PET-CT images co-segmentation of lung tumor using joint level set model

Medical imaging, used for both diagnosis and therapy planning, is evolving towards multi-modality acquisition protocols. Manual segmentation of 3D images is a tedious task and prone to inter- and inter-experts variability. Moreover, the automatic segmentation exploiting the characteristics of multi-modal images is still a difficult problem. Towards this end, Positron emission tomography (PET) and computed tomography (CT) are widely used. PET imaging has a high contrast but often leads to blurry tumor edges due to its limited spatial resolution, while CT imaging has a high resolution but a low contrast between a tumor and its surrounding normal soft tissues. Tumor segmentation from either a single PET or CT image is difficult. It is known that co-segmentation methods utilizing the complementary information between PET and CT can improve the segmentation accuracy. This complementary information can be either consistent or inconsistent in the image level. How to correctly localize tumor edges with the inconsistent information is one major challenge for co-segmentation. Aiming to solve this problem, a novel joint level set model is proposed to combine the evidences of PET and CT in a united energy form, achieving a co-segmentation in these two modalities. The convergence of the co- segmentation model corresponds to the most optimal tradeoff between the PET and CT. The different characteristics in these two imaging modalities are considered in the adaptive convergence process which starts mostly with the PET evidence to constrain the tumor location and stops mostly with the CT evidences to delineate boundary details. The adaptability of our proposed model is automatically realized by stepwise moderating the joint weights during the convergence process. The performance of the proposed model is validated on 20 nonsmall cell lung tumor PET-CT images. It achieves an average dice similarity coefficient (DSC) of 0.846±0.064 and positive predictive value (PPV) of 0.889±0.079, demonstrating the high accuracy of the proposed model for PET-CT images lung tumor co-segmentation.

[1]  Yong Yin,et al.  Topology polymorphism graph for lung tumor segmentation in PET-CT images , 2015, Physics in medicine and biology.

[2]  V. Grégoire,et al.  Gradient-based delineation of the primary GTV on FDG-PET in non-small cell lung cancer: a comparison with threshold-based approaches, CT and surgical specimens. , 2011, Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology.

[3]  Xiang Li,et al.  Deep Learning-Based Image Segmentation on Multimodal Medical Imaging , 2019, IEEE Transactions on Radiation and Plasma Medical Sciences.

[4]  Xinjian Chen,et al.  Random Walk and Graph Cut for Co-Segmentation of Lung Tumor on PET-CT Images , 2015, IEEE Transactions on Image Processing.

[5]  Anne Bol,et al.  A gradient-based method for segmenting FDG-PET images: methodology and validation , 2007, European Journal of Nuclear Medicine and Molecular Imaging.

[6]  Wei Lu,et al.  Tumor co-segmentation in PET/CT using multi-modality fully convolutional neural network , 2018, Physics in medicine and biology.

[7]  Linfang Xiao,et al.  Active contours driven by region-scalable fitting and optimized Laplacian of Gaussian energy for image segmentation , 2017, Signal Process..

[8]  Xinjian Chen,et al.  Automatic Segmentation and Quantification of White and Brown Adipose Tissues from PET/CT Scans , 2017, IEEE Transactions on Medical Imaging.

[9]  Isabelle Bloch,et al.  Joint variational segmentation of CT-PET data for tumoral lesions , 2010, 2010 IEEE International Symposium on Biomedical Imaging: From Nano to Macro.

[10]  Sasa Mutic,et al.  Concurrent multimodality image segmentation by active contours for radiotherapy treatment planning. , 2007, Medical physics.

[11]  Irène Buvat,et al.  Variational Segmentation of Vector-Valued Images With Gradient Vector Flow , 2014, IEEE Transactions on Image Processing.

[12]  Chunming Li,et al.  Implicit Active Contours Driven by Local Binary Fitting Energy , 2007, 2007 IEEE Conference on Computer Vision and Pattern Recognition.

[13]  Junjie Bai,et al.  Optimal Co-Segmentation of Tumor in PET-CT Images With Context Information , 2013, IEEE Transactions on Medical Imaging.

[14]  Hong-Bin Shen,et al.  Saliency driven region-edge-based top down level set evolution reveals the asynchronous focus in image segmentation , 2018, Pattern Recognit..

[15]  Zhong Yang,et al.  A New Iterative Triclass Thresholding Technique in Image Segmentation , 2014, IEEE Transactions on Image Processing.

[16]  Yu Zhao,et al.  Automated Whole-Body Bone Lesion Detection for Multiple Myeloma on 68Ga-Pentixafor PET/CT Imaging Using Deep Learning Methods , 2018, Contrast media & molecular imaging.

[17]  Xiaodong Wu,et al.  3D Alpha Matting Based Co-segmentation of Tumors on PET-CT Images , 2017, CMMI/RAMBO/SWITCH@MICCAI.

[18]  Yihua Tan,et al.  Variational PET/CT Tumor Co-Segmentation Integrated With PET Restoration , 2020, IEEE Transactions on Radiation and Plasma Medical Sciences.

[19]  I. Buvat,et al.  Comparative Assessment of Methods for Estimating Tumor Volume and Standardized Uptake Value in 18F-FDG PET , 2010, Journal of Nuclear Medicine.

[20]  Jérôme Lapuyade-Lahorgue,et al.  SPEQTACLE: An automated generalized fuzzy C-means algorithm for tumor delineation in PET. , 2015, Medical physics.

[21]  Xiaolei Huang,et al.  Automated tumor delineation using joint PET/CT information , 2007, SPIE Medical Imaging.

[22]  P Vera,et al.  Development of a generic thresholding algorithm for the delineation of 18FDG-PET-positive tissue: application to the comparison of three thresholding models , 2009, Physics in medicine and biology.

[23]  Tony F. Chan,et al.  Active contours without edges , 2001, IEEE Trans. Image Process..

[24]  Yong Yin,et al.  Lung Tumor Delineation Based on Novel Tumor-Background Likelihood Models in PET-CT Images , 2014, IEEE Transactions on Nuclear Science.