TY - GEN
T1 - Quantification of edematic effects in prostate brachytherapy interventions
AU - Hefny, Mohamed
AU - Abolmaesumi, Purang
AU - Karimaghaloo, Zahra
AU - Gobbi, David G.
AU - Ellis, Randy
AU - Fichtinger, Gabor
PY - 2008
Y1 - 2008
N2 - We present a quantitative model to analyze the detrimental effects of edema on the quality of prostate brachytherapy implants. We account for both tissue expansion and implant migration by mapping intra-operative ultrasound and post-implant CT. We pre-process the ultrasound with a phase congruency filter, and map it to the volume CT using a B-spline deformable mutual information similarity metric. To test the method, we implanted a standard training phantom with 48 seeds, imaged the phantom with ultrasound and CT and registered the two for ground truth. Edema was simulated by distorting the CT volume by known transformations. The objective was to match the distorted implant to the intra-operative ultrasound. Performance was measured relative to ground truth. We successfully mapped 100% of deformed seeds to ground truth under edematic expansion up to 40% of volume growth. Seed matching performance was 98% with random seed migration of 3mm superimposed on 10% edematic volume growth. This method promises to be clinically applicable as the first quantitative analysis tool to measure edematic implant deformations occurring between the operating room and post-operative CT imaging.
AB - We present a quantitative model to analyze the detrimental effects of edema on the quality of prostate brachytherapy implants. We account for both tissue expansion and implant migration by mapping intra-operative ultrasound and post-implant CT. We pre-process the ultrasound with a phase congruency filter, and map it to the volume CT using a B-spline deformable mutual information similarity metric. To test the method, we implanted a standard training phantom with 48 seeds, imaged the phantom with ultrasound and CT and registered the two for ground truth. Edema was simulated by distorting the CT volume by known transformations. The objective was to match the distorted implant to the intra-operative ultrasound. Performance was measured relative to ground truth. We successfully mapped 100% of deformed seeds to ground truth under edematic expansion up to 40% of volume growth. Seed matching performance was 98% with random seed migration of 3mm superimposed on 10% edematic volume growth. This method promises to be clinically applicable as the first quantitative analysis tool to measure edematic implant deformations occurring between the operating room and post-operative CT imaging.
UR - http://www.scopus.com/inward/record.url?scp=58949084219&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=58949084219&partnerID=8YFLogxK
U2 - 10.1007/978-3-540-85990-1-59
DO - 10.1007/978-3-540-85990-1-59
M3 - Conference contribution
AN - SCOPUS:58949084219
SN - 3540859896
SN - 9783540859895
VL - 5242 LNCS
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 493
EP - 500
BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
T2 - 11th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2008
Y2 - 6 September 2008 through 10 September 2008
ER -