Automatic lesion detection and segmentation of 18F-FET PET in gliomas: A full 3D U-Net convolutional neural network study
Top Cited Papers
Open Access
- 12 April 2018
- journal article
- research article
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 13 (4), e0195798
- https://doi.org/10.1371/journal.pone.0195798
Abstract
Introduction Amino-acids positron emission tomography (PET) is increasingly used in the diagnostic workup of patients with gliomas, including differential diagnosis, evaluation of tumor extension, treatment planning and follow-up. Recently, progresses of computer vision and machine learning have been translated for medical imaging. Aim was to demonstrate the feasibility of an automated F-18-fluoro-ethyl-tyrosine (F-18-FET) PET lesion detection and segmentation relying on a full 3D U-Net Convolutional Neural Network (CNN). Methods All dynamic F-18-FET PET brain image volumes were temporally realigned to the first dynamic acquisition, coregistered and spatially normalized onto the Montreal Neurological Institute template. Ground truth segmentations were obtained using manual delineation and thresholding (1.3 x background). The volumetric CNN was implemented based on a modified Keras implementation of a U-Net library with 3 layers for the encoding and decoding paths. Dice similarity coefficient (DSC) was used as an accuracy measure of segmentation. Results Thirty-seven patients were included (26 [70%] in the training set and 11 [30%] in the validation set). All 11 lesions were accurately detected with no false positive, resulting in a sensitivity and a specificity for the detection at the tumor level of 100%. After 150 epochs, DSC reached 0.7924 in the training set and 0.7911 in the validation set. After morphological dilatation and fixed thresholding of the predicted U-Net mask a substantial improvement of the DSC to 0.8231 (+ 4.1%) was noted. At the voxel level, this segmentation led to a 0.88 sensitivity [95% CI, 87.1 to, 88.2%] a 0.99 specificity [99.9 to 99.9%], a 0.78 positive predictive value: [76.9 to 78.3%], and a 0.99 negative predictive value [99.9 to 99.9%]. Conclusions With relatively high performance, it was proposed the first full 3D automated procedure for segmentation of F-18-FET PET brain images of patients with different gliomas using a U-Net CNN architecture.Keywords
This publication has 25 references indexed in Scilit:
- Machine learning for neuroirnaging with scikit-learnFrontiers in Neuroscience, 2014
- CBTRUS Statistical Report: Primary Brain and Central Nervous System Tumors Diagnosed in the United States in 2005-2009Neuro-Oncology, 2012
- Prediction of oligodendroglial histology and LOH 1p/19q using dynamic [18F]FET-PET imaging in intracranial WHO grade II and III gliomasNeuro-Oncology, 2012
- Molecular imaging of gliomas with PET: Opportunities and limitationsNeuro-Oncology, 2011
- FET PET for the evaluation of untreated gliomas: correlation of FET uptake and uptake kinetics with tumour gradingEuropean Journal of Nuclear Medicine and Molecular Imaging, 2007
- Intraobserver and Interobserver Variability in GTV Delineation on FDG-PET-CT Images of Head and Neck CancersInternational Journal of Radiation Oncology*Biology*Physics, 2007
- EANM procedure guidelines for brain tumour imaging using labelled amino acid analoguesEuropean Journal of Nuclear Medicine and Molecular Imaging, 2006
- O-(2-[18F]fluoroethyl)-l-tyrosine: uptake mechanisms and clinical applicationsNuclear Medicine and Biology, 2006
- Value of O-(2-[18F]fluoroethyl)-l-tyrosine PET for the diagnosis of recurrent gliomaEuropean Journal of Nuclear Medicine and Molecular Imaging, 2004
- Contrast Limited Adaptive Histogram Equalization image processing to improve the detection of simulated spiculations in dense mammogramsJournal of Digital Imaging, 1998