J. Alison Noble

International standards for fetal growth based on serial ultrasound measurements: the Fetal Growth Longitudinal Study of the INTERGROWTH-21st Project

BACKGROUND In 2006, WHO produced international growth standards for infants and children up to age 5 years on the basis of recommendations from a WHO expert committee. Using the same methods and conceptual approach, the Fetal Growth Longitudinal Study (FGLS), part of the INTERGROWTH-21(st) Project, aimed to develop international growth and size standards for fetuses. METHODS The multicentre, population-based FGLS assessed fetal growth in geographically defined urban populations in eight countries, in which most of the health and nutritional needs of mothers were met and adequate antenatal care was provided. We used ultrasound to take fetal anthropometric measurements prospectively from 14 weeks and 0 days of gestation until birth in a cohort of women with adequate health and nutritional status who were at low risk of intrauterine growth restriction. All women had a reliable estimate of gestational age confirmed by ultrasound measurement of fetal crown-rump length in the first trimester. The five primary ultrasound measures of fetal growth--head circumference, biparietal diameter, occipitofrontal diameter, abdominal circumference, and femur length--were obtained every 5 weeks (within 1 week either side) from 14 weeks to 42 weeks of gestation. The best fitting curves for the five measures were selected using second-degree fractional polynomials and further modelled in a multilevel framework to account for the longitudinal design of the study. FINDINGS We screened 13,108 women commencing antenatal care at less than 14 weeks and 0 days of gestation, of whom 4607 (35%) were eligible. 4321 (94%) eligible women had pregnancies without major complications and delivered live singletons without congenital malformations (the analysis population). We documented very low maternal and perinatal mortality and morbidity, confirming that the participants were at low risk of adverse outcomes. For each of the five fetal growth measures, the mean differences between the observed and smoothed centiles for the 3rd, 50th, and 97th centiles, respectively, were small: 2·25 mm (SD 3·0), 0·02 mm (3·0), and -2·69 mm (3·2) for head circumference; 0·83 mm (0·9), -0·05 mm (0·8), and -0·84 mm (1·0) for biparietal diameter; 0·63 mm (1·2), 0·04 mm (1·1), and -1·05 mm (1·3) for occipitofrontal diameter; 2·99 mm (3·1), 0·25 mm (3·2), and -4·22 mm (3·7) for abdominal circumference; and 0·62 mm (0·8), 0·03 mm (0·8), and -0·65 mm (0·8) for femur length. We calculated the 3rd, 5th 10th, 50th, 90th, 95th and 97th centile curves according to gestational age for these ultrasound measures, representing the international standards for fetal growth. INTERPRETATION We recommend these international fetal growth standards for the clinical interpretation of routinely taken ultrasound measurements and for comparisons across populations. FUNDING Bill & Melinda Gates Foundation.

2D+T acoustic boundary detection in echocardiography

In this paper we address the problem of spatio-temporal acoustic boundary detection in echocardiography. We propose a phase-based feature detection method to be used as the front end to higher-level 2D+T/3D+T reconstruction algorithms. We develop a 2D+T version of this algorithm and illustrate its performance on some typical echocardiogram sequences. We show how our temporal-based algorithm helps to reduce the number of spurious feature responses due to speckle and provides feature velocity estimates. Further, our approach is intensity-amplitude invariant. This makes it particularly attractive for echocardiographic segmentation, where choosing a single global intensity-based edge threshold is problematic.

Segmentation of ultrasound images: multiresolution 2D and 3D algorithm based on global and local statistics

In this paper, we propose a robust adaptive region segmentation algorithm for noisy images, within a Bayesian framework. A multiresolution implementation of the algorithm is performed using a wavelets basis and can be used to process both 2D and 3D data. In this work we focus on the adaptive character of the algorithm and we discuss how global and local statistics can be utilised in the segmentation process. We propose an improvement on the adaptivity by introducing an enhancement to control the adaptive properties of the segmentation process. This takes the form of a weighting function accounting for both local and global statistics, and is introduced in the minimisation. A new formulation of the segmentation problem allows us to control the effective contribution of each statistical component. The segmentation algorithm is demonstrated on synthetic data, 2D breast ultrasound data and on echocardiographic sequences ð2D þ TÞ. An evaluation of the performance of the proposed algorithm is also presented. � 2002 Elsevier Science B.V. All rights reserved.

Quantitative 3-Dimensional Echocardiography for Accurate and Rapid Cardiac Phenotype Characterization in Mice

Background—Insufficient techniques exist for rapid and reliable phenotype characterization of genetically manipulated mouse models of cardiac dysfunction. We developed a new, robust, 3-dimensional echocardiography (3D-echo) technique and hypothesized that this 3D-echo technique is as accurate as magnetic resonance imaging (MRI) and histology for assessment of left ventricular (LV) volume, ejection fraction, mass, and infarct size in normal and chronically infarcted mice. Methods and Results—Using a high-frequency, 7/15-MHz, linear-array ultrasound transducer, we acquired ECG and respiratory-gated, 500-&mgr;m consecutive short-axis slices of the murine heart within 4 minutes. The short-axis movies were reassembled off-line in a 3D matrix by using the measured platform locations to position each slice in 3D. Epicardial and endocardial heart contours were manually traced, and a B-spline surface was fitted to the delineated image curves to reconstruct the heart volumes. Excellent correlations were obtained between 3D-echo and MRI for LV end-systolic volumes (r=0.99, P<0.0001), LV end-diastolic volumes (r=0.99, P<0.0001), ejection fraction (r=0.99, P<0.0001), LV mass (r=0.94, P<0.0019), and infarct size (r=0.98, P<0.0001). Also, excellent correlations were found between the 3D-echo–derived LV mass and necropsy LV mass in normal mice (r=0.99, P<0.0001), as well as for 3D-echo–derived infarct size and histologically determined infarct size (r=0.99, P<0.0001) in mice with chronic heart failure. Bland-Altman analysis showed excellent limits of agreement between techniques for all measured parameters. Conclusion—This new, fast, and highly reproducible 3D-echo technique should be of widespread applicability for high-throughput murine cardiac phenotyping studies.

Learning to detect cells using non-overlapping extremal regions

Cell detection in microscopy images is an important step in the automation of cell based-experiments. We propose a machine learning-based cell detection method applicable to different modalities. The method consists of three steps: first, a set of candidate cell-like regions is identified. Then, each candidate region is evaluated using a statistical model of the cell appearance. Finally, dynamic programming picks a set of non-overlapping regions that match the model. The cell model requires few images with simple dot annotation for training and can be learned within a structured SVM framework. In the reported experiments, state-of-the-art cell detection accuracy is achieved for H&E stained histology, fluorescence, and phase-contrast images.

On the Choice of Band-Pass Quadrature Filters

Band-pass quadrature filters are extensively used in computer vision to estimate information from images such as: phase, energy, frequency and orientation,1 possibly at different scales and utilise this in further processing-tasks. The estimation is intrinsically noisy and depends critically on the choice of the quadrature filters. In this paper, we first study the mathematical properties of the quadrature filter pairs most commonly seen in the literature and then consider some new pairs derived from the classical feature detection literature. In the case of feature detection, we present the first attempt to design a quadrature pair based on filters derived for optimal edge/line detection. A comparison of the filters is presented in terms of feature detection performance, wherever possible, in the sense of Canny and in terms of phase stability. We conclude with remarks on how our analysis can aid in the choice of a filter pair for a given image processing task.

A novel ultrasound indentation system for measuring biomechanical properties of in vivo soft tissue

Technologies for soft tissue analysis are advancing at a rapid place. For instance, elastography, which provides soft tissue strain images, is starting to be tried in clinical practice as a tool for diagnosing cancer. Soft tissue deformation modeling and analysis is also an active area of research that has application in surgery planning and treatment. Typically, quantitative soft tissue analysis uses nominal values of soft tissue biomechanical properties. However, in practice, soft tissue properties can vary significantly between individuals. Hence, for soft tissue methodologies to reach their full potential as patient-specific techniques, there is a need to develop ways to efficiently measure soft tissue mechanical properties in vivo. This paper describes a prototype real-time ultrasound (US) indentation test system developed to meet this need. The system is based on the integration of a force sensor and an optical tracking system with a commercial US machine integrated with a suite of analysis methodologies. In a study on a single-layer phantom, we used the system to compare various methods of estimating linear elastic properties (via a theoretical approximation, 2-D finite element analysis, 3-D finite element analysis and a standard material-testing method). In a second study on a three-layer gelatin phantom, we describe a new finite-element-based inverse solution for recovering the Youngs moduli of each layer to show how the system can estimate properties of internal components of soft tissue. Finally, we show how the system can be used to derive a modified quasilinear viscoelastic (QVL) model on real breast tissue.

MAP MRF joint segmentation and registration of medical images

The problems of segmentation and registration are traditionally approached individually, yet the accuracy of one is of great importance in influencing the success of the other. In this paper, we aim to show that more accurate and robust results may be obtained through seeking a joint solution to these linked processes. The outlined approach applies Markov random fields in the solution of a maximum a posteriori model of segmentation and registration. The approach is applied to synthetic and real MRI data.

Random Forest Classification for Automatic Delineation of Myocardium in Real-Time 3D Echocardiography

Automatic delineation of the myocardium in real-time 3D echocardiography may be used to aid the diagnosis of heart problems such as ischaemia, by enabling quantification of wall thickening and wall motion abnormalities. Distinguishing between myocardial and non-myocardial tissue is, however, difficult due to low signal-to-noise ratio as well as the efficiency constraints imposed on any algorithmic solution by the large size of the data under consideration. In this paper, we take a machine learning approach treating this problem as a two-class 3D patch classification task. We demonstrate that solving such task using random forests , which are the discriminative classifiers developed recently in the machine learning community, allows to obtain accurate delineations in a matter of seconds (on a CPU) or even in real-time (on a GPU) for the entire 3D volume.

Velocity Estimation in Ultrasound Images: A Block Matching Approach

In this paper, we focus on velocity estimation in ultrasound images sequences. Ultrasound images present many difficulties in image processing because of the typically high level of noise found in them. Recently, Cohen and Dinstein have derived a new similarity measure, according to a simplified image formation model of ultrasound images, optimal in the maximum likelihood sense. This similarity measure is better for ultrasound images than others such as the sum-of-square differences or normalised cross-correlation because it takes into account the fact that the noise in an ultrasound image is multiplicative Rayleigh noise, and that displayed ultrasound images are log-compressed. In this work we investigate the use of this similarity measure in a block matching method. The underlying framework of the method is Singhs algorithm. New improvements are made both on the similarity measure and the Singh algorithm to provide better velocity estimates. A global optimisation scheme for algorithm parameter estimation is also proposed. We show that this optimisation makes an improvement of approximately 35% in comparison to the result obtained with the worst parameter set. Results on clinically acquired cardiac and breast ultrasound sequences, demonstrate the robustness of the method.