Synchrony Respiratory Tracking System Research Articles

Evaluation of Lung and Liver Tumor Dose Coverage Treated With the CyberKnife Synchrony System With Consideration of Measured Tracking Errors.

Lung and liver tumor dose coverage was evaluated for the CyberKnife synchrony respiratory tracking system (SRTS) with consideration of the motion tracking accuracy measured for motion patterns of individual patients. Seven treatment plans of six cases treated with the SRTS were evaluated. The motion phantom was moved with the motion data derived from the treatment log files. A laser emitted from the linac head to the moving phantom block was recorded with a webcam, and the tracking accuracy was evaluated. The dose volume histogram (DVH) of planning target volume (PTV) and gross tumor volume (GTV) were calculated by a pencil beam algorithm with shifting the beams with Gaussian random numbers mimicking the measured tracking errors. The tracking errors measured with the motion phantom in the lateral direction were within ±2 mm for 90% of beam-on time. The tracking errors in the longitudinal direction were within ±3.0 mm and ±1.1 mm for 90% and 50% of beam-on time, respectively. Although one case showed a decrease in the dose covering 95% of PTV (D95%) by 1.8%, the change in the dose covering 99% of GTV (D99%) was within 1%. This study evaluated the motion tracking errors of the SRTS by a motion phantom moved with the patients' respiration signal, and the impact of the tracking errors on the target coverage was calculated. Even for respiratory patterns with large maximum tracking errors, sufficient GTV coverage is achievable if the beam is accurately delivered for high percentage of beam-on time.

Dosimetric Evaluation of CyberKnife Synchrony System for Liver Tumors With Respiratory Phase Shifts.

This study evaluated the effects of the respiratory phase shifts between liver tumor and chest wall motions on the dose distribution for the CyberKnife Synchrony respiratory tracking system (SRTS). Eight patients who received stereotactic body radiotherapy for hepatocellular carcinoma or liver metastases were analyzed. Three-dimensional (3D) motion of the implanted fiducial markers and vertical motion of the sternal bone were derived from the four-dimensional computed tomography (4D-CT) images acquired with a 320-row area detector CT. For each patient, Gaussian random numbers were generated for the standard deviation of the tracking error calculated from the phase shift and a literature. For each voxel of the target, the dose delivered from each beam was calculated 100 times with the random 3D offsets representing the tracking error. The median respiratory phase shifts were 6.0% and 4.6% for the anterior-posterior (AP) and superior-inferior (SI) directions, respectively. The median motion tracking errors influenced by respiratory phase shifts were 1.21 mm and 0.96 mm for the AP and SI directions, respectively. The change in the dose covering 90% of the target (D90%) was within 1.1% when median phase shifts were considered. When evaluating the 90th percentile of the phase shifts, the D90% decreased up to 6.6%. We have developed a technique to estimate the impact of the respiratory phase shifts on the dose distribution of a liver tumor treated with the SRTS. The calculation of the respiratory phase shifts from the area-detector 4D-CT will be valuable to improve the tracking accuracy of the SRTS.

Study on Motion Management of Pancreatic Cancer Treated by CyberKnife.

PurposeWe investigated the movement characteristics of pancreas and the clinical accuracy of tracking pancreas with the Synchrony Respiratory Tracking System (SRTS) during the CyberKnife treatment. These data provide a clinical data basis for the expansion margins of pancreatic tumor target.Methods and MaterialsForty-two patients with pancreatic cancer treated by CyberKnife were retrospectively studied. The pancreatic displacement calculated from the x-ray images collected during the time interval between two consecutive movements constituted a data set.ResultsThe total mean motion amplitudes and standard deviations of pancreatic tumors in SI, LR, AP, and radial directions were 3.66 ± 1.71 mm, 0.97 ± 0.62 mm, 1.52 ± 1.02 mm, and 1.36 ± 0.49 mm, respectively. The overall mean correlation errors and standard deviations were 0.82 ± 0.46 mm, 0.47 ± 0.33 mm, 0.41 ± 0.24 mm, and 0.98 ± 0.37 mm, respectively. The overall mean prediction errors and standard deviations were 0.57 ± 0.14 mm, 0.62 ± 0.28 mm, 0.39 ± 0.17 mm, and 1.58 ± 0.36 mm, respectively. The correlation errors and prediction errors of pancreatic tumors at different anatomical positions in SI, LR, and AP directions were statistically significant (p < 0.05).ConclusionsThe tumor motion amplitude, the tumor location, and the treatment time are the main factors affecting the tracking accuracy. The pancreatic tumors at different anatomical locations should be treated differently to ensure sufficient dose coverage of the pancreatic target area.

Development of a tracking error prediction system for the CyberKnife Synchrony Respiratory Tracking System with use of support vector regression.

The accuracy of the CyberKnife Synchrony Respiratory Tracking System is dependent on the breathing pattern of a patient. Therefore, the tracking error in each patient must be determined. Support vector regression (SVR) can be used to easily identify the tracking error in each patient. This study aimed to develop a system with SVR that can predict tracking error according to a patient's respiratory waveform. Datasets of the respiratory waveforms of 93 patients were obtained. The feature variables were variation in respiration amplitude, tumor velocity, and phase shift between tumor and the chest wall, and the target variable was tracking error. A learning model was evaluated with tenfold cross-validation. We documented the difference between the predicted and actual tracking errors and assessed the correlation coefficient and coefficient of determination. The average difference and maximum difference between the actual and predicted tracking errors were 0.57 ± 0.63 mm and 2.1 mm, respectively. The correlation coefficient and coefficient of determination were 0.86 and 0.74, respectively. We developed a system for obtaining tracking error by using SVR. The accuracy of such a system is clinically useful. Moreover, the system can easily evaluate tracking error. We developed a system that can be used to predict the tracking error of SRTS in the CyberKnife Robotic Radiosurgery System using machine learning. The feature variables were the breathing parameters, and the target variable was the tracking error. We used support vector regression algorithm.

Comparison of modeling accuracy between Radixact® and CyberKnife® Synchrony® respiratory tracking system

Synchrony Respiratory Tracking system adapted from CyberKnife has been introduced in Radixact to compensate the tumor motion caused by respiration. This study aims to compare the modeling accuracy of the Synchrony system between Radixact and CyberKnife. Two Synchrony plans based on fiducial phantoms were created for CyberKnife and Radixact, respectively. Different respiratory motion traces were used to drive a motion platform to move along the superoinferior and left-right direction. The cycle time and the amplitude of target/surrogate motion of one selected motion trace were scaled to investigate the dependence of modeling accuracy on the motion characteristic. The predicted target position, the correlation error, potential difference (Radixact only) and standard error (CyberKnife only) were extracted from raw data or log files of the two systems. The modeling accuracy was evaluated by calculating the root-mean-square (RMS) error between the predicted target positions and the input motion trace. A threshold T95 within which 95% of the potential difference or the standard error lay was defined and evaluated. Except for the motion trace with a small amplitude and a good (linear) correlation between target and surrogate motion, Radixact showed smaller RMS errors than CyberKnife. The RMS error of both systems increased with the motion amplitude and showed a decreasing trend with the increasing cycle time. No correlation was found between the RMS error and the amplitude of surrogate motion. T95 could be a good estimator of modeling accuracy for CyberKnife rather than Radixact. The correlation error defined in Radixact were largely affected by the number of fiducial markers and the setup error. In general, the modeling accuracy of the Radixact Synchrony system is better than that of the CyberKnife Synchrony system under unfavorable conditions.

Cosmetic Outcomes of a Phase 1 Dose Escalation Study of 5-Fraction Stereotactic Partial Breast Irradiation for Early Stage Breast Cancer

Our purpose was to evaluate cosmetic changes after 5-fraction adjuvant stereotactic partial breast irradiation (S-PBI). Seventy-five women with in situ or invasive breast cancer stage 0, I, or II, with tumor size ≤3 cm, were enrolled after lumpectomy in a phase 1 dose escalation trial of S-PBI into cohorts receiving 30, 32.5, 35, 37.5, or 40 Gy in 5 fractions. Before S-PBI, 3 to 4 gold fiducial markers were placed in the lumpectomy cavity for tracking with the Synchrony respiratory tracking system. S-PBI was delivered with a CyberKnife robotic radiosurgery system. Patients and physicians evaluated global cosmesis using the Harvard Breast Cosmesis Scale. Eight independent panelists evaluated digital photography for global cosmesis and 10 subdomains at baseline and follow-up. McNemar tests were used to evaluate change in cosmesis, graded as excellent/good or fair/poor, from baseline to year 3. Wilcoxon signed rank tests were used to evaluate change in subdomains. Cohen's kappa (κ) statistic was used to estimate interobserver agreement (IOA) between raters, and Fleiss' κ was used to estimate IOA between panelists. Median cosmetic follow-up was 5, 5, 5, 4, and 3 years for the 30, 32.5, 35, 37.5, and 40 Gy cohorts. Most patients reported excellent/good cosmesis at both baseline (86.3%) and year 3 (89.8%). No dose cohort had significantly worsened cosmesis by year 3 on McNemar analysis. No cosmetic subdomain had significant worsening by year 3. IOA was fair for patient-physician (κ = 0.300, P < .001), patient-panel (κ = 0.295, P < .001), physician-panel (κ = 0.256, P < .001), and individual panelists (Fleiss κ = 0.327, P < .001). Dose escalation of S-PBI from 30 to 40 Gy in 5 fractions for early stage breast cancer was not associated with a detectable change in cosmesis by year 3. S-PBI is a promising modality for treatment of early stage breast cancer.

Factors affecting the accuracy of respiratory tracking of the image-guided robotic radiosurgery system.

To analyze the factors affecting the tracking accuracy of the CyberKnife Synchrony Respiratory Tracking System (SRTS). A dynamic motion phantom (motion phantom) reproduced the respiratory motions of each patient treated with the SRTS using a ball as the target. CyberKnife tracked the ball using the SRTS, and this process was recorded by a video camera mounted on the linear accelerator head. The tracking error was evaluated from the images captured by the video camera. Multiple regression analysis was used to identify factors affecting tracking accuracy from 91 cases. The median tracking error was 1.9mm (range 0.9-5.3mm). Four factors affected the tracking accuracy: the average absolute amplitude of the tumor motion in the cranio-caudal (CC) direction (p = 0.007), average position gap due to the phase shift between the internal tumor and external marker positions in the CC direction (p < 0.001), and average velocity of the tumor in the CC (p < 0.001) and anterior-posterior directions (p = 0.033). We identified factors that affected tracking accuracy. This information may assist the identification of suitable margins that should be added to each patient's clinical target volume.

Clinical efficacy of CyberKnife stereotactic body radiation therapy for hepatic hilar cancer

Objective To evaluate the efficacy and safety of CyberKnife stereotactic body radiation therapy (SBRT) in the treatment of hepatic hilar cancer. Methods A retrospective study was performed on the clinical data of 36 patients with hepatic hilar cancer who were admitted to our hospital and treated with CyberKnife SBRT from 2009 to 2015. In the 36 patients, 37 lesions were found with tumor diameters ranging from 1.5 to 5.5 mm (median diameter 3 cm). The Synchrony respiratory tracking system was used for 21 lesions in 20 patients, while the XSight spinal tracking system was used for 16 lesions in 16 patients. Local progression was evaluated based on contrast-enhanced computed tomography and/or magnetic resonance imaging. The Kaplan-Meier method was used to calculate local control (LC) and overall survival (OS) rates, and the log-rank test was used for survival comparison and univariate prognostic analysis. Results The median follow-up time was 12.7 months. The 1-and 2-year postoperative LC rates were 90% and 76%, respectively. The 1-year OS and progression-free survival (PFS) rates were 63% and 39%, respectively. The median OS and PFS times were 15.2 and 10 months, respectively. The incidence of grade 3 adverse reactions was 11%. Conclusions The CyberKnife SBRT is a safe and effective way to treat hepatic hilar cancer. Key words: Liver neoplasms/stereotactic radiotherapy; Stereotactic radiotherapy, CyberKnife; Treatment outcome

Effect of density heterogeneity on absorbed dose with CyberKnife Synchrony Respiratory Tracking System

Objective To measure the actual absorbed dose of the target in the QUASAR Respiratory Motion Phantom using the CyberKnife Synchrony Respiratory Tracking System, and to evaluate the effect of density heterogeneity on the absorbed dose of tumor gross target volume (GTV). Methods Nine groups were obtained by making different patterns of QUASAR phantom: rib thickness of 0, 20, and 50 mm, and motion amplitudes of 0, 10, and 15 mm. The nine groups were treated with static computed tomography (CT) in different time phases of four-dimensional CT (4DCT) plan, with the same beam and number of monitor units, and the 4D accumulated dose was calculated. The doses of static and 4D plans were calculated using Ray-tracing and Monte Carlo algorithms, and the absorbed doses of GTV in the nine groups were measured at the same time. Results There were a decrease in calculated absorbed dose of GTV and an increase in deviation between the planned and actual dose, with the increases in simulated rib thickness and motion amplitude. Conclusions The density heterogeneity has an impact on the absorbed dose of GTV. Both static CT and 4DCT plan can evaluate the absorbed dose of GTV in case of small rib thickness and motion amplitude, and 4DCT plan with Monte Carlo algorithm may be the optimal method for evaluation of the absorbed dose of GTV in case of large rib thickness and motion amplitude (deviation<3%) Key words: CyberKnife; Synchrony system; Tissue heterogeneous; Absorbed dose

Validation of a pretreatment delivery quality assurance method for the CyberKnife Synchrony system.

To evaluate the geometric and dosimetric accuracies of the CyberKnife Synchrony respiratory tracking system (RTS) and to validate a method for pretreatment patient-specific delivery quality assurance (DQA). An EasyCube phantom was mounted on the ExacTrac gating phantom, which can move along the superior-inferior (SI) axis of a patient to simulate a moving target. The authors compared dynamic and static measurements. For each case, a Gafchromic EBT3 film was positioned between two slabs of the EasyCube, while a PinPoint ionization chamber was placed in the appropriate space. There were three steps to their evaluation: (1) the field size, the penumbra, and the symmetry of six secondary collimators were measured along the two main orthogonal axes. Dynamic measurements with deliberately simulated errors were also taken. (2) The delivered dose distributions (from step 1) were compared with the planned ones, using the gamma analysis method. The local gamma passing rates were evaluated using three acceptance criteria: 3% local dose difference (LDD)/3 mm, 2%LDD/2 mm, and 3%LDD/1 mm. (3) The DQA plans for six clinical patients were irradiated in different dynamic conditions, to give a total of 19 cases. The measured and planned dose distributions were evaluated with the same gamma-index criteria used in step 2 and the measured chamber doses were compared with the planned mean doses in the sensitive volume of the chamber. (1) A very slight enlargement of the field size and of the penumbra was observed in the SI direction (on average <1 mm), in line with the overall average CyberKnife system error for tracking treatments. (2) Comparison between the planned and the correctly delivered dose distributions confirmed the dosimetric accuracy of the RTS for simple plans. The multicriteria gamma analysis was able to detect the simulated errors, proving the robustness of their method of analysis. (3) All of the DQA clinical plans passed the tests, both in static and dynamic conditions. No statistically significant differences were found between static and dynamic cases, confirming the high degree of accuracy of the Synchrony RTS. The presented methods and measurements verified the mechanical and dosimetric accuracy of the Synchrony RTS. Their method confirms the fact that the RTS, if used properly, is able to treat a moving target with great precision. By combining PinPoint ion chamber, EBT3 films, and gamma evaluation of dose distributions, their DQA method robustly validated the effectiveness of CyberKnife and Synchrony system.

Preparations for the implementation of Cyberknife Lung SBRT

Lung cancer is the fourth most common cancer in Ireland with over 2000 people diagnosed each year. Cyberknife Lung treatments allow for a high dose per fraction while greatly reducing normal tissue toxicity and treatment times. This is achieved using over one hundred non coplanar beams allowing for high dose conformality coupled with a steep dose falloff away from the tumour. With these properties Cyberknife allows for dose escalation and smaller margins to be used when compared to gantry based systems. The Synchrony Respiratory Motion tracking System correlated with live KV X-ray tumour tracking allows the Cyberknife to track the tumour in real-time. There are 2 primary methods for tumour tracking with this system (1) Fiducial based tumour tracking (2) algorithm based Soft tissue tracking. In this study we look at both methods and use E2E (End to End) tests and EBT3 Gafchromic films to quantify the error in both tumour tracking systems. A further enhancement of soft tissue tumour tracking known as LOT(Lung optimised treatment)is also evaluated in this study with a view towards deciding on appropriate clinical margins in situations where the tumour is visible in zero, one or both Cyberknife KV imagers. This preliminary study will focus on the tumour tracking and delivery accuracy of Lung SBRT treatments using Cyberknife. The equipment used for this study: Cyberknife VSI, Synchrony phantom, CIRS 4D phantom, Synchrony LED Vest, Fiducials, Mulitplan v4.6.1, CT scanner, and GafChromic films (EBT3).

Development of system using beam's eye view images to measure respiratory motion tracking errors in image-guided robotic radiosurgery system.

The accuracy of the CyberKnife Synchrony Respiratory Tracking System (SRTS) is considered to be patient‐dependent because the SRTS relies on an individual correlation between the internal tumor position (ITP) and the external marker position (EMP), as well as a prediction method to compensate for the delay incurred to adjust the position of the linear accelerator (linac). We aimed to develop a system for obtaining pretreatment statistical measurements of the SRTS tracking error by using beam's eye view (BEV) images, to enable the prediction of the patient‐specific accuracy. The respiratory motion data for the ITP and the EMP were derived from cine MR images obtained from 23 patients. The dynamic motion phantom was used to reproduce both the ITP and EMP motions. The CyberKnife was subsequently operated with the SRTS, with a CCD camera mounted on the head of the linac. BEV images from the CCD camera were recorded during the tracking of a ball target by the linac. The tracking error was measured at 15 Hz using in‐house software. To assess the precision of the position detection using an MR image, the positions of test tubes (determined from MR images) were compared with their actual positions. To assess the precision of the position detection of the ball, ball positions measured from BEV images were compared with values measured using a Vernier caliper. The SRTS accuracy was evaluated by determining the tracking error that could be identified with a probability of more than 95% (Ep95). The detection precision of the tumor position (determined from cine MR images) was <0.2 mm. The detection precision of the tracking error when using the BEV images was <0.2 mm. These two detection precisions were derived from our measurement system and were not obtained from the SRTS. The median of Ep95 was found to be 1.5 (range, 1.0–3.5) mm. The difference between the minimum and maximum Ep95 was 2.5 mm, indicating that this provides a better means of evaluating patient‐specific SRTS accuracy. A suitable margin, based on the predicted patient‐specific SRTS accuracy, can be added to the clinical target volume.PACS number: 87.53.Ly

Evaluation of Dose Uncertainty to the Target Associated With Real-Time Tracking Intensity-Modulated Radiation Therapy Using the CyberKnife Synchrony System.

We investigated the dose uncertainty caused by errors in real-time tracking intensity-modulated radiation therapy (IMRT) using the CyberKnife Synchrony Respiratory Tracking System (SRTS). Twenty lung tumors that had been treated with non-IMRT real-time tracking using CyberKnife SRTS were used for this study. After validating the tracking error in each case, we did 40 IMRT planning using 8 different collimator sizes for the 20 patients. The collimator size was determined for each planning target volume (PTV); smaller ones were one-half, and larger ones three-quarters, of the PTV diameter. The planned dose was 45 Gy in 4 fractions prescribed at 95% volume border of the PTV. Thereafter, the tracking error in each case was substituted into calculation software developed in house and randomly added in the setting of each beam. The IMRT planning incorporating tracking errors was simulated 1000 times, and various dose data on the clinical target volume (CTV) were compared with the original data. The same simulation was carried out by changing the fraction number from 1 to 6 in each IMRT plan. Finally, a total of 240 000 plans were analyzed. With 4 fractions, the change in the CTV maximum and minimum doses was within 3.0% (median) for each collimator. The change in D99 and D95 was within 2.0%. With decreases in the fraction number, the CTV coverage rate and the minimum dose decreased and varied greatly. The accuracy of real-time tracking IMRT delivered in 4 fractions using CyberKnife SRTS was considered to be clinically acceptable.

Retrospective evaluation of CTV to PTV margins using CyberKnife in patients with thoracic tumors.

The objectives of this study were to estimate global uncertainty for patients with thoracic tumors treated in our center using the CyberKnife VSI after placement of fiducial markers and to compare our findings with the standard CTV to PTV margins used to date. Datasets for 16 patients (54 fractions) treated with the CyberKnife and the Synchrony Respiratory Tracking System were analyzed retrospectively based on CT planning, tracking information, and movement data generated and saved in the logs files by the system. For each patient, we analyzed all the main uncertainty sources and assigned a value. We also calculated an expanded global uncertainty to ensure a robust estimation of global uncertainty and to enable us to determine the position of 95% of the CTV points with a 95% confidence level during treatment. Based on our estimation of global uncertainty and compared with our general margin criterion (5 mm in all three directions: superior/inferior [SI], anterior/posterior [AP], and lateral [LAT]), 100% were adequately covered in the LAT direction, as were 94% and 94% in the SI and AP directions. We retrospectively analyzed the main sources of uncertainty in the CyberKnife process patient by patient. This individualized approach enabled us to estimate margins for patients with thoracic tumors treated in our unit and compare the results with our standard 5 mm margin.PACS number: 87.55‐x

Correlation and prediction uncertainties in the CyberKnife Synchrony respiratory tracking system

The CyberKnife uses an online prediction model to improve radiation delivery when treating lung tumors. This study evaluates the prediction model used by the CyberKnife radiation therapy system in terms of treatment margins about the gross tumor volume (GTV). From the data log files produced by the CyberKnife synchrony model, the uncertainty in radiation delivery can be calculated. Modeler points indicate the tracked position of the tumor and Predictor points predict the position about 115 ms in the future. The discrepancy between Predictor points and their corresponding Modeler points was analyzed for 100 treatment model data sets from 23 de-identified lung patients. The treatment margins were determined in each anatomic direction to cover an arbitrary volume of the GTV, derived from the Modeler points, when the radiation is targeted at the Predictor points. Each treatment model had about 30 min of motion data, of which about 10 min constituted treatment time; only these 10 min were used in the analysis. The frequencies of margin sizes were analyzed and truncated Gaussian normal functions were fit to each direction's distribution. The standard deviation of each Gaussian distribution was then used to describe the necessary margin expansions in each signed dimension in order to achieve the desired coverage. In this study, 95% modeler point coverage was compared to 99% modeler coverage. Two other error sources were investigated: the correlation error and the targeting error. These were added to the prediction error to give an aggregate error for the CyberKnife during treatment of lung tumors. Considering the magnitude of 2sigma from the mean of the Gaussian in each signed dimension, the margin expansions needed for 95% modeler point coverage were 1.2 mm in the lateral (LAT) direction and 1.7 mm in the anterior-posterior (AP) direction. For the superior-inferior (SI) direction, the fit was poor; but empirically, the expansions were 3.5 mm. For 99% modeler point coverage, the AP margin was 3.6 mm and the lateral margin was 2.9 mm. The SI margins for 99% modeler point coverage were highly variable. The aggregate error at 95% was 6.9 mm in the SI direction, 4.6 mm in the AP direction, and 3.5 in the lateral direction. The Predictor points follow the Modeler points closely. Margins were found in each clinical direction that would provide 95% modeler point coverage for 95% of the models reviewed in this study. Similar margins were found in two clinical directions for 99% modeler point coverage in 95% of models. These results can offer guidance in the selection of CTV margins for treatment with the CyberKnife.

WE‐D‐204B‐02: Correlated 3D Respiratory Motion Prediction with Low‐Dimensional Feature‐Based Learning

Purpose: Current methods of respiratory motion prediction treat each dimension independently. In this study we develop and investigate a low dimensional inference methodology that takes advantage of the correlative relation among various physical dimensions for respiratory motion prediction. Method and Materials: Covariate‐response training pairs are constructed in the original 3D physical space, via state augmentation. Unlike existing methods where independent inference models are built for each of the x,y,z coordinates, we first map these high dimensional paired samples into a lower dimensional feature space, spanned by the principle components. In this feature space, a Gaussian kernel with block‐diagonal covariance is employed for learning the joint covariate‐response pdf, which upon marginalization yields the distribution for the projected prediction value. Finally, the estimate in the feature space is mapped back into the original physical space to recover the prediction representation in the (x,y,z) coordinate. The proposed method is compared against its counterpart with independent prediction, using 159 lung motion traces from the Synchrony respiratory tracking system. Both methods were tested for predicting 160ms and 570ms, spanning the range of measured DMLC tracing system latencies. Performance is quantified with 3D RMSE and 90% percentile 3D error. Results: Under fair setup conditions, the proposed method achieves better performance than its independent counterpart. The case‐wise 3D RMSE was reduced by 30–40%, with p‐values in the order of 10̂(−14) for both lookahead lengths. The 90% percentile 3D error was reduced from 1.8mm to 1.1mm for 160ms prediction and from 2.8mm to 2.0mm for 570ms prediction. Conclusion: We have proposed a general low dimensional feature learning methodology to perform correlated high dimensional inference. Combining it with the already effective kernel density estimation yields a highly efficient approach for real‐time respiratory motion prediction. Research supported by NIH/NCI R01 93626, Varian Medical Systems, and AAPM Research Seed Funding initiative.

A dosimetric comparison of stereotactic body radiation therapy techniques for lung cancer: robotic versus conventional linac‐based systems

The aim of this study is to compare the dosimetric characteristics of robotic and conventional linac‐based SBRT techniques for lung cancer, and to provide planning guidance for each modality. Eight patients who received linac‐based SBRT were retrospectively included in this study. A dose of 60 Gy given in three fractions was prescribed to each target. The Synchrony Respiratory Tracking System and a 4D dose calculation methodology were used for CyberKnife and linac‐based SBRT, respectively, to minimize respiratory impact on dose calculation. Identical image and contour sets were used for both modalities. While both modalities can provide satisfactory target dose coverage, the dose to GTV was more heterogeneous for CyberKnife than for linac planning/delivery in all cases. The dose to 1000 cc lung was well below institutional constraints for both modalities. In the high dose region, the lung dose depended on tumor size, and was similar between both modalities. In the low dose region, however, the quality of CyberKnife plans was dependent on tumor location. With anteriorly‐located tumors, the CyberKnife may deliver less dose to normal lung than linac techniques. Conversely, for posteriorly‐located tumors, CyberKnife delivery may result in higher doses to normal lung. In all cases studied, more monitor units were required for CyberKnife delivery for given prescription. Both conventional linacs and CyberKnife provide acceptable target dose coverage while sparing normal tissues. The results of this study provide a general guideline for patient and treatment modality selection based on dosimetric, tumor and normal tissue sparing considerations.PACS numbers: 87.53.Ly, 87.55.dk.

Cyberknife in the management of metastatic melanoma: A feasibility study.

e19033 Background: Cyberknife is a robotic radiosurgery device coupled with synchrony respiratory tracking system. It can deliver precisely high-dose radiotherapy with low fractionation to any part in the body, including a moving organ. Methods: Thirteen consecutive patients with wide spread metastases were treated to the most life-threatening lesions that included 17 sites: 6 patients with brain metastases treated to 8 sites, 2 spines to 3 sites and one to each of the following sites: lung, liver, intra-abdominal, nasopharynx, retro-orbital and floor of the orbit. Their ages ranged from 24-88 (median 61). The dose of radiation ranged from 2,000 to 3,500 cGy, carried out on outpatient over 1-5 fractions per patient. Results: All treatments were well tolerated with excellent patients' compliance. Their survival ranged from 2-15 months with a median of 4 months. Four patients had documented complete response to the treated lesions; one by biopsy and 3 radiologically. Two patients lived for 13 and 15 months (lung/spine). However, most of the patients succumbed to progression of other metastases. Duration of therapy ranged from 1-5 days. Conclusions: Cyberknife is a new effective therapy in melanoma. It is non-toxic, bloodless, precise local therapy administered on outpatient over a short period of time, and with no age limitation. It can reach areas that are not amenable to surgical resections to eliminate immediate life- threatening lesions. It can be administered while the patient continues to receive systemic therapy. No significant financial relationships to disclose.

Online prediction of respiratory motion: multidimensional processing with low-dimensional feature learning

Accurate real-time prediction of respiratory motion is desirable for effective motion management in radiotherapy for lung tumor targets. Recently, nonparametric methods have been developed and their efficacy in predicting one-dimensional respiratory-type motion has been demonstrated. To exploit the correlation among various coordinates of the moving target, it is natural to extend the 1D method to multidimensional processing. However, the amount of learning data required for such extension grows exponentially with the dimensionality of the problem, a phenomenon known as the ‘curse of dimensionality’. In this study, we investigate a multidimensional prediction scheme based on kernel density estimation (KDE) in an augmented covariate–response space. To alleviate the ‘curse of dimensionality’, we explore the intrinsic lower dimensional manifold structure and utilize principal component analysis (PCA) to construct a proper low-dimensional feature space, where kernel density estimation is feasible with the limited training data. Interestingly, the construction of this lower dimensional representation reveals a useful decomposition of the variations in respiratory motion into the contribution from semiperiodic dynamics and that from the random noise, as it is only sensible to perform prediction with respect to the former. The dimension reduction idea proposed in this work is closely related to feature extraction used in machine learning, particularly support vector machines. This work points out a pathway in processing high-dimensional data with limited training instances, and this principle applies well beyond the problem of target-coordinate-based respiratory-based prediction. A natural extension is prediction based on image intensity directly, which we will investigate in the continuation of this work. We used 159 lung target motion traces obtained with a Synchrony respiratory tracking system. Prediction performance of the low-dimensional feature learning-based multidimensional prediction method was compared against the independent prediction method where prediction was conducted along each physical coordinate independently. Under fair setup conditions, the proposed method showed uniformly better performance, and reduced the case-wise 3D root mean squared prediction error by about 30–40%. The 90% percentile 3D error is reduced from 1.80 mm to 1.08 mm for 160 ms prediction, and 2.76 mm to 2.01 mm for 570 ms prediction. The proposed method demonstrates the most noticeable improvement in the tail of the error distribution.

SU-FF-I-122: Assessment of Lung Tumors Treatment Accuracy Using CyberKnife Synchrony Model

Purpose: The Synchrony Respiratory Tracking System treats moving targets with much tighter margins. Due to the complex and erratic breathing motion of lungtumors, it is a challenging task to build a reliable internal/external correlation model and predictive algorithm. This study is to assess the treatment accuracy of lungtumors using the Synchrony system.Method and Materials: 21 lung patients with 97 fractions treated by a Synchrony system were studied. Available treatment data includes (i)tumor positions from sparsely acquired x‐ray image pairs; (ii) three surfaceLED markers motion continuously tracked by three Infrared detectors;(iii) the modeled motion calculated using the LED motion and the internal‐external correlation model and (iv) the predicted motion to compensate system latency. The correlation models can be validated by comparing the x‐ray image pairs determined tumor positions and the calculated modeled motion. The prediction precision can be evaluated using the modeled and predicted motion. The comparison of the x‐ray image pairs determined tumor positions and the predicted motion will assess the treatment effectiveness and provide feedback for the optimized margin expansion, computed using either a rolling ball method or un‐even margin expansions in X, Y, and Z directions. Results: Preliminary analysis showed the predictive model worked well with fast motion during inhale and exhale and decreased at the end of inhale and the end of exhale, especially at the reflection points. The fidelities of the correlation model changes over time, with correlation error up to 11.5mm in 3D distance for some fraction. The margin expansions required for different coverage percentages were calculated using both the rolling ball approach and un‐even expansion for each treatment direction. Conclusion: Evaluation of the correlation model, predictive results, and margin expansions were performed and the results are important to improve lungtumor treatments using the Synchrony System.