Work

Evaluation of Bone Invasion in Gingival Squamous Cell Carcinoma Using PET/CT Radiomics

Year: 2026

Type: article

Abstract: Background: Accurate preoperative assessment of bone invasion is essential for surgical planning in gingival squamous cell carcinoma (SCC). This study evaluated the diagnostic value of conventional 18F-FDG PET/CT-derived imaging parameters and PET-based radiomic features for predicting histopathologic bone invasion. Methods: This retrospective study included 68 patients with gingival SCC who underwent preoperative 18F-FDG PET/CT followed by primary surgical treatment. Bone invasion was confirmed histopathologically. Conventional PET parameters (MTD, SUVmax, SUVmean, MTV, and TLG) and PET-based radiomic features were extracted from the primary tumor. Group comparisons, receiver operating characteristic (ROC) analysis, and univariate logistic regression were performed. Results: Patients with bone invasion showed significantly higher MTD, SUVmax, MTV, and TLG than those without bone invasion (all p < 0.05), whereas SUVmean was not significantly different. Entropy, correlation, and zone entropy were significantly higher in the bone invasion-positive group. MTV showed the numerically highest AUC among conventional PET parameters (AUC = 0.748), whereas correlation (AUC = 0.725) and zone entropy (AUC = 0.714) showed the highest performance among radiomic features. Univariate logistic regression showed significant associations of MTD, MTV, TLG, entropy, correlation, and zone entropy with bone invasion. Conclusions: Conventional PET parameters and PET-based radiomic features were significantly associated with histopathologic bone invasion in gingival SCC. Radiomic features reflecting intratumoral heterogeneity may complement conventional metabolic parameters and improve preoperative assessment of bone invasion.

Source: Diagnostics

DOI: 10.3390/diagnostics16172685

Authors: Junwoo Kwark, Seunggon Jung, Min-Suk Kook, Hong-Ju Park, Junho Chang, Jaeyoung Ryu

Institutions: Chonnam National University, Chonnam National University Hospital, Chonnam National University Dental Hospital

Language: English

Topic: Radiomics and Machine Learning in Medical Imaging

Subfield: Radiology, Nuclear Medicine and Imaging

Field: Medicine

Domain: Health Sciences

SDG: Good health and well-being

Open Access status: gold

Estimated APC paid: $3,137