Browsing by Author "Chang, Zheng"

There 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.

Purpose: To investigate the feasibility of the Log-Demons deformable image registration (DIR) method to correct eddy current and Echo Planar Imaging (EPI) distortions while preserving diffusion tensor information.

Methods: A phantom MR scan was conducted using a diffusion phantom scan (Diffusion Phantom Model 128, High Precision Devices, Inc) on a clinical 3T scanner. The scan includes a standard T1-weighted scan and a 20‐direction diffusion tensor imaging (DTI) scan, which consists of one data set with b=0s/mm2 and twenty diffusion-weighted data sets with b=1,000s/mm2. A Log-Demons DIR algorithm was applied to the DTI images for eddy current and EPI distortion correction based on the b=0s/mm2 and T1 weighted data sets and compared the eddy current and EPI distortion corrections along the phase encoding direction by affine and demons DIR algorithms. The Log-Demons framework is optimized based on both similarity and regularization. The registered images were analyzed using Cross-correlation (CC) and mutual information (MI) to assess the performances of distortion corrections by the DIR methods. Quantitative deviations from the original data after correction were also evaluated using the mean, and root mean square error (RMSE) for thirteen regions of interest in the Apparent Diffusion Coefficient (ADC) and Fractional Anisotropy (FA) maps.

The Log-Demons DIR algorithm was then applied to the MASSIVE dataset, which provides diffusion-weighted volumes divided into four sets with both positive (+) and negative(-) diffusion gradient directions and both AP and PA phase encoding directions. The registered images were analyzed using the mutual information (MI) and the absolute mean difference of two images with opposing gradient directions to assess the performances of distortion corrections by the DIR methods. Images with opposing gradient directions were compared when comparing eddy current distortions and images with opposing phase encoding directions were compared for EPI distortions.

Results: In the phantom study, the MI and CC were improved by 2.15%,0.89%, and 39.39% compared to no correction, and affine, and demons algorithm respectively when correction for eddy current distortions. MI and CC were improved by 8.89%, 9.33%, and 9.20% compared to no correction, and affine, and demons algorithm respectively when correction for EPI distortions. Analysis of the tensor metrics using percent difference and the RMS of the ADC and FA found that the Log-Demons algorithm outperforms the other algorithms in terms of preserving diffusion information.

In the MASSIVE study, the Log-demons DIR method outperformed the demons algorithm in terms of MI but underperformed compared to the affine registration for both eddy current and EPI distortions corrections. The absolute mean difference was decreased by 2.94%, 0.44%, and 1.53% compared to no correction, and affine, and demons algorithm respectively when correcting for eddy current distortions, and decreased by 0.39%, 8.03%, and 13.19% compared to no correction, and affine, and demons algorithm respectively when correcting for EPI distortions.

Conclusion: This work indicates that the Log-Demons DIR algorithm is feasible to reduce eddy current and EPI distortions while preserving quantitative diffusion information. Although demonstrated with a DTI phantom study and brain study, this method could be extended for areas in which diffusion-weighted imaging is beneficial.

Purpose: Changes in patient volume, due to tumor shrinkage, dehydration, dysphagia and atrophy, could present issues in the accuracy of dosimetry throughout the course of treatment. The aim of this work is to study the dosimetric impacts of the volumetric changes during IMRT and to investigate the feasibilities of electronic portal imaging device (EPID) in predicting the impacts. Materials and Methods: An anthropomorphic head and neck phantom was used to represent two scenarios: symmetric and asymmetric volume loss. The phantom was simulated and planned according to the head and neck protocols used in our clinic. Dose volume histograms (DVH) were generated for each set up scenario and were used to calculate the integral dose expected at the coincident volume of the phantom. During treatment delivery, the EPID captured exit fluence of each beam at each level of bolus thickness. These images were quantitatively analyzed using gamma analysis with criteria of 3% and 3mm dose difference and distance-to-agreement respectively. Results: Comparing maximum to minimum volume in the symmetric situation with DVH generated in Eclipse show substantial fluctuations in dose. When comparing five layers of bolus material to zero layers of bolus material, the changes were most significant. The asymmetric volume change predicted dose fluctuations that were less significant than the symmetric phantom. As for gamma analysis, a quantitative evaluation of the integrated dose fluence, captured by the EPID, showed extreme variability in the images with five layers of bolus when compared to images with no bolus. Less significant variation was shown in layers of closer thicknesses, as expected. Conclusions: The phantom study indicates that volume loss could contribute to clinically considerable changes in the dose delivered to target and organs at risk. The proposed technique using EPID could provide valuable information about the variation of dose due to volumetric changes and might be potentially useful.

As an indispensable imaging modality, magnetic resonance imaging is a non-invasive and radiation-free procedure that could provide high soft-tissue contrast of the patients’ body. However, its time-consuming acquisition process shadows its medical application in clinic. It not only increases the risk of generating motion artifacts but also leads to high cost of MR scanning procesure. Based on both problem, we aim to reduce the MRI acquisition time by undersampling the k-space data while eliminate the aliasing artifacts caused by the undersampling process on the reconstructed images. This dissertation is a feasibility study to investigate how to accelerate the MRI acquisition process while preserving the image quality based on the deep learning methods. The study is unfolded by completing two tasks. The first task is to compare two fast MRI algorithms – the iterative reconstruction methods and deep learning methods – using the 25% Cartesian undersampling scheme in order to evaluate which one could better eliminate the aliasing artifacts on the reconstructed images. The second task aims to compare the Cartesian undersampling schemes against the golden-angle radial undersampling schemes with the undersampling ratios of one quarter, one-sixth, and one-eighth in order to investigate which schemes could better accelerate the MRI acquisition process.

As demonstrated by the comparison results of task one, the deep learning methods (DC-CNN) algorithm outperformed the iteration reconstruction (IR-TGV) algorithm for image reconstruction in terms of the image quality and computational efficiency. In task two, the comparison results indicated that the golden-angle radial undersampling schemes outperformed Cartesian undersampling schemes in terms of accelerating the MRI scanning process.

For further investigation, we still endeavor to develop a deep learning model that could better accelerate the MRI acquisition with the Cartesian undersampling schemes. In addition, such evaluation could be expanded to more anatomic sites, MRI sequences, and 3D MRI reconstructions.

IDH is a gene that heavily affects the treatment response of gliomas and is associated with patient’s prognosis. Convolutional neural networks (CNNs) showed great potential in predicting IDH mutations. However, these CNN models require time-intensive image-preprocessing before being used for predictions. There are two main purpose of this study. The first purpose is to investigate the feasibility of applying a novel convolutional neural network based on the Inception-ResNet to reduce image preprocessing steps and improve accuracy in IDH mutation prediction. The second purpose is to evaluate different data augmentation methods on brain studies.

MR images of 103 patients were selected from The Cancer Imaging Archive (TCIA). Intensity normalization of every individual slice is the only image pre-processing step. The T1w post-contrast, FLAIR, and T2w images at the same slice location were grouped together and considered as one training sample. This give rise to 209 IDH-mutant samples from 42 patients and 356 IDH-wild-type samples from 61 patients that were randomly selected to become training, validation, and test sets. To avoid overfitting in the model performance, data augmentation methods were applied individually to both training and validation sets in each training. The augmentation methods included duplication, noise addition, rotation, translation, cropping and mirroring. Images from one sample were fed to different input channels of Inception-ResNet, and the predictions were based on the extracted features and the patient’s age at diagnosis. Prediction accuracy was used to assess the performance of different augmentation methods.

With only intensity normalization, the proposed model using training sets augmented by rotation and noise addition achieved the IDH prediction accuracies of 91.8% and 91.8%, respectively. On the same training, validation and test sets, the proposed model trained on data augmented by duplication, cropping, translation, and mirroring gave accuracies of 81.6%, 79.6%, 83.7%, and 85.7%, respectively.

This work investigated the feasibility of the application of the novel convolutional neural network based on the Inception-ResNet on IDH mutation prediction, and high accuracies can be achieved with only intensity normalization as image preprocessing. Among all data augmentation methods, noise addition and rotation shows better performance and might suggest potential value for other clinical applications using machine learning algorithms.

Purpose: There are two goals of this study. The first goal of this study is to investigate the feasibility of using classic textural feature extraction in radiotherapy response assessment among a unique cohort of early stage breast cancer patients who received the single-dose preoperative radiotherapy. The second goal of this study is to investigate the clinical feasibility of using classic texture features as potential biomarkers which are supplementary to regional apparent diffusion coefficient in gynecological cancer radiotherapy response assessment.

Methods and Materials: For the breast cancer study, 15 patients with early stage breast cancer were enrolled in this retrospective study. Each patient received a single-fraction radiation treatment, and DWI and DCE-MRI scans were conducted before and after the radiotherapy. DWI scans were acquired using a spin-echo EPI sequence with diffusion weighting factors of b = 0 and b = 500 mm2/s, and the apparent diffusion coefficient (ADC) maps were calculated. DCE-MRI scans were acquired using a T1-weighted 3D SPGR sequence with a temporal resolution of about 1 minute. The contrast agent (CA) was intravenously injected with a 0.1 mmol/kg bodyweight dose at 2 ml/s. Two parameters, volume transfer constant (Ktrans) and kep were analyzed using the two-compartment Tofts pharmacokinetic model. For pharmacokinetic parametric maps and ADC maps, 33 textural features were generated from the clinical target volume (CTV) in a 3D fashion using the classic gray level co-occurrence matrix (GLCOM) and gray level run length matrix (GLRLM). Wilcoxon signed-rank test was used to determine the significance of each texture feature’s change after the radiotherapy. The significance was set to 0.05 with Bonferroni correction.

For the gynecological cancer study, 12 female patients with gynecologic cancer treated with fractionated external beam radiotherapy (EBRT) combined with high dose rate (HDR) intracavitary brachytherapy were studied. Each patient first received EBRT treatment followed by five fractions of HDR treatment. Before EBRT and before each fraction of brachytherapy, Diffusion Weighted MRI (DWI-MRI) and CT scans were acquired. DWI scans were acquired in sagittal plane utilizing a spin-echo echo-planar imaging sequence with weighting factors of b = 500 s/mm2 and b = 1000 s/mm2, one set of images of b = 0 s/mm2 were also acquired. ADC maps were calculated using linear least-square fitting method. Distributed diffusion coefficient (DDC) maps and stretching parameter α were calculated. For ADC and DDC maps, 33 classic texture features were generated utilizing the classic gray level run length matrix (GLRLM) and gray level co-occurrence matrix (GLCOM) from high-risk clinical target volume (HR-CTV). Wilcoxon signed-rank statistics test was applied to determine the significance of each feature’s numerical value change after radiotherapy. Significance level was set to 0.05 with multi-comparison correction if applicable.

Results: For the breast cancer study, regarding ADC maps calculated from DWI-MRI, 24 out of 33 CTV features changed significantly after the radiotherapy. For DCE-MRI pharmacokinetic parameters, all 33 CTV features of Ktrans and 33 features of kep changed significantly.

For the gynecological cancer study, regarding ADC maps, 28 out of 33 HR-CTV texture features showed significant changes after the EBRT treatment. 28 out of 33 HR-CTV texture features indicated significant changes after HDR treatments. The texture features that indicated significant changes after HDR treatments are the same as those after EBRT treatment. 28 out of 33 HR-CTV texture features showed significant changes after whole radiotherapy treatment process. The texture features that indicated significant changes for the whole treatment process are the same as those after HDR treatments.

Conclusion: Initial results indicate that certain classic texture features are sensitive to radiation-induced changes. Classic texture features with significant numerical changes can be used in monitoring radiotherapy effect. This might suggest that certain texture features might be used as biomarkers which are supplementary to ADC and DDC for assessment of radiotherapy response in breast cancer and gynecological cancer.

Abstract

The goal of modern radiotherapy is to precisely deliver a prescribed radiation dose to delineated target volumes that contain a significant amount of tumor cells while sparing the surrounding healthy tissues/organs. Precise delineation of treatment and avoidance volumes is the key for the precision radiation therapy. In recent years, considerable clinical and research efforts have been devoted to integrate MRI into radiotherapy workflow motivated by the superior soft tissue contrast and functional imaging possibility. Dynamic contrast-enhanced MRI (DCE-MRI) is a noninvasive technique that measures properties of tissue microvasculature. Its sensitivity to radiation-induced vascular pharmacokinetic (PK) changes has been preliminary demonstrated. In spite of its great potential, two major challenges have limited DCE-MRI’s clinical application in radiotherapy assessment: the technical limitations of accurate DCE-MRI imaging implementation and the need of novel DCE-MRI data analysis methods for richer functional heterogeneity information.

This study aims at improving current DCE-MRI techniques and developing new DCE-MRI analysis methods for particular radiotherapy assessment. Thus, the study is naturally divided into two parts. The first part focuses on DCE-MRI temporal resolution as one of the key DCE-MRI technical factors, and some improvements regarding DCE-MRI temporal resolution are proposed; the second part explores the potential value of image heterogeneity analysis and multiple PK model combination for therapeutic response assessment, and several novel DCE-MRI data analysis methods are developed.

I. Improvement of DCE-MRI temporal resolution. First, the feasibility of improving DCE-MRI temporal resolution via image undersampling was studied. Specifically, a novel MR image iterative reconstruction algorithm was studied for DCE-MRI reconstruction. This algorithm was built on the recently developed compress sensing (CS) theory. By utilizing a limited k-space acquisition with shorter imaging time, images can be reconstructed in an iterative fashion under the regularization of a newly proposed total generalized variation (TGV) penalty term. In the retrospective study of brain radiosurgery patient DCE-MRI scans under IRB-approval, the clinically obtained image data was selected as reference data, and the simulated accelerated k-space acquisition was generated via undersampling the reference image full k-space with designed sampling grids. Two undersampling strategies were proposed: 1) a radial multi-ray grid with a special angular distribution was adopted to sample each slice of the full k-space; 2) a Cartesian random sampling grid series with spatiotemporal constraints from adjacent frames was adopted to sample the dynamic k-space series at a slice location. Two sets of PK parameters’ maps were generated from the undersampled data and from the fully-sampled data, respectively. Multiple quantitative measurements and statistical studies were performed to evaluate the accuracy of PK maps generated from the undersampled data in reference to the PK maps generated from the fully-sampled data. Results showed that at a simulated acceleration factor of four, PK maps could be faithfully calculated from the DCE images that were reconstructed using undersampled data, and no statistically significant differences were found between the regional PK mean values from undersampled and fully-sampled data sets. DCE-MRI acceleration using the investigated image reconstruction method has been suggested as feasible and promising.

Second, for high temporal resolution DCE-MRI, a new PK model fitting method was developed to solve PK parameters for better calculation accuracy and efficiency. This method is based on a derivative-based deformation of the commonly used Tofts PK model, which is presented as an integrative expression. This method also includes an advanced Kolmogorov-Zurbenko (KZ) filter to remove the potential noise effect in data and solve the PK parameter as a linear problem in matrix format. In the computer simulation study, PK parameters representing typical intracranial values were selected as references to simulated DCE-MRI data for different temporal resolution and different data noise level. Results showed that at both high temporal resolutions (<1s) and clinically feasible temporal resolution (~5s), this new method was able to calculate PK parameters more accurate than the current calculation methods at clinically relevant noise levels; at high temporal resolutions, the calculation efficiency of this new method was superior to current methods in an order of 102. In a retrospective of clinical brain DCE-MRI scans, the PK maps derived from the proposed method were comparable with the results from current methods. Based on these results, it can be concluded that this new method can be used for accurate and efficient PK model fitting for high temporal resolution DCE-MRI.

II. Development of DCE-MRI analysis methods for therapeutic response assessment. This part aims at methodology developments in two approaches. The first one is to develop model-free analysis method for DCE-MRI functional heterogeneity evaluation. This approach is inspired by the rationale that radiotherapy-induced functional change could be heterogeneous across the treatment area. The first effort was spent on a translational investigation of classic fractal dimension theory for DCE-MRI therapeutic response assessment. In a small-animal anti-angiogenesis drug therapy experiment, the randomly assigned treatment/control groups received multiple fraction treatments with one pre-treatment and multiple post-treatment high spatiotemporal DCE-MRI scans. In the post-treatment scan two weeks after the start, the investigated Rényi dimensions of the classic PK rate constant map demonstrated significant differences between the treatment and the control groups; when Rényi dimensions were adopted for treatment/control group classification, the achieved accuracy was higher than the accuracy from using conventional PK parameter statistics. Following this pilot work, two novel texture analysis methods were proposed. First, a new technique called Gray Level Local Power Matrix (GLLPM) was developed. It intends to solve the lack of temporal information and poor calculation efficiency of the commonly used Gray Level Co-Occurrence Matrix (GLCOM) techniques. In the same small animal experiment, the dynamic curves of Haralick texture features derived from the GLLPM had an overall better performance than the corresponding curves derived from current GLCOM techniques in treatment/control separation and classification. The second developed method is dynamic Fractal Signature Dissimilarity (FSD) analysis. Inspired by the classic fractal dimension theory, this method measures the dynamics of tumor heterogeneity during the contrast agent uptake in a quantitative fashion on DCE images. In the small animal experiment mentioned before, the selected parameters from dynamic FSD analysis showed significant differences between treatment/control groups as early as after 1 treatment fraction; in contrast, metrics from conventional PK analysis showed significant differences only after 3 treatment fractions. When using dynamic FSD parameters, the treatment/control group classification after 1st treatment fraction was improved than using conventional PK statistics. These results suggest the promising application of this novel method for capturing early therapeutic response.

The second approach of developing novel DCE-MRI methods is to combine PK information from multiple PK models. Currently, the classic Tofts model or its alternative version has been widely adopted for DCE-MRI analysis as a gold-standard approach for therapeutic response assessment. Previously, a shutter-speed (SS) model was proposed to incorporate transcytolemmal water exchange effect into contrast agent concentration quantification. In spite of richer biological assumption, its application in therapeutic response assessment is limited. It might be intriguing to combine the information from the SS model and from the classic Tofts model to explore potential new biological information for treatment assessment. The feasibility of this idea was investigated in the same small animal experiment. The SS model was compared against the Tofts model for therapeutic response assessment using PK parameter regional mean value comparison. Based on the modeled transcytolemmal water exchange rate, a biological subvolume was proposed and was automatically identified using histogram analysis. Within the biological subvolume, the PK rate constant derived from the SS model were proved to be superior to the one from Tofts model in treatment/control separation and classification. Furthermore, novel biomarkers were designed to integrate PK rate constants from these two models. When being evaluated in the biological subvolume, this biomarker was able to reflect significant treatment/control difference in both post-treatment evaluation. These results confirm the potential value of SS model as well as its combination with Tofts model for therapeutic response assessment.

In summary, this study addressed two problems of DCE-MRI application in radiotherapy assessment. In the first part, a method of accelerating DCE-MRI acquisition for better temporal resolution was investigated, and a novel PK model fitting algorithm was proposed for high temporal resolution DCE-MRI. In the second part, two model-free texture analysis methods and a multiple-model analysis method were developed for DCE-MRI therapeutic response assessment. The presented works could benefit the future DCE-MRI routine clinical application in radiotherapy assessment.