Image Features of Head and Neck Squamous Cell Carcinoma on Hybrid FDG PET/MRI Regarding Human Papilloma Virus Infection Status

Objectives Human papilloma virus (HPV) infection is a significant causal and prognostic factor of head and neck squamous cell carcinoma (HNSCC). It has been reported that tissue characteristics are different by HPV infection status, particularly in metastatic lymph nodes. In this study, we investigated various image features of HNSCC on hybrid FDG PET/MRI, with regard to HPV infection.

Methods A total of 64 patients with HNSCC, who underwent FDG PET/MRI for pretreatment staging were retrospectively included. Tissues of primary tumors were tested for high-risk type HPV (HPV-HR) infection. For primary tumors and metastatic lymph nodes of the highest metabolism (if any), standardized uptake value (SUV), metabolic tumor volume (MTV), total lesion glycolysis (TLG), and apparent diffusion coefficient (ADC) were measured. Additionally, SUV-ADC product (SAP) in each voxel was calculated as a hybrid index for cellularity-adjusted metabolism. For heterogeneity, entropy and homogeneity were measured on FDG PET.

Results HPV-HR was positive in 31 patients (48.4%), and it was significantly related to poorer differentiation, larger size and more central necrosis of the metastatic nodes. However, there was no significant difference according to HPV-HR infection in age, sex, tumor size, node metastasis and stage. Image factors of maximal SUV, MTV, TLG, and SAP were significantly different in metastatic lymph nodes according to HPV-HR infection (P = 0.007, 0.010, 0.019, and 0.030, respectively), whereas they were not in primary tumors. Also, entropy and homogeneity were significantly different in metastatic lymph nodes (P = 0.015 and 0.011), whereas they were not in primary tumors.

Conclusions Various image features of metabolic activity were significantly different according to HPV-HR infection status in metastatic lymph nodes of HNSCC, whereas they were not in primary tumors. Further studies are warranted regarding pathophysiologic and prognostic significance of these image features in HNSCC.