Automatic tumour delineation in whole body PET/CT images

We propose a new method for automated delineation of tumour boundaries by using joint information from whole-body PET and diagnostic CT images. Due to varying levels of FDG uptake in different organs, we apply locally adaptive thresholds of SUV to acquire an initial estimate of hot spot locations and shape in PET images. The hot spot boundaries are further improved by applying Competition Diffusion (CD) and Mode-Seeking Region Growing (MSRG) algorithms. These hot spots seen in PET are then confirmed and more accurately segmented considering CT information, through the Joint Likelihood Ratio Test technique for probabilistic integration. Experiments show that the proposed multi-modal method achieves more accurate and reproducible tumour delineation than using PET or CT alone.

[1]  Umut Akdemir,et al.  Blob segmentation using joint space-intensity likelihood ratio test: application to 3D tumor segmentation , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).

[2]  Di Yan,et al.  Defining a radiotherapy target with positron emission tomography. , 2002, International journal of radiation oncology, biology, physics.

[3]  Johan Karlsson,et al.  Decision Support System for Lung Cancer using PET/CT Images , 2004 .

[4]  Lawrence R. Rabiner,et al.  A tutorial on hidden Markov models and selected applications in speech recognition , 1989, Proc. IEEE.

[5]  Matthew Turk,et al.  Automatic Hot Spot Detection and Segmentation in Whole Body FDG-PET Images , 2006, 2006 International Conference on Image Processing.

[6]  Dorin Comaniciu,et al.  Mean Shift: A Robust Approach Toward Feature Space Analysis , 2002, IEEE Trans. Pattern Anal. Mach. Intell..

[7]  Jean-François Daisne,et al.  Tumor volume in pharyngolaryngeal squamous cell carcinoma: comparison at CT, MR imaging, and FDG PET and validation with surgical specimen. , 2004, Radiology.

[8]  T. Kanuma,et al.  Diagnostic accuracy of FDG PET imaging for the detection of recurrent or metastatic gynecologic cancer , 2005, Annals of nuclear medicine.

[9]  Daniel P. Lopresti,et al.  Balancing the Role of Priors in Multi-Observer Segmentation Evaluation , 2009, J. Signal Process. Syst..

[10]  Michael Brady,et al.  A registration framework for the comparison of mammogram sequences , 2005, IEEE Transactions on Medical Imaging.

[11]  William M. Wells,et al.  Simultaneous truth and performance level estimation (STAPLE): an algorithm for the validation of image segmentation , 2004, IEEE Transactions on Medical Imaging.

[12]  Sung-Cheng Huang,et al.  Anatomy of SUV , 2000 .

[13]  Thomas Beyer,et al.  PET/CT scanners: a hardware approach to image fusion. , 2003, Seminars in nuclear medicine.

[14]  P. Grigsby,et al.  Measurement of tumor volume by PET to evaluate prognosis in patients with advanced cervical cancer treated by radiation therapy. , 2002, International journal of radiation oncology, biology, physics.

[15]  Michael Brady,et al.  Fusion of contrast-enhanced breast MR and mammographic imaging data , 2003, Medical Image Anal..

[16]  Toshiro Kubota,et al.  Reaction-diffusion systems for hypothesis propagation , 2000, Proceedings 15th International Conference on Pattern Recognition. ICPR-2000.

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