Apparent diffusion coefficients are closely related with high-risk human papilloma virus infection in cervical squamous cell carcinoma patients

Background Diffusion-weighted imaging (DWI) is a functional modality that indirectly represents the biological characteristics of cancer. However, the correlation of apparent diffusion coefficients (ADCs) to HPV type in patients with cervical cancer is still unknown. Purpose To explore the relationship between apparent diffusion coefficients calculated by DW-magnetic resonance imaging (MRI) and high-risk human papilloma virus (HR-HPV) infection in cervical squamous cell carcinoma (CSCC) patients. Material and Methods A total of 315 patients with histologically proven CSCC and who underwent pretreatment MRI were retrospectively analyzed. ADCs on the primary tumor were calculated. Receiver operating characteristic (ROC) curves were employed to determine the cut-off values of ADC(mean) and accuracy of prediction. Pearson correlation and multivariate logistic regression models were adopted to show the association of HR-HPV infection with clinical and imaging variables. Results ADC(mean) for HR-HPV+ CSCC was significantly lower than those for HR-HPV- CSCC. The area under the curve was 0.805 for predicting HR-HPV+ infection using mean ADCs. A cut-off value of 0.885 x 10(-3) mm(2)/s for mean ADCS could distinguish HR-HPV+ from HR-HPV- CSCC with 83.3% sensitivity, 73.9% specificity, and 80.8% accuracy. Regression analysis revealed that the HR-HPV infection was closely associated with tumor ADC value (odds ratio [OR] = 5.588, 95% confidence interval [CI] = 3.371-9.262), FIGO stage (OR = 2.789, 95% CI = 1.631-4.769), and histological grade (OR = 2.333, 95% CI = 1.676-3.247). Conclusion ADC value is related with the HR-HPV infection in patients with CSCC and DWI may be used for predicting HR-HPV infection in patients with CSCC.