Heterogeneity and SUVmax of 18F FDG PET/CT Predicts Outcomes of Salivary Gland Carcinoma of High Risk Histology

1433 Objectives: Heterogeneity and SUVmax of 18F FDG PET/CT Predicts Outcomes of Salivary Gland Carcinoma of High Risk Histology Studies have shown that maximum standard uptake values (SUVmax) may serve as a prognostic indicator in salivary gland carcinoma (SGC). SGC patients with high-risk histology suffered from poor prognosis. We intended to find whether the texture features of pretreatment 18F-FDG PET/CT images can provide additional prognostic information in SGC patients with high-risk histology. Methods: We retrospectively analyzed the pretreatment 18F-FDG PET/CT images of M0 high-risk histology SGC patients who had completed therapy. Fixed 40% percentage of SUVmax was used as a cutoff for tumor boundary. PET texture features were extracted using histogram analysis, normalized grey-level co-occurrence matrix, grey-level run length encoding method and grey-level size zone matrix. Receiver operating characteristic (ROC) curves were used to identify the optimal cutoff values for the PET parameters. Results: Eighty-six patients were enrolled. ROC curves and univariate Cox analyses revealed SUVmax, total lesion glycolysis, SUV-entropy, uniformity, entropy, run-percentage, high-intensity run emphasis and high-intensity zone emphasis were associated with overall survival (OS). Multivariate Cox analysis showed that PET parameters: SUVmax and SUV-entropy and clinical factors of performance status (PS), N2c-N3 stage were independently associated with OS, and disease-specific survival (DSS). A prognostic scoring system was derived. Patients with worse PS or N2c-N3 stage or presented with SUVmax more than 6.67 and SUV-entropy more than 2.50 experienced significantly worse OS and DSS. Conclusions: SUVmax and SUV-entropy were independent prognostic predictors in patients with high-risk histology SGC. We developed a scoring system that may serve as a risk-stratification strategy for guiding therapy.