Differentiating recurrent tumours from post-treatment changes in head and neck cancers: does diffusion-weighted MRI solve the eternal dilemma?

Abstract

To evaluate the utility of diffusion-weighted imaging (DWI) in differentiating post-treatment changes from tumour recurrence in head and neck cancers and to establish a threshold apparent diffusion coefficient (ADC) value to differentiate the two conditions. This was a prospective study of 80 treated head and neck cancer patients. The patient cohort consisted of a wide spectrum of head and neck sites, including the oral cavity, oropharynx, larynx, hypopharynx, paranasal sinuses, orbits, salivary glands, and infra-temporal fossa. Qualitative analysis of the diffusion images and quantitative analysis of the corresponding ADC maps was performed and the data were correlated with histopathological findings and clinical examinations. The mean ADC value of recurrent tumours in the present cohort was 0.932±0.19×10-3 mm2/s and the mean ADC value of lesions representing post-treatment changes was 1.394±0.32×10-3 mm2/s. A threshold ADC value of 1.2×10-3mm2/s used to differentiate post-treatment changes from recurrent head and neck cancers showed the highest combined sensitivity of 90.13%, specificity of 82.5%, accuracy of 86.4%, positive predictive value of 84.4%, negative predictive value of 88.9%, and mean kappa measurement of agreement of 72.8. Combined qualitative and quantitative analysis of DWI is a useful non-invasive technique to differentiate recurrent head and neck malignancies from post-treatment changes using a threshold ADC value.