Significance of first- and second-order textural [18F]FDG PET/CT parameters in the prognosis of head and neck squamous cell carcinoma patients
Background: This study aimed to evaluate the predictive value of first- and second-order textural [ 18 F]fluoro-2-deoxy-d-glucose positron emission tomography/computed tomography ( [ 18 F]FDG PET/CT ) features for overall survival (OS) in head and neck squamous cell carcinoma (HNSCC) and to analyse relationships between parameters and differences between T groups. Materials and methods: Eighty-three HNSCC patients were analysed. PET/CT metabolic and volumetric parameters included standardized uptake values (SUVs), volumes of interest (VOIs), metabolic tumour volume (MTV), total lesion glycolysis (TLG), and asphericity (ASP). Intratumoural heterogeneity was evaluated using first-order textural features — coefficient of variation (COV), skewness (SKE), kurtosis (K), and entropy (E) — and second-order textural features derived from neighbourhood grey-tone difference matrix (NGTDM), neighbourhood grey-level dependence matrix (NGLDM), grey-level distance zone matrix (GLDZM), and grey-level size zone matrix (GLSZM) matrices. Correlations among first-order parameters, OS, age, and T-stage were assessed using univariate and multivariate Cox regression and Kaplan–Meier analysis. Patients were grouped as T1/T2 (group 1), T3/T4 (group 2), and Tx [cancer of unknown primary (CUP); group 3]. Additionally, correlations between first- and second-order features were checked using Spearman’s rank correlation. Results: Metabolic tumour volume, TLG, entropy derived from PET (E PET ), and entropy derived from CT (E CT ) showed significant, moderate negative correlations with OS. Univariate analysis identified these parameters as significant OS predictors, while multivariate analysis indicated that EPET was not an independent prognostic factor. Significant differences in PET/CT parameters were found between group 1 and group 2 (8 parameters), and between group 2 and group 3 (10 parameters), with OS differing significantly between groups 1–2 and group 3. Conclusions : Metabolic tumour volume, TLG, E PET , and E CT serve as relevant prognostic indicators of OS in HNSCC. The observed differences in first-order textural features across T-stage groups highlight their potential value for tumour characterisation and classification.
Authors
- Paulina Cegła (ORCID: https://orcid.org/0000-0002-6676-3436)
- Michał Michalak (ORCID: https://orcid.org/0000-0002-2852-3984)
- Witold Cholewiński
- Wojciech Golusiński (ORCID: https://orcid.org/0000-0002-6075-3464)
- Ewa Majchrzak (ORCID: https://orcid.org/0000-0003-3555-2318)
- Anna Kubiak (ORCID: https://orcid.org/0000-0001-7363-5654)
- Ewelina Kaminska (ORCID: https://orcid.org/0009-0009-0635-8322)
- Agnieszka T. Boś-Liedke
- Maciej Kozak
Institutions
- Poznan University of Medical Sciences (PL)
- Greater Poland Cancer Center (PL)
- Adam Mickiewicz University in Poznań (PL)
Publication Details
- Journal
- Reports of Practical Oncology & Radiotherapy
- Published
- 2026-11-13
- DOI
- https://doi.org/10.5603/rpor.115418
- Primary Topic
- Radiomics and Machine Learning in Medical Imaging
- Type
- article
- Field-Weighted Citation Impact
- 0.00