ComBat Harmonization for MRI Radiomics: Impact on Nonbinary Tissue Classification by Machine Learning
Leithner, Doris; Nevin, Rachel B; Gibbs, Peter; Weber, Michael; Otazo, Ricardo; Vargas, H Alberto; Mayerhoefer, Marius E
OBJECTIVES/OBJECTIVE:The aims of this study were to determine whether ComBat harmonization improves multiclass radiomics-based tissue classification in technically heterogeneous MRI data sets and to compare the performances of 2 ComBat variants. MATERIALS AND METHODS/METHODS:One hundred patients who had undergone T1-weighted 3D gradient echo Dixon MRI (2 scanners/vendors; 50 patients each) were retrospectively included. Volumes of interest (2.5 cm3) were placed in 3 disease-free tissues with visually similar appearance on T1 Dixon water images: liver, spleen, and paraspinal muscle. Gray-level histogram (GLH), gray-level co-occurrence matrix (GLCM), gray-level run-length matrix (GLRLM), and gray-level size-zone matrix (GLSZM) radiomic features were extracted. Tissue classification was performed on pooled data from the 2 centers (1) without harmonization, (2) after ComBat harmonization with empirical Bayes estimation (ComBat-B), and (3) after ComBat harmonization without empirical Bayes estimation (ComBat-NB). Linear discriminant analysis with leave-one-out cross-validation was used to distinguish among the 3 tissue types, using all available radiomic features as input. In addition, a multilayer perceptron neural network with a random 70%:30% split into training and test data sets was used for the same task, but separately for each radiomic feature category. RESULTS:Linear discriminant analysis-based mean tissue classification accuracies were 52.3% for unharmonized, 66.3% for ComBat-B harmonized, and 92.7% for ComBat-NB harmonized data. For multilayer perceptron neural network, mean classification accuracies for unharmonized, ComBat-B-harmonized, and ComBat-NB-harmonized test data were as follows: 46.8%, 55.1%, and 57.5% for GLH; 42.0%, 65.3%, and 71.0% for GLCM; 45.3%, 78.3%, and 78.0% for GLRLM; and 48.1%, 81.1%, and 89.4% for GLSZM. Accuracies were significantly higher for both ComBat-B- and ComBat-NB-harmonized data than for unharmonized data for all feature categories (at P = 0.005, respectively). For GLCM (P = 0.001) and GLSZM (P = 0.005), ComBat-NB harmonization provided slightly higher accuracies than ComBat-B harmonization. CONCLUSIONS:ComBat harmonization may be useful for multicenter MRI radiomics studies with nonbinary classification tasks. The degree of improvement by ComBat may vary among radiomic feature categories, among classifiers, and among ComBat variants.
View-sharing for 4D magnetic resonance imaging with randomized projection-encoding enables improvements of respiratory motion imaging for treatment planning in abdominothoracic radiotherapy
Subashi, Ergys; Feng, Li; Liu, Yilin; Robertson, Scott; Segars, Paul; Driehuys, Bastiaan; Kelsey, Christopher R; Yin, Fang-Fang; Otazo, Ricardo; Cai, Jing
BACKGROUND AND PURPOSE/UNASSIGNED:The accuracy and precision of radiation therapy are dependent on the characterization of organ-at-risk and target motion. This work aims to demonstrate a 4D magnetic resonance imaging (MRI) method for improving spatial and temporal resolution in respiratory motion imaging for treatment planning in abdominothoracic radiotherapy. MATERIALS AND METHODS/UNASSIGNED:The spatial and temporal resolution of phase-resolved respiratory imaging is improved by considering a novel sampling function based on quasi-random projection-encoding and peripheral k-space view-sharing. The respiratory signal is determined directly from k-space, obviating the need for an external surrogate marker. The average breathing curve is used to optimize spatial resolution and temporal blurring by limiting the extent of data sharing in the Fourier domain. Improvements in image quality are characterized by evaluating changes in signal-to-noise ratio (SNR), resolution, target detection, and level of artifact. The method is validated in simulations, in a dynamic phantom, and in-vivo imaging. RESULTS/UNASSIGNED:Sharing of high-frequency k-space data, driven by the average breathing curve, improves spatial resolution and reduces artifacts. Although equal sharing of k-space data improves resolution and SNR in stationary features, phases with large temporal changes accumulate significant artifacts due to averaging of high frequency features. In the absence of view-sharing, no averaging and detection artifacts are observed while spatial resolution is degraded. CONCLUSIONS/UNASSIGNED:The use of a quasi-random sampling function, with view-sharing driven by the average breathing curve, provides a feasible method for self-navigated 4D-MRI at improved spatial resolution.
Dynamic contrast-enhanced MRI parametric mapping using high spatiotemporal resolution Golden-angle RAdial Sparse Parallel MRI and iterative joint estimation of the arterial input function and pharmacokinetic parameters
Mazaheri, Yousef; Kim, Nathanael; Lakhman, Yulia; Jafari, Ramin; Vargas, Alberto; Otazo, Ricardo
The aim of this work is to develop a data-driven quantitative dynamic contrast-enhanced (DCE) MRI technique using Golden-angle RAdial Sparse Parallel (GRASP) MRI with high spatial resolution and high flexible temporal resolution and pharmacokinetic (PK) analysis with an arterial input function (AIF) estimated directly from the data obtained from each patient. DCE-MRI was performed on 13 patients with gynecological malignancy using a 3-T MRI scanner with a single continuous golden-angle stack-of-stars acquisition and image reconstruction with two temporal resolutions, by exploiting a unique feature in GRASP that reconstructs acquired data with user-defined temporal resolution. Joint estimation of the AIF (both AIF shape and delay) and PK parameters was performed with an iterative algorithm that alternates between AIF and PK estimation. Computer simulations were performed to determine the accuracy (expressed as percentage error [PE]) and precision of the estimated parameters. PK parameters (volume transfer constant [Ktrans ], fractional volume of the extravascular extracellular space [ve ], and blood plasma volume fraction [vp ]) and normalized root-mean-square error [nRMSE] (%) of the fitting errors for the tumor contrast kinetic data were measured both with population-averaged and data-driven AIFs. On patient data, the Wilcoxon signed-rank test was performed to compare nRMSE. Simulations demonstrated that GRASP image reconstruction with a temporal resolution of 1â€‰s/frame for AIF estimation and 5â€‰s/frame for PK analysis resulted in an absolute PE of less than 5% in the estimation of Ktrans and ve , and less than 11% in the estimation of vp . The nRMSE (mean Â± SD) for the dual temporal resolution image reconstruction and data-driven AIF was 0.16â€‰Â±â€‰0.04 compared with 0.27â€‰Â±â€‰0.10 (pÂ <â€‰0.001) with 1â€‰s/frame using population-averaged AIF, and 0.23â€‰Â±â€‰0.07 with 5â€‰s/frame using population-averaged AIF (pÂ <â€‰0.001). We conclude that DCE-MRI data acquired and reconstructed with the GRASP technique at dual temporal resolution can successfully be applied to jointly estimate the AIF and PK parameters from a single acquisition resulting in data-driven AIFs and voxelwise PK parametric maps.
MR SIGnature MAtching (MRSIGMA) with retrospective self-evaluation for real-time volumetric motion imaging
Kim, Nathanael; Tringale, Kathryn R; Crane, Christopher; Tyagi, Neelam; Otazo, Ricardo
Multiband diffusion tensor imaging for presurgical mapping of motor and language pathways in patients with brain tumors
Jenabi, Mehrnaz; Young, Robert J; Moreno, Raquel; Gene, Madeleine; Cho, Nicholas; Otazo, Ricardo; Holodny, Andrei I; Peck, Kyung K
BACKGROUND AND PURPOSE/OBJECTIVE:Assessment of the essential white matter fibers of arcuate fasciculus and corticospinal tract (CST), required for preoperative planning in brain tumor patients, relies on the reliability of diffusion tensor imaging (DTI). The recent development of multiband DTI (mb-DTI) based on simultaneous multislice excitation could maintain the overall quality of tractography while not exceeding standard clinical care time. To address this potential, we performed quantitative analyses to evaluate tractography results of arcuate fasciculus and CST acquired by mb-DTI in brain tumor patients. METHODS:), and length (mm) in the whole brain, arcuate fasciculus, and CST. Additionally, three neuroradiologists performed a blinded visual assessment comparing s-DTI with mb-DTI. RESULTS:) in all patients with tumor involved in the arcuate fasciculus, CST, and whole brain tracts (p = .001). Visual assessment demonstrated that both techniques produced visually similar tracts. CONCLUSIONS:This study demonstrated the clinical potential and significant advantages of preoperative mb-DTI in brain tumor patients.
MRI-guided Radiation Therapy: An Emerging Paradigm in Adaptive Radiation Oncology
Otazo, Ricardo; Lambin, Philippe; Pignol, Jean-Philippe; Ladd, Mark E; Schlemmer, Heinz-Peter; Baumann, Michael; Hricak, Hedvig
Radiation therapy (RT) continues to be one of the mainstays of cancer treatment. Considerable efforts have been recently devoted to integrating MRI into clinical RT planning and monitoring. This integration, known as MRI-guided RT, has been motivated by the superior soft-tissue contrast, organ motion visualization, and ability to monitor tumor and tissue physiologic changes provided by MRI compared with CT. Offline MRI is already used for treatment planning at many institutions. Furthermore, MRI-guided linear accelerator systems, allowing use of MRI during treatment, enable improved adaptation to anatomic changes between RT fractions compared with CT guidance. Efforts are underway to develop real-time MRI-guided intrafraction adaptive RT of tumors affected by motion and MRI-derived biomarkers to monitor treatment response and potentially adapt treatment to physiologic changes. These developments in MRI guidance provide the basis for a paradigm change in treatment planning, monitoring, and adaptation. Key challenges to advancing MRI-guided RT include real-time volumetric anatomic imaging, addressing image distortion because of magnetic field inhomogeneities, reproducible quantitative imaging across different MRI systems, and biologic validation of quantitative imaging. This review describes emerging innovations in offline and online MRI-guided RT, exciting opportunities they offer for advancing research and clinical care, hurdles to be overcome, and the need for multidisciplinary collaboration.
MRSIGMA: Magnetic Resonance SIGnature MAtching for real-time volumetric imaging
Feng, Li; Tyagi, Neelam; Otazo, Ricardo
PURPOSE:To propose a real-time 3D MRI technique called MR SIGnature MAtching (MRSIGMA) for high-resolution volumetric imaging and motion tracking with very low imaging latency. METHODS:MRSIGMA consists of two steps: (1) offline learning of a database of possible 3D motion states and corresponding motion signature ranges and (2) online matching of new motion signatures acquired in real time with prelearned motion states. Specifically, the offline learning step (non-real-time) reconstructs motion-resolved 4D images representing different motion states and assigns a unique motion range to each state. The online matching step (real-time) acquires motion signatures only and selects one of the prelearned 3D motion states for each newly acquired signature, which generates 3D images efficiently in real time. The MRSIGMA technique was evaluated on 15 golden-angle stack-of-stars liver data sets, and the performance of respiratory motion tracking with the online-generated real-time 3D MRI was compared with the corresponding 2D projections acquired in real time. RESULTS: = 0.948) between motion displacement measured from the online-generated real-time 3D images and the 2D real-time projections. CONCLUSION:This proof-of-concept study demonstrates the feasibility of MRSIGMA for high-resolution real-time volumetric imaging, which shifts the acquisition and reconstruction burden to an offline learning step and leaves fast online matching for online imaging with very low imaging latency. The MRSIGMA technique can potentially be used for real-time motion tracking in MRI-guided radiation therapy.
Accelerating Prostate Diffusion-weighted MRI Using a Guided Denoising Convolutional Neural Network: Retrospective Feasibility Study
Kaye, Elena A; Aherne, Emily A; Duzgol, Cihan; HÃ¤ggström, Ida; Kobler, Erich; Mazaheri, Yousef; Fung, Maggie M; Zhang, Zhigang; Otazo, Ricardo; Vargas, Hebert A; Akin, Oguz
Purpose/UNASSIGNED:To investigate the feasibility of accelerating prostate diffusion-weighted imaging (DWI) by reducing the number of acquired averages and denoising the resulting image using a proposed guided denoising convolutional neural network (DnCNN). Materials and Methods/UNASSIGNED:-value DW image as a guidance input. Quantitative and qualitative reader evaluations were performed on the denoised hb DW images. A cumulative link mixed regression model was used to compare the readers' scores. The agreement between the apparent diffusion coefficient (ADC) maps (denoised vs reference) was analyzed using Bland-Altman analysis. Results/UNASSIGNED:/sec). Conclusion/UNASSIGNED:Â© RSNA, 2020.
Accelerated MR spectroscopic imaging-a review of current and emerging techniques
Bogner, Wolfgang; Otazo, Ricardo; Henning, Anke
Over more than 30 years in vivo MR spectroscopic imaging (MRSI) has undergone an enormous evolution from theoretical concepts in the early 1980s to the robust imaging technique that it is today. The development of both fast and efficient sampling and reconstruction techniques has played a fundamental role in this process. State-of-the-art MRSI has grown from a slow purely phase-encoded acquisition technique to a method that today combines the benefits of different acceleration techniques. These include shortening of repetition times, spatial-spectral encoding, undersampling of k-space and time domain, and use of spatial-spectral prior knowledge in the reconstruction. In this way in vivo MRSI has considerably advanced in terms of spatial coverage, spatial resolution, acquisition speed, artifact suppression, number of detectable metabolites and quantification precision. Acceleration not only has been the enabling factor in high-resolution whole-brain 1 H-MRSI, but today is also common in non-proton MRSI (31 P, 2 H and 13 C) and applied in many different organs. In this process, MRSI techniques had to constantly adapt, but have also benefitted from the significant increase of magnetic field strength boosting the signal-to-noise ratio along with high gradient fidelity and high-density receive arrays. In combination with recent trends in image reconstruction and much improved computation power, these advances led to a number of novel developments with respect to MRSI acceleration. Today MRSI allows for non-invasive and non-ionizing mapping of the spatial distribution of various metabolites' tissue concentrations in animals or humans, is applied for clinical diagnostics and has been established as an important tool for neuro-scientific and metabolism research. This review highlights the developments of the last five years and puts them into the context of earlier MRSI acceleration techniques. In addition to 1 H-MRSI it also includes other relevant nuclei and is not limited to certain body regions or specific applications.
Accelerated mono- and biexponential 3D-T1Ï relaxation mapping of knee cartilage using golden angle radial acquisitions and compressed sensing
Zibetti, Marcelo V W; Sharafi, Azadeh; Otazo, Ricardo; Regatte, Ravinder R
PURPOSE/OBJECTIVE:) mapping of knee cartilage. METHODS:-mapping of knee cartilage, including spatio-temporal finite differences, wavelets, dictionary from principal component analysis, and exponential decay models, and also low rank and low rank plus sparse models (L+S). Complex-valued fitting was used and Marchenko-Pastur principal component analysis filtering also tested. RESULTS:mapping, with median normalized absolute deviation below 10% up to AF of 6. CONCLUSION/CONCLUSIONS:mapping of knee cartilage, being it is a good alternative to Cartesian sampling for reducing scan time and/or improving image and mapping quality. The methods exponential decay models, spatio-temporal finite differences, and low rank obtained the best results for radial sampling patterns.