Keywords: Radiomics, Cancer

Motivation: Prediction of high-grade meningioma on preoperative MRI is essential in therapeutic planning and evaluation of prognosis. 

Goal(s): We seek to propose a data augmentation strategy to reduce class imbalance for model improvement.

Approach: In this study, we propose a dual-level augmentation strategy incorporating image-level augmentation and feature-level augmentation to tackle class-imbalance and improve the predictive performance of radiomics for meningioma grading on multisequence MRI.

Results: The radiomics models yields robust performance in 100 repetitions in 3-, 5-, and 10-fold cross-validation. In addition, our method significantly outperformed single-level augmentation (image or feature) or no augmentation in each cross-validation. 

Impact: As an effective and robust meningioma grading tool, our radiomics model has the potential to aid clinical decision making for a broader range of meningioma grades seen in practice, allowing for better radiomics-based pre-operative stratification and individualized patient management.

Keywords: Data Processing, Data Processing, Infant myelination

Motivation: While T1/T2-weighted ratio maps are important for the study of myelination, existing reconstruction tools can fail in infants and present difficulty to automated segmentation.

Goal(s): To provide a pipeline for obtaining accurate regional myelination measures of whole brain regions and white matter tracts.

Approach: We adapted existing T1/T2-weighted ratio pipelines to incorporate deep learning methods for segmentation and registration as well as a high-quality tract atlas.

Results: Our pipeline showed reduced errors and improved differentiation between 3- and 6-month-old infants from a South African longitudinal birth cohort study.

Impact: The improved T1/T2-weighted ratio contrast and detailed segmentations provided by our pipeline will enable study of specific myelination patterns during neurodevelopment, especially in populations exposed to risk factors for altered white matter maturation.

Keywords: Data Processing, MR Value

Motivation: There are no standard metrics to measure clinical MRI utilization. This gap contributes to a lack of best practices for MRI utilization optimization.

Goal(s): (1)Identify and standardize metrics relevant to MRI workflow; (2)demonstrate the use of these metrics to quantify MRI utilization; (3)identify opportunities for improvement in efficiency in a clinical MRI practice.

Approach: Timepoint and summary metrics were defined to characterize MRI workflow. Data from two MRI facilities were analyzed using these metrics to characterize MRI utilization.

Results: The defined standard set of metrics quantified MRI utilization and identified workflow areas that could be further optimized to improve MRI utilization.

Impact: It is currently difficult to communicate key timepoint metrics that characterize the utilization of clinical MRI exams. The development of standardized language to characterize MR exam timepoint metrics can facilitate best practices for workflow improvement. Such tools are essential to improve the value of MR.

Keywords: Data Processing, Quantitative Susceptibility mapping, Brain

Motivation: The contribution of physiologically induced dynamic magnetic susceptibility within the brain to the blood oxygen level dependent (BOLD) signal used for functional imaging has never been quantified.

Goal(s): Our goal was to measure the brain’s dynamic magnetic susceptibility signal as a function of the cardiac and respiratory cycles using quantitative susceptibility mapping (QSM).

Approach: We developed and evaluated an image processing framework to generate dynamic susceptibility maps over the course of cardiac pulsation and respiration.

Results: We demonstrated that QSM is sensitive to physiologically induced susceptibility changes and that our framework can be used to measure these dynamics.

Impact: The influence of physiologically induced dynamic magnetic susceptibility on BOLD signal could impact the derivation/interpretation of BOLD-based functional connectivity. Significant susceptibility changes could indicate a need to modify existing functional mapping approaches whereas the reverse would validate current techniques.

Keywords: Data Processing, Liver

Motivation: There is a lack of suitable tools for monitoring liver function to determine the optimal treatment strategy for patients.

Goal(s): To evaluate the value of B1-corrected VFA T1 mapping in predicting liver function.

Approach: T1 relaxation time at corresponding time-points in B1-corrected VFA T1 mapping were measured. We analyzed the correlation of T1 value with indocyanine green retention rate at 15min (ICG-R15) and conducted Receiver operating characteristic (ROC) to evaluate the albumin-bilirubin (ALBI) grade classification efficiency by T1 mapping parameters.

Results: T1-10min, T1-HBP and △T1% showed moderate correlation with ICG-R15. T1-5min had the best diagnostic efficacy in differentiating ALBI 1 and ALBI 2.

Impact: This study explored that T1 value of Gd-EOB-DTPA-enhanced MRI by B1-corrected VFA method can be used as noninvasive imaging indicators for estimation of liver function.

Keywords: Data Processing, Uterus

Motivation: Limitations of the FIGO classification of fibroid by MRI include interobserver variability, especially for large lesion, leading to distortion of the uterine landmarks.

Goal(s): We aim to evaluate the accuracy and reliability of FIGO type using organ-axial T2WI .

Approach: Organ- and body-axial T2WI were used to identify FIGO classification of fibroids. FIGO types assigned by each radiologist and operation outcome were compared. The association and correlation between FIGO types and cavity compression parameters were analysis.

Results: Organ-axial images showed excellent consistency (kappa=0.877, P=0.04) and accuracy (kappa=0.932, P=0.037) . Compression angle exhibited linear correlation with FIGO types with r of 0.68 (P<0.001) .

Impact: High resolution organ-axial T2 weighted MR could provide exceptional reliability and accuracy in identifying the FIGO classification for uterine fibroid and can be applied to assist in treatment planning and assessing response assessment.

Keywords: Segmentation, Segmentation, Dementia, Alzheimer's disease, Brain, White matter, Neuro

Motivation: Evaluating cognitive function-related white matter lesions (WML) conventionally requires 3D fluid attenuated inversion recovery (3D-FLAIR), which isn't always available. We aimed to explore the suitability of routinely acquired 3D-T1 weighted images (3D-T1WI) for WML assessment.

Goal(s): This study investigated whether 3D-T1WI could replace 3D-FLAIR in WML assessment. 

Approach: We compared the correlation coefficient, ICC, and DSC of WML volume between 3D-FLAIR and 3D-T1WI, as well as its correlation with cognitive scores.

Results: WML based on 3D-T1WI strongly correlated with WML based on 3D-FLAIR, with high ICC, DSC, and cognitive score associations, indicating the potential of 3D-T1WI for WML assessment alternative to 3D-FLAIR.

Impact: White matter lesions (WML) based on 3D-T1 weighted images (3D-T1WI) closely matched 3D-fluid attenuated inversion recovery (3D-FLAIR) in WML area, volume, and cognitive function associations. It is suggested 3D-T1WI is valuable alternative to 3D-FLAIR for WML volume assessment.

Keywords: Software Tools, Perfusion, Segmentation, DCE-MRI Perfusion, Analysis/Processing

Motivation: In order to extract quantitative measures from perfusion imaging, pharmacokinetic models such as the (extended) Tofts are utilized. These models require contrast agent concentration in plasma to be estimated in a robust way for model fitting.

Goal(s): Develop a method to measure the vascular input function in DCE-MRI brain scans based on superior sagittal sinus (SSS) values which is semi-automatic and robust to user input.

Approach: Given a user selected seed point inside the SSS, image segmentation in combination with voxel-wise intensity analysis is used.

Results: The method was shown to be robust towards user input in a small patient cohort. 

Impact: The presented method has the potential to improve the robustness of perfusion parameters, such as K_trans, v_e, and v_p, making their usage as quantitative imaging biomarkers more feasible through a more consistent vascular input function definition.

Keywords: Software Tools, Data Processing, analysis tools

Motivation: To benchmark multi-echo based DSC-MRI SAGE sequence, a compelling alternative to single echo acquisition analysis tool that is targeted for clinical translation.

Goal(s): Comparison of rCBV values derived from  the two software (IB Neuro X2, in-house algorithm).

Approach: We expanded a single-echo DSC-MRI brain tumor digital reference object (DRO) into an anthropomorphic phantom that recapitulates clinical DSC-MRI data structure and anatomy.

Results: The rCBV values obtained from both softwares are strongly correlated.

Impact: This study would potentially deliver a robust and reproducible post-processing tool for multi-site, multi-vendor SAGE-based DSC-MRI that can be used for clinical trials.

Keywords: Data Processing, Relaxometry, Ultrahigh field MRI

Motivation: Water T1 (wT1) has been shown to be a sensitive biomarker for muscle disease in the 3T system.

Goal(s): The aim of this work was to investigate the feasibility of simultaneous wT1/PDFF/R₂^* imaging in the 5T system.

Approach: Simultaneous multi-parametric imaging was achieved using chemical shift encoded multi-echo variable flip angle (CSE-VFA) sequences combined with B1 measurement techniques.

Results: Corrected by the preTFL approach, homogeneous wT1 maps are obtained even under B1 varying from 0.37 to 1.12.

Impact: In this work we have tested the feasibility of simultaneous quantification of wT1/PDFF/R₂^* in the 5T imaging system. This work has laid the foundation for subsequent research into the clinical significance of extracellular volume fraction.

Keywords: Software Tools, Data Processing, SaaS, PaaS, Cloud Computing, Cloud DICOM viewer

Motivation: Enable computationally expensive processing and dicom image review in cloud with low-cost solution amicable to low and unstable internet connectivity.

Goal(s): Utilize Cloud computing(SaaS) and cloud platform service(PaaS) to transfer data in secure and HIPPA compliant manner.

Approach: Fluwiz[5]  enables HIPAA compliant data transfer to Cloud where MR workflows, reconstruction and post processing algorithms can be implemented. Dicom images are viewed on a cloud based DICOM viewer (Orthanc) or pushed back to the local PACS.

Results: The data was pushed to cloud using Rasberry PI 3 edge gateway device and image is viewed using Orthanc’s web based image viewer StoneViewer[6].

Impact: Low cost solution to transfer the raw MR data to the cloud in secure and HIPPA compliant manner would help latest advancements in MR imaging to remote parts of the world where internet bandwidth itself can be challenging

Keywords: Data Processing, Oxygenation

Motivation: OEF as a biomarker requires SvO₂ quantities which can be estimated via susceptometry-based oximetry (SBO). The fundamental is based on the presence of deoxyhemoglobin which induces local field inhomogeneity. The conventional approach for SvO₂ estimation via SBO utilizes the phase difference accrued over time between intra- and extravascular compartments. 

Goal(s): Our goal was to demonstrate an alternative for SvO₂ estimation with a single-echo GRE image.

Approach: In this study, the k-space energy spectrum analysis was performed to quantify k-energy displacement for field inhomogeneity mapping. 

Results: The preliminary result shows a comparable SvO₂ value as that from the conventional approach. 

Impact: Susceptometry-based oximetry, based on field inhomogeneity, has become a metric for SvO₂ estimation. The KESA algorithm was applied to a single-echo GRE image to map field inhomogeneity. The comparable SvO₂ indicated the potential of KESA in SvO₂ estimation.

Keywords: Data Processing, Data Analysis, Speech, motion, atlas, dynamic MRI, cleft palate

Motivation: Characterizing velopharyngeal motion patterns in children with cleft palate is an important research topic. 

Goal(s): Utilizing recently improved dynamic MRI techniques, we aim to uncover distinctive deformation patterns in cleft palate speech from a statistical perspective. 

Approach: We propose a post-processing pipeline based on spatiotemporal atlases, manually segmented velopharyngeal labels, deformable registration, and principal component analysis. The speech dataset consisting of 17 normal controls and 4 patients was analyzed.

Results: The proposed method effectively captures and separates patient-specific deformation patterns within principal component’s feature spaces. Furthermore, it reveals the impact from different anatomical regions in cleft palate speech.

Impact: In practice, cleft palate patterns in speech MRI are too subtle for visual examination or conventional post-processing methods to reveal. Providing a solution to uncover such patterns is essential to help understand the anatomical and functional changes in this disorder.

Keywords: Software Tools, Software Tools, classification

Motivation: Our motivation is the challenge presented by the variability of DICOM metadata across different MRI scan manufacturers and protocols which can complicate scan type classification.

Goal(s): To provide a simple yet versatile tool for the classification of MRI scan types, which enhances classification accuracy through a refined methodology.

Approach: We use YAML-based declarative rules to process the arbitrary DICOM metadata, enabling nuanced categorization of MRI scans that can adapt to the mentioned variability.

Results: The ability to accurately map complex combinations of metadata characteristics to define scan types and intrinsic features, thereby achieving a classification process that is both precise and quick.

Impact: The fw-classification package simplifies the image classification workflow, minimizing potential for human error, and increasing throughput. This adaptable framework handles complex and heterogeneous metadata structures, which is necessary for robust classification across a variety of manufacturers and protocols.

Keywords: Segmentation, Segmentation

Motivation: The motivation for this research study is to better understand and characterize upper-airway collapse during sleep in patients with obstructive sleep apnea (OSA). The study aims to provide valuable insights into the dynamics of airway collapse.

Goal(s): The main goal is to quantitatively assess upper-airway collapse dynamics in both OSA and normal individuals, developing imaging phenotypes.

Approach: The research methodology involves data collection using MRI, manual analysis for quantitative imaging phenotypes, and presenting qualitative and quantitative findings.

Results: The study effectively visualizes airway collapse patterns in normal and OSA patients, develops quantitative imaging phenotypes, and distinguishes various collapse patterns.

Impact: The study's results provide researchers with new, quantitative insights into upper-airway collapse in sleep apnea. This may enable more precise diagnosis and treatment, stimulate further research into non-CPAP therapies, and improve the quality of care for patients with sleep apnea.

Keywords: Data Processing, Perfusion

Motivation: Least squares can provide biased and inefficient estimations of pharmokinetic parameters in certain circumstances.

Goal(s): To develop a more efficient method to extract pharmokinetic parameters as well as the limit on the precision of those parameters.

Approach: The likelihood function for multi-channel receive dynamic contrast enhanced MRI was used to develop a method to extract pharmokinetic parameters.

Results: The proposed maximum likelihood estimator method has substantially less bias than the typical least squares method. For the input parameters investigated the maximum likelihood estimated pharmacokinetic parameters had standard deviations within 10% or less of their fundamental lower bound.

Impact: The developed tool provides nearly efficient pharmokinetic estimations, thereby providing an improved method over the standard least squares approach, as well as demonstrating the limited utlility of future machine learning and other methods attempting to solve this problem.