Michelle Hilts | University of British Columbia (original) (raw)
Papers by Michelle Hilts
Purpose: To examine variability in target volume delineation for partial breast radiotherapy plan... more Purpose: To examine variability in target volume delineation for partial breast radiotherapy planning and evaluate characteristics associated with low interobserver concordance. Methods and Materials: Thirty patients who underwent planning CT for adjuvant breast radiotherapy formed the study cohort. Using a standardized scale to score seroma clarity and consensus contouring guidelines, three radiation oncologists independently graded seroma clarity and delineated seroma volumes for each case. Seroma geometric center coordinates, maximum diameters in three axes, and volumes were recorded. Conformity index (CI), the ratio of overlapping volume and encompassing delineated volume, was calculated for each case. Cases with CI #0.50 were analyzed to identify features associated with low concordance. Results: The median time from surgery to CT was 42.5 days. For geometric center coordinates, variations from the mean were 0.5-1.1 mm and standard deviations (SDs) were 0.5-1.8 mm. For maximum seroma dimensions, variations from the mean and SDs were predominantly <5 mm, with the largest SDs observed in the medial-lateral axis. The mean CI was 0.61 (range, 0.27-0.84). Five cases had CI #0.50. Conformity index was significantly associated with seroma clarity ( p < 0.001) and seroma volume ( p < 0.002). Features associated with reduced concordance included tissue stranding from the surgical cavity, proximity to muscle, dense breast parenchyma, and benign calcifications that may be mistaken for surgical clips. Conclusion: Variability in seroma contouring occurred in three dimensions, with the largest variations in the medial-lateral axis. Awareness of clinical features associated with reduced concordance may be applied toward training staff and refining contouring guidelines for partial breast radiotherapy trials. Ó 2007 Elsevier Inc.
Radiotherapy and Oncology, 2005
To demonstrate the benefit of deformable image registration (DIR) to define postimplant seromas i... more To demonstrate the benefit of deformable image registration (DIR) to define postimplant seromas in permanent breast seed implants (PBSIs) by adapting preimplant seromas to postimplant images. Planning and Day 0 CT scans from 15 consecutive PBSI patients were included. Seromas were contoured on the preimplant CTs. After implant, DIR (MIM Maestro; MIM Software, Inc., Cleveland, OH) was used to adapt preimplant seromas to postimplant scans. Resulting seromas were evaluated for accuracy and adjusted if required, and pre- and postimplant seroma volumes were recorded and compared. DIR was validated by assessing the spatial correspondence of surgical clips pre- and postimplant in eight cases. Retrospectively, rigid registration was also investigated. Surgical clip location corresponded between pre- and postimplant images with a median agreement of &amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;lt;1.0 mm. The seroma defined using DIR was an acceptable rendering of the postimplant seroma in 9 of 15 cases, and in the remaining six cases, minor alterations (mean, 1.0 ± 0.2 mm) were made. Mean change in seroma volumes between preimplant and postimplant imaging was -6.1 ± 26.4%. Rigid registration exhibited unacceptable variability, with poor concordance in postimplant seroma between different observers (median conformity index, 26.5%; range, 0-52.7%). Deformable image registration can be used to adapt the preimplant seroma to postimplant PBSI images and is recommended for implant quality assurance in PBSI. Using deformable registration avoids uncertainties associated with poor seroma visualization and interobserver variability when rigid registration is performed. The accurate deformation of implanted surgical clips demonstrates the reliability of deformable image registration for PBSI.
Purpose: To investigate postimplant dosimetry for different regions of the prostate gland in pati... more Purpose: To investigate postimplant dosimetry for different regions of the prostate gland in patients treated with transperineal 125 Iodine brachytherapy implants for low-and intermediate-risk prostate cancer. Methods and Materials: Two hundred eighty-four patients treated with permanent interstitial prostate brachytherapy comprised the study population. A nonuniform, urethral-sparing algorithm was used to plan all patients. Prostate contours were outlined on postimplant CT images. Prostate volumes were then divided into four quadrants: anterior-superior quadrant (ASQ), posterior-superior quadrant (PSQ), anterior-inferior quadrant (AIQ), and posterior-inferior quadrant (PIQ). Dose-volume histograms (DVHs) were calculated for the whole prostate and each quadrant.
Physics in Medicine and Biology, 2006
This study presents a new method of image signal-to-noise ratio (SNR) enhancement by utilizing a ... more This study presents a new method of image signal-to-noise ratio (SNR) enhancement by utilizing a newly developed 2D two-point maximum entropy regularization method (TPMEM). When utilized as an image filter, it is shown that 2D TPMEM offers unsurpassed flexibility in its ability to balance the complementary requirements of image smoothness and fidelity. The technique is evaluated for use in the
Journal of Applied Clinical Medical Physics, 2002
This study presents a 3D dose mapping of complex dose distributions using an x-ray computed tomog... more This study presents a 3D dose mapping of complex dose distributions using an x-ray computed tomography ͑CT͒ polymer gel dosimetry technique. Two polyacrylamide gels ͑PAGs͒ of identical composition were irradiated with the same four arc stereotactic treatment to maximum doses of 15 Gy ͑PAG1͒ and 8 Gy ͑PAG2͒. The PAGs were CT imaged using a previously defined protocol that involves image averaging and background subtraction to improve image quality. For comparison with the planned isodose distribution, the PAG images were converted to relative dose maps using a CT number-dose calibration curve or simple division. The PAG images were then co-registered with the planning CT images in the BrainLab ® treatment planning software which automatically provides reconstructed sagittal and coronal images for 3D evaluation of measured and planned dose. The hypointense high dose region in both sets of gel images agreed with the planned 80% isodose contour and was shifted by up to 1.5 and 3.0 mm in the axial and reconstructed planes, respectively. This demonstrates the ability of the CT gel technique to accurately localize the high dose region produced by the stereotactic treatment. The resulting agreement of the measured relative dose volume for PAG1 was within 3.0 mm for the 50% and 80% isodose surfaces. However, the dose contrast was too low in PAG2 to allow for accurate definition of measured relative dose surfaces. Thus, a PAG should be irradiated to higher doses if quantitative relative dose information is required. Unfortunately, this implies use of an additional PAG and its CT number dose response since doses greater than 8 -10 Gy fall outside the linear regions of the response.
Physics in Medicine and Biology, 2012
In this study a new x-ray CT polymer gel dosimetry (PGD) filtering technique is presented for the... more In this study a new x-ray CT polymer gel dosimetry (PGD) filtering technique is presented for the removal of (i) remnant ring and streak artefacts, and (ii) 'structured' noise in the form of minute, intrinsic gel density fluctuations. It is shown that the noise present within x-ray CT PGD images is not purely stochastic (pixel by pixel) in nature, but rather is 'structured', and hence purely stochastic-based noise-removal filters fail in removing this significant, unwanted noise component. The remnant artefact removal (RAR) technique is based on a class of signal stripping (i.e. baseline-estimation) algorithms typically used in the estimation of unwanted non-uniform baselines underlying spectral data. Here the traditional signal removal algorithm is recast, whereby the 'signal' that is removed is the structured noise and remnant artefacts, leaving the desired polymer gel dose distribution. The algorithm is extended to 2D and input parameters are optimized for PGD images. RAR filter results are tested on (i) synthetic images with measured gel background images added, in order to accurately represent actual noise present in PGD images, and (ii) PGD images of a three-field gel irradiation. RAR results are compared to a top-performing noise filter (adaptive mean, AM), used in previous x-ray CT PGD studies. It is shown that, in all cases, the RAR filter outperforms the AM filter, particularly in cases where either (i) a low-dose gel image has been acquired or (ii) the signal-to-noise ratio of the PG image is low, as in the case when a low number of image averages are acquired within a given experiment. Guidelines for the implementation of the RAR filter are given.
Physics in Medicine and Biology, 2012
This article reports on the dosimetric properties of a new N-isopropylacrylamide, high %T, polyme... more This article reports on the dosimetric properties of a new N-isopropylacrylamide, high %T, polymer gel formulation (19.5%T, 23%C), optimized for x-ray computed tomography (CT) polymer gel dosimetry (PGD). In addition, a new gel calibration technique is introduced together with an intensity-modulated radiation therapy (IMRT) treatment validation as an example of a clinical application of the new gel dosimeter. The dosimetric properties investigated include the temporal stability, spatial stability, batch reproducibility and dose rate dependence. The polymerization reaction is found to stabilize after 15 h post-irradiation. Spatial stability investigations reveal a small overshoot in response for gels imaged later than 36 h post-irradiation. Based on these findings, it is recommended that the new gel formulation be imaged between 15-36 h after irradiation. Intra-and inter-batch reproducibility are found to be excellent over the entire range of doses studied (0-28 Gy). A significant dose rate dependence is found for gels irradiated between 100-600 MU min −1 . Overall, the new gel is shown to have promising characteristics for CT PGD, however the implication of the observed dose rate dependence for some clinical applications remains to be determined. The new gel calibration method, based on pixel-by-pixel matching of dose and measured CT numbers, is found to be robust and to agree with the previously used region of interest technique. Pixel-by-pixel calibration is the new recommended standard for CT PGD. The dose resolution for the system was excellent, ranging from 0.2-0.5 Gy for doses between 0-20 Gy and 0.3-0.6 Gy for doses beyond 20 Gy. Comparison of the IMRT irradiation with planned doses yields excellent results: gamma pass rate (3%, 3 mm) of 99.3% at the isocentre slice and 93.4% over the entire treated volume.
Physics in Medicine and Biology, 2004
Due to a density change that occurs in irradiated polyacrylamide gel (PAG), x-ray computed tomogr... more Due to a density change that occurs in irradiated polyacrylamide gel (PAG), x-ray computed tomography (CT) has emerged as a feasible method of performing polymer gel dosimetry. However, applicability of the technique is currently limited by low sensitivity of the density change to dose. This work investigates the effect of PAG composition on the radiation induced density change and provides direction for future work in improving the sensitivity of CT polymer gel dosimetry. A model is developed that describes the PAG density change (delta(rho)gel) as a function of both polymer yield (%P) and an intrinsic density change, per unit polymer yield, that occurs on conversion of monomer to polymer (delta(rho)polymer). %P is a function of the fraction of monomer consumed and the weight fraction of monomer in the unirradiated gel (%T). Applying the model to experimental CT and Raman spectroscopic data, two important fundamental properties of the response of PAG density to dose (delta(rho)gel dose response) are discovered. The first property is that delta(rho)polymer)depends on PAG %C (cross-linking fraction of total monomer) such that low and high %C PAGs exhibit a higher deltarho(polymer)than do more intermediate %C PAGs. This relationship is opposite to the relationship of polymer yield to %C and is explained by the effect of %C on the type of polymer formed. The second property is that the delta(rho)gel dose response is linearly dependent on %T. From the model, the inference is that, at least for %T < or = 2%, monomer consumption and delta(rho)polymer depend solely on %C. In terms of optimizing CT polymer gel dosimetry for high sensitivity, these results indicate that delta(rho)polymer can be expected to vary with each polymer gel system and thus should be considered when choosing a polymer gel for CT gel dosimetry. However, delta(rho)polymerand %P cannot be maximized simultaneously and maximizing %P, by choosing gels with intermediate %C and high %T, is found to have the greatest impact on increasing the sensitivity of PAG density to dose. As such, future research into new gel formulations for high sensitivity CT polymer gel dosimetry should focus on gels that exhibit an intrinsic density change and maximizing polymer yield in these systems.
Physics in Medicine and Biology, 2006
Of the antioxidants used to scavenge oxygen in polymer gel dosimeters, tetrakis (hydroxymethyl) p... more Of the antioxidants used to scavenge oxygen in polymer gel dosimeters, tetrakis (hydroxymethyl) phosphonium chloride (THPC) has been shown to hold great promise due to its rapid oxygen scavenging abilities. In this study we (a) investigate the use of THPC as an antioxidant for polyacrylamide gel (PAGAT) dosimeters used in conjunction with x-ray computed tomography (CT) and (b) work to establish the reaction mechanisms of THPC with the polymer gel constituents. We establish the dose response reproducibility of PAGAT dosimeters when imaged with CT and show that PAGAT dosimeters exhibit highly reproducible dose responses for a range of irradiation times post gel manufacture (2-6 h) and CT imaging times post gel irradiation (1-5 days). The THPC concentration within the gel leading to a maximized dose response and minimized O(2) inhibition of polymerization is found to be approximately 4.5 mM. We further assess the stability of PAGAT dosimeters by investigating the reactions of THPC with the individual gel constituents. The importance of utilizing deionized water in polymer gel manufacture is noted. We show that, while THPC remains unreactive with acrylamide and bis-acrylamide under unirradiated conditions, THPC can react with gelatin to increase the cross-linking of the gelatin matrix in unirradiated dosimeters. THPC reactions with gelatin can lead to the lower observed dose sensitivity of PAGAT (approximately 0.36 +/- 0.04 H Gy(-1)) as compared to polyacrylamide gels manufactured under anoxic conditions (approximately 0.83 +/- 0.03 H Gy(-1)). The reactions of THPC which lead to O(2) scavenging, and potential reactions of THPC with other gel constituents, are proposed.
Physics in Medicine and Biology, 2010
The aim of this study was to verify the advanced inhomogeneous dose distribution produced by a vo... more The aim of this study was to verify the advanced inhomogeneous dose distribution produced by a volumetric arc therapy technique (RapidArc TM ) using 3D gel measurements and Monte Carlo (MC) simulations. The TPS (treatment planning system)-calculated dose distribution was compared with gel measurements and MC simulations, thus investigating any discrepancy between the planned dose delivery and the actual delivery. Additionally, the reproducibility of the delivery was investigated using repeated gel measurements. A prostate treatment plan was delivered to a 1.3 liter nPAG gel phantom using one single arc rotation and a target dose of 3.3 Gy. Magnetic resonance imaging of the gel was carried out using a 1.5 T scanner. The MC dose distributions were calculated using the VIMC-Arc code. The relative absorbed dose differences were calculated voxel-by-voxel, within the volume enclosed by the 90% isodose surface (VOI 90 ), for the TPS versus gel and TPS versus MC. The differences between the verification methods, MC versus gel, and between two repeated gel measurements were investigated in the same way. For all volume comparisons, the mean value was within 1% and the standard deviation of the differences was within 2.5% (1SD). A 3D gamma analysis between the dose matrices were carried out using gamma criteria 3%/3 mm and 5%/5 mm (% dose difference and mm distance to agreement) within the volume enclosed by the 50% isodose surface (VOI 50 ) and the 90% isodose surface (VOI 90 ), respectively. All comparisons resulted in very high pass rates. More than 95% of the TPS points were within 3%/3 mm of both the gel measurement and MC simulation, both inside VOI 50 and VOI 90 . 0031-9155/10/174885+14$30.00 © 2010 Institute of Physics and Engineering in Medicine Printed in the UK 4885 4886 S Ceberg et al Additionally, the repeated gel measurements showed excellent consistency, indicating reproducible delivery. Using MC simulations and gel measurements, this verification study successfully demonstrated that the RapidArc TM plan was both accurately calculated and delivered as planned.
Physics in Medicine and Biology, 2009
A factor currently limiting the clinical utility of x-ray CT polymer gel dosimetry is the overall... more A factor currently limiting the clinical utility of x-ray CT polymer gel dosimetry is the overall low dose sensitivity (and hence low dose resolution) of the system. Hence, active research remains in the investigation of polymer gel formulations with increased CT dose response. An ideal polymer gel dosimeter will exhibit a sensitive CT response which is linear over a suitable dose range, making clinical implementation reasonably straightforward. This study reports on the variations in rate and form of the CT dose response of irradiated polymer gels manufactured with glycerol, which is a co-solvent that permits dissolution of additional bisacrylamide above its water solubility limit (3% by weight). This study focuses on situations where the concentration of bisacrylamide is kept at or below its water solubility limit so that the influence of the co-solvent on the dose response can be explored separately from the effects of increased cross-linker concentration. CT imaging and Raman spectroscopy are used to construct dose-response curves for irradiated gels varying in (i) initial total monomer (%T) and (ii) initial co-solvent concentration. Results indicate that: (i) for a fixed glycerol concentration, gel response increases linearly with %T. Furthermore, the functional form of the dose response remains constant, in agreement with a previous model of polymer formation. (ii) Polymer gels with constant %T and increasing co-solvent concentrations also show enhanced CT response. In addition, the functional form of the response is altered in these gels as co-solvent concentration is increased. Raman data indicate that the fraction of bis-acrylamide incorporated into polymerization, as opposed to cyclization, increases as co-solvent concentration increases. The changes in functional form indicate varying polymer yields (per unit dose), akin to relative fractional monomer/cross-linker (i.e. %C) changes in earlier studies. These results are put into context of the model of polymer formation. The implications of these results on the clinical utility of polymer gels with co-solvent are highlighted.
Medical Physics, 2004
X-ray computed tomography (CT) has been established as a feasible method of performing dosimetry ... more X-ray computed tomography (CT) has been established as a feasible method of performing dosimetry using polyacrylamide gels (PAGs). A small density change occurs in PAG upon irradiation that provides contrast in PAG CT images. However, low dose resolution limits the clinical usefulness of the technique. This work investigates the potential of using image filtering techniques on PAG CT images in order to reduce image noise and improve dose resolution. CT image noise for the scanner and protocol used for the gel images is analyzed and found to be Gaussian distributed and independent of the contrast level in the images. As a result, several filters for reducing spatially invariant noise are investigated: mean, median, midpoint, adaptive mean, alpha-trimmed mean, sigma mean, and a relatively new filter called SUSAN (smallest univalue segment assimilating nucleus). All filters are applied, using 3x3, 5x5, and 7x7 pixel masks, to a CT image of a PAG irradiated with a stereotactic radiosurgery dose distribution. The dose resolution within 95% confidence (D(delta)95%) is calculated and compared for each filtered image, as well the unfiltered image. In addition, the ability of the filters to maintain the spatial integrity of the dose distribution is evaluated and compared. Results clearly indicate that the filters are not equal in their ability to improve D(delta)95% or in their effect on the spatial integrity of the dose distribution. In general, increasing mask size improves D(delta)95% but simultaneously degrades spatial dose information. The mean filter provides the greatest improvement in D(delta)95%, but also the greatest loss of spatial dose information. The SUSAN, mean adaptive, and alpha-trimmed mean filters all provide comparable, but slightly poorer dose resolution. In addition, the SUSAN and adaptive filters both excel at maintaining the spatial distribution of dose and overall are the best performing filters for this application. The midpoint filter, normally useful for Gaussian noise, is poor all-round, dramatically distorting the dose distribution for masks greater than 3x3. The median filter, a common edge preserving noise reduction filter, performs moderately well, but artificially increases high dose gradients. The sigma filter preserves the spatial distribution of dose very well but is least effective at improving dose resolution. In summary, dose resolution can be significantly improved in CT PAG dosimetry through postprocessing of CT images using spatial noise reduction filters. However, such filters are not equal in their ability to improve dose resolution or to maintain the spatial integrity of the dose distribution and an appropriate filter must be chosen depending on clinical demands of the application.
Medical Physics, 2008
X-ray computed tomography ͑CT͒ as a method of extracting 3D dose information from irradiated poly... more X-ray computed tomography ͑CT͒ as a method of extracting 3D dose information from irradiated polymer gel dosimeters is showing potential as a practical means to implement gel dosimetry in a radiation therapy clinic. However, the response of CT contrast to dose is weak and noise reduction is critical in order to achieve adequate dose resolutions with this method. Phantom design and CT imaging technique have both been shown to decrease image noise. In addition, image postprocessing using noise reduction filtering techniques have been proposed. This work evaluates in detail the use of the adaptive mean filter for reducing noise in CT gel dosimetry. Filter performance is systematically tested using both synthetic patterns mimicking a range of clinical dose distribution features as well as actual clinical dose distributions. Both low and high signal-to-noise ratio ͑SNR͒ situations are examined. For all cases, the effects of filter kernel size and the number of iterations are investigated. Results indicate that adaptive mean filtering is a highly effective tool for noise reduction CT gel dosimetry. The optimum filtering strategy depends on characteristics of the dose distributions and image noise level. For low noise images ͑SNR ϳ20͒, the filtered results are excellent and use of adaptive mean filtering is recommended as a standard processing tool. For high noise images ͑SNR ϳ5͒ adaptive mean filtering can also produce excellent results, but filtering must be approached with more caution as spatial and dose distortions of the original dose distribution can occur.
Medical Physics, 2007
This study reports on the effects of x-ray CT dose in CT imaged normoxic polyacrylamide ͑nPAG͒ ge... more This study reports on the effects of x-ray CT dose in CT imaged normoxic polyacrylamide ͑nPAG͒ gel dosimeters. The investigation is partitioned into three sections. First, the CT dose absorbed in nPAG is quantified under a range of typical gel CT imaging protocols. It is found that the maximum absorbed CT dose occurs for volumetric imaging and is in the range of 4.6± 0.2 cGy/image. This does scales linearly with image averaging. Second, using Raman spectroscopy, the response of nPAG to CT imaging photon energies ͑i.e., 120-140 kVp͒ is established and compared to the well known dose response of nPAG exposed to 6 MV photons. It is found that nPAG exhibits a weaker response ͑per unit dose͒ to 140-kVp incident photons as compared to 6 MV incident photons ͑slopes m 6 MV = −0.0374± 0.0006 Gy −1 and m 140 kVp = −0.016± 0.001 Gy −1 ͒. Finally, using the above data, an induced change in CT number ͑⌬N CT ͒ is calculated for nPAG imaged using a range of gel imaging protocols. It is found that under typical imaging protocols ͑120-140 kVp, 200 mAs, ϳ16-32 image averages͒ a ⌬N CT Ͻ 0.2 H is induced in active nPAG dosimeters. This ⌬N CT is below the current limit of detectability of CT nPAG polymer gel dosimetry. Under expanded imaging protocols ͑e.g., very high number of image averages͒ an induced ⌬N CT of ϳ0.5 H is possible. In these situations the additional polymerization occurring in nPAG due to the imaging process may need to be accounted for.
Journal of Physics: Conference Series, 2006
Journal of Physics: Conference Series, 2004
This work investigates factors which affect image noise in CT polymer gel dosimetry, discusses te... more This work investigates factors which affect image noise in CT polymer gel dosimetry, discusses techniques that can be used to further improve image noise and provides overall recommendations for the CT imaging of polymer gels.
International Journal of Radiation Oncology*Biology*Physics, 2002
Purpose: To investigate postimplant dosimetry for different regions of the prostate gland in pati... more Purpose: To investigate postimplant dosimetry for different regions of the prostate gland in patients treated with transperineal 125 Iodine brachytherapy implants for low-and intermediate-risk prostate cancer. Methods and Materials: Two hundred eighty-four patients treated with permanent interstitial prostate brachytherapy comprised the study population. A nonuniform, urethral-sparing algorithm was used to plan all patients. Prostate contours were outlined on postimplant CT images. Prostate volumes were then divided into four quadrants: anterior-superior quadrant (ASQ), posterior-superior quadrant (PSQ), anterior-inferior quadrant (AIQ), and posterior-inferior quadrant (PIQ). Dose-volume histograms (DVHs) were calculated for the whole prostate and each quadrant.
International Journal of Radiation Oncology*Biology*Physics, 2007
Purpose: To examine variability in target volume delineation for partial breast radiotherapy plan... more Purpose: To examine variability in target volume delineation for partial breast radiotherapy planning and evaluate characteristics associated with low interobserver concordance. Methods and Materials: Thirty patients who underwent planning CT for adjuvant breast radiotherapy formed the study cohort. Using a standardized scale to score seroma clarity and consensus contouring guidelines, three radiation oncologists independently graded seroma clarity and delineated seroma volumes for each case. Seroma geometric center coordinates, maximum diameters in three axes, and volumes were recorded. Conformity index (CI), the ratio of overlapping volume and encompassing delineated volume, was calculated for each case. Cases with CI #0.50 were analyzed to identify features associated with low concordance. Results: The median time from surgery to CT was 42.5 days. For geometric center coordinates, variations from the mean were 0.5-1.1 mm and standard deviations (SDs) were 0.5-1.8 mm. For maximum seroma dimensions, variations from the mean and SDs were predominantly <5 mm, with the largest SDs observed in the medial-lateral axis. The mean CI was 0.61 (range, 0.27-0.84). Five cases had CI #0.50. Conformity index was significantly associated with seroma clarity ( p < 0.001) and seroma volume ( p < 0.002). Features associated with reduced concordance included tissue stranding from the surgical cavity, proximity to muscle, dense breast parenchyma, and benign calcifications that may be mistaken for surgical clips. Conclusion: Variability in seroma contouring occurred in three dimensions, with the largest variations in the medial-lateral axis. Awareness of clinical features associated with reduced concordance may be applied toward training staff and refining contouring guidelines for partial breast radiotherapy trials. Ó 2007 Elsevier Inc.
Purpose: To examine variability in target volume delineation for partial breast radiotherapy plan... more Purpose: To examine variability in target volume delineation for partial breast radiotherapy planning and evaluate characteristics associated with low interobserver concordance. Methods and Materials: Thirty patients who underwent planning CT for adjuvant breast radiotherapy formed the study cohort. Using a standardized scale to score seroma clarity and consensus contouring guidelines, three radiation oncologists independently graded seroma clarity and delineated seroma volumes for each case. Seroma geometric center coordinates, maximum diameters in three axes, and volumes were recorded. Conformity index (CI), the ratio of overlapping volume and encompassing delineated volume, was calculated for each case. Cases with CI #0.50 were analyzed to identify features associated with low concordance. Results: The median time from surgery to CT was 42.5 days. For geometric center coordinates, variations from the mean were 0.5-1.1 mm and standard deviations (SDs) were 0.5-1.8 mm. For maximum seroma dimensions, variations from the mean and SDs were predominantly <5 mm, with the largest SDs observed in the medial-lateral axis. The mean CI was 0.61 (range, 0.27-0.84). Five cases had CI #0.50. Conformity index was significantly associated with seroma clarity ( p < 0.001) and seroma volume ( p < 0.002). Features associated with reduced concordance included tissue stranding from the surgical cavity, proximity to muscle, dense breast parenchyma, and benign calcifications that may be mistaken for surgical clips. Conclusion: Variability in seroma contouring occurred in three dimensions, with the largest variations in the medial-lateral axis. Awareness of clinical features associated with reduced concordance may be applied toward training staff and refining contouring guidelines for partial breast radiotherapy trials. Ó 2007 Elsevier Inc.
Radiotherapy and Oncology, 2005
To demonstrate the benefit of deformable image registration (DIR) to define postimplant seromas i... more To demonstrate the benefit of deformable image registration (DIR) to define postimplant seromas in permanent breast seed implants (PBSIs) by adapting preimplant seromas to postimplant images. Planning and Day 0 CT scans from 15 consecutive PBSI patients were included. Seromas were contoured on the preimplant CTs. After implant, DIR (MIM Maestro; MIM Software, Inc., Cleveland, OH) was used to adapt preimplant seromas to postimplant scans. Resulting seromas were evaluated for accuracy and adjusted if required, and pre- and postimplant seroma volumes were recorded and compared. DIR was validated by assessing the spatial correspondence of surgical clips pre- and postimplant in eight cases. Retrospectively, rigid registration was also investigated. Surgical clip location corresponded between pre- and postimplant images with a median agreement of &amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;lt;1.0 mm. The seroma defined using DIR was an acceptable rendering of the postimplant seroma in 9 of 15 cases, and in the remaining six cases, minor alterations (mean, 1.0 ± 0.2 mm) were made. Mean change in seroma volumes between preimplant and postimplant imaging was -6.1 ± 26.4%. Rigid registration exhibited unacceptable variability, with poor concordance in postimplant seroma between different observers (median conformity index, 26.5%; range, 0-52.7%). Deformable image registration can be used to adapt the preimplant seroma to postimplant PBSI images and is recommended for implant quality assurance in PBSI. Using deformable registration avoids uncertainties associated with poor seroma visualization and interobserver variability when rigid registration is performed. The accurate deformation of implanted surgical clips demonstrates the reliability of deformable image registration for PBSI.
Purpose: To investigate postimplant dosimetry for different regions of the prostate gland in pati... more Purpose: To investigate postimplant dosimetry for different regions of the prostate gland in patients treated with transperineal 125 Iodine brachytherapy implants for low-and intermediate-risk prostate cancer. Methods and Materials: Two hundred eighty-four patients treated with permanent interstitial prostate brachytherapy comprised the study population. A nonuniform, urethral-sparing algorithm was used to plan all patients. Prostate contours were outlined on postimplant CT images. Prostate volumes were then divided into four quadrants: anterior-superior quadrant (ASQ), posterior-superior quadrant (PSQ), anterior-inferior quadrant (AIQ), and posterior-inferior quadrant (PIQ). Dose-volume histograms (DVHs) were calculated for the whole prostate and each quadrant.
Physics in Medicine and Biology, 2006
This study presents a new method of image signal-to-noise ratio (SNR) enhancement by utilizing a ... more This study presents a new method of image signal-to-noise ratio (SNR) enhancement by utilizing a newly developed 2D two-point maximum entropy regularization method (TPMEM). When utilized as an image filter, it is shown that 2D TPMEM offers unsurpassed flexibility in its ability to balance the complementary requirements of image smoothness and fidelity. The technique is evaluated for use in the
Journal of Applied Clinical Medical Physics, 2002
This study presents a 3D dose mapping of complex dose distributions using an x-ray computed tomog... more This study presents a 3D dose mapping of complex dose distributions using an x-ray computed tomography ͑CT͒ polymer gel dosimetry technique. Two polyacrylamide gels ͑PAGs͒ of identical composition were irradiated with the same four arc stereotactic treatment to maximum doses of 15 Gy ͑PAG1͒ and 8 Gy ͑PAG2͒. The PAGs were CT imaged using a previously defined protocol that involves image averaging and background subtraction to improve image quality. For comparison with the planned isodose distribution, the PAG images were converted to relative dose maps using a CT number-dose calibration curve or simple division. The PAG images were then co-registered with the planning CT images in the BrainLab ® treatment planning software which automatically provides reconstructed sagittal and coronal images for 3D evaluation of measured and planned dose. The hypointense high dose region in both sets of gel images agreed with the planned 80% isodose contour and was shifted by up to 1.5 and 3.0 mm in the axial and reconstructed planes, respectively. This demonstrates the ability of the CT gel technique to accurately localize the high dose region produced by the stereotactic treatment. The resulting agreement of the measured relative dose volume for PAG1 was within 3.0 mm for the 50% and 80% isodose surfaces. However, the dose contrast was too low in PAG2 to allow for accurate definition of measured relative dose surfaces. Thus, a PAG should be irradiated to higher doses if quantitative relative dose information is required. Unfortunately, this implies use of an additional PAG and its CT number dose response since doses greater than 8 -10 Gy fall outside the linear regions of the response.
Physics in Medicine and Biology, 2012
In this study a new x-ray CT polymer gel dosimetry (PGD) filtering technique is presented for the... more In this study a new x-ray CT polymer gel dosimetry (PGD) filtering technique is presented for the removal of (i) remnant ring and streak artefacts, and (ii) 'structured' noise in the form of minute, intrinsic gel density fluctuations. It is shown that the noise present within x-ray CT PGD images is not purely stochastic (pixel by pixel) in nature, but rather is 'structured', and hence purely stochastic-based noise-removal filters fail in removing this significant, unwanted noise component. The remnant artefact removal (RAR) technique is based on a class of signal stripping (i.e. baseline-estimation) algorithms typically used in the estimation of unwanted non-uniform baselines underlying spectral data. Here the traditional signal removal algorithm is recast, whereby the 'signal' that is removed is the structured noise and remnant artefacts, leaving the desired polymer gel dose distribution. The algorithm is extended to 2D and input parameters are optimized for PGD images. RAR filter results are tested on (i) synthetic images with measured gel background images added, in order to accurately represent actual noise present in PGD images, and (ii) PGD images of a three-field gel irradiation. RAR results are compared to a top-performing noise filter (adaptive mean, AM), used in previous x-ray CT PGD studies. It is shown that, in all cases, the RAR filter outperforms the AM filter, particularly in cases where either (i) a low-dose gel image has been acquired or (ii) the signal-to-noise ratio of the PG image is low, as in the case when a low number of image averages are acquired within a given experiment. Guidelines for the implementation of the RAR filter are given.
Physics in Medicine and Biology, 2012
This article reports on the dosimetric properties of a new N-isopropylacrylamide, high %T, polyme... more This article reports on the dosimetric properties of a new N-isopropylacrylamide, high %T, polymer gel formulation (19.5%T, 23%C), optimized for x-ray computed tomography (CT) polymer gel dosimetry (PGD). In addition, a new gel calibration technique is introduced together with an intensity-modulated radiation therapy (IMRT) treatment validation as an example of a clinical application of the new gel dosimeter. The dosimetric properties investigated include the temporal stability, spatial stability, batch reproducibility and dose rate dependence. The polymerization reaction is found to stabilize after 15 h post-irradiation. Spatial stability investigations reveal a small overshoot in response for gels imaged later than 36 h post-irradiation. Based on these findings, it is recommended that the new gel formulation be imaged between 15-36 h after irradiation. Intra-and inter-batch reproducibility are found to be excellent over the entire range of doses studied (0-28 Gy). A significant dose rate dependence is found for gels irradiated between 100-600 MU min −1 . Overall, the new gel is shown to have promising characteristics for CT PGD, however the implication of the observed dose rate dependence for some clinical applications remains to be determined. The new gel calibration method, based on pixel-by-pixel matching of dose and measured CT numbers, is found to be robust and to agree with the previously used region of interest technique. Pixel-by-pixel calibration is the new recommended standard for CT PGD. The dose resolution for the system was excellent, ranging from 0.2-0.5 Gy for doses between 0-20 Gy and 0.3-0.6 Gy for doses beyond 20 Gy. Comparison of the IMRT irradiation with planned doses yields excellent results: gamma pass rate (3%, 3 mm) of 99.3% at the isocentre slice and 93.4% over the entire treated volume.
Physics in Medicine and Biology, 2004
Due to a density change that occurs in irradiated polyacrylamide gel (PAG), x-ray computed tomogr... more Due to a density change that occurs in irradiated polyacrylamide gel (PAG), x-ray computed tomography (CT) has emerged as a feasible method of performing polymer gel dosimetry. However, applicability of the technique is currently limited by low sensitivity of the density change to dose. This work investigates the effect of PAG composition on the radiation induced density change and provides direction for future work in improving the sensitivity of CT polymer gel dosimetry. A model is developed that describes the PAG density change (delta(rho)gel) as a function of both polymer yield (%P) and an intrinsic density change, per unit polymer yield, that occurs on conversion of monomer to polymer (delta(rho)polymer). %P is a function of the fraction of monomer consumed and the weight fraction of monomer in the unirradiated gel (%T). Applying the model to experimental CT and Raman spectroscopic data, two important fundamental properties of the response of PAG density to dose (delta(rho)gel dose response) are discovered. The first property is that delta(rho)polymer)depends on PAG %C (cross-linking fraction of total monomer) such that low and high %C PAGs exhibit a higher deltarho(polymer)than do more intermediate %C PAGs. This relationship is opposite to the relationship of polymer yield to %C and is explained by the effect of %C on the type of polymer formed. The second property is that the delta(rho)gel dose response is linearly dependent on %T. From the model, the inference is that, at least for %T < or = 2%, monomer consumption and delta(rho)polymer depend solely on %C. In terms of optimizing CT polymer gel dosimetry for high sensitivity, these results indicate that delta(rho)polymer can be expected to vary with each polymer gel system and thus should be considered when choosing a polymer gel for CT gel dosimetry. However, delta(rho)polymerand %P cannot be maximized simultaneously and maximizing %P, by choosing gels with intermediate %C and high %T, is found to have the greatest impact on increasing the sensitivity of PAG density to dose. As such, future research into new gel formulations for high sensitivity CT polymer gel dosimetry should focus on gels that exhibit an intrinsic density change and maximizing polymer yield in these systems.
Physics in Medicine and Biology, 2006
Of the antioxidants used to scavenge oxygen in polymer gel dosimeters, tetrakis (hydroxymethyl) p... more Of the antioxidants used to scavenge oxygen in polymer gel dosimeters, tetrakis (hydroxymethyl) phosphonium chloride (THPC) has been shown to hold great promise due to its rapid oxygen scavenging abilities. In this study we (a) investigate the use of THPC as an antioxidant for polyacrylamide gel (PAGAT) dosimeters used in conjunction with x-ray computed tomography (CT) and (b) work to establish the reaction mechanisms of THPC with the polymer gel constituents. We establish the dose response reproducibility of PAGAT dosimeters when imaged with CT and show that PAGAT dosimeters exhibit highly reproducible dose responses for a range of irradiation times post gel manufacture (2-6 h) and CT imaging times post gel irradiation (1-5 days). The THPC concentration within the gel leading to a maximized dose response and minimized O(2) inhibition of polymerization is found to be approximately 4.5 mM. We further assess the stability of PAGAT dosimeters by investigating the reactions of THPC with the individual gel constituents. The importance of utilizing deionized water in polymer gel manufacture is noted. We show that, while THPC remains unreactive with acrylamide and bis-acrylamide under unirradiated conditions, THPC can react with gelatin to increase the cross-linking of the gelatin matrix in unirradiated dosimeters. THPC reactions with gelatin can lead to the lower observed dose sensitivity of PAGAT (approximately 0.36 +/- 0.04 H Gy(-1)) as compared to polyacrylamide gels manufactured under anoxic conditions (approximately 0.83 +/- 0.03 H Gy(-1)). The reactions of THPC which lead to O(2) scavenging, and potential reactions of THPC with other gel constituents, are proposed.
Physics in Medicine and Biology, 2010
The aim of this study was to verify the advanced inhomogeneous dose distribution produced by a vo... more The aim of this study was to verify the advanced inhomogeneous dose distribution produced by a volumetric arc therapy technique (RapidArc TM ) using 3D gel measurements and Monte Carlo (MC) simulations. The TPS (treatment planning system)-calculated dose distribution was compared with gel measurements and MC simulations, thus investigating any discrepancy between the planned dose delivery and the actual delivery. Additionally, the reproducibility of the delivery was investigated using repeated gel measurements. A prostate treatment plan was delivered to a 1.3 liter nPAG gel phantom using one single arc rotation and a target dose of 3.3 Gy. Magnetic resonance imaging of the gel was carried out using a 1.5 T scanner. The MC dose distributions were calculated using the VIMC-Arc code. The relative absorbed dose differences were calculated voxel-by-voxel, within the volume enclosed by the 90% isodose surface (VOI 90 ), for the TPS versus gel and TPS versus MC. The differences between the verification methods, MC versus gel, and between two repeated gel measurements were investigated in the same way. For all volume comparisons, the mean value was within 1% and the standard deviation of the differences was within 2.5% (1SD). A 3D gamma analysis between the dose matrices were carried out using gamma criteria 3%/3 mm and 5%/5 mm (% dose difference and mm distance to agreement) within the volume enclosed by the 50% isodose surface (VOI 50 ) and the 90% isodose surface (VOI 90 ), respectively. All comparisons resulted in very high pass rates. More than 95% of the TPS points were within 3%/3 mm of both the gel measurement and MC simulation, both inside VOI 50 and VOI 90 . 0031-9155/10/174885+14$30.00 © 2010 Institute of Physics and Engineering in Medicine Printed in the UK 4885 4886 S Ceberg et al Additionally, the repeated gel measurements showed excellent consistency, indicating reproducible delivery. Using MC simulations and gel measurements, this verification study successfully demonstrated that the RapidArc TM plan was both accurately calculated and delivered as planned.
Physics in Medicine and Biology, 2009
A factor currently limiting the clinical utility of x-ray CT polymer gel dosimetry is the overall... more A factor currently limiting the clinical utility of x-ray CT polymer gel dosimetry is the overall low dose sensitivity (and hence low dose resolution) of the system. Hence, active research remains in the investigation of polymer gel formulations with increased CT dose response. An ideal polymer gel dosimeter will exhibit a sensitive CT response which is linear over a suitable dose range, making clinical implementation reasonably straightforward. This study reports on the variations in rate and form of the CT dose response of irradiated polymer gels manufactured with glycerol, which is a co-solvent that permits dissolution of additional bisacrylamide above its water solubility limit (3% by weight). This study focuses on situations where the concentration of bisacrylamide is kept at or below its water solubility limit so that the influence of the co-solvent on the dose response can be explored separately from the effects of increased cross-linker concentration. CT imaging and Raman spectroscopy are used to construct dose-response curves for irradiated gels varying in (i) initial total monomer (%T) and (ii) initial co-solvent concentration. Results indicate that: (i) for a fixed glycerol concentration, gel response increases linearly with %T. Furthermore, the functional form of the dose response remains constant, in agreement with a previous model of polymer formation. (ii) Polymer gels with constant %T and increasing co-solvent concentrations also show enhanced CT response. In addition, the functional form of the response is altered in these gels as co-solvent concentration is increased. Raman data indicate that the fraction of bis-acrylamide incorporated into polymerization, as opposed to cyclization, increases as co-solvent concentration increases. The changes in functional form indicate varying polymer yields (per unit dose), akin to relative fractional monomer/cross-linker (i.e. %C) changes in earlier studies. These results are put into context of the model of polymer formation. The implications of these results on the clinical utility of polymer gels with co-solvent are highlighted.
Medical Physics, 2004
X-ray computed tomography (CT) has been established as a feasible method of performing dosimetry ... more X-ray computed tomography (CT) has been established as a feasible method of performing dosimetry using polyacrylamide gels (PAGs). A small density change occurs in PAG upon irradiation that provides contrast in PAG CT images. However, low dose resolution limits the clinical usefulness of the technique. This work investigates the potential of using image filtering techniques on PAG CT images in order to reduce image noise and improve dose resolution. CT image noise for the scanner and protocol used for the gel images is analyzed and found to be Gaussian distributed and independent of the contrast level in the images. As a result, several filters for reducing spatially invariant noise are investigated: mean, median, midpoint, adaptive mean, alpha-trimmed mean, sigma mean, and a relatively new filter called SUSAN (smallest univalue segment assimilating nucleus). All filters are applied, using 3x3, 5x5, and 7x7 pixel masks, to a CT image of a PAG irradiated with a stereotactic radiosurgery dose distribution. The dose resolution within 95% confidence (D(delta)95%) is calculated and compared for each filtered image, as well the unfiltered image. In addition, the ability of the filters to maintain the spatial integrity of the dose distribution is evaluated and compared. Results clearly indicate that the filters are not equal in their ability to improve D(delta)95% or in their effect on the spatial integrity of the dose distribution. In general, increasing mask size improves D(delta)95% but simultaneously degrades spatial dose information. The mean filter provides the greatest improvement in D(delta)95%, but also the greatest loss of spatial dose information. The SUSAN, mean adaptive, and alpha-trimmed mean filters all provide comparable, but slightly poorer dose resolution. In addition, the SUSAN and adaptive filters both excel at maintaining the spatial distribution of dose and overall are the best performing filters for this application. The midpoint filter, normally useful for Gaussian noise, is poor all-round, dramatically distorting the dose distribution for masks greater than 3x3. The median filter, a common edge preserving noise reduction filter, performs moderately well, but artificially increases high dose gradients. The sigma filter preserves the spatial distribution of dose very well but is least effective at improving dose resolution. In summary, dose resolution can be significantly improved in CT PAG dosimetry through postprocessing of CT images using spatial noise reduction filters. However, such filters are not equal in their ability to improve dose resolution or to maintain the spatial integrity of the dose distribution and an appropriate filter must be chosen depending on clinical demands of the application.
Medical Physics, 2008
X-ray computed tomography ͑CT͒ as a method of extracting 3D dose information from irradiated poly... more X-ray computed tomography ͑CT͒ as a method of extracting 3D dose information from irradiated polymer gel dosimeters is showing potential as a practical means to implement gel dosimetry in a radiation therapy clinic. However, the response of CT contrast to dose is weak and noise reduction is critical in order to achieve adequate dose resolutions with this method. Phantom design and CT imaging technique have both been shown to decrease image noise. In addition, image postprocessing using noise reduction filtering techniques have been proposed. This work evaluates in detail the use of the adaptive mean filter for reducing noise in CT gel dosimetry. Filter performance is systematically tested using both synthetic patterns mimicking a range of clinical dose distribution features as well as actual clinical dose distributions. Both low and high signal-to-noise ratio ͑SNR͒ situations are examined. For all cases, the effects of filter kernel size and the number of iterations are investigated. Results indicate that adaptive mean filtering is a highly effective tool for noise reduction CT gel dosimetry. The optimum filtering strategy depends on characteristics of the dose distributions and image noise level. For low noise images ͑SNR ϳ20͒, the filtered results are excellent and use of adaptive mean filtering is recommended as a standard processing tool. For high noise images ͑SNR ϳ5͒ adaptive mean filtering can also produce excellent results, but filtering must be approached with more caution as spatial and dose distortions of the original dose distribution can occur.
Medical Physics, 2007
This study reports on the effects of x-ray CT dose in CT imaged normoxic polyacrylamide ͑nPAG͒ ge... more This study reports on the effects of x-ray CT dose in CT imaged normoxic polyacrylamide ͑nPAG͒ gel dosimeters. The investigation is partitioned into three sections. First, the CT dose absorbed in nPAG is quantified under a range of typical gel CT imaging protocols. It is found that the maximum absorbed CT dose occurs for volumetric imaging and is in the range of 4.6± 0.2 cGy/image. This does scales linearly with image averaging. Second, using Raman spectroscopy, the response of nPAG to CT imaging photon energies ͑i.e., 120-140 kVp͒ is established and compared to the well known dose response of nPAG exposed to 6 MV photons. It is found that nPAG exhibits a weaker response ͑per unit dose͒ to 140-kVp incident photons as compared to 6 MV incident photons ͑slopes m 6 MV = −0.0374± 0.0006 Gy −1 and m 140 kVp = −0.016± 0.001 Gy −1 ͒. Finally, using the above data, an induced change in CT number ͑⌬N CT ͒ is calculated for nPAG imaged using a range of gel imaging protocols. It is found that under typical imaging protocols ͑120-140 kVp, 200 mAs, ϳ16-32 image averages͒ a ⌬N CT Ͻ 0.2 H is induced in active nPAG dosimeters. This ⌬N CT is below the current limit of detectability of CT nPAG polymer gel dosimetry. Under expanded imaging protocols ͑e.g., very high number of image averages͒ an induced ⌬N CT of ϳ0.5 H is possible. In these situations the additional polymerization occurring in nPAG due to the imaging process may need to be accounted for.
Journal of Physics: Conference Series, 2006
Journal of Physics: Conference Series, 2004
This work investigates factors which affect image noise in CT polymer gel dosimetry, discusses te... more This work investigates factors which affect image noise in CT polymer gel dosimetry, discusses techniques that can be used to further improve image noise and provides overall recommendations for the CT imaging of polymer gels.
International Journal of Radiation Oncology*Biology*Physics, 2002
Purpose: To investigate postimplant dosimetry for different regions of the prostate gland in pati... more Purpose: To investigate postimplant dosimetry for different regions of the prostate gland in patients treated with transperineal 125 Iodine brachytherapy implants for low-and intermediate-risk prostate cancer. Methods and Materials: Two hundred eighty-four patients treated with permanent interstitial prostate brachytherapy comprised the study population. A nonuniform, urethral-sparing algorithm was used to plan all patients. Prostate contours were outlined on postimplant CT images. Prostate volumes were then divided into four quadrants: anterior-superior quadrant (ASQ), posterior-superior quadrant (PSQ), anterior-inferior quadrant (AIQ), and posterior-inferior quadrant (PIQ). Dose-volume histograms (DVHs) were calculated for the whole prostate and each quadrant.
International Journal of Radiation Oncology*Biology*Physics, 2007
Purpose: To examine variability in target volume delineation for partial breast radiotherapy plan... more Purpose: To examine variability in target volume delineation for partial breast radiotherapy planning and evaluate characteristics associated with low interobserver concordance. Methods and Materials: Thirty patients who underwent planning CT for adjuvant breast radiotherapy formed the study cohort. Using a standardized scale to score seroma clarity and consensus contouring guidelines, three radiation oncologists independently graded seroma clarity and delineated seroma volumes for each case. Seroma geometric center coordinates, maximum diameters in three axes, and volumes were recorded. Conformity index (CI), the ratio of overlapping volume and encompassing delineated volume, was calculated for each case. Cases with CI #0.50 were analyzed to identify features associated with low concordance. Results: The median time from surgery to CT was 42.5 days. For geometric center coordinates, variations from the mean were 0.5-1.1 mm and standard deviations (SDs) were 0.5-1.8 mm. For maximum seroma dimensions, variations from the mean and SDs were predominantly <5 mm, with the largest SDs observed in the medial-lateral axis. The mean CI was 0.61 (range, 0.27-0.84). Five cases had CI #0.50. Conformity index was significantly associated with seroma clarity ( p < 0.001) and seroma volume ( p < 0.002). Features associated with reduced concordance included tissue stranding from the surgical cavity, proximity to muscle, dense breast parenchyma, and benign calcifications that may be mistaken for surgical clips. Conclusion: Variability in seroma contouring occurred in three dimensions, with the largest variations in the medial-lateral axis. Awareness of clinical features associated with reduced concordance may be applied toward training staff and refining contouring guidelines for partial breast radiotherapy trials. Ó 2007 Elsevier Inc.