- Email: [email protected]
Utilization of Lung Regional Density Variation for Image Guided IMRT Treatment Planning: A 4D CT Study
Proceedings of the 51st Annual ASTRO Meeting regarding the parameter V90 (target volume included in the 90 per cent isodose) for PTV1, mean dose for both parotids and maximum dose for spinal cord. For plan verification we used classical film dosimetry together with measuring of absolute dose in a defined point within the IMRT phantom. Radiation treatment was performed on a linear accelerator Elekta Synergy with 6 MV photons. The plans created on PrecisePlan were used mainly for higher verification precision. For each treatment plan, we compared number of segments, number of monitoring units and treatment time for a single fraction. Statistical analysis was made using paired t-test with two-tailed p values. Results: Mean number of segments produced by PrecisePlan (ABIP) and Plato (BBIP) was 41 and 98, respectively (p \ 0,001). Similarly significant difference was observed comparing the mean number of monitor units - 448 for PrecisePlan, and 1591 for Plato (p \ 0,001). The number of MU is in fact the real irradiation time (‘‘beam on’’), thus the difference might be important for the operating life of radiation-generating parts of the linear accelerator. The number of MU, however, does not represent the real fraction time. To obtain this, we have to include the time necessary for gantry and collimator rotation as well as lamellae movement. The mean fraction times were 15 min. for PrecisePlan and 19 min. for Plato (p = 0,001). Conclusions: The optimization algorithm ABIP for step-and-shoot IMRT produces fewer segments, fewer monitor units and significantly shortens the total fraction time in comparison to the BBIP algorithm. Author Disclosure: J. Cvek, None; B. Otahal, None; P. Nevrelova, None; D. Feltl, None.
3063
Image Guided Radiotherapy (IGRT) for Prostate Cancer Comparing kV Imaging of Fiducial Markers with Cone Beam Computed Tomography (CBCT)
B. Barney1, R. J. Lee2, D. Handrahan2, W. T. Sause2 1
Mayo School of Graduate Medical Education, Rochester, MN, 2Intermountain Medical Center, Salt Lake City, UT
Purpose/Objective(s): To compare kilovoltage (kV) portal imaging using fiducial markers and cone beam CT (CBCT) for daily localization in the treatment of prostate cancer. Materials/Methods: From April 2007 to October 2008, 36 patients with prostate cancer received intensity-modulated radiotherapy with daily localization using implanted fiducials. Orthogonal kV portal imaging preceded all 1244 treatments. CBCT images with fiducials subtracted were obtained prior to 286 (23%) of these treatments. Shifts based on CBCT images were recorded. Shifts in anterior-posterior (AP), superior-inferior (SI) and left-right (LR) dimensions were made based on kV imaging of fiducials. Fiducial and CBCT shifts were compared using Bland-Altman limits of agreement. Mean and standard deviation of absolute differences were also compared. In both comparisons, a difference of less than 5mm (PTV margin) was considered acceptable. Patient subsets, including treatment start date, BMI (body mass index) and prostate size were also analyzed. Results: Of 286 treatments, 81 (28%) resulted in an absolute difference of . 5.0mm in one or more dimensions. The average absolute differences (in mm) in AP, SI, and LR dimensions were 3.4 ± 2.6, 3.1 ± 2.7 and 1.3 ± 1.6, respectively. With both imaging techniques deviations occurred most often in the posterior (fiducials: 78%, CBCT: 59%), superior (79%, 61%) and left (57%, 63%) directions. The Bland-Altman 95% CI (in mm) was (-4.0, 9.3) AP, (-9.0, 5.3) SI, and (-4.1, 3.9) LR. The percentage of shift agreements within +/- 5mm was 72.4, 72.7, and 97.2 in AP, SI, and LR directions, respectively. The percentage within +/- 3mm was 41.3 AP, 49.3 SI, and 87.4 LR. Generally, correlation between imaging techniques was not altered by study time, patient BMI, or prostate size. Conclusions: CBCT and kV fiducial imaging are similar for defining interfraction prostate shifts. With a PTV margin of 5mm, more than ˇ1/4 of CBCT and kV shifts differed enough to affect target coverage. This is even more pronounced with smaller margins (3mm). No one subset predicted imaging similarity, though large prostates (PTV . 200cm3) had a smaller average difference. kV fiducial imaging requires less daily physician input, is less time consuming and is our preferred method for prostate IGRT. Author Disclosure: B. Barney, None; R.J. Lee, None; D. Handrahan, None; W.T. Sause, None.
3064
Utilization of Lung Regional Density Variation for Image Guided IMRT Treatment Planning: A 4D CT Study
J. Li, M. Chan, K. Schupak, C. Burman Memorial Sloan-Kettering Cancer Center, Basking Ridge, NJ Purpose/Objective(s): Sparing of normal lung tissue during lung cancer radiotherapy is an important factor in 4D treatment planning. In the past, lung tissue sparing has been assumed to be directly related to the change of lung volume during breathing cycle. Although lung dose distribution can be significantly affected by lung density, little attention has been given to the dosimetric effect of changing lung regional densities (LRD) by respiratory motion. This study is to utilize 4D CT simulation to evaluate LRD changes induced by breathing cycles for image-guided radiotherapy treatment (IGRT) planning in lung cancer. Materials/Methods: 4D CT scans were acquired for 10 patients with lung cancer on a Philips Brilliance CT, using the respiratory signal from the Bellows on patient’ abdomen. All patients were in supine position. The generated CT images were sorted into ten phases. At each respiratory phase, nine regions of interest (ROI) for measurements of LRD were determined as follows: lung volume was first segmented into roughly three equal portions: superior, middle and inferior; and each portion were further divided into anterior, central, and posterior regions. Density was measured in the Hounsfield unit (HU) scale. For each patient, LRD variations from the breathing cycle in nine ROIs were compared and phases in which extreme change of density occurred were identified. To assess dosimetric consequences of LRD variations, IMRT plans were generated for different respiratory phases to simulate gated radiotherapy treatments. Results: There were significant density differences among nine ROIs for each patient. On average, the posterior region was 41.5% denser than the anterior region, and the superior portion only 5.7% denser than the inferior portion. During the breathing cycle, significant differences in LRD variation were observed comparing inferior and superior regions (p \ 0.01). LRD variation was more evident inferiorly. Mean lung density variations for inferior and superior regions were 18.1% and 7.2%, respectively. There was little difference in LRD variation between anterior and posterior regions. There was no clear phase relationship between the
S659
I. J. Radiation Oncology d Biology d Physics
S660
Volume 75, Number 3, Supplement, 2009
change of LRD and the change of lung volume: it depends on both patient breathing pattern and ROI. For gated IMRT treatment, mean dose to ipsilateral lung volume was reduced by up to 17% during the phase of low LRD if tumor was in inferior region, whereas up to 6.5% reduction if tumor was in superior region. Conclusions: To increase sparing of normal lung tissue, data of LRD variation should be incorporated into 4D treatment planning. For cancer treatments in the inferior region of lung, gated IMRT based on LRD variation can significantly reduce lung dose. 4D CT simulation provides a useful method to assess LRD variation induced by respiratory motion. Author Disclosure: J. Li, None; M. Chan, None; K. Schupak, None; C. Burman, None.
3065
Long-term Single Institutional Experience with the Use of Daily Image-guided Radiotherapy (IGRT) for the Treatment of Head and Neck Cancer
S. Won, J. Marsano, B. Li, R. Sreeraman, A. Michaud, J. Perks, J. A. Purdy, A. M. Chen University of California, Davis, Cancer Center, Department of Radiation Oncology, Sacramento, CA Purpose/Objective(s): The purpose of this study was to present a long-term single institutional experience with the use of daily image-guided radiotherapy (IGRT) for the treatment of head and neck cancer with intensity-modulated radiotherapy (IMRT). We also aimed to determine whether the percentage of weight loss over the course of treatment correlated with changes in positional alignment. Materials/Methods: Between 2005 and 2009, 124 patients with locally advanced head and neck cancer were treated at out our institution using IMRT with daily on-board imaging with megavoltage computed tomography (MVCT). The MVCT images were acquired over a longitudinal field of view that typically ranged from C7 to 2 cm superior to the base of skull using a coarse resolution in-slice thickness and nominal beam energy of 3.5 MV. After the MVCT images were reconstructed, they were fused with the treatment-planning CT images using automated registration bone presets and prepared for realignment if needed. These IGRT image sets were independently reviewed in the sagittal, coronal, and axial planes using the fusion split screen. We examined overall patient set-up shifts in the anterior-posterior (x), right-left (y), and superior-inferior (z) directions for IGRT, and the mean overall shift vector (MOSV) was subsequently calculated to be used as a measure of positional displacement. The percentage of weight loss during the entire course of treatment was determined from a retrospective review of patient records. The Pearson’s Correlation Coefficient, rho (r), was used to determine if a significant correlation existed between percentage weight loss and MOSV. Results: A total of 3,968 fractions were analyzed. The mean systematic alignment errors for all 124 patients were 0.5mm, 1.6mm, and 5.6mm in the x,- y-, z-dimensions with daily IGRT. The mean MOSV for all patients was 8.4mm with a standard deviation of 7.5mm. The mean percentage weight loss for all patients over the course of treatment was 7.2% with a standard deviation of 5.8%. When we compared MOSV to percent weight loss, we failed to detect any correlation between the two groups using the Pearson’s Correlation Coefficient (r=.05). The covariance (r^2) was determined to be 0.002. Conclusions: Although the benefits of daily IGRT have yet to be demonstrated with respect to the clinical endpoints of localregional control and overall survival for patients undergoing IMRT for head and neck cancer, this strategy appears to be a feasible and practical way of minimizing patient set-up error. Notably, the inability to identify any significant correlations between positional shifts as measured with daily IGRT and various weight loss parameters suggest that the random variation observed has causes more complex than simple weight loss. Author Disclosure: S. Won, None; J. Marsano, None; B. Li, None; R. Sreeraman, None; A. Michaud, None; J. Perks, None; J.A. Purdy, None; A.M. Chen, None.
3066
Development and Validation of an Automated IMRT Treatment Planning Algorithm
J. O. Deasy, V. Clark, Y. Chen, A. Apte, J. M. Michalski, J. Cui Washington University School of Medicine, St. Louis, MO Purpose/Objective(s): A key unmet need for adaptive radiotherapy treatment planning and improving radiotherapy workflow in general is a method that automatically generates high-quality IMRT plans. Automated treatment planning, if feasible, could also serve as a quality assurance ‘check’ compared to human treatment planning. We compared IMRT treatment plans, generated using hierarchical/prioritized optimization techniques (‘priopt’) and Monte Carlo dosimetry, against clinical Pinnacle treatment plans for definitive prostate treatments. Materials/Methods: Six randomly selected cases were replanned. The workflow consisted of: a fast, water-based dose calculation engine to generate the input beamlet influence matrix; a convex formulation of the treatment planning problem including ‘MOHx’ (mean-of-hottest x%) metrics as closely-correlated substitutes for dose-volume constraints; the MOSEK nonlinear solver; a leaf sequencing algorithm; a final Monte Carlo re-computation step (utilizing a commissioned head model that closely reproduces 6 MV Pinnacle IMRT results from our clinic); all embedded within the research system, CERR. Plans were compared with clinical Pinnacle treatment plans based on a list of planning metrics, and associated goals levels, elicited from the treating physician, including: Target max dose \ 110% prescription dose (constraint); Target max dose \ 107% prescription dose (preferred); Target D98 . prescription dose; Rectum V40 \ 35%; Rectum V65 \ 17%; Rectum V70 \ 25%; Bladder V40 \ 50%; Bladder V65 \ 25%; and Bladder V70 \ 25%. Results: Most plan metrics met associated goals for most plans. Priopt plan metrics, averaged over the six cases, were better than Pinnacle for 8/9 metrics based on the simplified dosimetry, and better for 4/9 metrics after leaf-sequencing and final Monte Carlo recalculation. Except for the presence of greater dose heterogeneity allowed in the priopt plans, metric differences were small (usually on the order of a few percent) and seemed unlikely to lead to differing patient outcomes. Conclusions: Remarkably, priopt metrics were very similar to those achieved in the clinic. The results suggest that human planners relying on the Pinnacle system are working at close to optimal levels. The results further suggest that this algorithm could work well as an automated tool that produces high-quality dose distributions, with the caveat that more work needs to be done to slightly
3063
Image Guided Radiotherapy (IGRT) for Prostate Cancer Comparing kV Imaging of Fiducial Markers with Cone Beam Computed Tomography (CBCT)
B. Barney1, R. J. Lee2, D. Handrahan2, W. T. Sause2 1
Mayo School of Graduate Medical Education, Rochester, MN, 2Intermountain Medical Center, Salt Lake City, UT
Purpose/Objective(s): To compare kilovoltage (kV) portal imaging using fiducial markers and cone beam CT (CBCT) for daily localization in the treatment of prostate cancer. Materials/Methods: From April 2007 to October 2008, 36 patients with prostate cancer received intensity-modulated radiotherapy with daily localization using implanted fiducials. Orthogonal kV portal imaging preceded all 1244 treatments. CBCT images with fiducials subtracted were obtained prior to 286 (23%) of these treatments. Shifts based on CBCT images were recorded. Shifts in anterior-posterior (AP), superior-inferior (SI) and left-right (LR) dimensions were made based on kV imaging of fiducials. Fiducial and CBCT shifts were compared using Bland-Altman limits of agreement. Mean and standard deviation of absolute differences were also compared. In both comparisons, a difference of less than 5mm (PTV margin) was considered acceptable. Patient subsets, including treatment start date, BMI (body mass index) and prostate size were also analyzed. Results: Of 286 treatments, 81 (28%) resulted in an absolute difference of . 5.0mm in one or more dimensions. The average absolute differences (in mm) in AP, SI, and LR dimensions were 3.4 ± 2.6, 3.1 ± 2.7 and 1.3 ± 1.6, respectively. With both imaging techniques deviations occurred most often in the posterior (fiducials: 78%, CBCT: 59%), superior (79%, 61%) and left (57%, 63%) directions. The Bland-Altman 95% CI (in mm) was (-4.0, 9.3) AP, (-9.0, 5.3) SI, and (-4.1, 3.9) LR. The percentage of shift agreements within +/- 5mm was 72.4, 72.7, and 97.2 in AP, SI, and LR directions, respectively. The percentage within +/- 3mm was 41.3 AP, 49.3 SI, and 87.4 LR. Generally, correlation between imaging techniques was not altered by study time, patient BMI, or prostate size. Conclusions: CBCT and kV fiducial imaging are similar for defining interfraction prostate shifts. With a PTV margin of 5mm, more than ˇ1/4 of CBCT and kV shifts differed enough to affect target coverage. This is even more pronounced with smaller margins (3mm). No one subset predicted imaging similarity, though large prostates (PTV . 200cm3) had a smaller average difference. kV fiducial imaging requires less daily physician input, is less time consuming and is our preferred method for prostate IGRT. Author Disclosure: B. Barney, None; R.J. Lee, None; D. Handrahan, None; W.T. Sause, None.
3064
Utilization of Lung Regional Density Variation for Image Guided IMRT Treatment Planning: A 4D CT Study
J. Li, M. Chan, K. Schupak, C. Burman Memorial Sloan-Kettering Cancer Center, Basking Ridge, NJ Purpose/Objective(s): Sparing of normal lung tissue during lung cancer radiotherapy is an important factor in 4D treatment planning. In the past, lung tissue sparing has been assumed to be directly related to the change of lung volume during breathing cycle. Although lung dose distribution can be significantly affected by lung density, little attention has been given to the dosimetric effect of changing lung regional densities (LRD) by respiratory motion. This study is to utilize 4D CT simulation to evaluate LRD changes induced by breathing cycles for image-guided radiotherapy treatment (IGRT) planning in lung cancer. Materials/Methods: 4D CT scans were acquired for 10 patients with lung cancer on a Philips Brilliance CT, using the respiratory signal from the Bellows on patient’ abdomen. All patients were in supine position. The generated CT images were sorted into ten phases. At each respiratory phase, nine regions of interest (ROI) for measurements of LRD were determined as follows: lung volume was first segmented into roughly three equal portions: superior, middle and inferior; and each portion were further divided into anterior, central, and posterior regions. Density was measured in the Hounsfield unit (HU) scale. For each patient, LRD variations from the breathing cycle in nine ROIs were compared and phases in which extreme change of density occurred were identified. To assess dosimetric consequences of LRD variations, IMRT plans were generated for different respiratory phases to simulate gated radiotherapy treatments. Results: There were significant density differences among nine ROIs for each patient. On average, the posterior region was 41.5% denser than the anterior region, and the superior portion only 5.7% denser than the inferior portion. During the breathing cycle, significant differences in LRD variation were observed comparing inferior and superior regions (p \ 0.01). LRD variation was more evident inferiorly. Mean lung density variations for inferior and superior regions were 18.1% and 7.2%, respectively. There was little difference in LRD variation between anterior and posterior regions. There was no clear phase relationship between the
S659
I. J. Radiation Oncology d Biology d Physics
S660
Volume 75, Number 3, Supplement, 2009
change of LRD and the change of lung volume: it depends on both patient breathing pattern and ROI. For gated IMRT treatment, mean dose to ipsilateral lung volume was reduced by up to 17% during the phase of low LRD if tumor was in inferior region, whereas up to 6.5% reduction if tumor was in superior region. Conclusions: To increase sparing of normal lung tissue, data of LRD variation should be incorporated into 4D treatment planning. For cancer treatments in the inferior region of lung, gated IMRT based on LRD variation can significantly reduce lung dose. 4D CT simulation provides a useful method to assess LRD variation induced by respiratory motion. Author Disclosure: J. Li, None; M. Chan, None; K. Schupak, None; C. Burman, None.
3065
Long-term Single Institutional Experience with the Use of Daily Image-guided Radiotherapy (IGRT) for the Treatment of Head and Neck Cancer
S. Won, J. Marsano, B. Li, R. Sreeraman, A. Michaud, J. Perks, J. A. Purdy, A. M. Chen University of California, Davis, Cancer Center, Department of Radiation Oncology, Sacramento, CA Purpose/Objective(s): The purpose of this study was to present a long-term single institutional experience with the use of daily image-guided radiotherapy (IGRT) for the treatment of head and neck cancer with intensity-modulated radiotherapy (IMRT). We also aimed to determine whether the percentage of weight loss over the course of treatment correlated with changes in positional alignment. Materials/Methods: Between 2005 and 2009, 124 patients with locally advanced head and neck cancer were treated at out our institution using IMRT with daily on-board imaging with megavoltage computed tomography (MVCT). The MVCT images were acquired over a longitudinal field of view that typically ranged from C7 to 2 cm superior to the base of skull using a coarse resolution in-slice thickness and nominal beam energy of 3.5 MV. After the MVCT images were reconstructed, they were fused with the treatment-planning CT images using automated registration bone presets and prepared for realignment if needed. These IGRT image sets were independently reviewed in the sagittal, coronal, and axial planes using the fusion split screen. We examined overall patient set-up shifts in the anterior-posterior (x), right-left (y), and superior-inferior (z) directions for IGRT, and the mean overall shift vector (MOSV) was subsequently calculated to be used as a measure of positional displacement. The percentage of weight loss during the entire course of treatment was determined from a retrospective review of patient records. The Pearson’s Correlation Coefficient, rho (r), was used to determine if a significant correlation existed between percentage weight loss and MOSV. Results: A total of 3,968 fractions were analyzed. The mean systematic alignment errors for all 124 patients were 0.5mm, 1.6mm, and 5.6mm in the x,- y-, z-dimensions with daily IGRT. The mean MOSV for all patients was 8.4mm with a standard deviation of 7.5mm. The mean percentage weight loss for all patients over the course of treatment was 7.2% with a standard deviation of 5.8%. When we compared MOSV to percent weight loss, we failed to detect any correlation between the two groups using the Pearson’s Correlation Coefficient (r=.05). The covariance (r^2) was determined to be 0.002. Conclusions: Although the benefits of daily IGRT have yet to be demonstrated with respect to the clinical endpoints of localregional control and overall survival for patients undergoing IMRT for head and neck cancer, this strategy appears to be a feasible and practical way of minimizing patient set-up error. Notably, the inability to identify any significant correlations between positional shifts as measured with daily IGRT and various weight loss parameters suggest that the random variation observed has causes more complex than simple weight loss. Author Disclosure: S. Won, None; J. Marsano, None; B. Li, None; R. Sreeraman, None; A. Michaud, None; J. Perks, None; J.A. Purdy, None; A.M. Chen, None.
3066
Development and Validation of an Automated IMRT Treatment Planning Algorithm
J. O. Deasy, V. Clark, Y. Chen, A. Apte, J. M. Michalski, J. Cui Washington University School of Medicine, St. Louis, MO Purpose/Objective(s): A key unmet need for adaptive radiotherapy treatment planning and improving radiotherapy workflow in general is a method that automatically generates high-quality IMRT plans. Automated treatment planning, if feasible, could also serve as a quality assurance ‘check’ compared to human treatment planning. We compared IMRT treatment plans, generated using hierarchical/prioritized optimization techniques (‘priopt’) and Monte Carlo dosimetry, against clinical Pinnacle treatment plans for definitive prostate treatments. Materials/Methods: Six randomly selected cases were replanned. The workflow consisted of: a fast, water-based dose calculation engine to generate the input beamlet influence matrix; a convex formulation of the treatment planning problem including ‘MOHx’ (mean-of-hottest x%) metrics as closely-correlated substitutes for dose-volume constraints; the MOSEK nonlinear solver; a leaf sequencing algorithm; a final Monte Carlo re-computation step (utilizing a commissioned head model that closely reproduces 6 MV Pinnacle IMRT results from our clinic); all embedded within the research system, CERR. Plans were compared with clinical Pinnacle treatment plans based on a list of planning metrics, and associated goals levels, elicited from the treating physician, including: Target max dose \ 110% prescription dose (constraint); Target max dose \ 107% prescription dose (preferred); Target D98 . prescription dose; Rectum V40 \ 35%; Rectum V65 \ 17%; Rectum V70 \ 25%; Bladder V40 \ 50%; Bladder V65 \ 25%; and Bladder V70 \ 25%. Results: Most plan metrics met associated goals for most plans. Priopt plan metrics, averaged over the six cases, were better than Pinnacle for 8/9 metrics based on the simplified dosimetry, and better for 4/9 metrics after leaf-sequencing and final Monte Carlo recalculation. Except for the presence of greater dose heterogeneity allowed in the priopt plans, metric differences were small (usually on the order of a few percent) and seemed unlikely to lead to differing patient outcomes. Conclusions: Remarkably, priopt metrics were very similar to those achieved in the clinic. The results suggest that human planners relying on the Pinnacle system are working at close to optimal levels. The results further suggest that this algorithm could work well as an automated tool that produces high-quality dose distributions, with the caveat that more work needs to be done to slightly