Meeting Banner
Abstract #1052

Multi-Scale Analysis of Apparent Diffusion Coefficient (ADC) Predicts Cervical Nodal Status in Patients with Head & Neck Squamous Cell Carcinoma

Shonit Punwani1, Pierpaolo Purpura1, Nikolaos Dikaios1, Heather Fitzke1, Alan Bainbridge2, David Price2, Scott Rice1, Simon Morley3, Timothy Beale3, Ruheena Mendes4, Martin Forster4, Dawn Carnell4, Thayalini Vaitilingam3, Nina Newton1, David Atkinson1, Steve Halligan1, Stuart Taylor1

1Centre for Medical Imaging, University College London, London, UK, United Kingdom; 2Medical Physics and Bioengineering, University College London Hospital, London, United Kingdom; 3Radiology, University College London Hospital, London, United Kingdom; 4Head and Neck Oncology, University College London Hospital, London, United Kingdom

The study assess multi-scale diffusion parameters (median volumetric nodal region of interest values, inter-voxel histogram distributions, and intra-voxel diffusion heterogeneity as assessed by the stretched exponential model) as classifiers of nodal status in patients with head and neck squamous cell carcinoma (SCC). Low b value (0, 50, 100) derived nodal ADC (perfusion sensitive) was the key parameter facilitating discrimination of metastatic from benign nodes in patients with head and neck SCC. The stretched exponential derived α value together with histogram features of ADC provide an accurate decision tree model for classification of nodal disease.

Abstract PDF

Keywords

additional ages alone applied arch ares aspiration assess automatically available axial benign bioengineering carcinoma cause cell cervical chosen classification classifiers coils college components compute consecutive contra contrast corded criteria curve dawn decision derived diagram diffusion discriminate disease ears ease election electro entire etched ether experienced expo exponential features female fine functional gram greater head heather histogram hospital importance include inclusion initial inter intra kingdom kurtosis least list lost lymph magnet manufacturer maps march martin matched media median medical mens metastases metastatic model mono multivariate near neck needle nodal node nodes occurs odes oncology package pairs para parametric patient patients performance physics plotted population ported positively predict prediction price primary prior prism proposed radiology receive recovery recurrence rent rice satisfying scale selection sensitive signed skewed skewness slice slow software solid squamous staging status stretched subsequent supine therapy thresh timothy tissue toge tomography trace trans tree ultrasound unilateral volume volumes volumetric whether workstation