Encrypted login | home

Program Information

Segmentation of Prostate CBCT Images with Implanted Calypso Transponders Using Double Haar Wavelet Transform

no image available
Y Liu

Y Liu1*, Z Saleh2 , Y Song3 , C Obcemea3 , M Chan4 , X Li5 , L Happersett6 , C Shi7 , X Qian8 , X Tang2 , (1) Shandong Communication and Media College, Jinan, Shandong, (2) Memorial Sloan Kettering Cancer Center, West Harrison, NY, (3) Memorial Sloan-Kettering Cancer Center, Sleepy Hollow, NY, (4) Memorial Sloan-Kettering Cancer Center, Basking Ridge, NJ, (5) Memorial Sloan Kettering Cancer Center, Rockville Centre, NY, (6) Memorial Sloan Kettering Cancer Center, New York, NY, (7) Saint Vincent Medical Center, Bridgeport, CT, (8) North Shore Long Island Jewish health System, North New Hyde Park, NY,


SU-F-J-27 (Sunday, July 31, 2016) 3:00 PM - 6:00 PM Room: Exhibit Hall

Purpose: Segmentation of prostate CBCT images is an essential step towards real-time adaptive radiotherapy. It is challenging For Calypso patients, as more artifacts are generated by the beacon transponders. We herein propose a novel wavelet-based segmentation algorithm for rectum, bladder, and prostate of CBCT images with implanted Calypso transponders.

Methods: Five hypofractionated prostate patients with daily CBCT were studied. Each patient had 3 Calypso transponder beacons implanted, and the patients were setup and treated with Calypso tracking system. Two sets of CBCT images from each patient were studied. The structures (i.e. rectum, bladder, and prostate) were contoured by a trained expert, and these served as ground truth. For a given CBCT, the moving window-based Double Haar transformation is applied first to obtain the wavelet coefficients. Based on a user defined point in the object of interest, a cluster algorithm based adaptive thresholding is applied to the low frequency components of the wavelet coefficients, and a Lee filter theory based adaptive thresholding is applied to the high frequency components. For the next step, the wavelet reconstruction is applied to the thresholded wavelet coefficients. A binary/segmented image of the object of interest is therefore obtained. DICE, sensitivity, inclusiveness and ΔV were used to evaluate the segmentation result.

Results: Considering all patients, the bladder has the DICE, sensitivity, inclusiveness, and ΔV ranges of [0.81-0.95], [0.76-0.99], [0.83-0.94], [0.02-0.21]. For prostate, the ranges are [0.77-0.93], [0.84-0.97], [0.68-0.92], [0.1-0.46]. For rectum, the ranges are [0.72-0.93], [0.57-0.99], [0.73-0.98], [0.03-0.42].

Conclusion: The proposed algorithm appeared effective segmenting prostate CBCT images with the present of the Calypso artifacts. However, it is not robust in two scenarios: 1) rectum with significant amount of gas; 2) prostate with very low contrast. Model based algorithm might improve the segmentation in these two scenarios.

Contact Email: