D. Song

WE‐C‐330A‐03: Seed Segmentation in C‐Arm Fluoroscopy for Brachytherapy Implant Reconstruction

Purpose: Intra‐operative dosimetry in prostate brachytherapy critically depends on discerning the 3‐D locations of implanted seeds. The accuracy of 3‐D seed reconstruction step is, in turn, limited by the accuracy with which the position and orientation of individual implanted seed in the fluoroscopic images can be found. A method for robustly segmenting the seeds in fluoroscopic images is proposed here. Methods and Materials: The process of determining the locations and orientations of implanted seeds is sub‐divided into three main steps. In the first step, the image is segmented by shape‐size based morphological approach to eliminate background noise and do away with non‐uniform brightness of the image, to get seed‐like regions. These regions are either single seeds or overlapping multiple seed clusters. In the second step, the regions are analyzed and classified definitively, in a two‐phase statistical process coupled with information extraction from original intensity image, into two classes: single seed and overlapping multiple seed cluster. In the third step, the region belonging to overlapping multiple seed cluster is resolved into its constituent individual seeds through a simple and novel technique. Results: The proposed algorithm was tested on a set of ten clinical fluoroscopic images. The algorithm correctly determines the seeds with overall average of 99.57%. The clusters are not correctly resolved only in two images (2 clusters each, 1.7% and 1.6% of total seeds in respective implants). One false positive (noise labeled as seed) each is reported in two images, both the cases being where the tip of catheter appears to be of the size and shape of seed. Conclusions: The algorithm builds on an existing framework of morphological processing and provides further improvements in classification and cluster resolution. The algorithm appears to be robust and accurate despite the poor resolution of clinical images.

TU‐EE‐A3‐03: Automated Segmentation of Radiographic Fiducials for C‐Arm Tracking

Purpose: Intraoperative quantitative C‐arm fluoroscopy guidance depends on discerning the relative pose of images (pose recovery). A possible method is to use radiographic fiducials visible in fluoro images [1,2]. We propose a robust and fast method for segmenting fiducials designed for brachytherapy applications. Methods and materials: The fiducial contains points, lines and ellipses made from BBs and wires[1]. The algorithm integrates the a‐priori knowledge of fiducials mechanical construction in a cleverly devised workflow. The BB segmentation is achieved using morphological top‐hat transform. This information serves as a heuristic input to line segmentation realized by a curve tracing algorithm which operates on edge image, followed by augmenting information from intensity image. Once the lines are segmented, this information feeds to the ellipse extraction step. For ellipse segmentation, intensity image is morphologically processed to eliminate background noise, followed by elimination of BB‐s and lines from the information obtained in prior steps. The resulting image consists of only ellipse segments. A fast variation of Hough transform is used to rectify the full ellipse from the segments. Results: The fiducial algorithm identified all the features (BBs, lines and ellipses) visible to human eye in all ten clinical images. Next the accuracy of fiducial segmentation was assessed numerically by feeding the results to the pose recovery algorithm of [1]. The fiducial was moved on an accurate mechanical platform (as ground truth) while the C‐arm was stationary. We reconstructed the relative poses with an accuracy of 1.2 mm in translation and 0.3 degrees in rotational based on the segmented fiducials. Conclusions: The algorithm makes effective use of a‐priori knowledge and combines the techniques of morphological segmentation, curve tracing, and Hough transform, resulting in a novel curve segmentation strategy.

WE-AB-BRA-12: Post-Implant Dosimetry in Prostate Brachytherapy by X-Ray and MRI Fusion

PURPOSE For post-implant dosimetric assessment after prostate brachytherapy, CT-MR fusion approach has been advocated due to the superior accuracy on both seeds localization and soft tissue delineation. However, CT deposits additional radiation to the patient, and seed identification in CT requires manual review and correction. In this study, we propose an accurate, low-dose, and cost-effective post-implant dosimetry approach based on X-ray and MRI. METHODS Implanted seeds are reconstructed using only three X-ray fluoroscopy images by solving a combinatorial optimization problem. The reconstructed seeds are then registered to MR images using an intensity-based points-to-volume registration. MR images are first pre-processed by geometric and Gaussian filtering, yielding smooth candidate seed-only images. To accommodate potential soft tissue deformation, our registration is performed in two steps, an initial affine followed by local deformable registrations. An evolutionary optimizer in conjunction with a points-to-volume similarity metric is used for the affine registration. Local prostate deformation and seed migration are then adjusted by the deformable registration step with external and internal force constraints. RESULTS We tested our algorithm on twenty patient data sets. For quantitative evaluation, we obtained ground truth seed positions by fusing the post-implant CT-MR images. Seeds were semi-automatically extracted from CT and manually corrected and then registered to the MR images. Target registration error (TRE) was computed by measuring the Euclidean distances from the ground truth to the closest registered X-ray seeds. The overall TREs (mean±standard deviation in mm) are 1.6±1.1 (affine) and 1.3±0.8 (affine+deformable). The overall computation takes less than 1 minute. CONCLUSION It has been reported that the CT-based seed localization error is ∼1.6mm and the seed localization uncertainty of 2mm results in less than 5% deviation of prostate D90. The average error of 1.3mm with our system outperforms the CT-based approach and is considered well within the clinically acceptable limit. Supported in part by NIH/NCI grant 5R01CA151395. The X-ray-based implant reconstruction method (US patent No. 8,233,686) was licensed to Acoustic MedSystems Inc.

SU-D-BRF-07: Ultrasound and Fluoroscopy Based Intraoperative Image-Guidance System for Dynamic Dosimetry in Prostate Brachytherapy

PURPOSE Prostate brachytherapy is a common treatment method for low-risk prostate cancer patients. Intraoperative treatment planning is known to improve the treatment procedure and the outcome. The current limitation of intraoperative treatment planning is the inability to localize the seeds in relation to the prostate. We developed an image-guidance system to fulfill this need to achieve intraoperative dynamic dosimetry in prostate brachytherapy. METHODS Our system is based on standard imaging equipments available in the operating room, including the transrectal ultrasound (TRUS) and the mobile C-arm. A simple fiducial is added to compute the C-arm pose. Three fluoroscopic images and an ultrasound volume of the seeds and the prostate are acquired and processed by four image processing algorithms: seed segmentation, fiducial detection with pose estimation, seed reconstruction, and seeds-to-TRUS registration. The updated seed positions allow the physician to assess the quality of implantation and dynamically adjust the treatment plan during the course of surgery to achieve improved exit dosimetry. RESULTS The system was tested on 10 phantoms and 37 patients. Seed segmentation resulted in a 1% false negative and 2% false positive rates. Fiducial detection with pose estimation resulted in a detection rate of 98%. Seed reconstruction had a mean reconstruction error of 0.4 mm. Seeds-to-TRUS registration had a mean registration error of 1.3 mm. The total processing time from image acquisition to registration was approximately 1 minute. CONCLUSION We present an image-guidance system for intraoperative dynamic dosimetry in prostate brachytherapy. Using standard imaging equipments and a simple fiducial, our system can be easily adopted in any clinics. Robust image processing algorithms enable accurate and fast computation of the delivered dose. Especially, the system enables detection of possible hot/cold spots during the surgery, allowing the physician to address these in the operating room rather than requiring additional therapy afterwards. This research was supported in part by the National Institutes of Health/National Cancer Institute (NIH/NCI) under grant 2R44CA099374 and grant 1R01CA151395, and in part by the Department of Defense (DoD) under grant W81XWH-05-1-0407.

SU‐C‐WAB‐06: Deformable Registration of Post‐Implant MRI to Intra‐Operative Ultrasound Images for Permanent Prostate Brachytherapy Treatment Assessment

PURPOSE Transrectal-ultrasound (TRUS) is the most common image modality used in permanent prostate brachytherapy (PPI), while MRI images can provide additional anatomical information. Our goals were to develop a novel method to register post-implant MRI to intra-operative ultrasound (US) images and demonstrate its potential usage for retrospective implant assessment. METHODS TRUS images of prostate and non-isocentric C-arm fluoroscopy (FL) images are captured intraoperatively right after implant. The reconstructed 3D seed cloud from FL images (seeds_FL) will be used as a bridge to register post MRI to US images. The Registration of Ultrasound and Fluoroscopy (RUF) images is done by an intensity-based point to volume algorithm. The day-one post CT and T2-MRI images are co-registered and the 3D seed cloud is segmented from CT (seeds_CT). The iterative-closest-point algorithm is used to compute rigid transformation between seeds_FL and seed_CT. MRI can be transferred to FL coordinate using registration of two seed clouds and then to US coordinate using RUF registration. A thin-plate-spline algorithm is used to deform the MRI contours and images according to deformation of two seed clouds. To demonstrate this registration method, post MRI images from ten patients were registered to US images. The prostate contours were compared between post US and MRI. The planning needle interferences with critical structure like neurovascular bundles (NVB) were investigated. RESULTS After registration, center of mass of both prostate and urethra were within 3mm between two contour sets. The anterior boundaries of prostate were often overestimated in post US contours. When planning needles were superimposed over contours, there were about 3-4 needles passing through or in close vicinity of NVB in all patients. CONCLUSION A novel method to register post-implant MRI to intra-operative US images was developed and demonstrated in evaluating accuracy of intra-operative US contours and accessing needle passages to NVB during PPI.

SU‐E‐T‐324: Evaluation of Prostate Volume and Shape Change after Permanent Prostate Brachytherapy Using Implanted Seed Displacement Analysis

Purpose: To investigate a novel method of characterizing prostate volume and shape changes after permanent prostate brachytherapy (PPB) by using the implanted seeds as a surrogate. Methods: In twenty‐one patients undergoing PPB, multiple 2D C‐arm based fluoroscopy (CF) were taken intraoperatively immediately after seed implantation (day 0) and CT scan was performed the next day (day 1). The 3D coordinates of seed locations for each patient were reconstructed from both CF and CT, respectively. The volume of seed cloud was calculated as the volume of its convex hull. The CT and CF seed clouds were registered using iterative closest point (ICP) algorithm. The boundary seed shifts in each dimension between CF and CT seeds were calculated and students t‐test was used to test if the seeds shifts were significantly different from zero. Results: Twelve patients show average of 7%±5% volume increase between day one CT seeds and day 0 CF seeds and nine patients show 9%±5% volume decrease. A positive correlation is found between seed‐calculated volume and intraoperative preplanning ultrasound volume. For patients with volume increase, there is significant expansion in both z dimension (SI) with p<0.001. For patients with volume decrease, there is shrinkage in both x dimension (LR) with p<0.05. An average 23±8 degree pitch angle around x (LR) axis was observed between two seed clouds, corresponding to the change of patient position from lithotomy position during the implant to supine position during CT.Conclusions: A method to use implanted seeds to determine prostate volume changes is demonstrated. The prostate volume can increase or decrease between end of implantation and day one CT, indicating edema either reaches its maximum intraoperatively, or still continues development after implantation. The ability to characterize immediate postoperative prostate volume and shape changes allows for correlation of intraoperative dynamic dosimetry to post‐implant CTdosimetry.

SU‐GG‐T‐254: A Quantitative Assessment of Safety Measures in a Radiation Oncology Clinic

Purpose: A key approach to improving safety performance in radiationoncology is the use of prospective safety measures. Here we quantify the statistical patterns of one such tool, failure mode and effects analysis (FMEA), and compare it to reported process deviation events from an in‐house reporting system. Materials and Methods: FMEA provides a methodology for prospectively identifying potential failure modes and ranking them by importance. Our analysis consisted of individual blinded scoring of each failure mode by approximately 10 staff members. The results were compared to a group consensus on a subset of failure modes, and also to data from our department deviation reporting system acquired over a three‐month time period just after the completion of the FMEA. Results: 159 different failure modes were identified. There was a high level of variability of FMEA scoring between people, with standard deviations of 76% and 34% for risk priority number and severity, respectively. Severity scores were well‐correlated with RPN scores (Pearson r=0.50). Severity and RPN scores from therapists were significantly lower than the scores from physicists and treatment planners (p<0.001). The mean severity scores for individuals were significantly different than the group consensus scores. Despite the large number of failure modes identified in the prospective analysis, only 10 of 24 (42%) of actual reported deviation events were identified in the FMEA. Conclusions: FMEA is a widely‐used tool for prospective safety improvement but is subject to significant variability in scoring and may not uncover some errors actual occurring in the clinic. Severity‐only scoring may provide a simplified alternative to FMEA, and the addition of an error reporting system may prove valuable.

SU‐GG‐T‐227: A Systematic Analysis of the External Beam Radiotherapy Process for Patient Safety

Purpose: We have applied failure mode and event analysis (FMEA) to review our external beam radiation therapy(EBRT) process with the intent of identifying possible failure modes and increasing the overall safety of the EBRT process. Method and Materials: Our review was initiated in December 2006 and included 12 members of the staff assisted by experts from Center for Innovation in Quality Patient Care (CIQPC) at our institution. Meeting weekly, the task group developed a visual map of the EBRT process, completed March, 2007. The group then performed a FMEA to identify those nodes or steps in the process that were most likely to exhibit a failure mode that would potentially harm a patient. Each failure mode was given a rank priority number (RPN) based upon tabulated survey scores for the severity, frequency, and detectibility of that failure. Results: We considered the top 15 failure modes with RPN scores between 75 and 160. The group then brainstormed solutions for the failure modes using accepted mistake‐proofing methods. Each solution was then scored for effectiveness (1 to 10) and feasibility (1 to 10). Solutions with both effectiveness and feasibility greater than 5 (27 total solutions) were considered for implementation into the clinic. We continue to evaluate the impact and effectiveness of these process modifications. Conclusion: FMEA is a widely‐used tool for improving safety and reliability and can be easily adapted for use in radiation oncology.

SU‐FF‐J‐62: Evolution of Tumor Volume and Motion in Non‐Small Cell Lung Cancer During Radiotherapy

Purpose: To assess changes in tumor volumes and motion trajectories of non‐small cell lungcancer patients over the course of radiotherapy.Materials and Methods: We acquired repeat CT scans at three time points through the treatment course. Scans were acquired on a Philips large‐bore 16‐slice scanner using either a respiration‐correlated 4D‐CT protocol or a breath‐hold protocol at moderate‐deep inspiration enforced with the Active Breathing Coordinator system. We utilized the Pinnacle treatment planning system to co‐register based on vertebral bodies. We contoured the lesions in each data set and measured lesion volumes using model‐based segmentation tools. Windowing parameters were kept constant for all scans. Using the contours at each respiratory phase we measured the excursion of the lesion and changes of the excursions from scan to scan. Results:Tumor sizes decreased through treatment by 5%, 46% and 48% in three patients analyzed here. The change in the average GTV excursions was (mean ± s.d. over patients): 3.2±4.3 mm (A/P), 0.4±0.6 mm (R/L) and −0.1±3.1 mm (S/I). The 3D vector excursion increased by 2.9±4.2 mm on average. The changes in motion extent are similar to the motion excursion themselves, and there appears to be a strong variability between patients. Conclusions: These preliminary data indicate noteworthy trends of tumor size and motion over the course of radiation therapy. The tumor volume decreases and there is indication that the tumor excursion increases. Further analysis is underway and will be presented. Such long term evolution has important implications for the design and delivery of radiotherapy to lungtumors.