J Chang

Panoramic cone beam computed tomography

PURPOSEnCone-beam computed tomography (CBCT) is the main imaging tool for image-guided radiotherapy but its functionality is limited by a small imaging volume and restricted image position (imaged at the central instead of the treatment position for peripheral lesions to avoid collisions). In this paper, the authors present the concept of panoramic CBCT, which can image patients at the treatment position with an imaging volume as large as practically needed.nnnMETHODSnIn this novel panoramic CBCT technique, the target is scanned sequentially from multiple view angles. For each view angle, a half scan (180°u2009+u2009θ(cone) where θ(cone) is the cone angle) is performed with the imaging panel positioned in any location along the beam path. The panoramic projection images of all views for the same gantry angle are then stitched together with the direct image stitching method (i.e., according to the reported imaging position) and full-fan, half-scan CBCT reconstruction is performed using the stitched projection images. To validate this imaging technique, the authors simulated cone-beam projection images of the Mathematical Cardiac Torso (MCAT) thorax phantom for three panoramic views. Gaps, repeated/missing columns, and different exposure levels were introduced between adjacent views to simulate imperfect image stitching due to uncertainties in imaging position or output fluctuation. A modified simultaneous algebraic reconstruction technique (modified SART) was developed to reconstruct CBCT images directly from the stitched projection images. As a gold standard, full-fan, full-scan (360° gantry rotation) CBCT reconstructions were also performed using projection images of one imaging panel large enough to encompass the target. Contrast-to-noise ratio (CNR) and geometric distortion were evaluated to quantify the quality of reconstructed images. Monte Carlo simulations were performed to evaluate the effect of scattering on the image quality and imaging dose for both standard and panoramic CBCT.nnnRESULTSnTruncated images with artifacts were observed for the CBCT reconstruction using projection images of the central view only. When the image stitching was perfect, complete reconstruction was obtained for the panoramic CBCT using the modified SART with the image quality similar to the gold standard (full-scan, full-fan CBCT using one large imaging panel). Imperfect image stitching, on the other hand, lead to (streak, line, or ring) reconstruction artifacts, reduced CNR, and/or distorted geometry. Results from Monte Carlo simulations showed that, for identical imaging quality, the imaging dose was lower for the panoramic CBCT than that acquired with one large imaging panel. For the same imaging dose, the CNR of the three-view panoramic CBCT was 50% higher than that of the regular CBCT using one big panel.nnnCONCLUSIONSnThe authors have developed a panoramic CBCT technique and demonstrated with simulation data that it can image tumors of any location for patients of any size at the treatment position with comparable or less imaging dose and time. However, the image quality of this CBCT technique is sensitive to the reconstruction artifacts caused by imperfect image stitching. Better algorithms are therefore needed to improve the accuracy of image stitching for panoramic CBCT.

Is there a tradeoff in using modified high tangent field radiation for treating an undissected node-positive axilla?

INTRODUCTIONnRecent data are changing axillary management in patients with 1 to 2 positive sentinel nodes. The proposed omission of completion axillary node dissection calls into question the need for axillary nodal irradiation. This study evaluates the difference in dose to the lung and heart and risk of radiation pneumonitis (RP) for patients treated with standard tangent fields (STF) compared with modified high tangent fields (MHTF).nnnMATERIALS AND METHODSnPlans of 30 patients treated with STF were evaluated. A second plan (MHTF) was developed to include axillary levels I (Ax1) and II (Ax2). Ax1 and Ax2 volumes were contoured based on the RTOG (Radiation Therapy Oncology Group) Atlas guidelines. Dose-volume histograms of the 2 plans were used to compare doses received by Ax1, Ax2, lung, and heart volumes. The risk of RP was calculated using normal tissue complication probability (NTCP) modeling.nnnRESULTSnThe D95 (dose to 95% of volume) received by Ax1 and Ax2 volumes increased from 16.38 Gy and 5.71 Gy for STF to 49.38 Gy and 48.08 Gy for MHTF, respectively. Mean lung dose increased from 5.40 Gy for STF to 9.47 Gy for MHTF. Mean ipsilateral lung V5, V10, and V20 values increased from 19%, 14%, and 10%, respectively, for STF, to 32%, 24%, and 18%, respectively, for MHTF. Mean heart dose increased from 1.98 Gy for STF to 3.93 Gy for MHTF. Mean heart V25 and V30 values increased from 2% and 1%, respectively, for STF, to 4% and 3%, respectively, for MHTF. NTCP for RP increased from near 0% for STF to 1% for MHTF.nnnCONCLUSIONnModified high tangent fields are necessary for definitive coverage of Ax1 and Ax2. This technique increases mean ipsilateral lung and heart doses as well as the V5, V10, and V20 of ipsilateral lung and the V25 and V30 of the heart. Risk of RP remains low by use of MHTF.

SU‐E‐J‐72: Shift Invariant Feature Transform (SIFT) Based Image Stitching for Panoramic Cone Beam CT (CBCT)

PURPOSEnPanoramic CBCT is a novel imaging technique that stitches together projection images from multiple views to increase the imaging volume to as large as practically needed using half (200 degree gantry rotation) scans only. In this study, we investigated a SIFT-based method to improve the accuracy of image stitching for panoramic CBCT.nnnMETHODSnPanoramic CBCT data were acquired for a torso phantom using two half scans: (A) the detector at the center position and (B) the detector shifted 148mm toward the right side of the phantom. For each projection image of scan A, a corresponding projection image of the same gantry angle was interpolated for scan B using images of two neighboring gantry angles. The SIFT method was applied to extract matching feature points in the overlapping region of the projection images of the same gantry angle from each scan. The rigid (translation and rotation) transform matrix derived from this analysis was then used to stitch together the two projection images for CBCT reconstruction using the simultaneous algebraic reconstruction technique (SART) method. Three overlapping (5-cm, 3-cm and 2-cm) sizes were used for the SIFT analysis to determine the optimal imaging parameters.nnnRESULTSnThe CBCT reconstructions using the SIFT stitching method showed significant artifact reduction compared with the direct stitching method. The success rate of matching (less than 0.5 degrees rotation and 2 pixels horizontal/vertical shifts) was close to 100% for 5-cm overlapping and gradually reduced as the overlapping size decreased. The minimum overlapping size we could reliably apply our method was 2 cm (93.8% success rate) below which significant operator involvement was required to determine the applicable rigid transform parameters.nnnCONCLUSIONnWe have demonstrated the effectiveness of the SIFT stitching method for panoramic CBCT reconstruction. We are currently improving the registration technique to further reduce the required overlapping size. This work was partially supported by a DOD grant DOD W81XWH1010862.

MO-FG-CAMPUS-JeP3-01: A Statistical Model for Analyzing the Rotational Error of Single Iso-Center Technique

PURPOSEnTo develop a generalized statistical model that incorporates the treatment uncertainty from the rotational error of single iso-center technique, and calculate the additional PTV (planning target volume) margin required to compensate for this error.nnnMETHODSnThe random vectors for setup and additional rotation errors in the three-dimensional (3D) patient coordinate system were assumed to follow the 3D independent normal distribution with zero mean, and standard deviations σx, σy, σz, for setup error and a uniform σR for rotational error. Both random vectors were summed, normalized and transformed to the spherical coordinates to derive the chi distribution with 3 degrees of freedom for the radical distance ρ. PTV margin was determined using the critical value of this distribution for 0.05 significant level so that 95% of the time the treatment target would be covered by ρ. The additional PTV margin required to compensate for the rotational error was calculated as a function of σx, σy, σz and σR.nnnRESULTSnThe effect of the rotational error is more pronounced for treatments that requires high accuracy/precision like stereotactic radiosurgery (SRS) or stereotactic body radiotherapy (SBRT). With a uniform 2mm PTV margin (or σx =σy=σz=0.7mm), a σR=0.32mm will decrease the PTV coverage from 95% to 90% of the time, or an additional 0.2mm PTV margin is needed to prevent this loss of coverage. If we choose 0.2 mm as the threshold, any σR>0.3mm will lead to an additional PTV margin that cannot be ignored, and the maximal σR that can be ignored is 0.0064 rad (or 0.37°) for iso-to-target distance=5cm, or 0.0032 rad (or 0.18°) for iso-to-target distance=10cm.nnnCONCLUSIONSnThe rotational error cannot be ignored for high-accuracy/-precision treatments like SRS/SBRT, particularly when the distance between the iso-center and target is large.

SU‐F‐T‐615: Comparison of Plan Quality for Linac‐Based Stereotactic Radiosurgery (SRS) Using Single‐ and Multi‐Isocenter Techniques

PURPOSEnTo compare the plan quality of linear accelerator (linac)-based stereotactic radiosurgery (SRS) using single-isocenter volumetric arc therapy (SI-VMAT), restricted single-isocenter dynamic-arc (RSI-DARC), and multi-isocenter DARC (MI-DARC) techniques.nnnMETHODSnFifteen SRS cases were randomly selected and re-planned using the SI-VMAT (Pinnacle), RSI-DARC (iPlanNet) and MI-DARC (iPlanNet). The number of planning target volumes (PTVs) for each plan ranged from 1 to 6. For SI-VMAT, a single isocenter and 3-4 VMAT beams are used for all PTVs, while for MI-DARC, each PTV has its own isocetner with 3 DARC beams. RSI-DARC uses one isocnter with 3-6 DARC beams to irradiate all PTVs within 2.5-cm radius. Both SI-DARC and RSI-DARC plans were optimized manually. The prescription dose was 20 Gy to each PTV. The maximal dose was 25 Gy for RSI-DARC and MI-DARC, but could not be controlled for SI-VMAT due to the nature of VMAT planning. Plan quality indexes including PTV coverage, mean dose of PTV (PTVmean) and tissue (Tmean), V12Gy, conformity index (CI), and V10Gy/VPTV were calculated and compared.nnnRESULTSnFull PTV coverage was achieved for all three techniques. Using the MI-DARC plans as the gold standard, the PTVmean of the SI-VMAT plans was 12.5%±8.3% (mean±standard deviation) higher, in comparison to 0.7%±1.4% for the RSI-DARC plans. Similar trend was observed for other indexes including V12Gy (39.4%±27.3% vs. 9.3%±7.8%), Tmean (35.0%±26.8% vs. 2.8%±3.4%), and V10Gy/VPTV (42.2%±31.5% vs. 9.9%±8.2%). CI is comparable (6.2%±14.2% vs. 6.3%±7.2%). Assuming the treatment time is proportional to the number of isocenters, the reduction of the treatment time in comparison to MI-DARC was 70% for SI-VMAT and 42% for RSI-DARC.nnnCONCLUSIONnAlthough the SI-VMAT can save a considerable amount of treatment time, the plan indexes also significantly deviates from the gold standard, MI-DARC. RSI-DARC, on the other hand, provides a good compromise between the treatment time and plan quality.

SU-E-J-237: Image Feature Based DRR and Portal Image Registration

PURPOSEnTwo-dimensional (2D) matching of the kV X-ray and digitally reconstructed radiography (DRR) images is an important setup technique for image-guided radiotherapy (IGRT). In our clinics, mutual information based methods are used for this purpose on commercial linear accelerators, but with often needs for manual corrections. This work proved the feasibility that feature based image transform can be used to register kV and DRR images.nnnMETHODSnThe scale invariant feature transform (SIFT) method was implemented to detect the matching image details (or key points) between the kV and DRR images. These key points represent high image intensity gradients, and thus the scale invariant features. Due to the poor image contrast from our kV image, direct application of the SIFT method yielded many detection errors. To assist the finding of key points, the center coordinates of the kV and DRR images were read from the DICOM header, and the two groups of key points with similar relative positions to their corresponding centers were paired up. Using these points, a rigid transform (with scaling, horizontal and vertical shifts) was estimated. We also artificially introduced vertical and horizontal shifts to test the accuracy of our registration method on anterior-posterior (AP) and lateral pelvic images.nnnRESULTSnThe results provided a satisfactory overlay of the transformed kV onto the DRR image. The introduced vs. detected shifts were fit into a linear regression. In the AP image experiments, linear regression analysis showed a slope of 1.15 and 0.98 with an R2 of 0.89 and 0.99 for the horizontal and vertical shifts, respectively. The results are 1.2 and 1.3 with R2 of 0.72 and 0.82 for the lateral image shifts.nnnCONCLUSIONnThis work provided an alternative technique for kV to DRR alignment. Further improvements in the estimation accuracy and image contrast tolerance are underway.

SU-E-T-377: Inaccurate Positioning Might Introduce Significant MapCheck Calibration Error in Flatten Filter Free Beams

PURPOSEnThis study investigates the calibration error of detector sensitivity for MapCheck due to inaccurate positioning of the device, which is not taken into account by the current commercial iterative calibration algorithm. We hypothesize the calibration is more vulnerable to the positioning error for the flatten filter free (FFF) beams than the conventional flatten filter flattened beams.nnnMETHODSnMapCheck2 was calibrated with 10MV conventional and FFF beams, with careful alignment and with 1cm positioning error during calibration, respectively. Open fields of 37cmx37cm were delivered to gauge the impact of resultant calibration errors. The local calibration error was modeled as a detector independent multiplication factor, with which propagation error was estimated with positioning error from 1mm to 1cm. The calibrated sensitivities, without positioning error, were compared between the conventional and FFF beams to evaluate the dependence on the beam type.nnnRESULTSnThe 1cm positioning error leads to 0.39% and 5.24% local calibration error in the conventional and FFF beams respectively. After propagating to the edges of MapCheck, the calibration errors become 6.5% and 57.7%, respectively. The propagation error increases almost linearly with respect to the positioning error. The difference of sensitivities between the conventional and FFF beams was small (0.11 ± 0.49%).nnnCONCLUSIONnThe results demonstrate that the positioning error is not handled by the current commercial calibration algorithm of MapCheck. Particularly, the calibration errors for the FFF beams are ~9 times greater than those for the conventional beams with identical positioning error, and a small 1mm positioning error might lead to up to 8% calibration error. Since the sensitivities are only slightly dependent of the beam type and the conventional beam is less affected by the positioning error, it is advisable to cross-check the sensitivities between the conventional and FFF beams to detect potential calibration errors due to inaccurate positioning. This work was partially supported by a DOD Grant No.; DOD W81XWH1010862.

SU‐E‐T‐206: Improving Radiotherapy Toxicity Based On Artificial Neural Network (ANN) for Head and Neck Cancer Patients

PURPOSE/OBJECTIVE(S)nThe aim of this study is to build the estimator of toxicity using artificial neural network (ANN) for head and neck cancer patients MATERIALS/METHODS: An ANN can combine variables into a predictive model during training and considered all possible correlations of variables. We constructed an ANN based on the data from 73 patients with advanced H&N cancer treated with external beam radiotherapy and/or chemotherapy at our institution. For the toxicity estimator we defined input data including age, sex, site, stage, pathology, status of chemo, technique of external beam radiation therapy (EBRT), length of treatment, dose of EBRT, status of post operation, length of follow-up, the status of local recurrences and distant metastasis. These data were digitized based on the significance and fed to the ANN as input nodes. We used 20 hidden nodes (for the 13 input nodes) to take care of the correlations of input nodes. For training ANN, we divided data into three subsets such as training set, validation set and test set. Finally, we built the estimator for the toxicity from ANN output.nnnRESULTSnWe used 13 input variables including the status of local recurrences and distant metastasis and 20 hidden nodes for correlations. 59 patients for training set, 7 patients for validation set and 7 patients for test set and fed the inputs to Matlab neural network fitting tool. We trained the data within 15% of errors of outcome. In the end we have the toxicity estimation with 74% of accuracy.nnnCONCLUSIONnWe proved in principle that ANN can be a very useful tool for predicting the RT outcomes for high risk H&N patients. Currently we are improving the results using cross validation.

Calibration of a detector array through beam profile reconstruction with error-locking.

An iterative method is proposed to calibrate radiation sensitivities of an arbitrary two‐dimensional (2D) array of detectors. The array is irradiated with a wide open‐field beam at the central position, as well as at laterally and longitudinal shifted positions; the 2D beam profile of the wide field is reconstructed iteratively from the ratios of shifted images to the central image. The propagation errors due to output variation and inaccurate array positioning are estimated and removed from the reconstructed beam profile by an error‐locking scheme with narrow open‐field irradiations. The beam profile is interpolated when necessary and then compared to raw detector responses to determine sensitivities. Two additional methods were implemented for comparison: 1) the commercial iterative calibration method for MapCHECK2 with translation and rotation operations; 2) a labor‐intensive noniterative method without the issue of error propagation. A MapCHECK2 2D detector array was used to validate the proposed method with the 6 MV photon beam from a Varian iX linear accelerator. All calibration methods were repeated three times. A total of 5, 9, and 29 irradiations were required to implement the commercial method, the proposed method and the noniterative method respectively. Moreover, a 5 mm positioning error was intentionally introduced into the calibration procedures of the commercial and the proposed method to test their robustness. Under the normal operation condition of the linear accelerator and with careful alignment of the MapCHECK2, the deviations of the calibrated sensitivities of the proposed method and commercial method with respect to the noniterative method were 0.30%±0.29% and 0.92%±0.63% respectively; when the 5 mm positioning error was presented, these two methods resulted in deviations of 0.40%±0.36% and 3.58%±1.94%, respectively. A patient study suggested that, due to this 5 mm positioning error, the mean DTA (dose to agreement) passing rate by the commercial method was 2.7% lower than that by the noniterative method, whereas the proposed method led to a comparable passing rate. It is evident from this study that the proposed iterative method leads to within 1% mean calibration results to established methods. It requires much fewer number of measurements than noniterative method and is more robust against the positioning error than the commercial iterative method. The method also eliminates the need of rotation operations and, therefore, is applicable to inline detector arrays without rotation function, such as electronic portal imager device (EPID). PACS number: 87.56.Fc

SU‐E‐T‐702: Performance of Multiple Conformity, Homogeneity and Dose Gradient Indices in SRS and SBRT Treatment Planning of the Spinal, Pelvic and Lung Lesions

Purpose: Presently, various forms of quality indices (QI) are used to compare and evaluate stereotactic treatment plans with no single QI accepted as universal. In this work we evaluate and compare the performance of 11 previously reported QI: 5 conformity (CI), 5 homogeneity (HI), and 1 dose gradient (DGI) indices as a function of target size, site, and mode of treatment. A novel unified quality function (UQF) which combines conformity, homogeneity and dose gradient is presented and investigated. Methods: 27 spine SRS, 8 pelvic SBRT and 34 lung SBRT cases were analyzed using an in‐house wxPython program which imports treatment plan data and calculates various quality indices, dose volume tables, displays 3D dose/CT images, cumulative and differential DVHs. Results: Mean spine, pelvis and lung coverages were 0.91, 0.93, 0.85 respectively. Mean maximum target dose was best for pelvis (108%), followed by the lung (115%) and the spine (118%). Mean minimum target dose was best for the lung (95%), followed by pelvis (94%) and spine (88%). Mean DGI were 4.6, 5.2 and 7.54 for the pelvis, spine and lung. UQF scores the best for pelvis (0.2), followed by spine (0.28) and the lung (0.42). IMRT spine plans had better mean coverages (0.94), and lower maximum target dose (114%), while VMAT plans had better minimum target dose (89%). Conclusion: Mean CI, HI and DGI show various sensitivities to the treatment site, size and the delivery type (VMAT or IMRT). All CI except coverage show large variations with volume < 130 cc while all 5 HI show little dependence on the PTV volume. DGI values fall sharply for the lung (from 15 to 6) and spine (from 10 to 5) as the PTV increases to about 100 cc, while pelvic DGI stays about 5 for all PTV sizes.