QA Results Research Articles

Pretreatment patient-specific quality assurance prediction based on 1D complexity metrics and 3D planning dose: classification, gamma passing rates, and DVH metrics

PurposeHighly modulated radiotherapy plans aim to achieve target conformality and spare organs at risk, but the high complexity of the plan may increase the uncertainty of treatment. Thus, patient-specific quality assurance (PSQA) plays a crucial role in ensuring treatment accuracy and providing clinical guidance. This study aims to propose a prediction model based on complexity metrics and patient planning dose for PSQA results.Materials and methodsPlanning dose, measurement-based reconstructed dose and plan complexity metrics of the 687 radiotherapy plans of patients treated in our institution were collected for model establishing. Global gamma passing rate (GPR, 3%/2mm,10% threshold) of 90% was used as QA criterion. Neural architecture models based on Swin-transformer were adapted to process 3D dose and incorporate 1D metrics to predict QA results. The dataset was divided into training (447), validation (90), and testing (150) sets. Evaluation of predictions was performed using mean absolute error (MAE) for GPR, planning target volume (PTV) HI and PTV CI, mean absolute percentage error (MAPE) for PTV D95, PTV D2 and PTV Dmean, and the area under the receiver operating characteristic (ROC) curve (AUC) for classification. Furthermore, we also compare the prediction results with other models based on either only 1D or 3D inputs.ResultsIn this dataset, 72.8% (500/687) plans passed the pretreatment QA under the criterion. On the testing set, our model achieves the highest performance, with the 1D model slightly surpassing the 3D model. The performance results are as follows (combine, 1D, and 3D transformer): The AUCs are 0.92, 0.88 and 0.86 for QA classification. The MAEs of prediction are 0.039, 0.046, and 0.040 for 3D GPR, 0.018, 0.021, and 0.019 for PTV HI, and 0.075, 0.078, and 0.084 for PTV CI. Specifically, for cases with 3D GPRs greater than 90%, the MAE could achieve 0.020 (combine). The MAPE of prediction is 1.23%, 1.52%, and 1.66% for PTV D95, 2.36%, 2.67%, and 2.45% for PTV D2, and 1.46%, 1.70%, and 1.71% for PTV Dmean.ConclusionThe model based on 1D complexity metrics and 3D planning dose could predict pretreatment PSQA results with high accuracy and the complexity metrics play a leading role in the model. Furthermore, dose-volume metric deviations of PTV could be predicted and more clinically valuable information could be provided.

Impact of the Acuros XB spatial discretization error on ArcCHECK VMAT QA for small-field SBRT.

To evaluate the impact of the Acuros XB spatial discretization errors on ArcCHECK volumetric modulated arc therapy (VMAT) QA for small-field SBRT plans. Eighteen SBRT VMAT arcs that failed the ArcCHECK VMAT QA were retrospectively analyzed. Plan verification doses were calculated using Eclipse Acuros XB, and absolute 3%/2mm gamma passing rates were calculated to compare ArcCHECK and MapCHECK2 with MapPHAN. Verification doses were recalculated using AAA in Eclipse and with the EGSnrc Monte Carlo package. In addition, error-reduced Acuros XB doses were calculated by subdividing the entire arc into several sub-arcs ("split-arc" method), with the angular ranges of the sub-arcs optimized to balance accuracy and efficiency. Relative gamma passing rates were calculated and compared for the four methods: (1) Acuros XB; (2) AAA; (3) EGSnrc Monte Carlo; and (4) the split-arc method. The absolute gamma passing rates were below 90% for ArcCHECK and above 95% for MapCHECK2. The averaged relative gamma passing rates were (1) 84.7% for clinical Acuros XB; (2) 96.8% for AAA; (3) 98.8% for EGSnrc Monte Carlo; and (4) 96.8% for the split-arc method with 60° sub-arc angle. Compared to the clinical Acuros XB, the split-arc method improved the relative gamma passing rate by 12.1% on average. No significant difference was found between AAA and the split-arc method (p>0.05). The Acuros XB spatial discretization errors can significantly impact the ArcCHECK VMAT QA results for small-field SBRT plans. The split-arc method may be used to improve the VMAT QA results.

BT-CKBQA: An efficient approach for Chinese knowledge base question answering

Knowledge Base Question Answering (KBQA), as an increasingly essential application, can provide accurate responses to user queries. ensuring that users obtain relevant information and make decisions promptly. The deep learning-based approaches have achieved satisfactory QA results by leveraging the neural network models. However, these approaches require numerous parameters, which increases the workload of tuning model parameters. To address this problem, we propose BT-CKBQA, a practical and highly efficient approach incorporating BM25 and Template-based predicate mapping for CKBQA. Besides, a concept lattice based approach is proposed for summarizing the knowledge base, which can largely improve the execution efficiency of QA with little loss of performance. Concretely, BT-CKBQA leverages the BM25 algorithm and custom dictionary to detect the subject of a question sentence. A template-based predicate generation approach is then proposed to generate candidate predicates. Finally, a ranking approach is provided with the joint consideration of character similarity and semantic similarity for predicate mapping. Extensive experiments are conducted over the NLPCC-ICCPOL 2016 and 2018 KBQA datasets, and the experimental results demonstrate the superiority of the proposed approach over the compared baselines. Particularly, the averaged F1-score result of BT-CKBQA for mention detection is up to 98.25%, which outperforms the best method currently available in the literature. For question answering, the proposed approach achieves superior results than most baselines with the F1-score value of 82.68%. Compared to state-of-the-art baselines, the execution efficiency and performance of QA per unit time can be improved with up to 56.39% and 44.06% gains, respectively. The experimental results for the diversification of questions indicate that the proposed approach performs better for diversified questions than domain-specific questions. The case study over a constructed COVID-19 knowledge base illustrates the effectiveness and practicability of BT-CKBQA.

Automatic Chinese knowledge-based question answering by the MGBA-LSTM-CNN model

The purpose of knowledge-based question answering (KBQA) is to accurately answer the questions raised by users through knowledge triples. Traditional Chinese KBQA methods rely heavily on artificial features, resulting in unsatisfactory QA results. To solve the above problems, this paper divides Chinese KBQA into two parts: entity extraction and attribute mapping. In the entity extraction stage, the improved Bi-LSTM-CNN-CRF model is used to identify the entity of questions and the Levenshtein distance method is used to resolve the entity link error. In the attribute mapping stage, according to the characteristics of questions and candidate attributes, the MGBA-LSTM-CNN model is proposed to encode questions and candidate attributes from the semantic level and word level, respectively, and splice them into new semantic vectors. Finally, the cosine distance is used to measure the similarity of the two vectors to find candidate attributes most similar to questions. The experimental results show that the system achieves good results in the Chinese question and answer data set.

OC-0116 Automated classification of pre-treatment QA results of VMAT plans with plan complexity metrics

OC-0116 Automated classification of pre-treatment QA results of VMAT plans with plan complexity metrics

Accuracy of the catalyst surface guidance system for patient monitoring during cranial SRS treatments.

The use of surface guided imaging in cranial stereotactic radiotherapy provides a non-ionising form of patient position verification that gives information on when patient position errors may require corrections to be applied. This work assessed the accuracy of Catalyst+ HD system for treatment geometries commonly used in cranial SRS. Average Catalyst reported error as a function of couch rotation agreed with measured kV and MV walkout within 0.5mm for the lateral and longitudinal directions. Change in Catalyst reported error with isocentre depth relative to the monitoring region of interest (ROI) from the surface was assessed and showed variation above 0.5mm, but within 1mm for isocentre depths ranging from 3 to 15cm from the phantom surface. Gantry occlusion of Catalyst cameras caused a shift in reported position error that was also dependent on isocentre depth relative to the monitoring ROI. SRS MapCHECK patients specific QA results showed improvement in gamma passing rates for a workflow where Catalyst reported errors above 0.5mm were corrected.

Clinical experience on patient-specific quality assurance for CBCT-based online adaptive treatment plan.

Ethos CBCT-based adaptive radiotherapy (ART) system can generate an online adaptive plan by re-optimizing the initial reference plan based on the patient anatomy at the treatment. The optimization process is fully automated without any room for human intervention. Due to the change in anatomy, the ART plan can be significantly different from the initial plan in terms of plan parameters such as the aperture shapes and number of monitor units (MUs). In this study, we investigated the feasibility of using calculation-based patient specific QA for ART plans in conjunction with measurement-based and calculation-based QA for initial plans to establish an action level for the online ART patient-specific QA. A cohort of 98 cases treated on CBCT-based ART system were collected for this study. We performed measurement-based QA using ArcCheck and calculation-based QA using Mobius for both the initial plan and the ART plan for analysis. For online the ART plan, Mobius calculation was conducted prior to the delivery, while ArcCheck measurement was delivered on the same day after the treatment. We first investigated the modulation factors (MFs) and MU numbers of the initial plans and ART plans, respectively. The γ passing rates of initial and ART plan QA were analyzed. Then action limits were derived for QA calculation and measurement for both initial and online ART plans, respectively, from 30 randomly selected patient cases, and were evaluated using the other 68 patient cases. The difference in MF between initial plan and ART-plan was 12.9% ± 12.7% which demonstrates their significant difference in plan parameters. Based on the patient QA results, pre-treatment calculation and measurement results are generally well aligned with ArcCheck measurement results for online ART plans, illustrating their feasibility as an indicator of failure in online ART QA measurements. Furthermore, using 30 randomly selected patient cases, the γ analysis action limit derived for initial plans and ART plans are 89.6% and 90.4% in ArcCheck QA (2%/2mm)andare 92.4% and 93.6% in Mobius QA(3%/2mm),respectively.According to the calculated action limits, the ArcCheck measurements for all the initial and ART plans passed QA successfully while the Mobius calculation action limits flagged seven and four failure cases respectively for initial plans and ART plans, respectively. An ART plan can be substantially different from the initial plan, and therefore a separate session of ART plan QA is needed to ensure treatment safety and quality. The pre-treatment QA calculation via Mobius can serve as a reliable indicator of failure in online ART plan QA. However, given that Ethos ART system is still relatively new, ArcCheck measurement of initial plan is still in practice. It may be skipped as we gain more experience and have better understanding of the system.

A Method to Simplify Patient-Specific QA of Single Isocenter Multiple Met SRS Plans

<h3>Purpose/Objective(s)</h3> Pre-treatment quality assurance of single isocenter multi-met stereotactic radiosurgery (SRS) plans is standardly performed by acquiring a dose measurement using a detector array or ion chamber and film. The major efficiency issue with QA of these plans is how many, and which of the multiple metastases (mets) should be measured. The purpose of this study was to develop a workflow using an independent 3D dose calculation (DC) and trajectory log (TL) QA software to efficiently determine which of the brain mets should be further evaluated by measurement. <h3>Materials/Methods</h3> 30 single isocenter multiple metastasis patient SRS plans with a total of 82 mets were retrospectively studied. The clinical SRS plans were generated with an automated dynamic conformal arc SRS planning system (TPS). All plans were sent to the DC software for calculation. The gamma analysis (3%, 1 mm), percent difference between TPS and DC calculated mean and max dose to PTV, and the dose difference of the centroid point of each PTV was recorded. All plans were delivered to a high-resolution diode array designed for SRS and measurements were used to perform a gamma analysis (3%, 1 mm). In addition, the trajectory logs of the QA delivery were used to calculate 3D dose distributions and the same metrics acquired from the DC software were calculated and recorded. Pearson correlation between DC and TL calculated metrics was calculated. <h3>Results</h3> The results of the dosimetric metrics calculated by the DC and TL software are presented in Table 1. Overall agreement between DC and TPS was excellent, with all mean percent differences within 1.1% and mean gamma pass rates greater than 99%. Furthermore, dosimetric metrics calculated by DC and TL were highly correlated, with the correlation coefficients for all metrics at least 0.96, which suggests DC results can be used to predict TL QA results. To use DC to screen which PTVs should be further evaluated, an action level was set such that if any metric fell outside the 95% confidence interval (mean + 1.96 × SD), a QA measurement would be performed. For all metrics, a ±5% action level was calculated to cover the 95% confidence interval. DC results for 6 PTVs exceeded the action level. Of these plans, the measured patient specific QA passed for all plans. While all measured QA passed, DC and TL QA appeared to be more sensitive than the diode array measurement in identifying potential plan quality issues. <h3>Conclusion</h3> An efficient workflow for patient specific QA of single isocenter multiple met SRS plans was developed. An independent 3D dose calculation is used to calculate agreement of dosimetry metrics compared to the TPS. If the percent difference of a metric for a PTV exceeds a 5% action level, it is selected for further QA using a dose measurement.

Improvement of IMRT QA prediction using imaging-based neural architecture search.

Machine learning (ML) has been used to predict the gamma passing rate (GPR) of intensity-modulated radiation therapy (IMRT) QA results. In this work, we applied a novel neural architecture search to automatically tune and search for the best deep neural networks instead of using hand-designed deep learning architectures. One hundred and eighty-two IMRT plans were created and delivered with portal dosimetry. A total of 1497 fields for multiple treatment sites were delivered and measured by portal imagers. Gamma criteria of 2%/2mm with a 5% threshold were used. Fluence maps calculated for each plan were used as inputs to a convolution neural network (CNN). Auto-Keras was implemented to search for the best CNN architecture for fluence image regression. The network morphism was adopted in the searching process, in which the base models were ResNet and DenseNet. The performance of this CNN approach was compared with tree-based ML models previously developed for this application, using the same dataset. The deep-learning-based approach had 98.3% of predictions within 3% of the measured 2%/2-mm GPRs with a maximum error of 3.1% and a mean absolute error of less than 1%. Our results show that this novel architecture search approach achieves comparable performance to the machine-learning-based approaches with handcrafted features. We implemented a novel CNN model using imaging-based neural architecture for IMRT QA prediction. The imaging-based deep-learning method does not require a manual extraction of relevant features and is able to automatically select the best network architecture.

Implementation of volumetric-modulated arc therapy for locally advanced breast cancer patients: Dosimetric comparison with deliverability consideration of planning techniques and predictions of patient-specific QA results via supervised machine learning

PurposeThe aim of this study was to implement a clinically deliverable VMAT planning technique dedicated to advanced breast cancer, and to predict failed QA using a machine learning (ML) model to optimize the QA workload. MethodsFor three planning methods (2A: 2-partial arc, 2AS: 2-partial arc with splitting, 4A: 4-partial arc), dosimetric results were compared with patient-specific QA performed with the electronic portal imaging device of the linac. A dataset was built with the pass/fail status of the plans and complexity metrics. It was divided into training and testing sets. An ML metamodel combining predictions from six base classifiers was trained on the training set to predict plans as ‘pass’ or ‘fail’. The predictive performances were evaluated using the unseen data of the testing set. ResultsThe dosimetric comparison highlighted that 4A was the highest dosimetric performant method but also the poorest performant in the QA process. 2AS spared the best heart, but provided the highest dose to the contralateral breast and lowest node coverage. 2A provides a dosimetric compromise between organ at risk sparing and PTV coverage with satisfactory QA results. The metamodel had a median predictive sensitivity of 73% and a median specificity of 91%. ConclusionsThe 2A method was selected to calculate clinically deliverable VMAT plans; however, the 2AS method was maintained when the heart was of particular importance and breath-hold techniques were not applicable. The metamodel provides promising predictive performance, and it is intended to be improved as a larger dataset becomes available.

P118 - Pre-treatment QA results of stereotactic plans against target volume and plan complexity

P118 - Pre-treatment QA results of stereotactic plans against target volume and plan complexity

A Unique Volumetric Modulated Arc Stereotactic Radiosurgery (SRS)Treatment Planning Approach for Volume Staged SRS to Large Arteriovenous Malformation

The management of large Arteriovenous Malformation AVMs (> 10cc) using volume staged Stereotactic Radiosurgery (VS SRS) usually requires more than 2 VS SRS procedures. VS SRS downsizes the treatment volume per stage allowing effective higher dose of SRS to relative smaller volumes of target to reduce risk of SRS induced brain necrosis. It is imperative to maintain the SRS treatment plan quality matrix of conformity and dose gradient at optimal values to minimize the adverse radiation effects to surrounding brain tissues. However, VS SRS plans can present a challenge at treatment planning given the angioarchitecture of staged volumes and total volume, delivery technique and modality limitations.A 37-year-old female patient has history of headaches, intraventricular hemorrhage, and cerebral arteriovenous malformation AVM in the right parietal lobe. This large AVM (16.5cc) scored 4 on Spetzler-Martin grading system. Patient was simulated with large board CT simulator using an SRS mask. 1 mm thickness of CT images with IV contrast per CTA protocol were obtained and imported into the TPS. MRA images data sets and Angio CBCT scans from arteriogram were registered to CT Simulation scans for contouring AMV nidus. The nidus was divided into 3 sub volumes of 8 cc, 4.5cc and 4 cc, respectively. The dose of 20 Gy was prescribed to entire volume of nidus and portions of total dose were delivered to sub volumes with interval of 2-4 weeks. The VMAT inverse treatment plans were generated for each volume/isocenter using non-coplanar arcs. to be delivered on HDMLC. Collimator and arc angles were configured geometrically to minimize dose spillage between staged volumes. Planning objectives including generalized uniform equivalent dose gEUD were assigned to planning volumes and OARs. Normal tissue objective was enabled, structure resolution was set to 1.25mm, and AAA calculation grid was set to 1mm. Individual and composite plans isodoses and DVHs were examined repetitively during the planning and optimization iterations to confirm meeting the acceptable VS SRS plan quality matrix or to adjust if needed. All three individual sub volume plans passed IMRT QA (> 90%, DTA 1mm, %DD 2%) RESULTS: The composite plan maximum dose was 33.3 Gy, minimum dose for total nidus volume was 94%, and 95.2%, 100%, 109% for three sub volumes of nidus, respectively. V95%was > 95% and D95% was > 95% for all three sub volumes and total volume in the composite plan. The mean dose was 27.7 Gy. CI was 0.63, PITV was 1.57, R50 was 4.36 while GI was 2.77, V12 was 52.9cc (3.9%).The proposed unique VMAT VS SRS for large AVM achieves adequate and comparable treatment planning quality parameters when evaluated with published literature covering VS SRS delivery with different treatment modalities such as robotic radiosurgery systems. Our preliminary results from previous treated patients using same planning technique showed good response rate and very low rate of brain necrosis.F. Wang: None. D. Soultan: None. H. Jiang: None. N. Demez: None.

Commissioning of and preliminary experience with a new fully integrated computed tomography linac.

PurposeA new medical linear accelerator (linac) platform integrated with helical computed tomography (CT), the uRT‐linac 506c, was introduced into clinical application in 2019 by United Imaging Healthcare (UIH) Co., Ltd. (Shanghai, China). It combines a Carm linac with a diagnostic‐quality 16‐slice CT imager, providing seamless workflow from simulation to treatment. The aim of this report is to assess the technical characteristics, commissioning results and preliminary experiences stemming from clinical usage.MethodsThe mechanical and imaging test procedures, commissioning data collection and TPS validation were summarized. CTIGRT accuracy was investigated with different loads and couch extensions. A series of end‐to‐end cases for different treatment sites and delivery techniques were tested preclinically to estimate the overall accuracy for the entire treatment scheme. The results of patient‐specific QA and machine stability during a one‐year operation are also reported.ResultsGantry/couch/collimator isocentricity was measured as 0.63 mm in radius. The TPS models were in agreement with the beam commissioning data within a deviation of 2%. An overall submillimeter accuracy was demonstrated for the CT‐IGRT process under all conditions. The absolute point dose difference for all the preclinical end‐to‐end tests was within 3%, and the gamma passing rate of the 2D dose distribution measured by EBT3 film was better than 90% (3% DD, 3 mm DTA and 10% threshold). Pretreatment QA of clinical cases resulted with better than 3% point dose difference and more than 99% gamma passing rate (3% DD/2 mm DTA/10% threshold) tested with Delta4. The output of the linac was mostly within 1% of variation in a one‐year operation.ConclusionThe commissioning results and clinical QA results show that the uRT‐linac 506c platform exhibits good and stable performance in mechanical and dosimetric accuracy. The integrated CT system provides an efficient workflow for image guidance with submillimeter localization precision, and will be a good starting point to proceed advanced adaptive radiotherapy.

A Beam Projection-Based Modified Gamma Analysis Scheme for Clinically Interpretable Pre-Treatment Dose Verification.

Purpose:To investigate a novel gamma analysis system for dose verification results in terms of clinical significance.Methods and Materials:The modified scheme redefined the computational domain of the conventional gamma analysis with the projections of beams and the regions of interest (ROI). We retrospectively studied 6 patients with the conventional and the modified gamma analysis schemes while compared their performances. The cold spots ratio of the planning target volume (PTV) and the hot spots ratio of the organs at risk (OAR) were also computed by the modified scheme to assess the clinical significance.Results:The result of the gamma passing rate in the modified method was conformable to that in the conventional method with a cut-off threshold of 5%. The cold spots ratio of PTV and hot spots ratio of OAR were able to be evaluated by the modified scheme. For an introduced 7.1% dose error, the discrimination ratio in gamma passing rate of the conventional method was lower than 2%, while it was improved to 5% by the modified method.Conclusions:The modified gamma analysis scheme had a comparable quality as the conventional scheme in terms of dose inspection. Besides, it could improve the clinical significance of the QA result and provide the assessment for ROI-specific discrepancy. The modified scheme could also be conveniently integrated into the conventional dose verification process, benefiting the less developed regions where high-end 3D dose verification devices are not affordable.

Commissioning of a Versa HDTM linear accelerator for three commercial treatment planning systems.

In a mixed‐vendor radiation oncology environment, it is advantageous if the department's treatment planning system (TPS) supports the linear accelerators of different vendors. In this publication beam data collection and modeling for the Versa HD linear accelerator in Monaco, Pinnacle, and Eclipse are discussed. In each TPS static field, Intensity‐Modulated Radiation Therapy (IMRT) step and shoot, and Volumetric‐Modulated Arc Therapy (VMAT) plans for flattened and flattening‐filter free photon beams of all available energies were evaluated for field sizes >3 × 3. To compare passing rates, identical beam model validation plans were calculated in each TPS. Eclipse, Monaco, and Pinnacle beam models passed validation measurements in homogeneous materials for a variety of treatment fields, including static, IMRT, and VMAT. In the case of Eclipse, the “dosimetric leaf gap” parameter was found to be critical for passing rates of VMAT plans. The source size parameter plays an important role as well for small fields. In the case of Pinnacle the multileaf collimator offset table needed to be optimized for better VMAT QA results. Each of the investigated treatment planning systems met the criteria to be used clinically in conjunction with Elekta Versa HD linear accelerators. It can be of great advantage to have the option to operate a TPS and linear accelerator from different vendors, as decisions surrounding linear accelerator or TPS purchases are very complicated and not just limited to technical considerations.

An electronic portal image device (EPID)-based multiplatform rapid daily LINAC QA tool.

PurposeTo develop an efficient and economic daily quality research tool (DQRT) for daily check of multiplatform linear accelerators (LINACs) with flattening filter (FF) and flattening filter‐free (FFF) photon beams by using an Electronic Portal Image Device (EPID).Materials and MethodsAfter EPID calibration, the monitored parameters were analyzed from a 10 cm × 10 cm open and 60° wedge portal images measured by the EPID with 100 MU exposure. Next, the repeatability of the EPID position accuracy, long‐term stability, and linearity between image gray value and exposure were verified. Output and beam quality stability of the 6‐MV FF and FFF beams measured by DQRT with the introduced setup errors of EPID were also surveyed. Besides, some test results obtained by DQRT were compared with those measured by FC65‐G and Matrixx. At last, the tool was evaluated on three LINACs (Synergy, VersaHD, TrueBeam) for 2 months with two popular commercial QA tools as references.ResultsThere are no differences between repeatability tests for all monitored parameters. Image grayscale values obtained by EPID and exposure show good linearity. Either 6 MV FF or FFF photon beam shows minimal impact to the results. The differences between FC65‐G, Matrixx and DQRT results are negligible. Monitor results of the two commercial tools are consistent with the DQRT results collected during the 2‐month period.ConclusionWith a shorter time and procedure, the DQRT is useful to daily QA works of LINACs, producing a QA result quality similarly to or more better than the traditional tools and giving richer contents to the QA results. For hospitals with limited QA time window available or lack of funding to purchase commercial QA tools, the proposed DQRT can provide an alternative and economic approach to accomplish the task of daily QA for LINACs.

Repeatability and Reproducibility of Total Retinal Blood Flow Measurements Using Bi-Directional Doppler OCT.

PurposeTo investigate the repeatability and reproducibility of total retinal blood flow measurements using a custom-built dual-beam bidirectional Doppler optical coherence tomography (OCT) system in healthy subjects.MethodsRepeatability and reproducibility were analyzed in 10 and 34 healthy subjects, respectively. For repeatability, measurements were taken twice within 30 minutes, for reproducibility, twice within two to five weeks. Two analysis approaches were compared for calculation of absolute blood velocities: a previously published approach resulting in values for total arterial (QA,abs) and total venous blood flow (QV,abs) and a novel approach taking into account that there is a fixed relation between the phase shift in the two OCT channels (QA,new, QV,new). Repeatability and reproducibility were quantified using intraclass correlation coefficients (ICCs).ResultsFor QA,abs and QV,abs, ICC values between 0.78 and 0.84 were obtained. QA,new and QV,new values revealed better repeatability and reproducibility as compared to the convential appoach. Repeatability ICCs for QA,new and QV,new were between 0.91 and 0.93, and reproducibility ICCs were between 0.87 and 0.91 indicating excellent reproducibility. Good agreement was observed between total retinal blood flow values as measured from retinal arteries and retinal veins.ConclusionsMeasurement of total retinal blood flow using dual-beam Doppler OCT shows excellent reproducibility, which can further be improved by using a novel algorithm for calculating blood velocities in retinal vessels.Translational RelevanceOur data indicate that dual-beam Doppler OCT can be used for longitudinal studies. Hence, quantitative retinal blood flow may be established as a biomarker for progression vascular eye diseases.

4089 Clinical Implementation of Monte Carlo Dose Calculation for Patient-Specific Radiotherapy Quality Assurance

OBJECTIVES/GOALS: The Monte Carlo dose calculation method is often considered the “gold standard” for patient dose calculations and can be as radiation dose measurements. Our study aims to develop a true Monte Carlo model that can be implemented in our clinic as part of our routine patient-specific quality assurance. METHODS/STUDY POPULATION: We have configured and validated a model of one of our linear accelerators used for radiation therapy treatments using the EGSnrc Monte Carlo simulation software. Measured dosimetric data was obtained from the linear accelerator and was used as the standard to compare the doses calculated with our model in EGSnrc. We will compare dose calculations between commercial treatment planning systems, the EGSnrc Monte Carlo model, and patient-specific measurements. We will implement the Monte Carlo model in our clinic for routine second-checks of patient plans, and to recalculate plans delivered to patients using machine log files. RESULTS/ANTICIPATED RESULTS: Our Monte Carlo model is within 1% agreement with our measured dosimetric data, and is an accurate representation of our linear accelerators used for patient treatments. With this high level of accuracy, we have begun simulating more complex patient treatment geometries, and expect the level of accuracy to be within 1% of measured data. We believe the Monte Carlo calculation based on machine log files will correlate with patient-specific QA analysis and results. The Monte Carlo model will be a useful tool in improving our patient-specific quality assurance protocol and can be utilized in further research. DISCUSSION/SIGNIFICANCE OF IMPACT: This work can be implemented directly in clinical practice to ensure patient doses are calculated as accurately as possible. These methods can be used by clinics who do not have access to more advanced dose calculation software, ensuring accuracy for all patients undergoing radiotherapy treatments.

Research on Using IMRT Plan for Preoperative Rectal Cancer Patients

Aims: Research the possibility of using IMRT for rectal cancer patients in preoperative radiotherapy. Methods and Material: The research object is the preoperative radiotherapy plan for rectal cancer patients. The research group made two plans (IMRT, 3DCRT) for each image series of 34 rectal cancer patients who have received preoperative radiotherapy in Hanoi Oncology Hospital; and then compared the dose distribution on PTV, bladder, intestine, femoral bones, the average MU, and QA results of two types of plan. Results: The 95% isodose line and 50% isodose of IMRT plan are closer than those of 3DCRT plan. The average dose of PTV in IMRT plan and 3DCRT plan are 5006 ± 23 cGy and 5036 ± 42 cGy, respectively. The HTCI and HI values of IMRT and 3D plan are 0.97 ± 0.026 and 5.37 ± 1.32; 1.00 ± 0.003 and 7.08 ± 0.88. About the dose of organ at risk: The maximum dose, average dose on the right, left femoral head in the IMRT plan are less than those values in the 3DCRT plan (6.2 Gy, 6 Gy, 7.4 Gy, 9 Gy, respectively). The maximum dose and average dose on the bladder of the IMRT plan are smaller than those values of the 3DCRT plan (5.3 Gy, 1.5 times, respectively). The maximum dose and average dose of intestine in the IMRT plan was less than those values in the 3DCRT plan (4.3 Gy, 1.54 times, respectively). The MU number of IMRT plan is 1.5 times bigger than that of 3DCRT plan. Gamma index of IMRT plan is better than that of 3DCRT plan (99% compared with 97%). Conclusions: Using IMRT plan in preoperative radiotherapy for rectal cancer patients can still ensure covered PTV as well as the 3D PLAN. Furthermore, the dose of PTV in the IMRT plan is more uniform than those in the 3D plan, and the dose effect on the OAR surrounding PTV is much lower than when using the 3D plan. When IMRT plan were used to treat the preoperative rectal cancer patients, the LINAC took more time than when using 3DCRT plan.

Social QA in non-CQA platforms

Community Question Answering (cQA) sites have emerged as platforms designed specifically for the exchange of questions and answers among communities of users. Although users tend to find good quality answers in cQA sites, there is evidence that they also engage in a significant volume of QA in other types of social sites, such as microblog platforms. Research indicates that users opt for these non-specific QA social networks because they contain up-to-date information on current events, also due to their rapid information propagation, and social trust. In this sense, we propose that microblog platforms can emerge as a novel, valuable source of information for QA information retrieval tasks. However, we have found that it is not straightforward to transfer existing approaches for automatically retrieving relevant answers in traditional cQA platforms for use in microblogs. This occurs because there are unique characteristics that differentiate microblog data from that of traditional cQA, such as noise and very short text length. In this work, we study (1) if microblog data can be used to automatically provide relevant answers for the QA task, and, in addition, (2) which features contribute the most for finding relevant answers for a particular query. In particular, we introduce a conversation (thread)-level document model, as well as a machine learning ranking framework for microblog QA. We validate our proposal by using factoid-QA as a proxy task, showing that Twitter conversations can indeed be used to automatically provide relevant results for QA. We are able to identify the importance of different features that contribute the most for QA ranking. In addition, we provide evidence that our method allows us to retrieve complex answers in the domain of non-factoid questions.