C. L. Knight

Evaluation and Comparison of Current Fetal Ultrasound Image Segmentation Methods for Biometric Measurements: A Grand Challenge

This paper presents the evaluation results of the methods submitted to Challenge US: Biometric Measurements from Fetal Ultrasound Images, a segmentation challenge held at the IEEE International Symposium on Biomedical Imaging 2012. The challenge was set to compare and evaluate current fetal ultrasound image segmentation methods. It consisted of automatically segmenting fetal anatomical structures to measure standard obstetric biometric parameters, from 2D fetal ultrasound images taken on fetuses at different gestational ages (21 weeks, 28 weeks, and 33 weeks) and with varying image quality to reflect data encountered in real clinical environments. Four independent sub-challenges were proposed, according to the objects of interest measured in clinical practice: abdomen, head, femur, and whole fetus. Five teams participated in the head sub-challenge and two teams in the femur sub-challenge, including one team who tackled both. Nobody attempted the abdomen and whole fetus sub-challenges. The challenge goals were two-fold and the participants were asked to submit the segmentation results as well as the measurements derived from the segmented objects. Extensive quantitative (region-based, distance-based, and Bland-Altman measurements) and qualitative evaluation was performed to compare the results from a representative selection of current methods submitted to the challenge. Several experts (three for the head sub-challenge and two for the femur sub-challenge), with different degrees of expertise, manually delineated the objects of interest to define the ground truth used within the evaluation framework. For the head sub-challenge, several groups produced results that could be potentially used in clinical settings, with comparable performance to manual delineations. The femur sub-challenge had inferior performance to the head sub-challenge due to the fact that it is a harder segmentation problem and that the techniques presented relied more on the femurs appearance.

Scientific basis for standardization of fetal head measurements by ultrasound: a reproducibility study

To compare the standard methods for ultrasound measurement of fetal head circumference (HC) and biparietal diameter (BPD) (outer‐to‐outer (BPDoo) vs outer‐to‐inner (BPDoi) caliper placement), and compare acquisition of these measurements in transthalamic (TT) vs transventricular (TV) planes.

Feature-based fuzzy connectedness segmentation of ultrasound images with an object completion step.

Highligts• Novel US segmentation approach based on the fuzzy connectedness framework.• Use of local phase and feature asymmetry to define affinity function.• Shape-based object completion step to detect and complete one or more gaps.• Novel regional entropy-based quantitative image quality assessment approach.• Method performs well across a variety of image qualities from clinical practice.

Oriented feature-based coupled ellipse fitting for soft tissue quantification in ultrasound images

Soft tissue quantification from ultrasound (US) images is a challenging task due to signal dropouts, missing boundaries, and presence of speckle. Typically, intensity and intensity gradient-based methods do not suffice to obtain a good representation of the object of interest. Prior work has shown that local phase, derived from the monogenic signal, extracts structural information from US images, being contrast invariant. An oriented edge map can be derived from feature asymmetry, resulting from combining different scales of the local phase, and local orientation. This paper proposes a novel feature-based approach, based on a modified Hough transform framework, for the detection of coupled ellipses for soft tissue quantification in US images using oriented edge maps, derived from the monogenic signal. Quantitative and qualitative results are illustrated on US images of the fetal arm across gestation, the object of interest being the adipose tissue layer, which is an indicator of fetal nutrition.

Ultrasound image segmentation using feature asymmetry and shape guided live wire

Ultrasound (US) is a versatile, low cost, real-time, widely available imaging modality. Manual segmentation for volumetric US measurements can be difficult and very time consuming, requiring slice-by-slice segmentations. However, automatic segmentation of ultrasound images can prove challenging due to the presence of speckle, attenuation, missing boundaries, signal dropouts, and artefacts. Semi-automatic segmentation techniques can improve the speed and accuracy of such measurements, taking advantage of clinical expertise while allowing user interaction. This paper presents a novel solution for interactive image segmentation on B-mode ultrasound images. The proposed method builds on the Live Wire framework and introduces two new sets of Live Wire costs, namely a Feature Asymmetry (FA) cost to localise edges and a weak shape constraint cost to aid the selection of appropriate boundaries in the presence of missing information or artefacts. The resulting semi-automatic segmentation method follows edges based on structural relevance rather than intensity gradients, adapting the method to ultrasound images, where the object boundaries are normally fuzzy. The new method is applied in the context of fetal arm adipose tissue quantification, the adipose tissue being an indicator of the fetal nutritional state. A quantitative and qualitative evaluation is performed with respect to related segmentation techniques. The method was tested on 48 manually segmented ultrasound images of the fetal arm across gestation, showing similar accuracy to the intensity-based Live Wire approach but superior repeatability while requiring significantly less time and user interaction.

Regularised feature-based fuzzy connectedness segmentation of ultrasound images for fetal soft tissue quantification across gestation

Ultrasound (US) image segmentation can be a challenging task due to signal dropouts, missing boundaries, and presence of speckle. Typically, purely intensity-based methods do not lead to a good segmentation of the structures of interest. Prior work has shown that local phase and feature asymmetry, derived from the monogenic signal, extract structural information from US images. This paper proposes a novel US segmentation approach based on the fuzzy connectedness framework. The approach uses local phase and feature asymmetry to define the so-called affinity function, which drives the segmentation algorithm, incorporates a shape-based object completion step, and regularises the result by mean curvature flow. Identifying new image-based biomarkers of fetal nutrition across gestation is essential to characterise the well-being of a fetus at an early stage. Results are presented on US images of the fetal arm taken at multiple gestational ages, the adipose tissue being an indicator of the fetal nutritional state.

OP17.05: Fetal arm fat: an in utero marker of body composition?

Objectives: To assess whether fetal arm fat area by 2D ultrasound in the last 4 weeks of pregnancy can be used as a surrogate marker of body composition. Methods: In this prospective study, fetal arm fat was measured using serial ultrasound scans at 4–6 weekly intervals throughout pregnancy in 106 optimally healthy women. Cross sectional fetal arm fat area was measured mid-humerus at 36–40 weeks of gestation. Newborn infants then underwent body composition analysis using air displacement plethysmography (PEA POD) to measure the percentage of whole body fat and fat-free mass. We compared prenatal findings (ultrasound fetal arm fat area) in babies whose postnatal whole body fat mass percentage was less than the 25th centile to those where this was above the 75th centile. Results: The percentage of body fat assessed by PEA POD in the 106 infants was normally distributed. There were 17 babies with body fat < 25th centile and 15 > 75th centile with a scan at ≥ 36+0/40 where arm fat measurements were possible. The average gestational age at birth was 39+6 and 40+2 weeks, respectively, with 12 and 13 day scan-birth intervals. As expected, the mean birthweight was lower in those with a lower percentage of fat mass (3052 g (2421–3632 g)) when compared to those with high fat mass (3870 g (3383–4460 g)). Prenatal mean fetal arm fat area was significantly smaller in those with low fat mass (566 mm2 (432–711 mm2)) compared to those with high fat mass (836 mm2 (596–1122 mm2), P < 0.001). Conclusions: Fetal arm fat area can be assessed by ultrasound at ≥ 36+0/40, and can distinguish those babies that are subsequently found to have high and low whole body fat mass. This provides preliminary evidence of the potential for this measurement as a tool to assess fetal nutritional status in utero.

OP17.04: Fetal arm fat: development throughout pregnancy in selectively screened optimally healthy patients

The AUC for the detection of late SGA was 0.652 (0.623–0.680) achieving detection rates of 14%, 23%, 32% and 70% for a FPR of 5, 10, 15 and 50% respectively. Conclusions: First-trimester integrated screening maybe useful to predict early SGA mainly because of its strong association with pre-eclampsia. The prediction of late SGA is poorer, but it might useful as part of first trimester strategies to select women requiring ultrasound in the third trimester.

P28.09: Novel automatic method for measuring fetal arm adipose tissue in 2D ultrasound images across gestation

Methods: The patients with morbidly adherent placenta assessed on antenatal ultrasound examination and confirmed at delivery in 2004–7 were identified from hospital records and a perinatal database. The antenatal sonographic appearance was compared with the clinical and histological findings at delivery to identify the sonographic features that were useful for the assessment of myometrial involvement. Results: Out of the 9 cases of morbidly adherent placenta identified (including 1 case of placenta percreta, 1 minor partial increta and 7 extensive increta), the criteria useful for assessment of the lateral placental/myometrial involvement were obliteration of retroplacental clear space, increased subplacental vascularity, placental-tissue interface disruption, and vessels crossing the interface disruption site; and for assessment of the depth, bladder serosa – uterine wall interface hypervascularity, vessels extending form placenta to bladder, placental-tissue interface disruption and vessels crossing placental-tissue interface disruption. Other useful signs included the myometrial vascular pattern and bulging of the lower uterine wall. Conclusions: Ultrasound is useful for assessment of the placental/myometrial involvement in morbidly adherent placenta. To assess the lateral extent and the depth of placental/myometrial involvement, different criteria in combinations may be required.

P27.07: Assessing fetal fractional arm adipose tissue volume using 3D ultrasound: the importance of measuring perpendicular to the humerus

using the VOCAL (Virtual Organ Computer-aided AnaLysis) imaging program which is integrated into the Voluson E8 ultrasound system. The intrauterine content was sent for microscopic examination. The findings were compared with corresponding volumes and findings from women with uncomplicated postpartum periods from a previous study. Results: In 17 of the 20 cases placental remnants were verified by the microscopic examination. Five of these women had a uterine volume that exceeded the largest interquartile volume observed in women with a normal puerperium and in 4 the volume exceeded the largest volume observed in the normal puerperium. The other 12 women had normal uterine volumes. Fifteen of the 17 women with placental remnants had a uterine cavity volume that exceeded the interquartile volume observed in women with a normal puerperium and in 13 the volume exceeded the largest volume observed in the normal puerperium. In all 9 cases examined 42 days or more after delivery the uterine cavity volume exceeded the largest volume observed in the normal puerperium. In all examined women an echogenic mass was seen in the uterine cavity. Conclusions: The uterine and especially the uterine cavity volumes are often larger in women with placental remnants compared to women with an uneventful postpartum course. This knowledge can be of help when placental remnants are to be verified or ruled out. A finding of an echogenic mass in the cavity has a strong correlation with the presence of placental remnants.