Automatic 3D Detection and Segmentation of Head and Neck Cancer from MRI Data

Abstract

A novel algorithm for automatic head and neck 3D tumour segmentation from magnetic resonance imaging (MRI) is presented. The proposed algorithm pre-processes the MRI data slices to enhance quality and reduce artefacts. An intensity standardisation process is performed between slices, followed by cancer region segmentation of central slice, to get the correct intensity range and rough location of tumour regions. Fourier interpolation is applied to create isotropic 3D MRI volume. A new location-constrained 3D level set method segments the tumour from the interpolated MRI volume. The proposed algorithm is tested on real MRI data. The results show that the novel 3D tumour volume extraction algorithm has an improved dice score and F-measure when compared to the previous 2D and 3D segmentation method.

Magnetic Fields and Cancer: Epidemiology, Cellular Biology, and Theranostics

“…The use of PET and MRI, separately or combined, has been successful for assessing the metastatic lymph nodes in patients with head and neck cancer and offers advantages in staging with regard to increased anatomical details and radiation dose reduction [316][317][318]. Diffusion-weighted imaging (DWI) and intravenous (IV) contrast T1 dynamic perfusion imaging are a valid support for the functional MRI of tumors of the head and neck [319][320][321][322] and algorithms for automatic head and neck 3D tumor segmentation from MRI have been developed [323]. For head and neck cancer, MRIguided radiotherapy achieves clinical outcomes that are comparable to contemporary series reporting on intensity-modulated radiotherapy (IMRT) [324] and the use of targeted 3 T MRI was found to be useful for defining the presence and extent of large nerve perineural spread in head and neck cancers [325].…”

Segmentation of Head and Neck Tumours Using Modified U-net

“…A variety of algorithms have been proposed for head and neck cancer or tissue segmentation, such as atlas-based techniques [3], training-based approaches [4,5], and Deformable model [6][7][8]. However, these methods cannot efficiently solve the automatic segmentation challenge.…”