Irregularity of pretherapeutic tumor FDG uptake as prognostic factor in head and neck carcinoma

Ziel/Aim:

FDG PET is widely used for diagnosis, staging, and therapy monitoring in patients with head and neck carcinoma. Quantitative analysis of FDG uptake in the tumor has the potential to complement visual reading for improvement of diagnostic accuracy. However, the prognostic value of SUV-based and some more sophisticated (semi-)quantitative measures is unsatisfactory in this cancer type. The aim of the present study was to propose and evaluate a new histogram-based measure of the irregularity of FDG uptake in the tumor as prognostic factor in head-and-neck carcinoma.

Methodik/Methods:

The retrospective analysis included 32 pts with head and neck tumor in whom whole-body FDG-PET/CT had been performed for primary staging (n=26) or recurrence diagnosis (n=6). All patients underwent therapy after PET/CT. The FDG image of the primary tumor was segmented fully automatically using the ROVER 3D segmentation tool which is based on thresholding at the volume-reproducible intensity threshold after subtraction of the local background. SUVmax, SUVmean, glycolytic volume (GV) and total mean glycolytic volume (MGV) were determined in addition to the novel measure of irregularity IRREG in the ROI. IRREG is defined as the deviation of the ROI's symmetry from a sphere symmetry, which can be computed as the third power of the ROI's surface divided by the second power of the ROI's volume. IRREG is normalized in such a way that for a spherical ROI IRREG is 1. Kaplan-Meier curves were obtained for all tested parameters with respect to both progression-free (PFS) and overall survival (OAS). Survival and PFS curves were separated by the best discrimination threshold by ROC analysis and compared by log-rank tests.

Ergebnisse/Results:

Four patients died during the follow-up, 11 further patients experienced tumor progression. Median PFS was 12.5 months. Median SUVmax/mean over the whole group was 10.9/5.9. Median glycolytic volumes were 16.0 ml and 85.6 ml for GV and MGV, respectively. Kaplan-Meier analysis revealed prognosticpower with respect to OAS for IRREG (p=0.016) and both glycolytic volumes (GV: p=0.0025, MGV: p=0.08). Statistical significance for the prediction of PFS was very high for IRREG (p=0.0003) and GV (p=0.0006), somewhat smaller for MGV (p=0.043). Higher tumor irregularity was associated with high risk with respect to both recurrence and survival. The SUV measures were not predictive, neither for OAS nor PFS.

Schlussfolgerungen/Conclusions:

The irregularity of FDG uptake of the primary tumor in the baseline FDG-PET as measured by the novel irregularity measure is predictive for tumor recurrence and survival in patients with head and neck carcinoma. Therefore, the irregularity measure is a promising marker for risk stratification in these patients. Reliable analysis of the independent contribution to risk stratification over other parameters such as the glycolytic volume requires further studies with larger patient samples.