Faculty Bibliography

BACKGROUND:Long-term follow-up using volumetric measurement could significantly assist in the management of vestibular schwannomas (VS). Manual segmentation of VS from MRI for treatment planning and follow-up assessment is labor-intensive and time-consuming. This study aims to develop a deep learning technique to fully automatically segment VS from MRI. METHODS:This study retrospectively analyzed MRI data of 737 patients who received gamma knife radiosurgery for VS. Treatment planning T1-weighted isotropic MR and manually contoured gross tumor volumes (GTV) were used for model development. A 3D convolutional neural network (CNN) was built on ResNet blocks. Spatial attenuation and deep supervision modules were integrated in each decoder level to enhance the training for the small tumor volume on brain MRI. The model was trained and tested on 587 and 150 patient data, respectively, from this institution (n = 495) and a publicly available dataset (n = 242). The model performance were assessed by the Dice similarity coefficient (DSC), 95% Hausdorff distance (HD95), average symmetric surface (ASSD) and relative absolute volume difference (RAVD) of the model segmentation results against the GTVs. RESULTS:Measured on combined testing data from two institutions, the proposed method achieved mean DSC of 0.91 ± 0.08, ASSD of 0.3 ± 0.4 mm, HD95 of 1.3 ± 1.6 mm, and RAVD of 0.09 ± 0.15. The DSCs were 0.91 ± 0.09 and 0.92 ± 0.06 on 100 testing patients of this institution and 50 of the public data, respectively. CONCLUSIONS:A CNN model was developed for fully automated segmentation of VS on T1-Weighted isotropic MRI. The model achieved good performance compared with physician clinical delineations on a sizeable dataset from two institutions. The proposed method potentially facilitates clinical workflow of radiosurgery for VS patient management.

PURPOSE/UNASSIGNED:Cardiac radioablation is an emerging therapy for recurrent ventricular tachycardia. Electrophysiology (EP) data, including electroanatomic maps (EAM) and electrocardiographic imaging (ECGI), provide crucial information for defining the arrhythmogenic target volume. The absence of standardized workflows and software tools to integrate the EP maps into a radiation planning system limits their use. This study developed a comprehensive software tool to enable efficient utilization of the mapping for cardiac radioablation treatment planning. METHODS AND MATERIALS/UNASSIGNED:After the scar area is outlined on the mapping surface, the tool extracts and extends the annotated patch into a closed surface and converts it into a structure set associated with the anatomic images. The tool then exports the structure set and the images as The Digital Imaging and Communications in Medicine Standard in Radiotherapy for a radiation treatment planning system to import. Overlapping the scar structure on simulation CT, a transmural target volume is delineated for treatment planning. RESULTS/UNASSIGNED:The tool has been used to transfer Ensite NavX EAM data into the Varian Eclipse treatment planning system in radioablation on 2 patients with ventricular tachycardia. The ECGI data from CardioInsight was retrospectively evaluated using the tool to derive the target volume for a patient with left ventricular assist device, showing volumetric matching with the clinically used target with a Dice coefficient of 0.71. CONCLUSIONS/UNASSIGNED:HeaRTmap smoothly fuses EP information from different mapping systems with simulation CT for accurate definition of radiation target volume. The efficient integration of EP data into treatment planning potentially facilitates the study and adoption of the technique.

PURPOSE/OBJECTIVE:Stereotactic radiosurgery (SRS) has become an important modality in the treatment of brain metastases. The purpose of this study is to investigate the potential of radiomic features from planning magnetic resonance (MR) images and dose maps to predict local failure after SRS for brain metastases. MATERIALS/METHODS/METHODS:Twenty-eight patients who received Gamma Knife (GK) radiosurgery for brain metastases were retrospectively reviewed in this IRB-approved study. 179 irradiated tumors included 42 that locally failed within one-year follow-up. Using SRS tumor volumes, radiomic features were calculated on T1-weighted contrast-enhanced MR images acquired for treatment planning and planned dose maps. 125 radiomic features regarding tumor shape, dose distribution, MR intensities and textures were extracted for each tumor. Logistic regression with automatic feature selection was built to predict tumor progression from local control after SRS. Feature selection and model evaluation using receiver operating characteristic (ROC) curves were performed in a nested cross validation (CV) scheme. The associations between selected radiomic features and treatment outcomes were statistically assessed by univariate analysis. RESULTS:The logistic model with feature selection achieved ROC AUC of 0.82 ± 0.09 on 5-fold CV, providing 83% sensitivity and 70% specificity for predicting local failure. A total of 10 radiomic features including 1 shape feature, 6 MR images and 3 dose distribution features were selected. These features were significantly associated with treatment outcomes (p < 0.05). The model was validated on independent holdout data with an AUC of 0.78. CONCLUSIONS:Radiomic features from planning MR images and dose maps provided prognostic information in SRS for brain metastases. A model built on the radiomic features shows promise for early prediction of tumor local failure after treatment, potentially aiding in personalized care for brain metastases.

PURPOSE/OBJECTIVE:Owing to histologic complexities of brain tumors, its diagnosis requires the use of multimodalities to obtain valuable structural information so that brain tumor subregions can be properly delineated. In current clinical workflow, physicians typically perform slice-by-slice delineation of brain tumor subregions, which is a time-consuming process and also more susceptible to intra- and inter-rater variabilities possibly leading to misclassification. To deal with this issue, this study aims to develop an automatic segmentation of brain tumor in MR images using deep learning. METHOD/METHODS:In this study, we develop a context deep-supervised U-Net to segment brain tumor subregions. A context block which aggregates multiscale contextual information for dense segmentation was proposed. This approach enlarges the effective receptive field of convolutional neural networks, which, in turn, improves the segmentation accuracy of brain tumor subregions. We performed the fivefold cross-validation on the Brain Tumor Segmentation Challenge (BraTS) 2020 training dataset. The BraTS 2020 testing datasets were obtained via BraTS online website as a hold-out test. For BraTS, the evaluation system divides the tumor into three regions: whole tumor (WT), tumor core (TC), and enhancing tumor (ET). The performance of our proposed method was compared against two state-of-the-arts CNN networks in terms of segmentation accuracy via Dice similarity coefficient (DSC) and Hausdorff distance (HD). The tumor volumes generated by our proposed method were compared with manually contoured volumes via Bland-Altman plots and Pearson analysis. RESULTS:The proposed method achieved the segmentation results with a DSC of 0.923 ± 0.047, 0.893 ± 0.176, and 0.846 ± 0.165 and a 95% HD95 of 3.946 ± 7.041, 3.981 ± 6.670, and 10.128 ± 51.136 mm on WT, TC, and ET, respectively. Experimental results demonstrate that our method achieved comparable to significantly (p < 0.05) better segmentation accuracies than other two state-of-the-arts CNN networks. Pearson correlation analysis showed a high positive correlation between the tumor volumes generated by proposed method and manual contour. CONCLUSION/CONCLUSIONS:Overall qualitative and quantitative results of this work demonstrate the potential of translating proposed technique into clinical practice for segmenting brain tumor subregions, and further facilitate brain tumor radiotherapy workflow.

Purpose: This study investigated the accuracy of two dose calculation algorithms (AAA version 15.6.05 and AcurosXB (AXB) version 15.6.05) in lung, fat, and bone regions of a lung phantom using 2D Gamma analysis.
Method(s): A commercial thorax phantom (CIRS) was marked and CT simulated with 1.5 mm slice thickness then planned in Eclipse v15.6 (Varian) using an oblique 3x3 cm field traversing soft tissue, lung, and spine regions. Dose was calculated using AAA and AXB algorithms at 1.0 mm resolution for four different beam energies with the same number of monitor units: 6X, 6X-FFF, 10X and 10X-FFF. Each field was separately delivered to the phantom with Gafchromic EBT3 film placed in the axial plane through the beam isocenter. Film alignment was performed using built-in phantom pins as registration marks. Gamma analysis was performed in SNC Patient 8.2 (Sun Nuclear Corp.) software using distanceto- agreement (DTA) to compare AAA and AXB calculated doses against one another and against measured dose.
Result(s): Considerable dosimetric differences occurred between AAA and AXB especially at tissue interfaces and in the beam penumbra within the lung. About 10% and 30% of analyzed points had dose differences larger than 3% for 6X and 10X, respectively. The differences were a few percentages smaller for non-flattened beams. In comparison of calculation with measurement, the 3%/2mm-DTA passing rate for AAA was 3.7% higher than AXB for the 6X beam (3.9% for 6X-FFF), but -15.9% lower for 10X (- 24.9% for 10X-FFF).
Conclusion(s): Dose calculation algorithm accuracy was assessed in a heterogeneous thorax phantom using 2D gamma analysis and Gafchromic film. Large differences in the passing rate, especially at 10X(FFF), suggest to verify that clinical plan evaluation metrics such as coverage and dose constrains are not compromised by the choice of dose calculation algorithm

PURPOSE/OBJECTIVE:The treatment couch position of a patient in external beam radiation therapy (EBRT) is usually acquired during initial treatment setup. This procedure has shown potential failure modes leading to near misses and adverse events in radiation treatment. This study aims to develop a method to automatically determine the couch position before setting up a patient for initial treatment. METHODS:The Qfix couch-tops (kVue and DoseMax) have embedded reference marks (BBs) indicating its index levels and couch centerline. With the ESAPI, a C# script was programmed to automatically find the couch-top and embedded BBs in the planning CT and derive the treatment couch position according to treatment isocenter of a plan. Couch positions of EBRT plans with the kVue couch-top and SBRT plans using the DoseMax were calculated using the script. The calculation was evaluated by comparing calculated positions with couch coordinates captured during the initial treatment setup after image guidance. The calculations were further compared with daily treatment couch positions post image-guided adjustment for each treatment fraction. RESULTS:For plans using the kVue couch-top for various treatment sites, the median (5-95 percentiles) differences between calculated and captured couch positions were 0.1 (-0.2 - 0.9), 0.5 (-1.1-2.0), 0.10 (-1.3-1.3) cm in the vertical, longitudinal, and lateral direction respectively. For the DoseMax couch-top, the median differences were 0.1 (-0.2-0.7), 0.2 (-0.3-1.1), and 0.2 (-0.7-0.9) cm in respective direction. The calculated positions were within 1 and 2 cm from the mean fraction positions for 95% patients on DoseMax and kVue couch-top respectively. CONCLUSIONS:A method that automatically and accurately calculates treatment couch position from simulation CT was implemented in Varian Eclipse for Qfix couch-tops. This technique increases the efficiency of patient setup and enhances patient safety by reducing the risks of positioning errors.

PURPOSE/OBJECTIVE:The purpose of this study is to quantify dosimetric changes throughout the delivery of oropharyngeal cancer treatment and to investigate the application of statistical process control (SPC) for the management of significant deviations during the course of radiotherapy. METHODS:Thirteen oropharyngeal cancer patients with daily cone beam computed tomography (CBCT) were retrospectively reviewed. Cone beam computed tomography images of every other fraction were imported to the Velocity software and registered to planning CT using the 6 DOF (degrees of freedom) couch shifts generated during patient setup. Using Velocity "Adaptive Monitoring" module, the setup-corrected CBCT was matched to planning CT using a deformable registration. Volumes and dose metrics at each fraction were calculated and rated with plan values to evaluate interfractional dosimetric variations using a SPC framework. T-tests between plan and fraction volumes were performed to find statistically insignificant fractions. Average upper and lower process capacity limits (UCL, LCL) of each dose metric were derived from these fractions using conventional SPC guidelines. RESULTS:Gross tumor volume (GTV) and organ at risk (OAR) volumes in the first 13 fractions had no significant changes from the pretreatment planning CT. The GTV and the parotid glands subsequently decreased by 10% at the completion of treatment. There were 3-4% increases in parotid mean doses, but no significant differences in dose metrics of GTV and other OARs. The changes were organ and patient dependent. Control charts for various dose metrics were generated to assess the metrics at each fraction for individual patient. CONCLUSIONS:Daily CBCT could be used to monitor dosimetric variations of targets and OARs resulting from volume changes and tissue deformation in oropharyngeal cancer radiotherapy. Treatment review with the guidance of a SPC tool allows for an objective and consistent clinical decision to apply adaptive radiotherapy.

PURPOSE/OBJECTIVE:Ultrasound (US)-guided high dose rate (HDR) prostate brachytherapy requests the clinicians to place HDR needles (catheters) into the prostate gland under transrectal US (TRUS) guidance in the operating room. The quality of the subsequent radiation treatment plan is largely dictated by the needle placements, which varies upon the experience level of the clinicians and the procedure protocols. Real-time plan dose distribution, if available, could be a vital tool to provide more subjective assessment of the needle placements, hence potentially improving the radiation plan quality and the treatment outcome. However, due to low signal-to-noise ratio (SNR) in US imaging, real-time multi-needle segmentation in 3D TRUS, which is the major obstacle for real-time dose mapping, has not been realized to date. In this study, we propose a deep learning-based method that enables accurate and real-time digitization of the multiple needles in the 3D TRUS images of HDR prostate brachytherapy. METHODS:A deep learning model based on the U-Net architecture was developed to segment multiple needles in the 3D TRUS images. Attention gates were considered in our model to improve the prediction on the small needle points. Furthermore, the spatial continuity of needles was encoded into our model with total variation (TV) regularization. The combined network was trained on 3D TRUS patches with the deep supervision strategy, where the binary needle annotation images were provided as ground truth. The trained network was then used to localize and segment the HDR needles for a new patient's TRUS images. We evaluated our proposed method based on the needle shaft and tip errors against manually defined ground truth and compared our method with other state-of-art methods (U-Net and deeply supervised attention U-Net). RESULTS:Our method detected 96% needles of 339 needles from 23 HDR prostate brachytherapy patients with 0.290 ± 0.236 mm at shaft error and 0.442 ± 0.831 mm at tip error. For shaft localization, our method resulted in 96% localizations with less than 0.8 mm error (needle diameter is 1.67 mm), while for tip localization, our method resulted in 75% needles with 0 mm error and 21% needles with 2 mm error (TRUS image slice thickness is 2 mm). No significant difference is observed (P = 0.83) on tip localization between our results with the ground truth. Compared with U-Net and deeply supervised attention U-Net, the proposed method delivers a significant improvement on both shaft error and tip error (P < 0.05). CONCLUSIONS:We proposed a new segmentation method to precisely localize the tips and shafts of multiple needles in 3D TRUS images of HDR prostate brachytherapy. The 3D rendering of the needles could help clinicians to evaluate the needle placements. It paves the way for the development of real-time plan dose assessment tools that can further elevate the quality and outcome of HDR prostate brachytherapy.