Learning to map 2D ultrasound images into 3D space with minimal human annotation

Neural implicit surface reconstruction of freehand 3D ultrasound volume with geometric constraints

Three-dimensional (3D) freehand ultrasound (US) is a widely used imaging modality that allows non-invasive imaging of medical anatomy without radiation exposure. Surface reconstruction of US volume is vital to acquire the accurate anatomical structures needed for modeling, registration, and visualization. However, traditional methods cannot produce a high-quality surface due to image noise. Despite improvements in smoothness, continuity, and resolution from deep learning approaches, research on surface reconstruction in freehand 3D US is still limited. This study introduces FUNSR, a self-supervised neural implicit surface reconstruction method to learn signed distance functions (SDFs) from US volumes. In particular, FUNSR iteratively learns the SDFs by moving the 3D queries sampled around volumetric point clouds to approximate the surface, guided by two novel geometric constraints: sign consistency constraint and on-surface constraint with adversarial learning. Our approach has been thoroughly evaluated across four datasets to demonstrate its adaptability to various anatomical structures, including a hip phantom dataset, two vascular datasets and one publicly available prostate dataset. We also show that smooth and continuous representations greatly enhance the visual appearance of US data. Furthermore, we highlight the potential of our method to improve segmentation performance, and its robustness to noise distribution and motion perturbation.

Automatic multi-view pose estimation in focused cardiac ultrasound

Focused cardiac ultrasound (FoCUS) is a valuable point-of-care method for evaluating cardiovascular structures and function, but its scope is limited by equipment and operator's experience, resulting in primarily qualitative 2D exams. This study presents a novel framework to automatically estimate the 3D spatial relationship between standard FoCUS views. The proposed framework uses a multi-view U-Net-like fully convolutional neural network to regress line-based heatmaps representing the most likely areas of intersection between input images. The lines that best fit the regressed heatmaps are then extracted, and a system of nonlinear equations based on the intersection between view triplets is created and solved to determine the relative 3D pose between all input images. The feasibility and accuracy of the proposed pipeline were validated using a novel realistic in silico FoCUS dataset, demonstrating promising results. Interestingly, as shown in preliminary experiments, the estimation of the 2D images' relative poses enables the application of 3D image analysis methods and paves the way for 3D quantitative assessments in FoCUS examinations.

Sensorless volumetric reconstruction of fetal brain freehand ultrasound scans with deep implicit representation

Three-dimensional (3D) ultrasound imaging has contributed to our understanding of fetal developmental processes by providing rich contextual information of the inherently 3D anatomies. However, its use is limited in clinical settings, due to the high purchasing costs and limited diagnostic practicality. Freehand 2D ultrasound imaging, in contrast, is routinely used in standard obstetric exams, but inherently lacks a 3D representation of the anatomies, which limits its potential for more advanced assessment. Such full representations are challenging to recover even with external tracking devices due to internal fetal movement which is independent from the operator-led trajectory of the probe. Capitalizing on the flexibility offered by freehand 2D ultrasound acquisition, we propose ImplicitVol to reconstruct 3D volumes from non-sensor-tracked 2D ultrasound sweeps. Conventionally, reconstructions are performed on a discrete voxel grid. We, however, employ a deep neural network to represent, for the first time, the reconstructed volume as an implicit function. Specifically, ImplicitVol takes a set of 2D images as input, predicts their locations in 3D space, jointly refines the inferred locations, and learns a full volumetric reconstruction. When testing natively-acquired and volume-sampled 2D ultrasound video sequences collected from different manufacturers, the 3D volumes reconstructed by ImplicitVol show significantly better visual and semantic quality than the existing interpolation-based reconstruction approaches. The inherent continuity of implicit representation also enables ImplicitVol to reconstruct the volume to arbitrarily high resolutions. As formulated, ImplicitVol has the potential to integrate seamlessly into the clinical workflow, while providing richer information for diagnosis and evaluation of the developing brain.

Ultrasound Plane Pose Regression: Assessing Generalized Pose Coordinates in the Fetal Brain

In obstetric ultrasound (US) scanning, the learner's ability to mentally build a three-dimensional (3D) map of the fetus from a two-dimensional (2D) US image represents a significant challenge in skill acquisition. We aim to build a US plane localization system for 3D visualization, training, and guidance without integrating additional sensors. This work builds on top of our previous work, which predicts the six-dimensional (6D) pose of arbitrarily oriented US planes slicing the fetal brain with respect to a normalized reference frame using a convolutional neural network (CNN) regression network. Here, we analyze in detail the assumptions of the normalized fetal brain reference frame and quantify its accuracy with respect to the acquisition of transventricular (TV) standard plane (SP) for fetal biometry. We investigate the impact of registration quality in the training and testing data and its subsequent effect on trained models. Finally, we introduce data augmentations and larger training sets that improve the results of our previous work, achieving median errors of 2.97 mm and 6.63° for translation and rotation, respectively.

Semi-Supervised Representation Learning for Segmentation on Medical Volumes and Sequences

Benefiting from the massive labeled samples, deep learning-based segmentation methods have achieved great success for two dimensional natural images. However, it is still a challenging task to segment high dimensional medical volumes and sequences, due to the considerable efforts for clinical expertise to make large scale annotations. Self/semi-supervised learning methods have been shown to improve the performance by exploiting unlabeled data. However, they are still lack of mining local semantic discrimination and exploitation of volume/sequence structures. In this work, we propose a semi-supervised representation learning method with two novel modules to enhance the features in the encoder and decoder, respectively. For the encoder, based on the continuity between slices/frames and the common spatial layout of organs across subjects, we propose an asymmetric network with an attention-guided predictor to enable prediction between feature maps of different slices of unlabeled data. For the decoder, based on the semantic consistency between labeled data and unlabeled data, we introduce a novel semantic contrastive learning to regularize the feature maps in the decoder. The two parts are trained jointly with both labeled and unlabeled volumes/sequences in a semi-supervised manner. When evaluated on three benchmark datasets of medical volumes and sequences, our model outperforms existing methods with a large margin of 7.3% DSC on ACDC, 6.5% on Prostate, and 3.2% on CAMUS when only a few labeled data is available. Further, results on the M&M dataset show that the proposed method yields improvement without using any domain adaption techniques for data from unknown domain. Intensive evaluations reveal the effectiveness of representation mining, and superiority on performance of our method. The code is available at https://github.com/CcchenzJ/BootstrapRepresentation.

Evolving the Era of 5D Ultrasound? A Systematic Literature Review on the Applications for Artificial Intelligence Ultrasound Imaging in Obstetrics and Gynecology

Artificial intelligence (AI) has gained prominence in medical imaging, particularly in obstetrics and gynecology (OB/GYN), where ultrasound (US) is the preferred method. It is considered cost effective and easily accessible but is time consuming and hindered by the need for specialized training. To overcome these limitations, AI models have been proposed for automated plane acquisition, anatomical measurements, and pathology detection. This study aims to overview recent literature on AI applications in OB/GYN US imaging, highlighting their benefits and limitations. For the methodology, a systematic literature search was performed in the PubMed and Cochrane Library databases. Matching abstracts were screened based on the PICOS (Participants, Intervention or Exposure, Comparison, Outcome, Study type) scheme. Articles with full text copies were distributed to the sections of OB/GYN and their research topics. As a result, this review includes 189 articles published from 1994 to 2023. Among these, 148 focus on obstetrics and 41 on gynecology. AI-assisted US applications span fetal biometry, echocardiography, or neurosonography, as well as the identification of adnexal and breast masses, and assessment of the endometrium and pelvic floor. To conclude, the applications for AI-assisted US in OB/GYN are abundant, especially in the subspecialty of obstetrics. However, while most studies focus on common application fields such as fetal biometry, this review outlines emerging and still experimental fields to promote further research.

Long-term Dependency for 3D Reconstruction of Freehand Ultrasound Without External Tracker

OBJECTIVE

Reconstructing freehand ultrasound in 3D without any external tracker has been a long-standing challenge in ultrasound-assisted procedures. We aim to define new ways of parameterising long-term dependencies, and evaluate the performance.

METHODS

First, long-term dependency is encoded by transformation positions within a frame sequence. This is achieved by combining a sequence model with a multi-transformation prediction. Second, two dependency factors are proposed, anatomical image content and scanning protocol, for contributing towards accurate reconstruction. Each factor is quantified experimentally by reducing respective training variances.

RESULTS

1) The added long-term dependency up to 400 frames at 20 frames per second (fps) indeed improved reconstruction, with an up to 82.4% lowered accumulated error, compared with the baseline performance. The improvement was found to be dependent on sequence length, transformation interval and scanning protocol and, unexpectedly, not on the use of recurrent networks with long-short term modules; 2) Decreasing either anatomical or protocol variance in training led to poorer reconstruction accuracy. Interestingly, greater performance was gained from representative protocol patterns, than from representative anatomical features.

CONCLUSION

The proposed algorithm uses hyperparameter tuning to effectively utilise long-term dependency. The proposed dependency factors are of practical significance in collecting diverse training data, regulating scanning protocols and developing efficient networks.

SIGNIFICANCE

The proposed new methodology with publicly available volunteer data and code for parametersing the long-term dependency, experimentally shown to be valid sources of performance improvement, which could potentially lead to better model development and practical optimisation of the reconstruction application.

Ultrasound Frame-to-Volume Registration via Deep Learning for Interventional Guidance

Fusing intraoperative 2-D ultrasound (US) frames with preoperative 3-D magnetic resonance (MR) images for guiding interventions has become the clinical gold standard in image-guided prostate cancer biopsy. However, developing an automatic image registration system for this application is challenging because of the modality gap between US/MR and the dimensionality gap between 2-D/3-D data. To overcome these challenges, we propose a novel US frame-to-volume registration (FVReg) pipeline to bridge the dimensionality gap between 2-D US frames and 3-D US volume. The developed pipeline is implemented using deep neural networks, which are fully automatic without requiring external tracking devices. The framework consists of three major components, including one) a frame-to-frame registration network (Frame2Frame) that estimates the current frame's 3-D spatial position based on previous video context, two) a frame-to-slice correction network (Frame2Slice) adjusting the estimated frame position using the 3-D US volumetric information, and three) a similarity filtering (SF) mechanism selecting the frame with the highest image similarity with the query frame. We validated our method on a clinical dataset with 618 subjects and tested its potential on real-time 2-D-US to 3-D-MR fusion navigation tasks. The proposed FVReg achieved an average target navigation error of 1.93 mm at 5-14 fps. Our source code is publicly available at https://github.com/DIAL-RPI/Frame-to-Volume-Registration.

Prescriptive Method for Optimizing Cost of Data Collection and Annotation in Machine Learning of Clinical Ultrasound

Machine learning in medical ultrasound faces a major challenge: the prohibitive costs of producing and annotating clinical data. Optimizing the data collection and annotation will improve model training efficiency, reducing project cost and times. This paper prescribes a 2-phase method for cost optimization based on iterative accuracy/sample size predictions, and active learning for annotation optimization.

METHODS

Using public breast, fetal, and lung ultrasound datasets we can: Optimize data collection by statistically predicting accuracy for a desired dataset size; and optimize labeling efficiency using Active Learning, where predictions with lowest certainty were labelled manually using feedback. A practical case study on BUSI data was used to demonstrate the method prescribed in this work.

RESULTS

With small data subsets, ~10%, dataset size vs. final accuracy relations can be predicted with diminishing results after 50% usage. Manual annotation was reduced by ~10% using active learning to focus the annotation.

CONCLUSION

This led to cost reductions of 50%-66%, depending on requirements and initial cost model, on BUSI dataset with a negligible accuracy drop of 3.75% from theoretical maximums.Clinical Relevance- This work provides methodology to optimize dataset size and manual data labelling, this allows generation of cost-effective datasets, of interest to all, but particularly for financially limited trials and feasibility studies, Reducing the time burden on annotating clinicians.

Computational methods for the analysis of early-pregnancy brain ultrasonography: a systematic review

BACKGROUND

Early screening of the brain is becoming routine clinical practice. Currently, this screening is performed by manual measurements and visual analysis, which is time-consuming and prone to errors. Computational methods may support this screening. Hence, the aim of this systematic review is to gain insight into future research directions needed to bring automated early-pregnancy ultrasound analysis of the human brain to clinical practice.

METHODS

We searched PubMed (Medline ALL Ovid), EMBASE, Web of Science Core Collection, Cochrane Central Register of Controlled Trials, and Google Scholar, from inception until June 2022. This study is registered in PROSPERO at CRD42020189888. Studies about computational methods for the analysis of human brain ultrasonography acquired before the 20th week of pregnancy were included. The key reported attributes were: level of automation, learning-based or not, the usage of clinical routine data depicting normal and abnormal brain development, public sharing of program source code and data, and analysis of the confounding factors.

FINDINGS

Our search identified 2575 studies, of which 55 were included. 76% used an automatic method, 62% a learning-based method, 45% used clinical routine data and in addition, for 13% the data depicted abnormal development. None of the studies shared publicly the program source code and only two studies shared the data. Finally, 35% did not analyse the influence of confounding factors.

INTERPRETATION

Our review showed an interest in automatic, learning-based methods. To bring these methods to clinical practice we recommend that studies: use routine clinical data depicting both normal and abnormal development, make their dataset and program source code publicly available, and be attentive to the influence of confounding factors. Introduction of automated computational methods for early-pregnancy brain ultrasonography will save valuable time during screening, and ultimately lead to better detection, treatment and prevention of neuro-developmental disorders.

FUNDING

The Erasmus MC Medical Research Advisor Committee (grant number: FB 379283).

Automatic measurements of fetal intracranial volume from 3D ultrasound scans

Three-dimensional fetal ultrasound is commonly used to study the volumetric development of brain structures. To date, only a limited number of automatic procedures for delineating the intracranial volume exist. Hence, intracranial volume measurements from three-dimensional ultrasound images are predominantly performed manually. Here, we present and validate an automated tool to extract the intracranial volume from three-dimensional fetal ultrasound scans. The procedure is based on the registration of a brain model to a subject brain. The intracranial volume of the subject is measured by applying the inverse of the final transformation to an intracranial mask of the brain model. The automatic measurements showed a high correlation with manual delineation of the same subjects at two gestational ages, namely, around 20 and 30 weeks (linear fitting R2(20 weeks) = 0.88, R2(30 weeks) = 0.77; Intraclass Correlation Coefficients: 20 weeks=0.94, 30 weeks = 0.84). Overall, the automatic intracranial volumes were larger than the manually delineated ones (84 ± 16 vs. 76 ± 15 cm3; and 274 ± 35 vs. 237 ± 28 cm3), probably due to differences in cerebellum delineation. Notably, the automated measurements reproduced both the non-linear pattern of fetal brain growth and the increased inter-subject variability for older fetuses. By contrast, there was some disagreement between the manual and automatic delineation concerning the size of sexual dimorphism differences. The method presented here provides a relatively efficient way to delineate volumes of fetal brain structures like the intracranial volume automatically. It can be used as a research tool to investigate these structures in large cohorts, which will ultimately aid in understanding fetal structural human brain development.

SVoRT: Iterative Transformer for Slice-to-Volume Registration in Fetal Brain MRI

Volumetric reconstruction of fetal brains from multiple stacks of MR slices, acquired in the presence of almost unpredictable and often severe subject motion, is a challenging task that is highly sensitive to the initialization of slice-to-volume transformations. We propose a novel slice-to-volume registration method using Transformers trained on synthetically transformed data, which model multiple stacks of MR slices as a sequence. With the attention mechanism, our model automatically detects the relevance between slices and predicts the transformation of one slice using information from other slices. We also estimate the underlying 3D volume to assist slice-to-volume registration and update the volume and transformations alternately to improve accuracy. Results on synthetic data show that our method achieves lower registration error and better reconstruction quality compared with existing state-of-the-art methods. Experiments with real-world MRI data are also performed to demonstrate the ability of the proposed model to improve the quality of 3D reconstruction under severe fetal motion.

A guiding approach of Ultrasound scan for accurately obtaining standard diagnostic planes of fetal brain malformation

BACKGROUND

Standard planes (SPs) are crucial for the diagnosis of fetal brain malformation. However, it is very time-consuming and requires extensive experiences to acquire the SPs accurately due to the large difference in fetal posture and the complexity of SPs definitions.

OBJECTIVE

This study aims to present a guiding approach that could assist sonographer to obtain the SPs more accurately and more quickly.

METHODS

To begin with, sonographer uses the 3D probe to scan the fetal head to obtain 3D volume data, and then we used affine transformation to calibrate 3D volume data to the standard body position and established the corresponding 3D head model in 'real time'. When the sonographer uses the 2D probe to scan a plane, the position of current plane can be clearly show in 3D head model by our RLNet (regression location network), which can conduct the sonographer to obtain the three SPs more accurately. When the three SPs are located, the sagittal plane and the coronal planes can be automatically generated according to the spatial relationship with the three SPs.

RESULTS

Experimental results conducted on 3200 2D US images show that the RLNet achieves average angle error of the transthalamic plane was 3.91±2.86°, which has a obvious improvement compared other published data. The automatically generated coronal and sagittal SPs conform the diagnostic criteria and the diagnostic requirements of fetal brain malformation.

CONCLUSIONS

A guiding scanning method based deep learning for ultrasonic brain malformation screening is firstly proposed and it has a pragmatic value for future clinical application.