Browsing by Author "Yin, Fang-Fang"
Results Per Page
Sort Options
Item Open Access A Comparative Study of Radiomics and Deep-Learning Approaches for Predicting Surgery Outcomes in Early-Stage Non-Small Cell Lung Cancer (NSCLC)(2022) Zhang, HaozhaoPurpose: To compare radiomics and deep-learning (DL) methods for predicting NSCLC surgical treatment failure. Methods: A cohort of 83 patients undergoing lobectomy or wedge resection for early-stage NSCLC from our institution was studied. There were 7 local failures and 16 non-local failures (regional and/or distant). Gross tumor volumes (GTV) were contoured on pre-surgery CT datasets after 1mm3 isotropic resolution resampling. For the radiomics analysis, 92 radiomics features were extracted from the GTV and z-score normalizations were performed. The multivariate association between the extracted features and clinical endpoints were investigated using a random forest model following 70%-30% training-test split. For the DL analysis, both 2D and 3D model designs were executed using two different deep neural networks as transfer learning problems: in 2D-based design, 8x8cm2 axial fields-of-view(FOVs) centered within the GTV were adopted for VGG-16 training; in 3D-based design, 8x8x8 cm3 FOVs centered within the GTV were adopted for U-Net’s encoder path training. In both designs, data augmentation (rotation, translation, flip, noise) was included to overcome potential training convergence problems due to the imbalanced dataset, and the same 70%-30% training-test split was used. The performances of the 3 models (Radiomics, 2D-DL, 3D-DL) were tested to predict outcomes including local failure, non-local failure, and disease-free survival. Sensitivity/specificity/accuracy/ROC results were obtained from their 20 trained versions. Results: The radiomics models showed limited performances in all three outcome prediction tasks. The 2D-DL design showed significant improvement compared to the radiomics results in predicting local failure (ROC AUC = 0.546±0.056). The 3D-DL design achieved the best performance for all three outcomes (local failure ROC AUC = 0.768 ± 0.051, non-local failure ROC AUC = 0.683±0.027, disease-free ROC AUC = 0.694±0.042) with statistically significant improvements from radiomics/2D-DL results. Conclusions: 3D-DL execution outperformed the 2D-DL in predicting clinical outcomes after surgery for early-stage NSCLC. By contrast, classic radiomics approach did not achieve satisfactory results.
Item Open Access A positioning QA procedure for 2D/2D (kV/MV) and 3D/3D (CT/CBCT) image matching for radiotherapy patient setup.(Journal of applied clinical medical physics, 2009-10-06) Guan, Huaiqun; Hammoud, Rabih; Yin, Fang-FangA positioning QA procedure for Varian's 2D/2D (kV/MV) and 3D/3D (planCT/CBCT) matching was developed. The procedure was to check: (1) the coincidence of on-board imager (OBI), portal imager (PI), and cone beam CT (CBCT)'s isocenters (digital graticules) to a linac's isocenter (to a pre-specified accuracy); (2) that the positioning difference detected by 2D/2D (kV/MV) and 3D/3D(planCT/CBCT) matching can be reliably transferred to couch motion. A cube phantom with a 2 mm metal ball (bb) at the center was used. The bb was used to define the isocenter. Two additional bbs were placed on two phantom surfaces in order to define a spatial location of 1.5 cm anterior, 1.5 cm inferior, and 1.5 cm right from the isocenter. An axial scan of the phantom was acquired from a multislice CT simulator. The phantom was set at the linac's isocenter (lasers); either AP MV/R Lat kV images or CBCT images were taken for 2D/2D or 3D/3D matching, respectively. For 2D/2D, the accuracy of each device's isocenter was obtained by checking the distance between the central bb and the digital graticule. Then the central bb in orthogonal DRRs was manually moved to overlay to the off-axis bbs in kV/MV images. For 3D/3D, CBCT was first matched to planCT to check the isocenter difference between the two CTs. Manual shifts were then made by moving CBCT such that the point defined by the two off-axis bbs overlay to the central bb in planCT. (PlanCT can not be moved in the current version of OBI1.4.) The manual shifts were then applied to remotely move the couch. The room laser was used to check the accuracy of the couch movement. For Trilogy (or Ix-21) linacs, the coincidence of imager and linac's isocenter was better than 1 mm (or 1.5 mm). The couch shift accuracy was better than 2 mm.Item Open Access A Radiomics Machine Learning Model for Post-Radiotherapy Overall Survival Prediction of Non-Small Cell Lung Cancer (NSCLC)(2023) Zhang, RihuiPurpose: To predict post-radiotherapy overall survival group of NSCLC patients based on clinical information and radiomics analysis of simulation CT. Materials/Methods: A total of 258 non-adenocarcinoma patients who received radical radiotherapy or chemo-radiation were studied: 45/50/163 patients were identified as short(0-6mos)/mid(6-12mos)/long(12+mos) survival groups, respectively. For each patient, we first extracted 76 radiomics features within the gross tumor volume(GTV) identified in the simulation CT; these features were combined with patient clinical information (age, overall stage, and GTV volume) as a patient-specific feature vector, which was utilized by a 2-step machine learning model for survival group prediction. This model first identifies patients with long survival prediction via a supervised binary classifier; for those with otherwise prediction, a 2nd classifier further generates short/mid survival prediction. Two machine learning classifiers, explainable boosting machine(EBM) and balanced random forest(BRF), were interrogated as a comparison study. During the model training, all patients were divided into training/test sets by an 8:2 ratio, and 100-fold random sampling were applied to the training set with a 7:1 validation ratio. Model performances were evaluated by the sensitivity, accuracy, and ROC results. Results: The model with EBM demonstrated an overall ROC AUC (0.58±0.04) with limited sensitivities in short (0.02±0.04) and mid group (0.11±0.08) predictions due to imbalanced data sample distribution. In contrast, the model with BRF improved short/mid group sensitivities to 0.32±0.11/0.29±0.16, respectively, but the improvement of ROC AUC (0.60±0.04) is limited. Nevertheless, both EBM (0.46±0.04) and BRF (0.57±0.04) approaches achieved limited overall accuracy; a noticeable overlap was found in their feature lists with top 10 feature weight rankings. Conclusion: The proposed two-step machine learning model with BRF classifier possesses a better performance than the one with EBM classifier in the post-radiotherapy survival group prediction of NSCLC. Future works, preferably in the joint use of deep learning, are in demand to further improve the prediction results.
Item Open Access A robust deformable image registration enhancement method based on radial basis function.(Quantitative imaging in medicine and surgery, 2019-07) Liang, Xiao; Yin, Fang-Fang; Wang, Chunhao; Cai, JingBackground:To develop and evaluate a robust deformable image registration (DIR) enhancement method based on radial basis function (RBF) expansion. Methods:To improve DIR accuracy using sparsely available measured displacements, it is crucial to estimate the motion correlation between the voxels. In the proposed method, we chose to derive this correlation from the initial displacement vector fields (DVFs), and represent it in the form of RBF expansion coefficients of the voxels. The method consists of three steps: (I) convert an initial DVF to a coefficient matrix comprising expansion coefficients of the Wendland's RBF; (II) modify the coefficient matrix under the guidance of sparely distributed landmarks to generate the post-enhancement coefficient matrix; and (III) convert the post-enhancement coefficient matrix to the post-enhancement DVF. The method was tested on five DIR algorithms using a digital phantom. 3D registration errors were calculated for comparisons between the pre-/post-enhancement DVFs and the ground-truth DVFs. Effects of the number and locations of landmarks on DIR enhancement were evaluated. Results:After applying the DIR enhancement method, the 3D registration errors per voxel (unit: mm) were reduced from pre-enhancement to post-enhancement by 1.3 (2.4 to 1.1, 54.2%), 0.0 (0.9 to 0.9, 0.0%), 6.1 (8.2 to 2.1, 74.4%), 3.2 (4.7 to 1.5, 68.1%), and 1.7 (2.9 to 1.2, 58.6%) for the five tested DIR algorithms respectively. The average DIR error reduction was 2.5±2.3 mm (percentage error reduction: 51.1%±29.1%). 3D registration errors decreased inverse-exponentially as the number of landmarks increased, and were insensitive to the landmarks' locations in relation to the down-sampling DVF grids. Conclusions:We demonstrated the feasibility of a robust RBF-based method for enhancing DIR accuracy using sparsely distributed landmarks. This method has been shown robust and effective in reducing DVF errors using different numbers and distributions of landmarks for various DIR algorithms.Item Open Access Accelerating Brain DTI and GYN MRI Studies Using Neural Network(2021) Yan, YuhaoThere always exists a demand to accelerate the time-consuming MRI acquisition process. Many methods have been proposed to achieve this goal, including deep learning method which appears to be a robust tool compared to conventional methods. While many works have been done to evaluate the performance of neural networks on standard anatomical MR images, few attentions have been paid to accelerating other less conventional MR image acquisitions. This work aims to evaluate the feasibility of neural networks on accelerating Brain DTI and Gynecological Brachytherapy MRI. Three neural networks including U-net, Cascade-net and PD-net were evaluated. Brain DTI data was acquired from public database RIDER NEURO MRI while cervix gynecological MRI data was acquired from Duke University Hospital clinic data. A 25% Cartesian undersampling strategy was applied to all the training and test data. Diffusion weighted images and quantitative functional maps in Brain DTI, T1-spgr and T2 images in GYN studies were reconstructed. The performance of the neural networks was evaluated by quantitatively calculating the similarity between the reconstructed images and the reference images, using the metric Total Relative Error (TRE). Results showed that with the architectures and parameters set in this work, all three neural networks could accelerate Brain DTI and GYN T2 MR imaging. Generally, PD-net slightly outperformed Cascade-net, and they both outperformed U-net with respect to image reconstruction performance. While this was also true for reconstruction of quantitative functional diffusion weighted maps and GYN T1-spgr images, the overall performance of the three neural networks on these two tasks needed further improvement. To be concluded, PD-net is very promising on accelerating T2-weighted-based MR imaging. Future work can focus on adjusting the parameters and architectures of the neural networks to improve the performance on accelerating GYN T1-spgr MR imaging and adopting more robust undersampling strategy such as radial undersampling strategy to further improve the overall acceleration performance.
Item Open Access Accuracy and efficiency of image-guided radiation therapy (IGRT) for preoperative partial breast radiosurgery.(Journal of radiosurgery and SBRT, 2020-01) Yoo, Sua; O'Daniel, Jennifer; Blitzblau, Rachel; Yin, Fang-Fang; Horton, Janet KObjective
To analyze and evaluate accuracy and efficiency of IGRT process for preoperative partial breast radiosurgery.Methods
Patients were initially setup with skin marks and 5 steps were performed: (1) Initial orthogonal 2D kV images, (2) pre-treatment 3D CBCT images, (3) verification orthogonal 2D kV images, (4) treatment including mid-treatment 2D kV images (for the final 15 patients only), and (5) post-treatment orthogonal 2D kV or 3D CBCT images. Patient position was corrected at each step to align the biopsy clip and to verify surrounding soft tissue positioning.Results
The mean combined vector magnitude shifts and standard deviations at the 5 imaging steps were (1) 0.96 ± 0.69, (2) 0.33 ± 0.40, (3) 0.05 ± 0.12, (4) 0.15 ± 0.17, and (5) 0.27 ± 0.24 in cm. The mean total IGRT time was 40.2 ± 13.2 minutes. Each step was shortened by 2 to 5 minutes with improvements implemented. Overall, improvements in the IGRT process reduced the mean total IGRT time by approximately 20 minutes. Clip visibility was improved by implementing oblique orthogonal images.Conclusion
Multiple imaging steps confirmed accurate patient positioning. Appropriate planning and imaging strategies improved the effectiveness and efficiency of the IGRT process for preoperative partial breast radiosurgery.Item Embargo Advanced Deep Learning Methods for Brain Metastasis Post-SRS Outcome Management(2023) Zhao, JingtongPurpose: The purpose of this study is to develop and validate two deep learning (DL) models for the management of brain metastasis (BM) patients treated with stereotactic radiosurgery (SRS). The first model is a radiomics-integrated deep learning (RIDL) model, which aims to distinguish between radionecrosis and tumor recurrence in patients with post-SRS radiographic progression. The second model, a novel dose-incorporated deep ensemble learning (DEL) model, aims to accurately predict local failure outcomes in brain metastasis patients following SRS.Materials/Methods: A total of 51 patients with post-SRS radiographic progression (37 radionecrosis, 14 recurrence) and 114 BMs (including 26 BMs that developed biopsy-confirmed local failure post-SRS) from 85 patients were included in this study. For the first aim, a radiomics-integrated deep learning (RIDL) model was developed using three steps: 1) 184 radiomics features (RFs) were extracted from the SRS planning target volume (PTV) and 60% isodose volume (V60%); 2) a deep neural network (DNN) mimicking the encoding path of U-net was trained for radionecrosis or recurrence prediction using the 3D MR volume. Prior to the binary prediction output, latent variables in the DNN were extracted as 512 deep features (DFs); and 3) all extracted features were synthesized as a multi-dimensional input for support vector machine (SVM) execution. Key features with higher linear kernel weighting factors were identified by clustering analysis and were utilized by SVM to predict radionecrosis or recurrence. During model training, 50 model versions were acquired with random validation sample assignments following an 8:2 training/test ratio, and sensitivity, specificity, accuracy, and ROC were evaluated and compared with results from a radiomics-only and a DNN-only prediction model. For the second aim, a novel dose-incorporated deep ensemble learning (DEL) model was developed. The DEL design included four VGG-19 deep encoder networks, and each sub-network utilized a different variable type as input for BM outcome prediction. The DEL's outcome was synthesized from the four sub-network results via logistic regression. For each BM, four variables were obtained, including three with different curvatures during spherical projection and one with the original planar images. The proposed DEL model was developed using an 8:2 ratio for training/test assignment, and 10 model versions were acquired with random validation sample assignments. The DEL model performance was compared based on ROC analysis to a single VGG-19 encoder and to DEL models with the same projection designs, which used T1-CE MRI as the only input. Results: The RIDL model demonstrated superior performance compared to radiomics-based and DNN-only prediction models for distinguishing radionecrosis from tumor recurrence in brain metastasis patients with post-SRS radiographic progression. The RIDL model achieved the best prediction accuracy (0.643±0.059) and sensitivity (0.650±0.122) results with 32 identified key features (3 RFs+29 DFs), and it also demonstrated superior ROC results (AUC=0.688±0.035). In addition, for patients with NSCLC primary disease, the RF joint energy extracted from V60% and one DF correlated with ALK/EGFR mutations, respectively. Moreover, the DEL model achieved an excellent ROC AUC=0.84±0.03 with high sensitivity (0.78±0.08), specificity (0.81±0.09), and accuracy (0.80±0.06) results. This outperformed the MRI-only single VGG-19 encoder (sensitivity:0.35±0.01, AUC:0.64±0.08) and the MRI-only DEL (sensitivity:0.60±0.09, AUC:0.68±0.06) models. Conclusions: The RIDL model successfully differentiates brain metastasis radionecrosis from recurrence using a single post-SRS MR scan. Integration of clinical and treatment-related features is warranted to develop a comprehensive clinico-radiomic model. Additionally, the dose-incorporated DEL model design demonstrated robust and promising performance. It could potentially improve other radiotherapy outcome models and warrant further evaluation.
Item Open Access An Ensemble Approach to Knowledge-Based Intensity-Modulated Radiation Therapy Planning.(Frontiers in oncology, 2018-01) Zhang, Jiahan; Wu, Q Jackie; Xie, Tianyi; Sheng, Yang; Yin, Fang-Fang; Ge, YaorongKnowledge-based planning (KBP) utilizes experienced planners' knowledge embedded in prior plans to estimate optimal achievable dose volume histogram (DVH) of new cases. In the regression-based KBP framework, previously planned patients' anatomical features and DVHs are extracted, and prior knowledge is summarized as the regression coefficients that transform features to organ-at-risk DVH predictions. In our study, we find that in different settings, different regression methods work better. To improve the robustness of KBP models, we propose an ensemble method that combines the strengths of various linear regression models, including stepwise, lasso, elastic net, and ridge regression. In the ensemble approach, we first obtain individual model prediction metadata using in-training-set leave-one-out cross validation. A constrained optimization is subsequently performed to decide individual model weights. The metadata is also used to filter out impactful training set outliers. We evaluate our method on a fresh set of retrospectively retrieved anonymized prostate intensity-modulated radiation therapy (IMRT) cases and head and neck IMRT cases. The proposed approach is more robust against small training set size, wrongly labeled cases, and dosimetric inferior plans, compared with other individual models. In summary, we believe the improved robustness makes the proposed method more suitable for clinical settings than individual models.Item Open Access An Exploratory Radiomics Approach to Quantifying Pulmonary Function in CT Images(Scientific Reports, 2019-12) Lafata, Kyle J; Zhou, Zhennan; Liu, Jian-Guo; Hong, Julian; Kelsey, Chris R; Yin, Fang-FangItem Open Access An Investigation of Machine Learning Methods for Delta-radiomic Feature Analysis(2018) Chang, YushiBackground: Radiomics is a process of converting medical images into high-dimensional quantitative features and the subsequent mining these features for providing decision support. It is conducted as a potential noninvasive, low-cost, and patient-specific routine clinical tool. Building a predictive model which is reliable, efficient, and accurate is a vital part for the success of radiomics. Machine learning method is a powerful tool to achieve this. Feature extraction strongly affects the performance. Delta-feature is one way of feature extraction methods to reflect the spatial variation in tumor phenotype, hence it could provide better treatment-specific assessment.
Purpose: To compare the performance of using pre-treatment features and delta-features for assessing the brain radiosurgery treatment response, and to investigate the performance of different combinations of machine learning methods for feature selection and for feature classification.
Materials and Methods: A cohort of 12 patients with brain treated by radiosurgery was included in this research. The pre-treatment, one-week post-treatment, and two-month post-treatment T1 and T2 FLAIR MR images were acquired. 61 radiomic features were extracted from the gross tumor volume (GTV) of each image. The delta-features from pre-treatment to two post-treatment time points were calculated. With leave-one-out sampling, pre-treatment features and the two sets of delta-features were separately input into a univariate Cox regression model and a machine learning model (L1-regularized logistic regression [L1-LR], random forest [RF] or neural network [NN]) for feature selection. Then a machine learning method (L1-LR, L2-regularized logistic regression [L2-LR], RF, NN, kernel support vector machine [Kernel-SVM], linear-SVM, or naïve bayes [NB]) was used to build a classification model to predict overall survival. The performance of each model combination and feature type was estimated by the area under receiver operating characteristic (ROC) curve (AUC).
Results: The AUC of one-week delta-features was significantly higher than that of pre-treatment features (p-values < 0.0001) and two-month delta-features (p-value= 0.000). The model combinations of L1-LR for feature selection and RF for classification as well as RF for feature selection and NB for classification based on one-week delta-features presented the highest AUC values (both AUC=0.944).
Conclusions: This work potentially implied that the delta-features could be better in predicting treatment response than pre-treatment features, and the time point of computing the delta-features was a vital factor in assessment performance. Analyzing delta-features using a suitable machine learning approach is potentially a powerful tool for assessing treatment response.
Item Open Access An investigation of machine learning methods in delta-radiomics feature analysis.(PloS one, 2019-01) Chang, Yushi; Lafata, Kyle; Sun, Wenzheng; Wang, Chunhao; Chang, Zheng; Kirkpatrick, John P; Yin, Fang-FangPURPOSE:This study aimed to investigate the effectiveness of using delta-radiomics to predict overall survival (OS) for patients with recurrent malignant gliomas treated by concurrent stereotactic radiosurgery and bevacizumab, and to investigate the effectiveness of machine learning methods for delta-radiomics feature selection and building classification models. METHODS:The pre-treatment, one-week post-treatment, and two-month post-treatment T1 and T2 fluid-attenuated inversion recovery (FLAIR) MRI were acquired. 61 radiomic features (intensity histogram-based, morphological, and texture features) were extracted from the gross tumor volume in each image. Delta-radiomics were calculated between the pre-treatment and post-treatment features. Univariate Cox regression and 3 multivariate machine learning methods (L1-regularized logistic regression [L1-LR], random forest [RF] or neural networks [NN]) were used to select a reduced number of features, and 7 machine learning methods (L1-LR, L2-LR, RF, NN, kernel support vector machine [KSVM], linear support vector machine [LSVM], or naïve bayes [NB]) was used to build classification models for predicting OS. The performances of the total 21 model combinations built based on single-time-point radiomics (pre-treatment, one-week post-treatment, and two-month post-treatment) and delta-radiomics were evaluated by the area under the receiver operating characteristic curve (AUC). RESULTS:For a small cohort of 12 patients, delta-radiomics resulted in significantly higher AUC than pre-treatment radiomics (p-value<0.01). One-week/two-month delta-features resulted in significantly higher AUC (p-value<0.01) than the one-week/two-month post-treatment features, respectively. 18/21 model combinations were with higher AUC from one-week delta-features than two-month delta-features. With one-week delta-features, RF feature selector + KSVM classifier and RF feature selector + NN classifier showed the highest AUC of 0.889. CONCLUSIONS:The results indicated that delta-features could potentially provide better treatment assessment than single-time-point features. The treatment assessment is substantially affected by the time point for computing the delta-features and the combination of machine learning methods for feature selection and classification.Item Open Access Association of pre-treatment radiomic features with lung cancer recurrence following stereotactic body radiation therapy.(Physics in medicine and biology, 2019-01-08) Lafata, Kyle J; Hong, Julian C; Geng, Ruiqi; Ackerson, Bradley G; Liu, Jian-Guo; Zhou, Zhennan; Torok, Jordan; Kelsey, Chris R; Yin, Fang-FangThe purpose of this work was to investigate the potential relationship between radiomic features extracted from pre-treatment x-ray CT images and clinical outcomes following stereotactic body radiation therapy (SBRT) for non-small-cell lung cancer (NSCLC). Seventy patients who received SBRT for stage-1 NSCLC were retrospectively identified. The tumor was contoured on pre-treatment free-breathing CT images, from which 43 quantitative radiomic features were extracted to collectively capture tumor morphology, intensity, fine-texture, and coarse-texture. Treatment failure was defined based on cancer recurrence, local cancer recurrence, and non-local cancer recurrence following SBRT. The univariate association between each radiomic feature and each clinical endpoint was analyzed using Welch's t-test, and p-values were corrected for multiple hypothesis testing. Multivariate associations were based on regularized logistic regression with a singular value decomposition to reduce the dimensionality of the radiomics data. Two features demonstrated a statistically significant association with local failure: Homogeneity2 (p = 0.022) and Long-Run-High-Gray-Level-Emphasis (p = 0.048). These results indicate that relatively dense tumors with a homogenous coarse texture might be linked to higher rates of local recurrence. Multivariable logistic regression models produced maximum [Formula: see text] values of [Formula: see text], and [Formula: see text], for the recurrence, local recurrence, and non-local recurrence endpoints, respectively. The CT-based radiomic features used in this study may be more associated with local failure than non-local failure following SBRT for stage I NSCLC. This finding is supported by both univariate and multivariate analyses.Item Open Access Automatic Planning of Whole Breast Radiation Therapy Using Machine Learning Models.(Frontiers in Oncology, 2019-01) Sheng, Yang; Li, Taoran; Yoo, Sua; Yin, Fang-Fang; Blitzblau, Rachel; Horton, Janet K; Ge, Yaorong; Wu, Q JackiePurpose: To develop an automatic treatment planning system for whole breast radiation therapy (WBRT) based on two intensity-modulated tangential fields, enabling near-real-time planning. Methods and Materials: A total of 40 WBRT plans from a single institution were included in this study under IRB approval. Twenty WBRT plans, 10 with single energy (SE, 6MV) and 10 with mixed energy (ME, 6/15MV), were randomly selected as training dataset to develop the methodology for automatic planning. The rest 10 SE cases and 10 ME cases served as validation. The auto-planning process consists of three steps. First, an energy prediction model was developed to automate energy selection. This model establishes an anatomy-energy relationship based on principle component analysis (PCA) of the gray level histograms from training cases' digitally reconstructed radiographs (DRRs). Second, a random forest (RF) model generates an initial fluence map using the selected energies. Third, the balance of overall dose contribution throughout the breast tissue is realized by automatically selecting anchor points and applying centrality correction. The proposed method was tested on the validation dataset. Non-parametric equivalence test was performed for plan quality metrics using one-sided Wilcoxon Signed-Rank test. Results: For validation, the auto-planning system suggested same energy choices as clinical-plans in 19 out of 20 cases. The mean (standard deviation, SD) of percent target volume covered by 100% prescription dose was 82.5% (4.2%) for auto-plans, and 79.3% (4.8%) for clinical-plans (p > 0.999). Mean (SD) volume receiving 105% Rx were 95.2 cc (90.7 cc) for auto-plans and 83.9 cc (87.2 cc) for clinical-plans (p = 0.108). Optimization time for auto-plan was <20 s while clinical manual planning takes between 30 min and 4 h. Conclusions: We developed an automatic treatment planning system that generates WBRT plans with optimal energy selection, clinically comparable plan quality, and significant reduction in planning time, allowing for near-real-time planning.Item Open Access Automatic Pulmonary Nodule Detection and Localization from Biplanar Chest Radiographs Using Convolutional Neural Network(2019) Hu, Shen-ChiangChest x-ray (CXR) is the most common examination in pulmonary nodule detection and an automatic nodule detection algorithm is desirable. Currently, convolutional neural network (CNN) is widely applied in CXR. However, there is a lack of dataset with clear nodule annotation, also the small size of pulmonary nodules hampers its performance,
finally, there is no study of lung nodule detection utilizing end-to-end CNN model and lateral CXR images. In this study, coronal and lateral CXR images were generated from CT phantom for training separately, and U-Net architecture CNN models were implemented with modifying a number of convolutional layers, adding shortcut connection, using weighted loss function and the impact of these modifications was evaluated on model performance. Finally, the models were tested on a test set under the condition of different nodule diameter, number, and location. In CT phantom dataset, U-Net trained with the residual unit and weighted loss showed the capability in detecting 5 mm nodules and increased training speed. Overall, model trained with coronal images provided better detection result than using lateral images, but their outputs could be combined to obtain nodule localization information in 3D. The number of nodules and adjacency of nodules has no prominent effect on detection, however, models were prone to failure when the nodule was too small (< 5 mm), was close to the edges of the lung, or was overlapped with moderate to the high-density anatomic structure.
Item Embargo Benign and Malignant Lymph Nodes Classification in Non-Small Cell Lung Cancer via Machine Learning Model(2024) Ge, JingyuObjective:To develop a machine learning model that integrates deep learning image features and radiomics features to classify lymph nodes as benign or malignant in Non-Small Cell Lung Cancer (NSCLC). Methods: The dataset comprises contrast-enhanced CT scans from 541 lung cancer patients before surgery, collected at a Shanghai Hospital between July 2015 and December 2017 under an IRB study. It includes 1,237 lymph nodes, identified from preoperative CT scans due to enlargement and confirmed as non-small cell lung cancer (NSCLC) via surgical pathology. Lymph node classification into malignant or benign categories utilized in postoperative pathological reports. Our method employs a dual radiomic feature extraction strategy. The deep image features (DIF) were derived from the final convolutional layer of a pre-trained VGG-16 encoder network to characterize the lymph node’s image texture. A total of nine 2D shape-based radiomic features (RF) are extracted based on the Py-radiomics calculation toolbox to characterize lymph node morphological information. And ninety-two handcrafted radiomic features (HRF) are extracted. The extracted DIF, RF, and HRF were combined and fed into a Random Forest classifier for the benign and malignant lymph node classification. The random forest classifier was trained following an 8:2 train/test split ratio and evaluated using Area Under the Curve (AUC), Receiver Operating Characteristic (ROC), and p-value, and 5-fold cross-validation was also employed to objectively evaluate model performance.
Results: The mean AUC for the Random Forest classifier using only 2D shape features is 0.691, while mean AUC for the classifier employing only DIF is 726. Utilizing both DIF and HRF for classification resulted in an average AUC of 0.724, whereas integrating RF with DIF achieved superior classification efficacy, boasting the highest average AUC of 0.746. All results were considered statistically significant with a p-value of less than 0.05. Conclusion: The combination of image texture analysis refers to DIF with morphological information offers an enhanced characterization ability to classify lymph nodes as benign or malignant from CT images for lung NSCLC patients.
Item Embargo CBCT image enhancement for improving accuracy of radiomics analysis and soft tissue target localization(2023) Zhang, ZeyuCone-beam computed tomography (CBCT) is one of the most commonly used image modalities in radiation therapy. It provides valuable information for target localization and outcome prediction throughout treatment courses. However, CBCT images suffer from various artifacts caused by scattering, beam hardening, undersampling, system hardware instability, and motions of the patient, which severely degrade the CBCT image quality. In addition, CBCT images have extremely poor soft-tissue contrast, making it almost impossible to accurately localize tumors in the soft tissue, such as liver tumors.
This dissertation presents the improvements of CBCT image quality for better outcome prediction and target localization by developing the deep learning and finite element based image enhancement model.
A deep learning based CBCT image enhancement model was developed to improve the radiomic feature accuracy. The model was trained based on 4D CBCT of ten patients and tested on three patients with different tumor sizes. The results show that 4D CBCT image quality can substantially affect the accuracy of the radiomic features, and the degree of impact is feature-dependent. The deep learning model was able to enhance the anatomical details and edge information in the 4D-CBCT as well as removing other image artifacts. This enhancement of image quality resulted in reduced errors for most radiomic
features. The average reduction of radiomics errors for 3 patients are 20.0%, 31.4%, 36.7%, 50.0%, 33.6% and 11.3% for histogram, GLCM, GLRLM, GLSZM, NGTDM and Wavelet features. And the error reduction was more significant for patients with larger tumors. To further improve the results, a patient-specific based training model has been developed. The model was trained based on the augmentation dataset of a single patient and tested on the different 4D CBCT of the same patient. Compared with a group-based model, the patient-specific training model further improved the accuracy of radiomic features, especially for features with large errors in the group-based model. For example, the 3D whole-body and ROI loss-based patient-specific model reduces the errors of the first-order median feature by 83.67%, the wavelet LLL feature maximum by 91.98%, and the wavelet HLL skewness feature by 15.0% on average for the four patients tested.
In addition, a patient-specific deep learning model is proposed to generate synthetic magnetic resonance imaging (MRI) from CBCT to improve tumor localization. A key innovation is using patient-specific CBCT-MRI image pairs to train a deep learning model to generate synthetic MRI from CBCT. Specifically, patient planning CT was deformably registered to prior MRI, and then used to simulate CBCT with simulated projections and Feldkamp, Davis, and Kress reconstruction. These CBCT-MRI images were augmented using translations and rotations to generate enough patient-specific training data. A U-Net-based deep learning model was developed and trained to generate synthetic MRI from CBCT in the liver, and then tested on a different CBCT dataset.
Synthetic MRIs were quantitatively evaluated against ground-truth MRI. On average, the synthetic MRI achieved 28.01, 0.025, and 0.929 for peak signal-to-noise ratio, mean square error, and structural similarity index, respectively, outperforming CBCT images on the three patients tested. To further improve the robustness of synthetic MRI generation, we developed an organ specific biomechanical model. This model registers the pretreatment MRI images to onboard CBCT images based on the organ contours, and combines the MRI organ with CBCT body to the generate hybrid MRI/CBCT. 48 registration cases were performed, which includes 18 Monte Carlo simulated cases and 30 real patient cases. We identified tumor landmarks of hybrid MRI/CBCT, onboard CBCT and planning CT, and calculated errors of landmark locations of two CBCT images. The errors were calculated based on the landmark differences of two CBCT images and ground truth planning CT. The results show that the tumor landmark localization accuracy around tumor is improved by 54.2 ± 22.2 %.
Item Open Access Clinical assessment and characterization of a dual tube kilovoltage X-ray localization system in the radiotherapy treatment room.(Journal of applied clinical medical physics, 2008-01-13) Lee, Sung-Woo; Jin, Jian-Yue; Guan, Huaiqun; Martin, Flavious; Kim, Jae Ho; Yin, Fang-FangINTRODUCTION:Although flat-panel based X-ray imaging has been well tested in diagnostic radiology, its use as an image-guided-radiotherapy (IGRT) system in a treatment room is new and requires systematic assessment. MATERIALS AND METHODS:BrainLab Novalis IGRT system was used for this study. It consists of two floor mounted kV X-ray tubes projecting obliquely into two flat-panel detectors mounted on the ceiling. The system automatically fuses the 2D localization images with 3D simulation CT image to provide positioning guidance. The following characteristics of the system were studied: (1) Coincidence of the isocenters between the IGRT and Linac; (2) Image quality; (3) Exposure; (4) Linearity, uniformity and repeatability. RESULTS:(1) Localization accuracy and coincidence of the isocenters between the IGRT and Linac was better than 1-mm. (2) The spatial resolution was quantified using the relative modulation-transfer-function with f50=0.7-0.9 lp/mm. The variation of contrast-noise-ratio with technical settings was measured. (3) The maximal exposure of an image was less than 95 mR. An empirical relation between the exposure and the X-ray technical setting was derived. (4) The linearity, uniformity and repeatability of the system generally meet the requirements. CONCLUSION:The system can be safely and reliably used as a target localization device.Item Open Access Clinical Experience With Machine Learning-Based Automated Treatment Planning for Whole Breast Radiation Therapy.(Advances in radiation oncology, 2021-03) Yoo, Sua; Sheng, Yang; Blitzblau, Rachel; McDuff, Susan; Champ, Colin; Morrison, Jay; O'Neill, Leigh; Catalano, Suzanne; Yin, Fang-Fang; Wu, Q JackiePurpose
The machine learning-based automated treatment planning (MLAP) tool has been developed and evaluated for breast radiation therapy planning at our institution. We implemented MLAP for patient treatment and assessed our clinical experience for its performance.Methods and materials
A total of 102 patients of breast or chest wall treatment plans were prospectively evaluated with institutional review board approval. A human planner executed MLAP to create an auto-plan via automation of fluence maps generation. If judged necessary, a planner further fine-tuned the fluence maps to reach a final plan. Planners recorded the time required for auto-planning and manual modification. Target (ie, breast or chest wall and nodes) coverage and dose homogeneity were compared between the auto-plan and final plan.Results
Cases without nodes (n = 71) showed negligible (<1%) differences for target coverage and dose homogeneity between the auto-plan and final plan. Cases with nodes (n = 31) also showed negligible difference for target coverage. However, mean ± standard deviation of volume receiving 105% of the prescribed dose and maximum dose were reduced from 43.0% ± 26.3% to 39.4% ± 23.7% and 119.7% ± 9.5% to 114.4% ± 8.8% from auto-plan to final plan, respectively, all with P ≤ .01 for cases with nodes (n = 31). Mean ± standard deviation time spent for auto-plans and additional fluence modification for final plans were 12.1 ± 9.3 and 13.1 ± 12.9 minutes, respectively, for cases without nodes, and 16.4 ± 9.7 and 26.4 ± 16.4 minutes, respectively, for cases with nodes.Conclusions
The MLAP tool has been successfully implemented for routine clinical practice and has significantly improved planning efficiency. Clinical experience indicates that auto-plans are sufficient for target coverage, but improvement is warranted to reduce high dose volume for cases with nodal irradiation. This study demonstrates the clinical implementation of auto-planning for patient treatment and the significant importance of integrating human experience and feedback to improve MLAP for better clinical translation.Item Open Access Comparisons of volumetric modulated arc therapy (VMAT) quality assurance (QA) systems: sensitivity analysis to machine errors.(Radiation Oncology (London, England), 2016-11-07) Liang, Bin; Liu, Bo; Zhou, Fugen; Yin, Fang-Fang; Wu, QiuwenIn volumetric modulated arc therapy (VMAT), gantry angles, dose rate and the MLC positions vary with the radiation delivery. The quality assurance (QA) system should be able to catch the planning and machine errors. The aim of this study was to investigate the sensitivity of three VMAT QA systems to machine errors.Several types of potential linac machine errors unique to VMAT delivery were simulated in sinusoidal function of gantry angle, including gantry angle itself, MLC position and linac output. Two commercial QA systems, ArcCheck and Delta4, and an in-house developed EPID technique were compared in this study. Fifteen full arcs from head and neck plans were selected and modified to include five magnitudes of each type of error, resulting in measurements and γ analyses of 240 arcs on each system. Both qualitative and quantitative comparisons were performed using receiver operating characteristic (ROC), γ pass rate gradient, and overlap histogram methods.In ROC analysis, the area under curve (AUC) represents the sensitivity and increases with the error magnitude. Using the criteria of 2 %/2 mm/2° (angle to agreement, ATA, only for EPID) and keeping AUC > 0.95, the minimum error detectable of ArcCheck, Delta4 and EPID are (2, 3, 3)° in gantry angle and (4, 2, 3) mm in MLC positions for the head and neck plans. No system is sensitive to the simulated output error, the AUC values were all below 0.70 even with 5 % output error. The γ gradient for gantry angle, MLC position and output errors are (-5.1, -2.6, -3.6)%/°, (-2.6, -7.1, -3.3)%/mm and (-0.2, -0.2, -0.3)%/% for ArcCheck, Delta4 and EPID, respectively. Therefore, these two analyses are consistent and support the same conclusion. The ATA parameter in EPID technique can be adjusted to tune its sensitivity.We found that ArcCheck is more sensitive to gantry angle error and Delta4 is more sensitive to MLC position error. All three systems are not sensitive to the simulated output error. With additional analysis parameter, the EPID technique can be tuned to have optimal sensitivity and is able to perform QA for full field size with highest resolution. In addition, ROC analysis avoids the choice of γ pass rate threshold and is more robust compared with other analysis methods.Item Open Access Computed tomography dose index and dose length product for cone-beam CT: Monte Carlo simulations.(Journal of applied clinical medical physics, 2011-01-19) Kim, Sangroh; Song, Haijun; Samei, Ehsan; Yin, Fang-Fang; Yoshizumi, Terry TDosimetry in kilovoltage cone beam computed tomography (CBCT) is a challenge due to the limitation of physical measurements. To address this, we used a Monte Carlo (MC) method to estimate the CT dose index (CTDI) and the dose length product (DLP) for a commercial CBCT system. As Dixon and Boone showed that CTDI concept can be applicable to both CBCT and conventional CT, we evaluated weighted CT dose index (CTDI(w)) and DLP for a commercial CBCT system. Two extended CT phantoms were created in our BEAMnrc/EGSnrc MC system. Before the simulations, the beam collimation of a Varian On-Board Imager (OBI) system was measured with radiochromic films (model: XR-QA). The MC model of the OBI X-ray tube, validated in a previous study, was used to acquire the phase space files of the full-fan and half-fan cone beams. Then, DOSXYZnrc user code simulated a total of 20 CBCT scans for the nominal beam widths from 1 cm to 10 cm. After the simulations, CBCT dose profiles at center and peripheral locations were extracted and integrated (dose profile integral, DPI) to calculate the CTDI per each beam width. The weighted cone-beam CTDI (CTDI(w,l)) was calculated from DPI values and mean CTDI(w,l) (CTDI(w,l)) and DLP were derived. We also evaluated the differences of CTDI(w) values between MC simulations and point dose measurements using standard CT phantoms. In results, it was found that CTDI(w,600) was 8.74 ± 0.01 cGy for head and CTDI(w,900) was 4.26 ± 0.01 cGy for body scan. The DLP was found to be proportional to the beam collimation. We also found that the point dose measurements with standard CT phantoms can estimate the CTDI within 3% difference compared to the full integrated CTDI from the MC method. This study showed the usability of CTDI as a dose index and DLP as a total dose descriptor in CBCT scans.