Generalisability of fetal ultrasound deep learning models to low-resource imaging settings in five African countries.

Most artificial intelligence (AI) research and innovations have concentrated in high-income countries, where imaging data, IT infrastructures and clinical expertise are plentiful. However, slower progress has been made in limited-resource environments where medical imaging is needed. For example, in Sub-Saharan Africa, the rate of perinatal mortality is very high due to limited access to antenatal screening. In these countries, AI models could be implemented to help clinicians acquire fetal ultrasound planes for the diagnosis of fetal abnormalities. So far, deep learning models have been proposed to identify standard fetal planes, but there is no evidence of their ability to generalise in centres with low resources, i.e. with limited access to high-end ultrasound equipment and ultrasound data. This work investigates for the first time different strategies to reduce the domain-shift effect arising from a fetal plane classification model trained on one clinical centre with high-resource settings and transferred to a new centre with low-resource settings. To that end, a classifier trained with 1792 patients from Spain is first evaluated on a new centre in Denmark in optimal conditions with 1008 patients and is later optimised to reach the same performance in five African centres (Egypt, Algeria, Uganda, Ghana and Malawi) with 25 patients each. The results show that a transfer learning approach for domain adaptation can be a solution to integrate small-size African samples with existing large-scale databases in developed countries. In particular, the model can be re-aligned and optimised to boost the performance on African populations by increasing the recall to [Formula: see text] and at the same time maintaining a high precision across centres. This framework shows promise for building new AI models generalisable across clinical centres with limited data acquired in challenging and heterogeneous conditions and calls for further research to develop new solutions for the usability of AI in countries with fewer resources and, consequently, in higher need of clinical support.

Validation of an autonomous artificial intelligence-based diagnostic system for holistic maculopathy screening in a routine occupational health checkup context.

PURPOSE: This study aims to evaluate the ability of an autonomous artificial intelligence (AI) system for detection of the most common central retinal pathologies in fundus photography. METHODS: Retrospective diagnostic test evaluation on a raw dataset of 5918 images (2839 individuals) evaluated with non-mydriatic cameras during routine occupational health checkups. Three camera models were employed: Optomed Aurora (field of view - FOV 50º, 88% of the dataset), ZEISS VISUSCOUT 100 (FOV 40º, 9%), and Optomed SmartScope M5 (FOV 40º, 3%). Image acquisition took 2 min per patient. Ground truth for each image of the dataset was determined by 2 masked retina specialists, and disagreements were resolved by a 3rd retina specialist. The specific pathologies considered for evaluation were "diabetic retinopathy" (DR), "Age-related macular degeneration" (AMD), "glaucomatous optic neuropathy" (GON), and "Nevus." Images with maculopathy signs that did not match the described taxonomy were classified as "Other." RESULTS: The combination of algorithms to detect any abnormalities had an area under the curve (AUC) of 0.963 with a sensitivity of 92.9% and a specificity of 86.8%. The algorithms individually obtained are as follows: AMD AUC 0.980 (sensitivity 93.8%; specificity 95.7%), DR AUC 0.950 (sensitivity 81.1%; specificity 94.8%), GON AUC 0.889 (sensitivity 53.6% specificity 95.7%), Nevus AUC 0.931 (sensitivity 86.7%; specificity 90.7%). CONCLUSION: Our holistic AI approach reaches high diagnostic accuracy at simultaneous detection of DR, AMD, and Nevus. The integration of pathology-specific algorithms permits higher sensitivities with minimal impact on its specificity. It also reduces the risk of missing incidental findings. Deep learning may facilitate wider screenings of eye diseases.

Assessment of Radiomics and Deep Learning for the Segmentation of Fetal and Maternal Anatomy in Magnetic Resonance Imaging and Ultrasound.

Recent advances in fetal imaging open the door to enhanced detection of fetal disorders and computer-assisted surgical planning. However, precise segmentation of womb's tissues is challenging due to motion artifacts caused by fetal movements and maternal respiration during acquisition. This work aims to efficiently segment different intrauterine tissues in fetal magnetic resonance imaging (MRI) and 3D ultrasound (US). First, a large set of ninety-four radiomic features are extracted to characterize the mother uterus, placenta, umbilical cord, fetal lungs, and brain. The optimal features for each anatomy are identified using both K-best and Sequential Forward Feature Selection techniques. These features are then fed to a Support Vector Machine with instance balancing to accurately segment the intrauterine anatomies. To the best of our knowledge, this is the first time that "Radiomics" is expanded from classification tasks to segmentation purposes to deal with challenging fetal images. In addition, we evaluate several state-of-the-art deep learning-based segmentation approaches. Validation is extensively performed on a set of 60 axial MRI and 3D US images from pathological and clinical cases. Our results suggest that combining the selected 10 radiomic features per anatomy along with DeepLabV3+ or BiSeNet architectures for MRI, and PSPNet or Tiramisu for 3D US, can lead to the highest fetal / maternal tissue segmentation performance, robustness, informativeness, and heterogeneity. Therefore, this work opens new avenues for advancement of segmentation techniques and, in particular, for improved fetal surgical planning.

Convexity shape constraints for retinal blood vessel segmentation and foveal avascular zone detection.

Diabetic retinopathy (DR) has become a major worldwide health problem due to the increase in blindness among diabetics at early ages. The detection of DR pathologies such as microaneurysms, hemorrhages and exudates through advanced computational techniques is of utmost importance in patient health care. New computer vision techniques are needed to improve upon traditional screening of color fundus images. The segmentation of the entire anatomical structure of the retina is a crucial phase in detecting these pathologies. This work proposes a novel framework for fast and fully automatic blood vessel segmentation and fovea detection. The preprocessing method involved both contrast limited adaptive histogram equalization and the brightness preserving dynamic fuzzy histogram equalization algorithms to enhance image contrast and eliminate noise artifacts. Afterwards, the color spaces and their intrinsic components were examined to identify the most suitable color model to reveal the foreground pixels against the entire background. Several samples were then collected and used by the renowned convexity shape prior segmentation algorithm. The proposed methodology achieved an average vasculature segmentation accuracy exceeding 96%, 95%, 98% and 94% for the DRIVE, STARE, HRF and Messidor publicly available datasets, respectively. An additional validation step reached an average accuracy of 94.30% using an in-house dataset provided by the Hospital Sant Joan of Reus (Spain). Moreover, an outstanding detection accuracy of over 98% was achieved for the foveal avascular zone. An extensive state-of-the-art comparison was also conducted. The proposed approach can thus be integrated into daily clinical practice to assist medical experts in the diagnosis of DR.

Segmentation of the placenta and its vascular tree in Doppler ultrasound for fetal surgery planning.

PURPOSE: Twin-to-twin transfusion syndrome (TTTS) is a serious condition that occurs in about 10-15% of monochorionic twin pregnancies. In most instances, the blood flow is unevenly distributed throughout the placenta anastomoses leading to the death of both fetuses if no surgical procedure is performed. Fetoscopic laser coagulation is the optimal therapy to considerably improve co-twin prognosis by clogging the abnormal anastomoses. Notwithstanding progress in recent years, TTTS surgery is highly risky. Computer-assisted planning of the intervention can thus improve the outcome. METHODS: In this work, we implement a GPU-accelerated random walker (RW) algorithm to detect the placenta, both umbilical cord insertions and the placental vasculature from Doppler ultrasound (US). Placenta and background seeds are manually initialized in 10-20 slices (out of 245). Vessels are automatically initialized in the same slices by means of Otsu thresholding. The RW finds the boundaries of the placenta and reconstructs the vasculature. RESULTS: We evaluate our semiautomatic method in 5 monochorionic and 24 singleton pregnancies. Although satisfactory performance is achieved on placenta segmentation (Dice ≥ 84.0%), some vascular connections are still neglected due to the presence of US reverberation artifacts (Dice ≥ 56.9%). We also compared inter-user variability and obtained Dice coefficients of ≥ 76.8% and ≥ 97.42% for placenta and vasculature, respectively. After a 3-min manual initialization, our GPU approach speeds the computation 10.6 times compared to the CPU. CONCLUSIONS: Our semiautomatic method provides a near real-time user experience and requires short training without compromising the segmentation accuracy. A powerful approach is thus presented to rapidly plan the fetoscope insertion point ahead of TTTS surgery.

TTTS-STgan: Stacked Generative Adversarial Networks for TTTS Fetal Surgery Planning Based on 3D Ultrasound.

Twin-to-twin transfusion syndrome (TTTS) is characterized by an unbalanced blood transfer through placental abnormal vascular connections. Prenatal ultrasound (US) is the imaging technique to monitor monochorionic pregnancies and diagnose TTTS. Fetoscopic laser photocoagulation is an elective treatment to coagulate placental communications between both twins. To locate the anomalous connections ahead of surgery, preoperative planning is crucial. In this context, we propose a novel multi-task stacked generative adversarial framework to jointly learn synthetic fetal US generation, multi-class segmentation of the placenta, its inner acoustic shadows and peripheral vasculature, and placenta shadowing removal. Specifically, the designed architecture is able to learn anatomical relationships and global US image characteristics. In addition, we also extract for the first time the umbilical cord insertion on the placenta surface from 3D HD-flow US images. The database consisted of 70 US volumes including singleton, mono- and dichorionic twins at 17-37 gestational weeks. Our experiments show that 71.8% of the synthesized US slices were categorized as realistic by clinicians, and that the multi-class segmentation achieved Dice scores of 0.82 ± 0.13, 0.71 ± 0.09, and 0.72 ± 0.09, for placenta, acoustic shadows, and vasculature, respectively. Moreover, fetal surgeons classified 70.2% of our completed placenta shadows as satisfactory texture reconstructions. The umbilical cord was successfully detected on 85.45% of the volumes. The framework developed could be implemented in a TTTS fetal surgery planning software to improve the intrauterine scene understanding and facilitate the location of the optimum fetoscope entry point.

Deep Q-CapsNet Reinforcement Learning Framework for Intrauterine Cavity Segmentation in TTTS Fetal Surgery Planning.

Fetoscopic laser photocoagulation is the most effective treatment for Twin-to-Twin Transfusion Syndrome, a condition affecting twin pregnancies in which there is a deregulation of blood circulation through the placenta, that can be fatal to both babies. For the purposes of surgical planning, we design the first automatic approach to detect and segment the intrauterine cavity from axial, sagittal and coronal MRI stacks. Our methodology relies on the ability of capsule networks to successfully capture the part-whole interdependency of objects in the scene, particularly for unique class instances (i.e., intrauterine cavity). The presented deep Q-CapsNet reinforcement learning framework is built upon a context-adaptive detection policy to generate a bounding box of the womb. A capsule architecture is subsequently designed to segment (or refine) the whole intrauterine cavity. This network is coupled with a strided nnU-Net feature extractor, which encodes discriminative feature maps to construct strong primary capsules. The method is robustly evaluated with and without the localization stage using 13 performance measures, and directly compared with 15 state-of-the-art deep neural networks trained on 71 singleton and monochorionic twin pregnancies. An average Dice score above 0.91 is achieved for all ablations, revealing the potential of our approach to be used in clinical practice.

TTTS-GPS: Patient-specific preoperative planning and simulation platform for twin-to-twin transfusion syndrome fetal surgery.

Twin-to-twin transfusion syndrome (TTTS) is a serious condition that may occur in pregnancies when two or more fetuses share the same placenta. It is characterized by abnormal vascular connections in the placenta that cause blood to flow unevenly between the babies. If left untreated, perinatal mortality occurs in 90% of cases, whilst neurological injuries are still present in TTTS survivors. Minimally invasive fetoscopic laser surgery is the standard and optimal treatment for this condition, but is technically challenging and can lead to complications. Acquiring and maintaining the required surgical skills need consistent practice, and a steep learning curve. An accurate preoperative planning is thus vital for complex TTTS cases. To this end, we propose the first TTTS fetal surgery planning and simulation platform. The soft tissue of the mother, the uterus, the umbilical cords, the placenta and its vascular tree are segmented and registered automatically from magnetic resonance imaging and 3D ultrasound using computer vision and deep learning techniques. The proposed state-of-the-art technology is integrated into a flexible C++ and MITK-based application to provide a full exploration of the intrauterine environment by simulating the fetoscope camera as well as the laser ablation, determining the correct entry point, training doctors' movements and trajectory ahead of operation, which allows improving upon current practice. A comprehensive usability study is reported. Experienced surgeons rated highly our TTTS planner and simulator, thus being a potential tool to be implemented in real and complex TTTS surgeries.

Fully automatic 3D reconstruction of the placenta and its peripheral vasculature in intrauterine fetal MRI.

Recent advances in fetal magnetic resonance imaging (MRI) open the door to improved detection and characterization of fetal and placental abnormalities. Since interpreting MRI data can be complex and ambiguous, there is a need for robust computational methods able to quantify placental anatomy (including its vasculature) and function. In this work, we propose a novel fully-automated method to segment the placenta and its peripheral blood vessels from fetal MRI. First, a super-resolution reconstruction of the uterus is generated by combining axial, sagittal and coronal views. The placenta is then segmented using 3D Gabor filters, texture features and Support Vector Machines. A uterus edge-based instance selection is proposed to identify the support vectors defining the placenta boundary. Subsequently, peripheral blood vessels are extracted through a curvature-based corner detector. Our approach is validated on a rich set of 44 control and pathological cases: singleton and (normal / monochorionic) twin pregnancies between 25-37 weeks of gestation. Dice coefficients of 0.82 â€¯±â€¯ 0.02 and 0.81 â€¯±â€¯ 0.08 are achieved for placenta and its vasculature segmentation, respectively. A comparative analysis with state of the art convolutional neural networks (CNN), namely, 3D U-Net, V-Net, DeepMedic, Holistic3D Net, HighRes3D Net and Dense V-Net is also conducted for placenta localization, with our method outperforming all CNN approaches. Results suggest that our methodology can aid the diagnosis and surgical planning of severe fetal disorders.

Segmentation and classification in MRI and US fetal imaging: Recent trends and future prospects.

Fetal imaging is a burgeoning topic. New advancements in both magnetic resonance imaging and (3D) ultrasound currently allow doctors to diagnose fetal structural abnormalities such as those involved in twin-to-twin transfusion syndrome, gestational diabetes mellitus, pulmonary sequestration and hypoplasia, congenital heart disease, diaphragmatic hernia, ventriculomegaly, etc. Considering the continued breakthroughs in utero image analysis and (3D) reconstruction models, it is now possible to gain more insight into the ongoing development of the fetus. Best prenatal diagnosis performances rely on the conscious preparation of the clinicians in terms of fetal anatomy knowledge. Therefore, fetal imaging will likely span and increase its prevalence in the forthcoming years. This review covers state-of-the-art segmentation and classification methodologies for the whole fetus and, more specifically, the fetal brain, lungs, liver, heart and placenta in magnetic resonance imaging and (3D) ultrasound for the first time. Potential applications of the aforementioned methods into clinical settings are also inspected. Finally, improvements in existing approaches as well as most promising avenues to new areas of research are briefly outlined.

Improving the performance of empirical mode decomposition via Tsallis entropy: Application to Alzheimer EEG analysis.

Alzheimer is a degenerative disorder that attacks neurons, resulting in loss of memory, thinking, language skills, and behavioral changes. Computer-aided detection methods can uncover crucial information recorded by electroencephalograms. A systematic literature search presents the wavelet transform as a frequently used technique in Alzheimer's detection. However, it requires a defined basis function considered a significant problem. In this work, the concept of empirical mode decomposition is introduced as an alternative to process Alzheimer signals. The performance of empirical mode decomposition heavily relies on a parameter called threshold. In our previous works, we found that the existing thresholding techniques were not able to highlight relevant information. The use of Tsallis entropy as a thresholder is evaluated through the combination of empirical mode decomposition and neural networks. Thanks to the extraction of better features that boost the classification accuracy, the proposed approach outperforms the state-of-the-art in terms of peak signal to noise ratio and root mean square error. Hence, our methodology is more likely to succeed than methods based on other landmarks such as Bayes, Normal and Visu shrink. We finally report an accuracy rate of 80%, while the aforementioned techniques only yield performances of 65%, 60% and 40%, respectively.

Fetal MRI Synthesis via Balanced Auto-Encoder Based Generative Adversarial Networks.

Machine learning approaches for image analysis require large amounts of training imaging data. As an alternative, the use of realistic synthetic data reduces the high cost associated to medical image acquisition, as well as avoiding confidentiality and privacy issues, and consequently allows the creation of public data repositories for scientific purposes. Within the context of fetal imaging, we adopt an auto-encoder based Generative Adversarial Network for synthetic fetal MRI generation. The proposed architecture features a balanced power of the discriminator against the generator during training, provides an approximate convergence measure, and enables fast and robust training to generate high-quality fetal MRI in axial, sagittal and coronal planes. We demonstrate the feasibility of the proposed approach quantitatively and qualitatively by segmenting relevant fetal structures to assess the anatomical fidelity of the simulation, and performing a clinical verisimilitude study distinguishing the simulated data from the real images. The results obtained so far are promising, which makes further investigation on this new topic worthwhile.

Fuzzy-entropy threshold based on a complex wavelet denoising technique to diagnose Alzheimer disease.

The presence of irregularities in electroencephalographic (EEG) signals entails complexities during the Alzheimer's disease (AD) diagnosis. In addition, the uncertainty presented on EEG raises major issues in the improvement of the classification rate. The multi-resolution analysis through an optimum threshold will likely achieve better results in distinguishing AD and normal EEG signals. Hence, a fuzzy-entropy concept defined in a complex multi-resolution wavelet has been proposed to obtain the most appropriate threshold. First, the complex coefficients are fuzzified using a Gaussian membership function. Afterwards, the ability of the proposed fuzzy-entropy threshold has been compared with traditional thresholds in complex wavelet domain. Experimental results show that the authors' methodology produces a higher signal-to-noise ratio and a lower root-mean-square error than traditional approaches. Moreover, a neural network scheme is performed along several features to classify AD from normal EEG signals obtaining a specificity of 87.5%.