Prospective Investigation of (18)FDG-PET/MRI with Intravoxel Incoherent Motion Diffusion-Weighted Imaging to Assess Survival in Patients with Oropharyngeal or Hypopharyngeal Carcinoma

SIMPLE SUMMARY: Intravoxel incoherent motion (IVIM) diffusion-weighted imaging (DWI) is an advanced MRI technique from which both diffusion and perfusion parameters of tumor tissues can be extracted. The question as to whether (18)F-FDG PET/MRI using the IVIM-DWI technique may provide prognostic information in patients with head-and-neck cancers remains unanswered. This prospective study was therefore designed to evaluate the prognostic significance of IVIM, dynamic contrast-enhanced MRI (DCE-MRI), and PET parameters derived from integrated PET/MRI in patients with oropharyngeal or hypopharyngeal squamous cell carcinomas (OHSCC). Our results indicated that metabolic tumor volume on (18)F-FDG PET and D* derived from IVIM-DWI were independent risk factors for overall survival, whereas K(trans) and iAUC derived from DCE-MRI were significant predictors of recurrence-free survival. Combining PET/MRI biomarkers with clinical risk factors aided in formulating prognostic models, which demonstrated higher accuracy in predicting overall or recurrence-free survival over the traditional TNM staging system. ABSTRACT: To prospectively investigate the prognostic value of (18)F-FDG PET/MRI in patients with oropharyngeal or hypopharyngeal squamous cell carcinomas (OHSCC) treated by chemoradiotherapy. The study cohort consisted of patients with OHSCC who had undergone integrated PET/MRI prior to chemoradiotherapy or radiotherapy. Imaging parameters derived from intravoxel incoherent motion (IVIM), dynamic contrast-enhanced MRI (DCE-MRI), and (18)F-FDG PET were analyzed in relation to overall survival (OS) and recurrence-free survival (RFS). In multivariable analysis, T classification (p < 0.001), metabolic tumor volume (p = 0.013), and pseudo-diffusion coefficient (p = 0.008) were identified as independent risk factors for OS. The volume transfer rate constant (p = 0.015), initial area under the curve (p = 0.043), T classification (p = 0.018), and N classification (p = 0.018) were significant predictors for RFS. The Harrell’s c-indices of OS and RFS obtained from prognostic models incorporating clinical and PET/MRI predictors were significantly higher than those derived from the traditional TNM staging system (p = 0.001). The combination of clinical risk factors with functional parameters derived from IVIM and DCE-MRI plus metabolic PET parameters derived from (18)F-FDG PET in integrated PET/MRI outperformed the information provided by traditional TNM staging in predicting the survival of patients with OHSCC.