Compatibility of AC electromagnetic localization with brachytherapy seeds and surgical clips

Abstracts / Brachytherapy 6 (2007) 77e118 independent of source strength, considerable cost savings can be achieved by using high activity sources with a low density implant. Fewer sources may translate into fewer needles, decreased perineal trauma and oedema, and reduced time in the operating room, compounding the savings and potentially reducing the morbidity. This study compares post implant dosimetry between high density source implants (HDI) and low density source implants (LDI). Methods and Materials: Dosimetric analysis of the whole prostate (V200, V150, V100, D90, D80, contiguous V200 and V150, external index), prostate quadrants (V200, V150, V100, D90), rectum (V150, V120, V100, V80, V60) and deviated surrogate urethra (V200, V150, V120, V100, V80) was performed on 39 consecutive prostate brachytherapy LDI and 39 volume matched HDI over the same time period. The distinction between low source density and high source density was based on differing prescribed dose using 125-Iodine sources. All cases were contoured by two independent blinded observers. Repeated measures analysis of variance was used to look at the effects of treatment arm, observer and their interaction. Results: Whole prostate (WP) volume did not differ significantly between the treatment arms, mean of 25.4 cc for LDI and 26.6 cc for HDI. The median number of sources was 80 in the LDI group and 100 for HDI group. There was no significant difference in any of the measured postimplant dosimetric parameters for the WP or quadrants, surrogate urethra or rectum. Conclusions: A 20% decrease in source density, which is mathematically equivalent to a 20% increase in source strength, has no measurable effect on CT-based postimplant dosimetry. This study has shown that neither dose homogeneity nor conformality is compromised with a lower source density. Higher strength sources have the potential for considerable cost saving and reduced morbidity. PD-9 Presentation Time: 11:10 AM 3D prostate segmentation in TRUS images using image warping and ellipsoid fitting Sara Badiei, M.A.Sc. (Candidate)1 Septimiu E. Salcudean, Ph.D.1 Jim Varah, Ph.D.2 Nawaid Usmani, M.D.3 Nick Chng, Ph.D. (Candidate)3 Alexandra Kruk, R.T.T.3 Ingrid Spadinger, Ph.D.3 William J. Morris, M.D.3 1 Electrical Engineering, University of British Columbia (UBC); 2 Computer Science, UBC; 3Radiation Oncology, BC Cancer Agency, Vancouver, BC, Canada. Purpose: In this study we evaluate a 3D semi-automatic prostate segmentation algorithm using manual segmentation by 3 blinded observers as the gold standard. For a given prostate volume, our hypotheses are: 1) The repeatability of the semi-automatic segmentation is at least as good as the intra-observer variability in manual segmentation. 2) The difference between manual segmentation and semiautomatic segmentation will be less than or equal to inter-observer variability in manual segmentation. 3) If we apply the dosimetric plan that was designed for the manually segmented PTV volume to the semiautomatically segmented PTV volume, the resulting dose coverage and conformality will be as good as that of the manually segmented prostate volume for which the plan was designed. Methods and Materials: This 3D segmentation algorithm is the extension of a successful 2D segmentation algorithm in which we used image warping, Star Kalman edge detection and ellipse (ellipsoid in 3D) fitting to find the prostate boundary [1]. In 3D, this algorithm only requires 3 initialization points to segment the entire volume. Ten transrectal ultrasound (TRUS) prostate volume studies were chosen from the Brachytherapy database at the BC Cancer Agency and all identifying information was removed. Two trained non-expert and one expert observer manually segmented all 10 cases and then initialized all 10 cases (for the segmentation algorithm) independently. The cases were presented to each observer randomly to minimize subjective bias. Two weeks later, all observers manually segmented and initialized 5 of the original 10 cases to evaluate intraobserver variability. Results: The 3D segmentation results will be finalized and presented at the conference. Previously, the 2D segmentation algorithm was tested on 17 midgland TRUS images [1]. Distance-based metrics between manual and semi-automatic contours showed a mean absolute difference of

95

0.67  0.18 mm, which is significantly lower than inter-observer variability. Area-based metrics showed an average sensitivity greater than 97% and average accuracy greater than 93%. Even without code optimization, the 2D algorithm was almost two times faster than manual segmentation. Conclusions: The proposed 3D segmentation algorithm returns shapes that are easy to plan on and has potential for real-time dosimetry applications. As well the authors expect improved performance and repeatability over manual segmentation. [1] Badiei S, Salcudean SE, Varah J, Morris WJ. Prostate segmentation in 2D ultrasound images using image warping and ellipse fitting. MICCAI conference. Springer LNCS 2006;4191:17e24. PD-10 Presentation Time: 11:15 AM Compatibility of AC electromagnetic localization with brachytherapy seeds and surgical clips Robyn Vera, D.O.1 Peter D. Grimm, D.O.2 Tricia Zeller, R.N., M.H.A.2 Timothy Mate, M.D.2 Richard S. Eidens, Ph.D.3 1The University of Texas M.D. Anderson Cancer Center, Houston, TX; 2Seattle Prostate Institute/ Swedish Hospital Medical Center, Seattle, WA; 3Calypso Medical Technologies, Seattle, WA. Purpose: The CalypsoÒ 4D Localization System, used for continuous prostate tracking during EBRT, is based on AC electromagnetics and small implanted wireless transponders (BeaconÒ transponders). Large amounts of conductive metal nearby might affect localization performance. The effects of permanent brachytherapy implant seeds and surgical clips on the Calypso system were examined in a phantom. Methods and Materials: Utilizing a prostate phantom and a standard template, dummy RAPIDÔ strand (Oncura, Plymouth Meeting, PA) and free seeds were inserted to uniformly surround three transponders in a pattern that mimicked a very large prostate implant (150 seeds). The Calypso detection array was positioned 20 cm above the phantom. Transponder location data were continuously collected over an 8-minute period, a typical EBRT fraction time. Without disturbing the phantom or transponder positions, seeds were removed and transponder positions were again measured for 8 minutes as the control. Transponder positions with and without brachytherapy seeds were compared to assess any possible interference by the seeds. In a similar fashion, surgical clips were placed in the phantom and tested for the possible interaction with the transponder’s electromagnetic localization system. Data were collected and compared with and without the surgical clips. Results: Permanent brachytherapy seeds and surgical clips both had an insignificant effect on the ability of the electromagnetic localization system to detect the transponder position. Detected variance of transponder localization in the presence of either metal was within micrometers of the no-metal control. Conclusions: Data collected during this study verify permanent brachytherapy seeds or surgical clips do not appear to interfere with the Calypso target localization system. This study suggests a potential use of the Calypso technology to provide setup and continuous target localization during EBRT after prostate seed implantation or radical prostatectomy. Further clinical investigation is indicated. TZ, TM, and RSE have a financial interest in Calypso Medical Technologies. PD-11 Presentation Time: 11:20 AM Penile bulb dosimetrydDoes prostate brachytherapy technique matter? Jenny P. Nobes, Sara J. Khaksar, Abdul Ismail, Maria A. Hawkins, Stephen E.M. Langley, Robert W. Laing Oncology, St Luke’s Cancer Centre, Guildford, United Kingdom. Purpose: Erectile dysfunction (ED) following prostate brachytherapy (BXT) has been reported to be related to dose received by the penile bulb. To minimise penile bulb dose whilst preserving prostate dosimetry we have developed a modified technique for I125 seed BXT using both