Automatic segmentation of 15 critical anatomical labels and measurements of cardiac axis and cardiothoracic ratio in fetal four chambers using nnU-NetV2.

BACKGROUND: Accurate segmentation of critical anatomical structures in fetal four-chamber view images is essential for the early detection of congenital heart defects. Current prenatal screening methods rely on manual measurements, which are time-consuming and prone to inter-observer variability. This study develops an AI-based model using the state-of-the-art nnU-NetV2 architecture for automatic segmentation and measurement of key anatomical structures in fetal four-chamber view images. METHODS: A dataset, consisting of 1,083 high-quality fetal four-chamber view images, was annotated with 15 critical anatomical labels and divided into training/validation (867 images) and test (216 images) sets. An AI-based model using the nnU-NetV2 architecture was trained on the annotated images and evaluated using the mean Dice coefficient (mDice) and mean intersection over union (mIoU) metrics. The model's performance in automatically computing the cardiac axis (CAx) and cardiothoracic ratio (CTR) was compared with measurements from sonographers with varying levels of experience. RESULTS: The AI-based model achieved a mDice coefficient of 87.11% and an mIoU of 77.68% for the segmentation of critical anatomical structures. The model's automated CAx and CTR measurements showed strong agreement with those of experienced sonographers, with respective intraclass correlation coefficients (ICCs) of 0.83 and 0.81. Bland-Altman analysis further confirmed the high agreement between the model and experienced sonographers. CONCLUSION: We developed an AI-based model using the nnU-NetV2 architecture for accurate segmentation and automated measurement of critical anatomical structures in fetal four-chamber view images. Our model demonstrated high segmentation accuracy and strong agreement with experienced sonographers in computing clinically relevant parameters. This approach has the potential to improve the efficiency and reliability of prenatal cardiac screening, ultimately contributing to the early detection of congenital heart defects.

Prenatal Diagnosis of Cerebellar Cortical Dysplasia: Case Report.

This was a study of 12 cerebellar cortical dysplasias (CCDs) fetuses, these cases were characterized by a disorder of cerebellar fissures. Historically, CCD diagnosis was primarily performed using postnatal imaging. Unique to this study was the case series of CCD for prenatal diagnosis using prenatal ultrasound, as well as we found that AXIN1 and FOXC1 mutations may be related to CCD.

An Easy and Effective Method for Evaluating the Position of Conus Medullaris: Counting the Number of Vertebral Ossification Center Below the End of Conus Medullaris.

OBJECTIVE: This study aimed to ascertain the conus medullaris position by counting the number of ossification centers in the vertebral bodies below the conus medullaris endpoint (N) and assess its utility in screening for closed spinal dysraphism and tethered cord syndrome. METHODS: A total of 900 normal fetuses and 146 fetuses with closed spinal dysraphism or tethered cord syndrome were included in this study. The N values were tallied and compared along the spinal longitudinal plane. The receiver operating characteristic curve was utilized, and the cut-off value of N was analyzed. RESULTS: The counting of N was successfully performed in 856 normal and 146 abnormal fetuses. In the normal group, an increase in N with gestational age was observed. Specifically, in the subgroup of 17-20 wk fetuses, N was ≥6 in 117 out of 131 cases. This figure increased to 211 out of 213 in 21-24 wk and 512 out of 512 in 25-41 wk, respectively. Cases with N ≥7 accounted for 715 out of 856 fetuses in the 17-41 wk range. In the abnormal group, N was less than 7 in 152 out of 163 fetuses, showing statistical differences between the two groups. With a cut-off value of 6.5, specificity and sensitivity reached 93.3% and 83.5%. CONCLUSIONS: The counting of N was found to be a straightforward and efficient method for evaluating the position of the conus medullaris.

Cleft palate in fetuses: feasibility of early diagnosis by Crystal and Realistic Vue rendering 3D ultrasound technology in the first trimester.

Objectives: This study aimed to evaluate the feasibility of direct visualization of a normal fetal palate and detect cleft palate in the first trimester with a novel three-dimensional ultrasound (3D US) technique, Crystal and Realistic Vue (CRV) rendering technology. Methods: Two-dimensional (2D) images and 3D volumes of healthy and cleft palate fetuses at 11-13+6 weeks were obtained prospectively. 2D ultrasound views included the coronal view of the retronasal triangle and the midsagittal view of the face. 3D-CRV views were analyzed by multiplanar mode display. The pregnancy outcomes of all fetuses were determined during the follow-up period. Results: In our study, 124 fetuses were recruited, including 100 healthy fetuses and 24 cleft palate fetuses. The cleft palate with lip was observed in 23 fetuses (bilateral in 15, unilateral in 6, median in 2), and one cleft palate was only found in the abnormal group. The bilateral (n = 12) and median (n = 2) cleft palates with lips and the cleft palate alone (n = 1) were associated with other anatomical or chromosomal abnormalities, and one unilateral cleft palate with cleft lip had concomitant NT thickening. In the cleft palate fetus group, 16 fetuses suffered intrauterine death, which was associated with other structural or chromosomal abnormalities in 14 fetuses, seven cases were terminated after consultation, and one was delivered at term. The coronal view of the retronasal triangle and the midsagittal view was easily obtained in all fetuses. 3D-CRV images of palatal parts were clearly obtained in all cases. Unilateral, bilateral, and median cleft palates with cleft lips were visually demonstrated and classified by the 3D-CRV technique. Conclusion: It is feasible to identify the palate by 3D-CRV in the first trimester in both healthy and cleft palate fetuses. Together with 2D ultrasonography as a complementary diagnostic tool, 3D-CRV is helpful in classifying the cleft palate with a reasonable degree of certainty.

Sonographic demonstration of the sulci and gyri on the convex surface in normal fetuses using 3D-ICRV rendering technology.

PURPOSE: To demonstrate morphological alteration of the sulci and gyri on the convex surface in normal fetuses using innovative three-dimensional inversion and Crystalvue and Realisticvue (3D-ICRV) rendering technology. MATERIALS AND METHODS: 3D fetal brain volumes were collected from low-risk singleton pregnancies between 15+0 and 35+6 gestational weeks. Volumes were acquired from the transthalamic axial plane by transabdominal ultrasonography and were then post-processed with Crystalvue, Realisticvue rendering software and inversion mode. Volume quality was assessed. The anatomic definition of the sulci and gyri was determined according to location and orientation. The morphology alteration and sulcus display rates were recorded in sequential order of gestational weeks. Follow-up data were collected in all cases. RESULTS: 294 of 300 fetuses (294 brain volumes) (98%) with qualified fetal brain volumes were included (n=294, median 27 gestational weeks). 6 fetuses with unsatisfactory 3D-ICRV image quality were excluded. The morphology of the sulci and gyri on the brain convex surface could be demonstrated clearly on 3D-ICRV images. The Sylvian fissure was the first structure to be recognized. From 25 to 30 weeks, other sulci and gyri became visible. An ascending trend in the display rate of the sulci was found in this period. Follow-up showed no detectable anomalies. CONCLUSION: 3D-ICRV rendering technology is different from traditional 3D ultrasound. It can provide vivid and intuitive prenatal visualization of the sulci and gyri on the brain surface. Moreover, it may offer new ideas for neurodevelopment exploration.

A New Parameter to Evaluate Fetal Sylvian Fissure by Transabdominal 2-D Ultrasound.

OBJECTIVE: Evaluation of the inclination direction and degree of the Sylvian fissure plateau has not been reported. We aimed to evaluate the Sylvian fissure plateau by Sylvian fissure plateau angle (SFPA) in axial views at 23-28 wk gestation. METHODS: A prospective ultrasound evaluation of 180 normal and 3 abnormal singleton pregnant women was conducted at 23-28 wk gestation. All cases were assessed in three axial planes of the fetal brain (the transthalamic, transventricular and transcerebellar plane) using transabdominal 2-D images. The SFPAs of all cases were measured between the brain midline and a line drawn along the Sylvian fissure plateau. Intraclass correlation coefficients (ICCs) were used to assess the intra- and inter-observer repeatability of SFPA measurements. RESULTS: The SFPAs in normal cases in the transthalamic, transventricular and transcerebellar planes were all above y = 0, while in abnormal cases were below y = 0. However, there was no major difference between the angles measured on the transthalamic and transventricular planes (p = 0.365). There was a major difference between the SFPAs on the transcerebellar plane and transthalamic/transventricular plane (p < 0.05). The intra- and inter-observer ICCs were excellent at 0.971 (95% confidence interval [CI]: 0.945-0.984) and 0.936 (95% CI: 0.819-0.979), respectively. CONCLUSION: The SFPAs of the normal cases in three axial views were stable at 23-28 wk gestation, suggesting that 0° may be a good cut-off value for evaluating abnormal SFPA. Findings offer a potential method by which the SFPA < 0°, as shown in three abnormal cases described herein, can be evaluated prenatally and thus serve as another tool for malformations of cortical development assessment, especially for frontoobitalopercula dysplasia. We recommend use of SFPA of the transthalamic plane to evaluate the Sylvian fissure in clinical work.

Fetal Ultrasound Standard Plane Detection With Coarse-to-Fine Multi-Task Learning.

The ultrasound standard plane plays an important role in prenatal fetal growth parameter measurement and disease diagnosis in prenatal screening. However, obtaining standard planes in a fetal ultrasound video is not only laborious and time-consuming but also depends on the clinical experience of sonographers to a certain extent. To improve the acquisition efficiency and accuracy of the ultrasound standard plane, we propose a novel detection framework that utilizes both the coarse-to-fine detection strategy and multi-task learning mechanism for feature-fused images. First, traditional manually-designed features and deep learning-based features are fused to obtain low-level shared features, which can enhance the model's feature expression ability. Inspired by the process of human recognition, ultrasound standard plane detection is divided into a coarse process of plane type classification and a fine process of standard-or-not detection, which is implemented via an end-to-end multi-task learning network. The region-of-interest area is also recognised in our detection framework to suppress the influence of a variable maternal background. Extensive experiments are conducted on three ultrasound planes of the first-class fetal examination, i.e., the femur, thalamus, and abdomen ultrasound images. The experiment results show that our method outperforms competing methods in terms of accuracy, which demonstrates the efficacy of the proposed method and can reduce the workload of sonographers in prenatal screening.

Prenatal Diagnosis of Congenital Cataract: Sonographic Features and Perinatal Outcome in 41 Cases.

PURPOSE: To describe the prenatal ultrasonographic characteristics and perinatal outcomes of congenital cataract. MATERIALS AND METHODS: We analyzed congenital cataract diagnosed prenatally at four referral centers between August 2004 and February 2019. The diagnosis was confirmed by postnatal ophthalmologic evaluation of liveborn infants or autopsy for terminated cases. Maternal demographics, genetic testing results, prenatal ultrasound images, and perinatal outcomes were abstracted. RESULTS: Total of 41 cases of congenital cataract diagnosed prenatally among 788 751 women undergoing anatomic survey. Based on the sonographic characteristics, 16/41 (39.0 %) had a dense echogenic structure, 15/41 (36.6 %) had a hyperechogenic spot and 10/41 (24.4 %) had the "double ring" sign. 17/41 (41.5 %) were isolated, and 24/41 (58.5 %) had associated intraocular and extraocular findings. Microphthalmia, cardiac abnormalities, and central nervous system abnormalities were the most common associated abnormalities. Regarding potential etiology, 6 cases had a known family history of congenital cataract, 4 cases had confirmed congenital rubella infection, and 2 cases had aneuploidy. 31/41 (75.6 %) elected termination and 10/41 (24.4 %) elected to continue their pregnancy. Among the 10 cases, one case died, one case was lost to follow-up, and the remaining 8 cases were referred for ophthalmologist follow-up and postnatal surgery. CONCLUSION: Once fetal cataracts are detected, a detailed fetal anatomy survey to rule out associated abnormalities and a workup to identify the potential etiology are recommended. Prenatal diagnosis of congenital cataracts provides vital information for counseling and subsequent management.

A new classification of congenital abnormalities of UPVS: sonographic appearances, screening strategy and clinical significance.

The umbilical-portal venous system (UPVS) plays an important role in embryonic development, as well as a significant blood circulation system to ensure the normal blood supply of fetal heart and brain and other vital organs. Congenital anomalies of UPVS contain many subtypes with a broad spectrum of manifestations and prognoses. Furthermore, because of fetal small lumen of UPVS, the sonographic evaluation remains difficult in utero. Appreciation of normal embryology and anatomy of UPVS is essential to an understanding of sonographic characteristics of anomalies of UPVS and fetal sequential changes. Through reviewing previous references and our experience with congenital abnormalities of UPVS, a new comprehensive classification is proposed. The new classification identifies three types of congenital abnormalities of UPVS based on morphological abnormalities and shunts. The embryology and etiology, sonographic, clinical and prognostic characteristics of each subtype of the new classification are described in detail. Knowledge of congenital abnormalities of UPVS can give sonographers a clue and aid prenatal sonographic diagnosis. The purpose of this article is to help the sonographers to understand the new classification of congenital abnormalities of UPVS, master the sonographic characteristics of each subtype and prenatal ultrasonographic screening strategy, and guide subsequent appropriate counseling and management.

Two fetuses in one family of arterial tortuosity syndrome: prenatal ultrasound diagnosis.

BACKGROUND: Arterial tortuosity syndrome (ATS) is a rare autosomal recessive connective tissue disorder chiefly characterized by elongated and tortuosity of the large and medium sized arteries and anomalies of the vascular elastic fibers. Here we reported cases of brother about ATS from the same family on the prenatal ultrasound diagnosis. Reports of this case are rare in antenatally and we draw the vessel simulated diagram to display visually. CASE PRESENTATION: Prenatal ultrasound scanning at 29 weeks of gestation of the first fetus showed obvious tortuous and elongated of the aortic arch, ductus arteriosus, left and right pulmonary arteries, carotid and subclavian arteries. Three months after delivery, Contrast-enhanced computed tomography images (CTA) were performed to clearly display vascular abnormalities consistent with prenatal diagnosis of ultrasound. Whole exome sequencing (WES) was performed eight months after birth, two heterozygous variants of SLC2A10 gene was detected in newborn and their father and mother, respectively. Prenatal ultrasound scan at 22 weeks of gestation of the second fetus showed similar cardiovascular imaging. After birth the siblings have facial characteristic features gradually as aging. No surgical intervention was performed in the siblings follow up 19 months. CONCLUSIONS: The key points of prenatal ultrasound diagnosis of ATS are the elongation and tortuosity of the large and medium sized arteries. Genetic counseling is the process of providing individuals and families with information on the nature, inheritance, and implications of genetic disorders to help them make informed medical and personal decisions.

A Prenatal Ultrasound Scanning Approach: One-Touch Technique in Second and Third Trimesters.

This study was aimed at evaluating the performance of the innovative technique Smart Fetus (SF) developed to recognize the planes and obtain the basic biometric measurements of fetuses automatically. This prospective study included 1005 uncomplicated singleton pregnancies undergoing routine examinations. For every pregnancy, planes, including the transverse section of the thalami, transverse section of the abdomen and longitudinal section of the femur, were acquired, and standard biometric measurements, including biparietal diameter, head circumference, abdominal circumference and femur length, were obtained using SF and traditional ultrasound technique (TUT). The accuracy, reproducibility and time required for the analysis of SF were compared with those of TUT. In 998 of 1005 cases (99.30%), SF successfully acquired the sections and made all measurements. The agreement between the techniques was high for all measurements. The time to obtain sections and measure biometric parameters or solely measure biometric parameters was significantly shorter with SF than with TUT. No significant differences were found in SF repeated measurements obtained by two independent observers. The SF technique helped in the acquisition of reliable standard sections and biometric measurements and saved time. It might serve as a novel ultrasound scanning approach and improve workflow efficiency.

CacheTrack-YOLO: Real-Time Detection and Tracking for Thyroid Nodules and Surrounding Tissues in Ultrasound Videos.

To accurately detect and track the thyroid nodules in a video is a crucial step in the thyroid screening for identification of benign and malignant nodules in computer-aided diagnosis (CAD) systems. Most existing methods just perform excellent on static frames selected manually from ultrasound videos. However, manual acquisition is labor-intensive work. To make the thyroid screening process in a more natural way with less labor operations, we develop a well-designed framework suitable for practical applications for thyroid nodule detection in ultrasound videos. Particularly, in order to make full use of the characteristics of thyroid videos, we propose a novel post-processing approach, called Cache-Track, which exploits the contextual relation among video frames to propagate the detection results into adjacent frames to refine the detection results. Additionally, our method can not only detect and count thyroid nodules, but also track and monitor surrounding tissues, which can greatly reduce the labor work and achieve computer-aided diagnosis. Experimental results show that our method performs better in balancing accuracy and speed.

Fetal Open and Closed Spina Bifida on a Routine Scan at 11 Weeks to 13 Weeks 6 Days.

OBJECTIVES: This study aimed to determine the sensitivity of a first-trimester routine scan in detecting spina bifida (SB) and evaluating the first-trimester intracranial signs. METHODS: This retrospective study was a review of a prospectively collected database. All cases of SB diagnosed in a tertiary center from 2008 to 2015 were identified. The ultrasound images and medical records were reviewed. All cases of SB diagnosed prenatally were confirmed at birth or autopsy. RESULTS: A total of 24 cases of SB were diagnosed from 53,349 pregnancy cases. Except for 10 cases with a body stalk anomaly, craniorachischisis, or iniencephaly, 7 cases with open spina bifida (OSB) and 7 cases with closed spina bifida (CSB) were analyzed. The first-trimester detection rates were 100% (7 of 7) for OSB and 28.5% (2 of 7) for CSB. Eight cases were highly suspected of SB in the first trimester because of an abnormal appearance of the posterior brain; 3 were false-positive cases. Two isolated cases of OSB had first-trimester intracranial signs. An obliterated cisterna magna (CM) showed the highest sensitivity for OSB but low specificity. Two cases of OSB had no discernible landmark of intracranial translucency and the CM, and 4 showed normal intracranial translucency with an obliterated CM. All CSB cases were coupled with a normal hind brain except for 2 cases. CONCLUSIONS: A first-trimester routine scan has high sensitivity in screening for OSB. The CM may be the most sensitive intracranial sign.

Routine first-trimester ultrasound screening using a standardized anatomical protocol.

BACKGROUND: First-trimester ultrasound scans were introduced to China for nearly 20 years. The ability of first-trimester ultrasound screening to detect different malformations was variable. A recent systematic review concluded that the use of a standardized anatomic protocol was the most crucial factor to improve the sensitivity of first-trimester ultrasound screening for anomalies. Standardized sectional scans have long been used for routine anatomy screening during the second trimester. However, during the first trimester, most of the previous studies have described the observation of anatomic structures but have not specified clearly the standard sectional views. OBJECTIVE: We aimed to determine the performance of routine first-trimester scans using a standardized anatomic protocol for detecting structural abnormalities in China. STUDY DESIGN: This was a large retrospective study involving 59,063 sequential unselected pregnancies. Scans at 11 to 13+6 weeks were performed in a single center during a 7-year span. All fetuses were examined following a predefined protocol for standardized views. RESULTS: From October 2008 to December 2015, first-trimester scans were performed in 53,349 pregnant women with available outcome. Of these, there were 1578 (3%) pregnancies that presented with at least 1 fetal structural abnormality. The detection rate for first-trimester screening was 43.1% (95% confidence interval, 40.6%-45.5%). Routine first-trimester scans detected 95.6% of abdominal wall defects, 66.3% of nervous system defects, 33.8% of limbs and skeleton malformations, 30.8% of facial abnormalities, 21.2% of urogenital abnormalities, 18.4% of thoracic and lung abnormalities, and 4.1% of gastrointestinal tract abnormalities. During the first trimester, 37.7% of cardiac defects were identified and included 57.9% of major cardiac defects and 2.6% of mild cardiac defects. A robust high detection rate for anencephaly, exencephaly, cephalocele, holoprosencephaly, exomphalos, gastroschisis, Pentalogy of Cantrell, sirenomelia, and body stalk anomaly was achieved during routine first-trimester scans. CONCLUSION: A standardized anatomic protocol is advised when performing routine first-trimester ultrasound screening. It is recommended that screening for severe structural abnormalities should be extended to the first trimester.

A novel technique to assess fetal corpus callosum by two-dimensional axial plane.

OBJECTIVE: The definition of new normal values of the corpus callosum (CC) in axial sonographic scans and evaluation of their feasibility in diagnosing abnormal CC. METHODS: A cross-sectional study assessed CC from 20-gestational-week to full-term. CC observations across three axial planes (the largest CC length plane, trans-genu-and-splenium plane, and trans-body plane) were developed. The largest CC length, genu and splenium thickness, and body width and thickness were compared with compound scatter plots. Ultrasonographic features of normal and abnormal CC were described and the feasibility of the new approach studied. Intra-class correlation coefficient (ICC) was used for assessing the intra- and inter-observer agreements. RESULTS: Six hundred seventy normal and 42 abnormal fetuses from 20-gestational-week to full-term were studied. The mean normal and abnormal group maternal ages were 30.46 ± 4.36 years and 29.69 ± 4.49 years (p = 0.269). The success rate in obtaining satisfactory axial planes reached 100% but only 13.9% for sagittal plane in the normal group. The success rate of abnormal cases obtaining satisfactory axial planes was 100% and 59.5% by sagittal plane (p < 0.05). The compound scatter plots of abnormal and normal groups showed that the largest CC length and body width were significantly lower in normal fetuses, and the thickness of the genu and splenium with CC hypoplasia was significantly lower than normal fetuses. The intra- and inter-observer agreements were reproducible (all ICC > 0.850). CONCLUSIONS: The feasibility of incorporating an evaluation of CC into routine anatomical screening was demonstrated. Additionally, a focused examination of the craniocerebral axial planes exploring CC at the time of central nervous system scanning might facilitate CC anomaly detection. KEY POINTS: • Three axial planes with direct CC measurements can detect CC anomalies more accurately compared with indirect CC signs. Besides, this method is simpler, more convenient, and time-saving compared with the sagittal plane. • Assessing fetal CC on the axial plane helps clinicians to diagnose fetuses with abnormal CC. • A prospective single-center study showed that our new technique provides enough diagnostic confidence.

Hybrid attention for automatic segmentation of whole fetal head in prenatal ultrasound volumes.

BACKGROUND AND OBJECTIVE: Biometric measurements of fetal head are important indicators for maternal and fetal health monitoring during pregnancy. 3D ultrasound (US) has unique advantages over 2D scan in covering the whole fetal head and may promote the diagnoses. However, automatically segmenting the whole fetal head in US volumes still pends as an emerging and unsolved problem. The challenges that automated solutions need to tackle include the poor image quality, boundary ambiguity, long-span occlusion, and the appearance variability across different fetal poses and gestational ages. In this paper, we propose the first fully-automated solution to segment the whole fetal head in US volumes. METHODS: The segmentation task is firstly formulated as an end-to-end volumetric mapping under an encoder-decoder deep architecture. We then combine the segmentor with a proposed hybrid attention scheme (HAS) to select discriminative features and suppress the non-informative volumetric features in a composite and hierarchical way. With little computation overhead, HAS proves to be effective in addressing boundary ambiguity and deficiency. To enhance the spatial consistency in segmentation, we further organize multiple segmentors in a cascaded fashion to refine the results by revisiting context in the prediction of predecessors. RESULTS: Validated on a large dataset collected from 100 healthy volunteers, our method presents superior segmentation performance (DSC (Dice Similarity Coefficient), 96.05%), remarkable agreements with experts (-1.6±19.5 mL). With another 156 volumes collected from 52 volunteers, we ahieve high reproducibilities (mean standard deviation 11.524 mL) against scan variations. CONCLUSION: This is the first investigation about whole fetal head segmentation in 3D US. Our method is promising to be a feasible solution in assisting the volumetric US-based prenatal studies.

A Generic Quality Control Framework for Fetal Ultrasound Cardiac Four-Chamber Planes.

Quality control/assessment of ultrasound (US) images is an essential step in clinical diagnosis. This process is usually done manually, suffering from some drawbacks, such as dependence on operator's experience and extensive labors, as well as high inter- and intra-observer variation. Automatic quality assessment of US images is therefore highly desirable. Fetal US cardiac four-chamber plane (CFP) is one of the most commonly used cardiac views, which was used in the diagnosis of heart anomalies in the early 1980s. In this paper, we propose a generic deep learning framework for automatic quality control of fetal US CFPs. The proposed framework consists of three networks: (1) a basic CNN (B-CNN), roughly classifying four-chamber views from the raw data; (2) a deeper CNN (D-CNN), determining the gain and zoom of the target images in a multi-task learning manner; and (3) the aggregated residual visual block net (ARVBNet), detecting the key anatomical structures on a plane. Based on the output of the three networks, overall quantitative score of each CFP is obtained, so as to achieve fully automatic quality control. Experiments on a fetal US dataset demonstrated our proposed method achieved a highest mean average precision (mAP) of 93.52% at a fast speed of 101 frames per second (FPS). In order to demonstrate the adaptability and generalization capacity, the proposed detection network (i.e., ARVBNet) has also been validated on the PASCAL VOC dataset, obtaining a highest mAP of 81.2% when input size is approximately 300 × 300.

Multi-task learning for quality assessment of fetal head ultrasound images.

It is essential to measure anatomical parameters in prenatal ultrasound images for the growth and development of the fetus, which is highly relied on obtaining a standard plane. However, the acquisition of a standard plane is, in turn, highly subjective and depends on the clinical experience of sonographers. In order to deal with this challenge, we propose a new multi-task learning framework using a faster regional convolutional neural network (MF R-CNN) architecture for standard plane detection and quality assessment. MF R-CNN can identify the critical anatomical structure of the fetal head and analyze whether the magnification of the ultrasound image is appropriate, and then performs quality assessment of ultrasound images based on clinical protocols. Specifically, the first five convolution blocks of the MF R-CNN learn the features shared within the input data, which can be associated with the detection and classification tasks, and then extend to the task-specific output streams. In training, in order to speed up the different convergence of different tasks, we devise a section train method based on transfer learning. In addition, our proposed method also uses prior clinical and statistical knowledge to reduce the false detection rate. By identifying the key anatomical structure and magnification of the ultrasound image, we score the ultrasonic plane of fetal head to judge whether it is a standard image or not. Experimental results on our own-collected dataset show that our method can accurately make a quality assessment of an ultrasound plane within half a second. Our method achieves promising performance compared with state-of-the-art methods, which can improve the examination effectiveness and alleviate the measurement error caused by improper ultrasound scanning.

Congenital abnormalities of the aortic arch: revisiting the 1964 Stewart classification.

The traditional classification of congenital aortic arch abnormalities was described by James Stewart and colleagues in 1964. Since that time, advances in diagnostic imaging technology have led to better delineation of the vasculature anatomy and the identification of previously unrecognized and unclassified anomalies. In this manuscript, we review the existing literature and propose a series of modifications to the original Stewart classification of congenital aortic arch abnormalities to incorporate this new knowledge. In brief, we propose the following modifications: (1) In Group I, we further divide subgroup B into left arch atretic and right arch atretic; (2) In Group II, we add three more subgroups, including aberrant right innominate artery, "isolated" right innominate artery (RIA), "isolated" right carotid artery with aberrant right subclavian artery; (3) In Groups I, II, and III, we add a subgroup of absence of both ductus arteriosus; and (4) In Group IV, we add three subgroups, including circumflex retro-esophageal aorta arch, persistent V aortic arch, and anomalous origin of pulmonary artery from ascending aorta.

Towards Automated Semantic Segmentation in Prenatal Volumetric Ultrasound.

Volumetric ultrasound is rapidly emerging as a viable imaging modality for routine prenatal examinations. Biometrics obtained from the volumetric segmentation shed light on the reformation of precise maternal and fetal health monitoring. However, the poor image quality, low contrast, boundary ambiguity, and complex anatomy shapes conspire toward a great lack of efficient tools for the segmentation. It makes 3-D ultrasound difficult to interpret and hinders the widespread of 3-D ultrasound in obstetrics. In this paper, we are looking at the problem of semantic segmentation in prenatal ultrasound volumes. Our contribution is threefold: 1) we propose the first and fully automatic framework to simultaneously segment multiple anatomical structures with intensive clinical interest, including fetus, gestational sac, and placenta, which remains a rarely studied and arduous challenge; 2) we propose a composite architecture for dense labeling, in which a customized 3-D fully convolutional network explores spatial intensity concurrency for initial labeling, while a multi-directional recurrent neural network (RNN) encodes spatial sequentiality to combat boundary ambiguity for significant refinement; and 3) we introduce a hierarchical deep supervision mechanism to boost the information flow within RNN and fit the latent sequence hierarchy in fine scales, and further improve the segmentation results. Extensively verified on in-house large data sets, our method illustrates a superior segmentation performance, decent agreements with expert measurements and high reproducibilities against scanning variations, and thus is promising in advancing the prenatal ultrasound examinations.