Q Wu

Tradeoffs of integrating real-time tracking into IGRT for prostate cancer treatment

This study investigated the integration of the Calypso real-time tracking system, based on implanted ferromagnetic transponders and a detector array, into the current process for image-guided radiation treatment (IGRT) of prostate cancer at our institution. The current IGRT process includes magnetic resonance imaging (MRI) for prostate delineation, CT simulation for treatment planning, daily on-board kV and CBCT imaging for target alignment, and MRI/MRS for post-treatment assessment. This study assesses (1) magnetic-field-induced displacement and radio-frequency (RF)-induced heating of transponders during MRI at 1.5 T and 3 T, and (2) image artifacts caused by transponders and the detector array in phantom and patient cases with the different imaging systems. A tissue-equivalent phantom mimicking prostate tissue stiffness was constructed and implanted with three operational transponders prior to phantom solidification. The measurements show that the Calypso system is safe with all the imaging systems. Transponder position displacements due to the MR field are minimal (<1.0 mm) for both 1.5 T and 3 T MRI scanners, and the temperature variation due to MRI RF heating is <0.2 degrees C. The visibility of transponders and bony anatomy was not affected on the OBI kV and CT images. Image quality degradation caused by the detector antenna array is observed in the CBCT image. Image artifacts are most significant with the gradient echo sequence in the MR images, producing null signals surrounding the transponders with radii approximately 1.5 cm and length approximately 4 cm. Thus, Calypso transponders can preclude the use of MRI/MRS in post-treatment assessment. Modifications of the clinical flow are required to accommodate and minimize the substantial MRI artifacts induced by the Calypso transponders.

Volumetric-Modulated Arc Therapy: Effective and Efficient End-to-End Patient-Specific Quality Assurance

PURPOSE To explore an effective and efficient end-to-end patient-specific quality-assurance (QA) protocol for volumetric modulated arc radiotherapy (VMAT) and to evaluate the suitability of a stationary radiotherapy QA device (two-dimensional [2D] ion chamber array) for VMAT QA. METHODS AND MATERIALS Three methods were used to analyze 39 VMAT treatment plans for brain, spine, and prostate: ion chamber (one-dimensional absolute, n = 39), film (2D relative, coronal/sagittal, n = 8), and 2D ion chamber array (ICA, 2D absolute, coronal/sagittal, n = 39) measurements. All measurements were compared with the treatment planning system dose calculation either via gamma analysis (3%, 3- to 4-mm distance-to-agreement criteria) or absolute point dose comparison. The film and ion chamber results were similarly compared with the ICA measurements. RESULTS Absolute point dose measurements agreed well with treatment planning system computed doses (ion chamber: median deviation, 1.2%, range, -0.6% to 3.3%; ICA: median deviation, 0.6%, range, -1.8% to 2.9%). The relative 2D dose measurements also showed good agreement with computed doses (>93% of pixels in all films passing gamma, >90% of pixels in all ICA measurements passing gamma). The ICA relative dose results were highly similar to those of film (>90% of pixels passing gamma). The coronal and sagittal ICA measurements were statistically indistinguishable by the paired t test with a hypothesized mean difference of 0.1%. The ion chamber and ICA absolute dose measurements showed a similar trend but had disparities of 2-3% in 18% of plans. CONCLUSIONS After validating the new VMAT implementation with ion chamber, film, and ICA, we were able to maintain an effective yet efficient patient-specific VMAT QA protocol by reducing from five (ion chamber, film, and ICA) to two measurements (ion chamber and single ICA) per plan. The ICA (Matrixx®, IBA Dosimetry) was validated for VMAT QA, but ion chamber measurements are recommended for absolute dose comparison until future developments correct the ICA angular dependence.

Adaptive radiation therapy: technical components and clinical applications.

In current standard radiation therapy process, patient anatomy is represented by the snapshot of computed tomographic images at the simulation for treatment planning. However, patient anatomy during the treatment course is not static, and the changes can be in the order of centimeters. The goal of the adaptive radiation therapy (ART) is to measure and account these variations in the treatment process, so that the optimal planned dose distribution is the same as the final delivered dose distribution. The field of the ART is rapidly evolving. The implementation of the ART principle is built on technical components in 3 main areas: image guidance, dose verification, and plan adaptation. The purpose of this review was to present different ART methods currently developed and used by different investigators.

Analysis of image quality for real-time target tracking using simultaneous kV-MV imaging.

Real-time tracking can provide high accuracy localization for a moving target and minimize the effect of motion. Simultaneous kV-MV imaging has been proposed as a real-time tracking technique by utilizing the existing kV on-board imager (OBI) and the MV electronic portal device (EPID) mounted on the linear accelerator. The orthogonal pair of kV-MV images acquired simultaneously can provide 3-D localization in real-time. However, the kV and MV beams cross shooting the target interfere with each other with beam scattering, which affects the quality of images. The success of this modality heavily relies on the image quality, especially the visibility of the target, which was investigated in this study. The kV and MV images were acquired for a gold implant marker that was used as a surrogate of the target and placed in an IMRT thorax phantom, a dynamic phantom, and a pelvis phantom to test the image quality in different situations. Contrast-to-noise ration (CNR) was used to quantitatively describe the visibility of the target in the image. CNR can be obtained by statistical calculation from image processing and physics analysis with ion chamber measurement. The difference is described by contrast detection efficiency (CDE). By comparing the ratio (R) of CNR with and without the MV beam on, the MV beam scatter was found to have dramatically reduced the target visibility in the kV images (R=0.47), which was supported by an independent physics analysis that treats beam scatter as a noise. In contrast, the kV scatter effect on the MV images was minor (R=0.93). The effect of tumor motion was visible but tolerable for the target tracking purpose. CNR varied with different tumor sites and was lower for the pelvis than the thorax. Different kV imaging parameters such as kVp, mAs, and exposure time ms were tested for different cases. Considering a threshold of 1.0 CNR as a measure for the target visibility, a range of CNR from 1.3 to 4.2 was reached with appropriate tuning of those imaging parameters. This study has shown that CNR is a key parameter that can be used for assessing the visibility of the target in digital imaging and the quality of kV/MV images. It has also been shown that reasonable target visibility can be obtained using simultaneous kV-MV imaging for real-time target tracking.

Dosimetric comparison of 3D conformal, IMRT, and V-MAT techniques for accelerated partial-breast irradiation (APBI).

The purpose is to dosimetrically compare the following 3 delivery techniques: 3-dimensional conformal radiation therapy (3D-CRT), intensity-modulated arc therapy (IMRT), and volumetric-modulated arc therapy (V-MAT) in the treatment of accelerated partial-breast irradiation (APBI). Overall, 16 patients with T1/2N0 breast cancer were treated with 3D-CRT (multiple, noncoplanar photon fields) on the RTOG 0413 partial-breast trial. These cases were subsequently replanned using static gantry IMRT and V-MAT technology to understand dosimetric differences among these 3 techniques. Several dosimetric parameters were used in plan quality evaluation, including dose conformity index (CI) and dose-volume histogram analysis of normal tissue coverage. Quality assurance studies including gamma analysis were performed to compare the measured and calculated dose distributions. The IMRT and V-MAT plans gave more conformal target dose distributions than the 3D-CRT plans (p < 0.05 in CI). The volume of ipsilateral breast receiving 5 and 10Gy was significantly less using the V-MAT technique than with either 3D-CRT or IMRT (p < 0.05). The maximum lung dose and the ipsilateral lung volume receiving 10 (V10) or 20Gy (V20) were significantly less with both V-MAT and IMRT (p < 0.05). The IMRT technique was superior to 3D-CRT and V-MAT of low dose distributions in ipsilateral lung (p < 0.05 in V5 and D5). The total mean monitor units (MUs) for V-MAT (621.0 ± 111.9) were 12.2% less than those for 3D-CRT (707.3 ± 130.9) and 46.5% less than those for IMRT (1161.4 ± 315.6) (p < 0.05). The average machine delivery time was 1.5 ± 0.2 minutes for the V-MAT plans, 7.0 ± 1.6 minutes for the 3D-CRT plans, and 11.5 ± 1.9 minutes for the IMRT plans, demonstrating much less delivery time for V-MAT. Based on this preliminary study, V-MAT and IMRT techniques offer improved dose conformity as compared with 3D-CRT techniques without increasing dose to the ipsilateral lung. In terms of MU and delivery time, V-MAT is significantly more efficient for APBI than for conventional 3D-CRT and static-beam IMRT.

SU‐E‐T‐553: Measurement of Incident Electron Spots On TrueBeam

PURPOSE Lateral dimensions of the incident electron beam (spot sizes) are essential for accurate modelling of x-ray fields, especially for small field output factors and penumbrae. We present spot sizes measured on the Varian TrueBeam linac and propose an explanation for the observed shapes. METHODS Spot were measured on 3 TrueBeam linacs using the 4X, 6X, 6FFF, 8X, 10X, 10FFF, and 15X beams. Two methods were used: a spot camera, and an edge CT reconstruction based on images of a thin plate with edge intersecting the beam at different collimator rotations. 2-dimensional Gaussian functions were fit to the data, with the spot orientation used as a fitting parameter. RESULTS Spots were consistent for the 3 linacs. Sigmas of the fitted Gaussians were between 0.47 and 1.05 mm for all the beams. The 6X, 6FFF, and 15X beams had a distinct shape: the major axes were 30-40% larger than the minor axes, and the major axes were oriented 40 degrees from the inplane direction. The remaining spots were more symmetric, with major axes at most 15% larger than minor, and a similar orientation. The spot shapes could be explained by positing a helical electron motion at the entrance to the bend magnet, and using the known (non-achromatic) transport through the magnet. Fringe fields of the bend magnet, solenoid, or steering coils could produce the helical motion; the precise source is not known. Spots measured with the edge technique were 10% smaller in size than those measured with the camera. The 10FFF spots were 10% smaller than the 10X spots, as measured with the camera, but the 6X and 6FFF spots were similar to each other. CONCLUSION Representative spots for TrueBeam are presented. Values may be used for input to modelling of fluence output, for example using Monte Carlo methods. Several of the authors are employees of Varian Medical Systems, as noted in the affiliations.

Outlier identification in radiation therapy knowledge‐based planning: A study of pelvic cases

Purpose: The purpose of this study was to apply statistical metrics to identify outliers and to investigate the impact of outliers on knowledge‐based planning in radiation therapy of pelvic cases. We also aimed to develop a systematic workflow for identifying and analyzing geometric and dosimetric outliers. Methods: Four groups (G1–G4) of pelvic plans were sampled in this study. These include the following three groups of clinical IMRT cases: G1 (37 prostate cases), G2 (37 prostate plus lymph node cases) and G3 (37 prostate bed cases). Cases in G4 were planned in accordance with dynamic‐arc radiation therapy procedure and include 10 prostate cases in addition to those from G1. The workflow was separated into two parts: 1. identifying geometric outliers, assessing outlier impact, and outlier cleaning; 2. identifying dosimetric outliers, assessing outlier impact, and outlier cleaning. G2 and G3 were used to analyze the effects of geometric outliers (first experiment outlined below) while G1 and G4 were used to analyze the effects of dosimetric outliers (second experiment outlined below).A baseline model was trained by regarding all G2 cases as inliers. G3 cases were then individually added to the baseline model as geometric outliers. The impact on the model was assessed by comparing leverages of inliers (G2) and outliers (G3). A receiver‐operating‐characteristic (ROC) analysis was performed to determine the optimal threshold. The experiment was repeated by training the baseline model with all G3 cases as inliers and perturbing the model with G2 cases as outliers.A separate baseline model was trained with 32 G1 cases. Each G4 case (dosimetric outlier) was subsequently added to perturb the model. Predictions of dose‐volume histograms (DVHs) were made using these perturbed models for the remaining 5 G1 cases. A Weighted Sum of Absolute Residuals (WSAR) was used to evaluate the impact of the dosimetric outliers. Results: The leverage of inliers and outliers was significantly different. The Area‐Under‐Curve (AUC) for differentiating G2 (outliers) from G3 (inliers) was 0.98 (threshold: 0.27) for the bladder and 0.81 (threshold: 0.11) for the rectum. For differentiating G3 (outlier) from G2 (inlier), the AUC (threshold) was 0.86 (0.11) for the bladder and 0.71 (0.11) for the rectum. Significant increase in WSAR was observed in the model with 3 dosimetric outliers for the bladder (P < 0.005 with Bonferroni correction), and in the model with only 1 dosimetric outlier for the rectum (P < 0.005). Conclusions: We established a systematic workflow for identifying and analyzing geometric and dosimetric outliers, and investigated statistical metrics for outlier detection. Results validated the necessity for outlier detection and clean‐up to enhance model quality in clinical practice.

MO‐D‐BRB‐09: Treatment Delivery QA for Online Adaptive Radiation Therapy Based on Dynamic Machine Information (DMI): A Feasibility Study

PURPOSE To implement a quality assurance (QA) system for the treatment delivery of online adaptive radiation therapy utilizing Dynamic Machine Information (DMI). MATERIALS AND METHODS DMI provides the expected/actual MLC leaf-positions, delivered MU, and beam-on status every 50ms during delivery. In this study a stream of DMI inputs is simulated by playing back Dynalog information recorded while delivering a test fluence map (FM). Based on these DMI inputs, the QA system performs three levels of monitoring/verification on the plan delivery process: (1) Following each input, actual and expected FMs delivered up to the current MLC position is dynamically updated using corresponding MLC positions in the DMI. The magnitude and frequency of pixel-by-pixel fluence differences between these two FMs are calculated and visualized in histograms.(2) At each control point, actual MLC positions are verified against the treatment plan for potential errors in data transfer between the treatment planning system (TPS) and the MLC controller.(3) Both (1) and (2) can signal beam-hold with a user-specified error tolerance.(4) After treatment, delivered dose is reconstructed in TPS based on DMI data during delivery, and compared to planned dose. RESULTS (1) Efficiency: Average latency from DMI input to the completion of fluence difference calculation is <1ms.(2) Efficacy: For test FM, transient error in leaf positions is (-0.07±0.28)mm; cumulative errors in delivered fluence is (0.003±0.183)% of the maximal fluence. The system can also identify data transfer errors between TPS and MLC controller. Off-line dose reconstruction and evaluation show <0.5% dosimetric discrepancy from planned dose distribution for the test FM. CONCLUSION This QA system is capable of identifying MLC position/fluence errors in near real-time, and assessing dosimetric impact of the treatment delivery process. It is thus a valuable tool for clinical implementation of online adaptive radiation therapy. (Research partially supported by Varian) Research partially supported by Varian Medical Systems.

MO‐D‐BRC‐07: Reducing Artifacts in Cone‐Beam CT Images Caused by the Presence of An Array Used for Tracking Transponders during Radiotherapy

Purpose: To compare two methods for reducing artifacts in cone‐beam CT(CBCT)images caused by the presence of an electromagnetic array used for tracking transponders during radiotherapy. Methods: A pelvis phantom was imaged using a gantry mounted CBCT system at two dose levels typical for prostate imaging. Both dose protocols acquired approximately 670 projections at 125 kVp. Tube currents and pulse widths were 80 mA/13 ms and 63 mA/10 ms for standard and low dose scans. The phantom was imaged with and without the array in position. The two methods used to reduce artifacts were: 1) filtering projection images prior to reconstruction, and 2) halving the dose per projection and doubling the number of projections. The filtering method replaced pixel values with neighboring values when pixels differed from neighboring pixels by a threshold percentage. 18, 22, 26 and 30% thresholds were investigated. Artifact reduction was quantified by computing difference images between the ‘corrected’ reconstructions and the image reconstructions for the phantom without the presence of the array. The filtering method was also applied to a prostate patient data set. Results: Average pixel values of the difference images were reduced by 6.4 and 7.3% for the 18% threshold filtered reconstructions for low and high doses, respectively, compared with difference images between the unaltered reconstructions with and without the panel. Reductions were 4.4 and 4.9% for other thresholds for low and high doseimages. 8.6 and 19.6% reductions were calculated for low and high doseimages using increased projection numbers. Improved soft tissuevisibility was noted in the filtered patient reconstructions. Conclusions: Both filtering and increasing projection numbers reduced artifacts caused by the presence of an electromagnetic array in CBCTimaging. Quantitative and visual analyses suggested that greater artifact reduction was achieved with increased projection numbers.

SU-E-T-229: Machine Learning Methods for Knowledge Based Treatment Planning of Prostate Cancer

PURPOSE To evaluate the accuracy of the dose prediction models constructed with machine learning techniques, Support Vector Machine (SVM) and Artificial Neural Network (ANN) for the prediction of dose volume histogram (DVH) of organs-at-risk (OAR) in IMRT, compared to the model constructed by stepwise multiple regression (MR), and to investigate the number of prior plans required for the models to produce reliable predictions. METHODS IMRT plans from 102 prostate cases were randomly divided into two datasets for training and testing, respectively. The testing dataset contains a fixed number of 20 cases, while the number of cases in the training dataset varied from 5 to 80. Models were constructed with SVM, ANN, or MR to formulate the dependence of the OAR DVH on patient anatomical features including the Distance to Target Histogram (DTH), PTV and OAR volumes and their overlap, among other volumetric or spatial information. The D50 (Dose value at 50% volume) and the mean square of difference between D50 of clinical and predicted DVH were calculated for each modeling technique at each specific training dataset number. RESULTS The mean square of difference of D50 between clinical and predicted DVH decreases with the number of cases in the training dataset, and reaches stable beyond 30 for MR. With the 80 case training dataset, for the bladder model, the SVM predicted 70% D50 values within 10% error and the ANN predicted 85%, compared to 85% with multiple regression. For the rectum model, the numbers are SVM 80%, ANN 70%, and MR 85%. CONCLUSION The machine learning techniques SVM and ANN are comparable to MR for producing OAR DVH prediction of the prostate cancer. The minimal number of training cases is around 30. Partially supported by NIH/NCI under grant #R21CA161389 and a master research grant by Varian Medical System.