Hui Jing Yu

Atlas-based liver segmentation and hepatic fat-fraction assessment for clinical trials

Automated assessment of hepatic fat-fraction is clinically important. A robust and precise segmentation would enable accurate, objective and consistent measurement of hepatic fat-fraction for disease quantification, therapy monitoring and drug development. However, segmenting the liver in clinical trials is a challenging task due to the variability of liver anatomy as well as the diverse sources the images were acquired from. In this paper, we propose an automated and robust framework for liver segmentation and assessment. It uses single statistical atlas registration to initialize a robust deformable model to obtain fine segmentation. Fat-fraction map is computed by using chemical shift based method in the delineated region of liver. This proposed method is validated on 14 abdominal magnetic resonance (MR) volumetric scans. The qualitative and quantitative comparisons show that our proposed method can achieve better segmentation accuracy with less variance comparing with two other atlas-based methods. Experimental results demonstrate the promises of our assessment framework.

Central endoscopy reads in inflammatory bowel disease clinical trials: The role of the imaging core lab

Clinical trials in inflammatory bowel disease (IBD) are evolving at a rapid pace by employing central reading for endoscopic mucosal assessment in a field that was, historically, largely based on assessments by local physicians. This transition from local to central reading carries with it numerous technical, operational, and scientific challenges, many of which can be resolved by imaging core laboratories (ICLs), a concept that has a longer history in clinical trials in a number of diseases outside the realm of gastroenterology. For IBD trials, ICLs have the dual goals of providing objective, consistent assessments of endoscopic findings using central-reading paradigms whilst providing important expertise with regard to operational issues and regulatory expectations. This review focuses on current approaches to using ICLs for central endoscopic reading in IBD trials.

Accurate thigh inter-muscular adipose quantification using a data-driven and sparsity-constrained deformable model

The thigh inter-muscular adipose tissue (IMAT) quantification plays a critical role in various medical analysis tasks, e.g., the analysis of physical performance or the diagnose of knee osteoarthritis. In recent years, several techniques have been proposed to perform automated thigh tissues quantification. However, nobody has provided effective methods to track fascia lata, which is an important anatomic trail to distinguish between subcutaneous adipose tissue (SAT) and I-MAT in thigh. As a result, the estimation of IMAT may not be accurate for subjects with pathological conditions. On the other hand, tissue prior information, e.g., intensity, orientation and scale, becomes critical to infer and refine the fascia lata boundary from image appearance cues. In this paper, we propose a novel data-driven and sparsity-constrained de-formable model to obtain accurate fascia lata labeling. The model deformation is driven by the target points on fascia lata detected by a local discriminative classifier in a narrowband fashion. By using a sparsity-constrained optimization, the deformation is solved with errors and outliers suppression. The proposed approach has been evaluated on a set of 3D MR thigh volumes. In a comparison with another state-of-art framework, our approach produces superior performance.

A DXA Whole Body Composition Cross-Calibration Experience: Evaluation With Humans, Spine, and Whole Body Phantoms

New densitometer installation requires cross-calibration for accurate longitudinal assessment. When replacing a unit with the same model, the International Society for Clinical Densitometry recommends cross-calibrating by scanning phantoms 10 times on each instrument and states that spine bone mineral density (BMD) should be within 1%, whereas total body lean, fat, and %fat mass should be within 2% of the prior instrument. However, there is limited validation that these recommendations provide adequate total body cross-calibration. Here, we report a total body cross-calibration experience with phantoms and humans. Cross-calibration between an existing and new Lunar iDXA was performed using 3 encapsulated spine phantoms (GE [GE Lunar, Madison, WI], BioClinica [BioClinica Inc, Princeton, NJ], and Hologic [Hologic Inc, Bedford, MA]), 1 total body composition phantom (BioClinica), and 30 human volunteers. Thirty scans of each phantom and a total body scan of human volunteers were obtained on each instrument. All spine phantom BMD means were similar (within 1%; <-0.010 g/cm2 bias) between the existing and new dual-energy X-ray absorptiometry unit. The BioClinica body composition phantom (BBCP) BMD and bone mineral content (BMC) values were within 2% with biases of 0.005 g/cm2 and -3.4 g. However, lean and fat mass and %fat differed by 4.6%-7.7% with biases of +463 g, -496 g, and -2.8%, respectively. In vivo comparison supported BBCP data; BMD and BMC were within ∼2%, but lean and fat mass and %fat differed from 1.6% to 4.9% with biases of +833 g, -860 g, and -1.1%. As all body composition comparisons exceeded the recommended 2%, the new densitometer was recalibrated. After recalibration, in vivo bias was lower (<0.05%) for lean and fat; -23 and -5 g, respectively. Similarly, BBCP lean and fat agreement improved. In conclusion, the BBCP behaves similarly, but not identical, to human in vivo measurements for densitometer cross-calibration. Spine phantoms, despite good BMD and BMC agreement, did not detect substantial lean and fat differences observed using BBCP and in vivo assessments. Consequently, spine phantoms are inadequate for dual-energy X-ray absorptiometry whole body composition cross-calibration.

An Automated and Robust Framework for Quantification of Muscle and Fat in the Thigh

The tissue quantification in the thigh (e.g. cross-sectional areas of adipose tissue and muscle) is important, since their quantities reflect adverse metabolic effects and muscle function. Traditional manual analysis is time-consuming and operator-dependent, especially in the case of multi-slices or 3D datasets. In clinical trials, there are a large amount of datasets acquired from magnetic resonance imaging (MRI) or X-ray computed tomography (CT) that requires automatic labeling of individual tissues. Since most segmentation algorithms are not suited for different modalities, we present an automatic and robust framework for the quantitative assessment of muscle and fat tissues on 3D MR or CT data. In our framework, a variational Bayesian Gaussian mixture model is used to cluster regions of interest in images into adipose tissues (fat and marrow), muscle, bone and background. The identification of each cluster is based on marrow detection. Furthermore, we use a combination of parametric and geodesic active contour models to distinguish different adipose tissues in 3D images. To validate our proposed framework, we have conducted preliminary experiments on five volumetric mid-thigh axial datasets of MR and CT images from clinical trials.

ACCURACY OF BMAS HIPPOCAMPUS SEGMENTATION USING THE HARMONIZED HIPPOCAMPAL PROTOCOL

 Hippocampal volume (HCV) measured with MRI has been widely used as a key biomarker for both improving subject selection and monitoring treatment efficacy in Alzheimer’s Disease (AD) studies. However various hippocampal protocols exist in the literature, each including a different set of subfields and sub-structures, potentially leading to confusion and additional complexity for direct comparison and consistency in reporting study results.

Medical Imaging Modalities

Medical imaging is now utilized extensively in clinical trials for eligibility, efficacy, and safety evaluations. The uses of imaging span from a qualitative assessment of disease findings to quantitative assessments, each resting on diagnosis of the condition or change in the severity of the condition. This introductory chapter is designed for the novice with a limited or no background in radiological techniques and aims to briefly review the different imaging techniques, technology, terminology, and optimal imaging uses.

Automatic Liver Segmentation and Hepatic Fat Fraction Assessment in MRI

Automated assessment of hepatic fat fraction is clinically important. A robust and precise segmentation would enable accurate, objective and consistent measurement of liver fat fraction for disease quantification, therapy monitoring and drug development. However, segmenting the liver in clinical trials is a challenging task due to the variability of liver anatomy as well as the diverse sources the images were acquired from. In this paper, we propose an automated and robust framework for liver segmentation and assessment. It uses single statistical atlas registration to initialize a robust deformable model to get fine segmentation. Fat fraction map is computed by using chemical shift based method in the delineated region of liver. This proposed method is validated on 14 abdominal magnetic resonance (MR) volumetric scans. The qualitative and quantitative comparisons show that our proposed method can achieve better segmentation accuracy with less variance comparing with an automatic graph cut method. Experimental results demonstrate the promises of our assessment framework.

Evaluation of Quantitative Computed Tomography Cortical Hip Quadrant in a Clinical Trial With Rosiglitazone: A Potential New Study Endpoint.

Quantitative computed tomography (QCT) measurements have been used extensively to ascertain information about bone quality and density due to the 3-dimensional information provided and the ability to segment out trabecular and cortical bones. QCT imaging helps to improve our understanding of the role that each bone compartment plays in the pathogenesis and prognosis of fracture. This study was conducted to explore longitudinal changes in femoral neck (FN) cortical bone structure using both volumetric bone mineral density (vBMD) and cortical shell thickness assessments via QCT in a double-blind, randomized, multicenter clinical trial in postmenopausal women with type 2 diabetes mellitus. This study also examined whether treatment-associated changes in the cortical bone vBMD and thickness in femoral neck quadrants could be evaluated. Subjects were randomized to rosiglitazone (RSG) or metformin (MET) for 52 wk followed by 24 wk of open-label MET. A subset of 87 subjects underwent QCT scans of the hip at baseline, after 52 wk of double-blind treatment, and after 24 wk of treatment with MET using standard full-body computed tomography scanners. All scans were evaluated and analyzed centrally. Cortical vBMD at the FN was precisely segmented from trabecular bone and used to assess a possible therapeutic effect on this bone compartment. QCT analysis showed reductions in adjusted mean percentage change in vBMD and in absolute cortical thickness occurred with RSG treatment from baseline to week 52, whereas changes with MET were generally minimal. The reductions observed during RSG treatment for 1 yr appeared to partially reverse during the open-label MET phase from weeks 52 to 76. The femoral neck quadrant may provide utility as a potential endpoint in clinical trials for the understanding of the therapeutic effect of new entities on cortical bone vs trabecular bone; however, further clinical validation is needed. TRIAL REGISTRATION The protocol (GSK study number AVD111179) was registered on ClinicalTrials.gov as NCT00679939.

A detection-driven and sparsity-constrained deformable model for fascia lata labeling and thigh inter-muscular adipose quantification

A deformable model for robust reconstruction of fascia lata surface is proposed.The models deformation is driven through a discriminative detector.A sparsity-constrained optimization is used to suppress detection outliers. Quantification of the thigh inter-muscular adipose tissue (IMAT) plays a critical role in various medical data analysis tasks, e.g., the analysis of physical performance or the diagnosis of knee osteoarthritis. Several techniques have been proposed to perform automated thigh tissues quantification. However, none of them has provided an effective method to track fascia lata, which is an important anatomic trail to distinguish between subcutaneous adipose tissue (SAT) and IMAT in the thigh. As a result, the estimates of IMAT may not be accurate due to the unclear appearance cues, complicated anatomic, or pathological characteristics of the fascia lata. Thus, prior tissue information, e.g., intensity, orientation and scale, becomes critical to infer the fascia lata location from magnetic resonance (MR) images. In this paper, we propose a novel detection-driven and sparsity-constrained deformable model to obtain accurate fascia lata labeling. The models deformation is driven by the detected control points on fascia lata through a discriminative detector in a narrow-band fashion. By using a sparsity-constrained optimization, the deformation is solved from errors and outliers suppression. The proposed approach has been evaluated on a set of 3D MR thigh volumes. In a comparison with the state-of-the-art framework, our approach produces superior performance.