Prognostic Value of Volumetric and Voxel-Based [ 18 F]FDG-PET Metrics in Head and Neck Cancer
Andrej Doma, Tina Škerl, Andrej Studen, Primož StrojanIntroduction
In head and neck cancer (HNC), the prognostic value of [ 18 F]FDG PET/CT SUVmax is limited. Volumetric parameters such as metabolic tumor volume (MTV) and lean body mass-normalized SUV (SUL) may improve risk stratification, particularly in patients with higher body mass index.
Methods
We conducted a retrospective cohort study of 28 patients with histologically confirmed non-metastatic squamous cell HNC undergoing pretreatment [ 18 F]FDG PET/CT and definitive (chemo)radiotherapy. Primary tumor (P), neck nodal (N), and total (T) MTV, SUVmax, and SULmax were measured. Associations with 2.5-year overall survival (OS2.5), local (LOC), regional (REG), and distant (M+) failure were assessed using logistic regression, receiver operating characteristic (ROC) analysis, and Kaplan–Meier curves. Statistical analyses were performed in R (v3.5); p < 0.05 considered statistically significant.
Results
The median age was 55 years (IQR, 45–65), and most patients were female (78.6%). Primary tumor sites were nasopharynx (57.1%) or oropharynx (39.3%); one patient had an unknown primary tumor. All tumors were squamous cell carcinomas. Median follow-up was 37.6 months and 11 patients (39%) died. MTV-T showed the strongest association with OS2.5 (AUC 0.85, threshold 14 ml, p = 0.003), followed by SULmax-N (AUC 0.78, p = 0.025), SUVmax-N, and MTV-P (both AUC 0.77, p = 0.03). For distant failure, SULmax-N and SUVmax-N were the strongest predictors (both AUC 0.90, p = 0.0012), outperforming volumetric parameters. Voxel-based primary tumor uptake metrics did not predict local or regional failure, which were associated with T and N stage and MTV-P. SUL metrics showed high correlation with body weight-normalized SUVs (r=0.99–1.00) but no discriminatory advantage.
Conclusion
Volumetric PET parameters outperformed voxel-based uptake metrics for overall survival prediction, while nodal SULmax and SUVmax better predicted distant failure. Lean body mass normalization did not improve prognostic performance. These findings support PET-derived volumetric and metabolic intensity parameters for risk stratification and require validation in larger cohorts.