Search

Purpose Postimplant analysis in permanent breast seed implant (PBSI) is performed at inconsistent times subsequent to seed implantation across cancer centers, creating challenges in the interpretation of dosimetric data and ultimately the correlation with clinical outcomes. The purpose of this study is to determine the most appropriate time postimplant to perform this analysis. Methods and Materials Nine patients treated at our institution with PBSI were included in this analysis. Each underwent 4 postimplant CT scans: 0, 15, 30, and 60 days postimplant. A model of the accumulated dose was created by deformably registering the Day 15, 30, and 60 postimplant CT scans and dose matrices to the Day 0 scan, scaling for seed decay. The results from this model were compared to each individual postplan by integral comparison of dose–volume histogram curves for a dose evaluation volume. Results The Day 30 postplan showed the best agreement with the accumulated dose model and the smallest interpatient variability across the patient cohort. The mean (±SD) for the dose evaluation volume V 90 , V 100 , V 150 , and V 200 for the accumulated dose model was 90 ± 7%, 86 ± 8%, 66 ± 14%, and 41 ± 16%, respectively. Conclusions Based on the results of this patient cohort, we recommend that postimplant dosimetric analysis for PBSI be performed approximately 30 days following the implant. [ABSTRACT FROM AUTHOR]

Purpose Permanent breast seed implant (PBSI) is a developing brachytherapy technique for the treatment of early-stage breast cancer. In current practice, PBSI uses manual planning strategies to generate clinical treatment plans. In this work, a simulated annealing-based algorithm is developed to demonstrate the first application of inverse optimization for PBSI. Methods and Materials Target, skin, and chest wall muscle contours, exported from a treatment planning system in digital imaging and communications in medicine format, are used as inputs. To optimize, the user defines the dose–volume histogram objectives for the target and specifies a relative weighting for target and skin constraints. A 10-patient cohort of previously treated patients was planned by using the inverse optimization algorithm. Plan quality was compared to the clinically treated manually generated plans using the V 90% , V 100% , V 150% , and V 200% for the planning target volume (PTV), V 90% and D 0.2 cc for skin dose, and PTV conformity indices. Results For each of the 10 patients, patient-wise paired differences between inverse and manual plans were analyzed and presented in box plots. Comparing inverse and manual planning techniques, a statistical difference was not observed ( p > 0.05) in PTV coverage criteria ( V 90% , V 100% ) and dose to skin 2mm . A statistical difference was observed in the inverse plans as a reduction of the V 150% (mean of 6.2%) and increase in conformity index of the 20%, 50%, 90%, and 100% isodose lines. Conclusions This work presents the first application of inverse optimization used to generate PBSI treatment plans. A 10-patient cohort previously treated with PBSI was retrospectively planned for comparison with the clinically treated manually generated plans. [ABSTRACT FROM AUTHOR]

Purpose To develop an anthropomorphic breast phantom for use in credentialing of permanent breast seed implant brachytherapy. Methods and Materials A representative external contour and target volume was used as the basis of mold manufacturing for anthropomorphic breast phantom development. Both target and normal tissue were composed of gel-like materials that provide suitable computed tomography and ultrasound contrast for brachytherapy delivery. The phantoms were evaluated for consistency in construction (target location) and Hounsfield unit (computed tomography contrast). For both target and normal tissue, the speed of sound was measured and compared to the image reconstruction algorithm's expectation value. Five phantoms were imaged preimplant and postimplant to assess interphantom similarity as well as to evaluate the uncertainty in quantifying seed position. Results The average Hounsfield units of the target and normal tissue gels is −146 ± 5 and 23 ± 1, respectively. The average speed of sound of the target and normal tissue gels is 1485 ± 7 m/s and 1558 ± 9 m/s, respectively, resulting in an estimated 0.4 mm uncertainty in image guidance. The registration/deformation uncertainty was determined to be 0.8 mm. The standard combined uncertainty in assessing seed position spatial accuracy, also including a 0.9 mm estimate based on literature for seed localization, is estimated to be 1.3 mm. Conclusions The development of the anthropomorphic breast phantom and evaluation of both the consistency as well as overall seed position uncertainty illustrates the suitability of this phantom for use in brachytherapy end-to-end delivery and implant accuracy evaluation. When evaluating a user's implant accuracy, we estimate a standard combined uncertainty of 1.3 mm. [ABSTRACT FROM AUTHOR]