open Access status: gold
Estimated APC paid: $2,382
OpenAlex ID: W7236106495
FWCI:
Cites: 38
Cited by:
Related to: 10
Parts: 4