Purpose: Cone-beam computed tomography (CBCT) is a new image-guided radiation therapy (IGRT) technique for patient alignment in radiotherapy. The CBCT x-ray volume imaging system from Elekta allows for a variety of alignment methods. The aim of this study is to assess the accuracy of soft-tissue-based automatic alignment as compared with manual alignment using intraprostatic fiducials. Methods and Materials: All patients were treated on an Elekta Synergy S linear accelerator with kilovoltage CBCT. All alignments were performed using the x-ray volume imaging system and associated software. Automatic alignment with gray-value-based registration and manual alignment to fiducial markers were performed. Transitional corrections along each axis as well as 3-dimensional vectors were compared with evaluate the accuracy of gray-value-based registration compared with fiducials. Results: The distribution of the 3-dimensional vectors between gray-value and fiducial registrations demonstrated notable differences. The mean summed vector was 0.75 cm, with a standard deviation (SD) of 0.52 cm and range from 0.04 to 2.06 cm. There was minimal difference along the lateral direction, with a mean ± SD of −0.02 cm ± 0.13 cm. However, there were large discrepancies along the superior-inferior and anterior-posterior direction alignments, with mean ± SD values of −0.55 ± 0.48 cm and −0.31 ± 0.43 cm, respectively. Conclusions: CBCT with soft-tissue-based automatic corrections is not an accurate alignment compared with manual alignment to fiducial markers for prostate IGRT. We have concluded that a daily manual alignment to fiducials is one of the most reliable methods to maintain accuracy in prostate IGRT.
Treatment of the wrong body part due to incorrect setup is among the leading types of errors in radiotherapy. The purpose of this paper is to report an efficient automatic patient safety system (PSS) to prevent gross setup errors. The system consists of a pair of charge-coupled device (CCD) cameras mounted in treatment room, a single infrared reflective marker (IRRM) affixed on patient or immobilization device, and a set of in-house developed software. Patients are CT scanned with a CT BB placed over their surface close to intended treatment site. Coordinates of the CT BB relative to treatment isocenter are used as reference for tracking. The CT BB is replaced with an IRRM before treatment starts. PSS evaluates setup accuracy by comparing real-time IRRM position with reference position. To automate system workflow, PSS synchronizes with the record-and-verify (R&V) system in real time and automatically loads in reference data for patient under treatment. Special IRRMs, which can permanently stick to patient face mask or body mold throughout the course of treatment, were designed to minimize therapist's workload. Accuracy of the system was examined on an anthropomorphic phantom with a designed end-to-end test. Its performance was also evaluated on head and neck as well as abdominalpelvic patients using cone-beam CT (CBCT) as standard. The PSS system achieved a seamless clinic workflow by synchronizing with the R&V system. By permanently mounting specially designed IRRMs on patient immobilization devices, therapist intervention is eliminated or minimized. Overall results showed that the PSS system has sufficient accuracy to catch gross setup errors greater than 1 cm in real time. An efficient automatic PSS with sufficient accuracy has been developed to prevent gross setup errors in radiotherapy. The system can be applied to all treatment sites for independent positioning verification. It can be an ideal complement to complex image-guidance systems due to its advantages of continuous tracking ability, no radiation dose, and fully automated clinic workflow.
Recent knowledge on the effects of cardiac toxicity warrants greater precision for left-sided breast radiotherapy. Different breath-hold (BH) maneuvers (abdominal vs thoracic breathing) can lead to chest wall positional variations, even though the patient's tidal volume remains consistent. This study aims to investigate the feasibility of using optical tracking for real-time quality control of active breathing coordinator (ABC)-assisted deep inspiration BH (DIBH).An in-house optical tracking system (OTS) was used to monitor ABC-assisted DIBH. The stability and localization accuracy of the OTS were assessed with a ball-bearing phantom. Seven patients with left-sided breast cancer were included. A free-breathing (FB) computed tomography (CT) scan and an ABC-assisted BH CT scan were acquired for each patient. The OTS tracked an infrared (IR) marker affixed over the patient's xiphoid process to measure the positional variation of each individual BH. Using the BH within which the CT scan was performed as the reference, the authors quantified intra- and interfraction BH variations for each patient. To estimate the dosimetric impact of BH variations, the authors studied the positional correlation between the marker and the left breast using the FB CT and BH CT scans. The positional variations of 860 BHs as measured by the OTS were retrospectively incorporated into the original treatment plans to evaluate their dosimetric impact on breast and cardiac organs [heart and left anterior descending (LAD) artery].The stability and localization accuracy of the OTS was within 0.2 mm along each direction. The mean intrafraction variation among treatment BHs was less than 2.8 mm in all directions. Up to 12.6 mm anteroposterior undershoot, where the patient's chest wall displacement of a BH is less than that of a reference BH, was observed with averages of 4.4, 3.6, and 0.1 mm in the anteroposterior, craniocaudal, and mediolateral directions, respectively. A high positional correlation between the marker and the breast was found in the anteroposterior and craniocaudal directions with respective Pearson correlation values of 0.95 and 0.93, but no mediolateral correlation was found. Dosimetric impact of BH variations on breast coverage was negligible. However, the mean heart dose, mean LAD dose, and max LAD dose were estimated to increase from 1.4/7.4/18.6 Gy (planned) to 2.1/15.7/31.0 Gy (delivered), respectively.In ABC-assisted DIBH, large positional variation can occur in some patients, due to their different BH maneuvers. The authors' study has shown that OTS can be a valuable tool for real-time quality control of ABC-assisted DIBH.
Purpose Ionization chambers are the detectors of choice for photon beam profile scanning. However, they introduce significant volume averaging effect (VAE) that can artificially broaden the penumbra width by 2–3 mm. The purpose of this study was to examine the feasibility of photon beam profile deconvolution (the elimination of VAE from ionization chamber‐measured beam profiles) using a three‐layer feedforward neural network. Methods Transverse beam profiles of photon fields between 2 × 2 and 10 × 10 cm 2 were collected with both a CC13 ionization chamber and an EDGE diode detector on an Elekta Versa HD accelerator. These profiles were divided into three datasets (training, validation and test) to train and test a three‐layer feedforward neural network. A sliding window was used to extract input data from the CC13‐measured profiles. The neural network produced the deconvolved value at the center of the sliding window. The full deconvolved profile was obtained after the sliding window was moved over the measured profile from end to end. The EDGE‐measured beam profiles were used as reference for the training, validation, and test. The number of input neurons, which equals the sliding window width, and the number of hidden neurons were optimized with a parametric sweeping method. A total of 135 neural networks were fully trained with the Levenberg–Marquardt backpropagation algorithm. The one with the best overall performance on the training and validation dataset was selected to test its generalization ability on the test dataset. The agreement between the neural network‐deconvolved profiles and the EDGE‐measured profiles was evaluated with two metrics: mean squared error (MSE) and penumbra width difference (PWD). Results Based on the two‐dimensional MSE plots, the optimal combination of sliding window width of 15 and 5 hidden neurons was selected for the final neural network. Excellent agreement was achieved between the neural network‐deconvolved profiles and the reference profiles in all three datasets. After deconvolution, the mean PWD reduced from 2.43 ± 0.26, 2.44 ± 0.36, and 2.46 ± 0.29 mm to 0.15 ± 0.15, 0.04 ± 0.03, and 0.14 ± 0.09 mm for the training, validation, and test dataset, respectively. Conclusions We demonstrated the feasibility of photon beam profile deconvolution with a feedforward neural network in this work. The beam profiles deconvolved with a three‐layer neural network had excellent agreement with diode‐measured profiles.
Accurately localizing lung tumor localization is essential for high-precision radiation therapy techniques such as stereotactic body radiation therapy (SBRT). Since direct monitoring of tumor motion is not always achievable due to the limitation of imaging modalities for treatment guidance, placement of fiducial markers on the patient's body surface to act as a surrogate for tumor position prediction is a practical alternative for tracking lung tumor motion during SBRT treatments. In this work, the authors propose an innovative and robust model to solve the multimarker position optimization problem. The model is able to overcome the major drawbacks of the sparse optimization approach (SOA) model.The principle-component-analysis (PCA) method was employed as the framework to build the authors' statistical prediction model. The method can be divided into two stages. The first stage is to build the surrogate tumor matrix and calculate its eigenvalues and associated eigenvectors. The second stage is to determine the "best represented" columns of the eigenvector matrix obtained from stage one and subsequently acquire the optimal marker positions as well as numbers. Using 4-dimensional CT (4 DCT) and breath hold CT imaging data, the PCA method was compared to the SOA method with respect to calculation time, average prediction accuracy, prediction stability, noise resistance, marker position consistency, and marker distribution.The PCA and SOA methods which were both tested were on all 11 patients for a total of 130 cases including 4 DCT and breath-hold CT scenarios. The maximum calculation time for the PCA method was less than 1 s with 64 752 surface points, whereas the average calculation time for the SOA method was over 12 min with 400 surface points. Overall, the tumor center position prediction errors were comparable between the two methods, and all were less than 1.5 mm. However, for the extreme scenarios (breath hold), the prediction errors for the PCA method were not only smaller, but were also more stable than for the SOA method. Results obtained by imposing a series of random noises to the surrogates indicated that the PCA method was much more noise resistant than the SOA method. The marker position consistency tests using various combinations of 4 DCT phases to construct the surrogates suggested that the marker position predictions of the PCA method were more consistent than those of the SOA method, in spite of surrogate construction. Marker distribution tests indicated that greater than 80% of the calculated marker positions fell into the high cross correlation and high motion magnitude regions for both of the algorithms.The PCA model is an accurate, efficient, robust, and practical model for solving the multimarker position optimization problem to predict lung tumor motion during SBRT treatments. Due to its generality, PCA model can also be applied to other imaging guidance system whichever using surface motion as the surrogates.
Advances in linear accelerator technology have led to the emergence of conformal radiotherapy, in which complex beam configurations can be utilized to maximize radiation dose to the treatment volume, while minimizing dose to adjacent critical structures. The beam configuration is optimized in terms of position, collimation and intensity in treatment planning. The adequacy of the treatment rests on the verification of the localized treatment volume with respect to the radiation beams. Verification is performed using megavoltage portal imaging on the treatment accelerator or linac. Errors affecting treatment delivery can arise either in the construction of the custom collimation, or in positioning the patient on the linac. The collimation shape is verified using edge detection with respect to a reference point defined by both a user and a user independent template method. Image registration of anatomical landmarks using Chamfer matching is used to compute positioning errors of the patient.
