Search

Clinical adoption of deep learning models has been hindered, in part, because the black-box nature of neural networks leads to concerns regarding their trustworthiness and reliability. These concerns are particularly relevant in the field of neuroimaging due to the complex brain phenotypes and inter-subject heterogeneity often encountered. The challenge can be addressed by interpretable deep learning (iDL) methods that enable the visualisation and interpretation of the inner workings of deep learning models. This study systematically reviewed the literature on neuroimaging applications of iDL methods and critically analysed how iDL explanation properties were evaluated. Seventy-five studies were included, and ten categories of iDL methods were identified. We also reviewed five properties of iDL explanations that were analysed in the included studies: biological validity, robustness, continuity, selectivity, and downstream task performance. We found that the most popular iDL approaches used in the literature may be sub-optimal for neuroimaging data, and we discussed possible future directions for the field.

Congenital Heart Disease (CHD) is a group of cardiac malformations present already during fetal life, representing the prevailing category of birth defects globally. Our aim in this study is to aid 3D fetal vessel topology visualisation in aortic arch anomalies, a group which encompasses a range of conditions with significant anatomical heterogeneity. We present a multi-task framework for automated multi-class fetal vessel segmentation from 3D black blood T2w MRI and anomaly classification. Our training data consists of binary manual segmentation masks of the cardiac vessels' region in individual subjects and fully-labelled anomaly-specific population atlases. Our framework combines deep learning label propagation using VoxelMorph with 3D Attention U-Net segmentation and DenseNet121 anomaly classification. We target 11 cardiac vessels and three distinct aortic arch anomalies, including double aortic arch, right aortic arch, and suspected coarctation of the aorta. We incorporate an anomaly classifier into our segmentation pipeline, delivering a multi-task framework with the primary motivation of correcting topological inaccuracies of the segmentation. The hypothesis is that the multi-task approach will encourage the segmenter network to learn anomaly-specific features. As a secondary motivation, an automated diagnosis tool may have the potential to enhance diagnostic confidence in a decision support setting. Our results showcase that our proposed training strategy significantly outperforms label propagation and a network trained exclusively on propagated labels. Our classifier outperforms a classifier trained exclusively on T2w volume images, with an average balanced accuracy of 0.99 (0.01) after joint training. Adding a classifier improves the anatomical and topological accuracy of all correctly classified double aortic arch subjects., Comment: Accepted for publication at the Journal of Machine Learning for Biomedical Imaging (MELBA) https://melba-journal.org/2023:015

Fetal Magnetic Resonance Imaging at low field strengths is emerging as an exciting direction in perinatal health. Clinical low field (0.55T) scanners are beneficial for fetal imaging due to their reduced susceptibility-induced artefacts, increased T2* values, and wider bore (widening access for the increasingly obese pregnant population). However, the lack of standard automated image processing tools such as segmentation and reconstruction hampers wider clinical use. In this study, we introduce a semi-automatic pipeline using quantitative MRI for the fetal body at low field strength resulting in fast and detailed quantitative T2* relaxometry analysis of all major fetal body organs. Multi-echo dynamic sequences of the fetal body were acquired and reconstructed into a single high-resolution volume using deformable slice-to-volume reconstruction, generating both structural and quantitative T2* 3D volumes. A neural network trained using a semi-supervised approach was created to automatically segment these fetal body 3D volumes into ten different organs (resulting in dice values > 0.74 for 8 out of 10 organs). The T2* values revealed a strong relationship with GA in the lungs, liver, and kidney parenchyma (R^2>0.5). This pipeline was used successfully for a wide range of GAs (17-40 weeks), and is robust to motion artefacts. Low field fetal MRI can be used to perform advanced MRI analysis, and is a viable option for clinical scanning., Comment: Accepted by MICCAI 2023

Magnetic resonance imaging of whole fetal body and placenta is limited by different sources of motion affecting the womb. Usual scanning techniques employ single-shot multi-slice sequences where anatomical information in different slices may be subject to different deformations, contrast variations or artifacts. Volumetric reconstruction formulations have been proposed to correct for these factors, but they must accommodate a non-homogeneous and non-isotropic sampling, so regularization becomes necessary. Thus, in this paper we propose a deep generative prior for robust volumetric reconstructions integrated with a diffeomorphic volume to slice registration method. Experiments are performed to validate our contributions and compare with a state of the art method in a cohort of $72$ fetal datasets in the range of $20-36$ weeks gestational age. Results suggest improved image resolution and more accurate prediction of gestational age at scan when comparing to a state of the art reconstruction method. In addition, gestational age prediction results from our volumetric reconstructions compare favourably with existing brain-based approaches, with boosted accuracy when integrating information of organs other than the brain. Namely, a mean absolute error of $0.618$ weeks ($R^2=0.958$) is achieved when combining fetal brain and trunk information., Comment: 23 pages, 15 figures, 1 table

Tracking microsctructural changes in the developing brain relies on accurate inter-subject image registration. However, most methods rely on either structural or diffusion data to learn the spatial correspondences between two or more images, without taking into account the complementary information provided by using both. Here we propose a deep learning registration framework which combines the structural information provided by T2-weighted (T2w) images with the rich microstructural information offered by diffusion tensor imaging (DTI) scans. We perform a leave-one-out cross-validation study where we compare the performance of our multi-modality registration model with a baseline model trained on structural data only, in terms of Dice scores and differences in fractional anisotropy (FA) maps. Our results show that in terms of average Dice scores our model performs better in subcortical regions when compared to using structural data only. Moreover, average sum-of-squared differences between warped and fixed FA maps show that our proposed model performs better at aligning the diffusion data.

The use of convolutional neural networks (CNNs) for classification tasks has become dominant in various medical imaging applications. At the same time, recent advances in interpretable machine learning techniques have shown great potential in explaining classifiers' decisions. Layer-wise relevance propagation (LRP) has been introduced as one of these novel methods that aim to provide visual interpretation for the network's decisions. In this work we propose the application of 3D CNNs with LRP for the first time for neonatal T2-weighted magnetic resonance imaging (MRI) data analysis. Through LRP, the decisions of our trained classifier are transformed into heatmaps indicating each voxel's relevance for the outcome of the decision. Our resulting LRP heatmaps reveal anatomically plausible features in distinguishing preterm neonates from term ones.

Accurately estimating and correcting the motion artifacts are crucial for 3D image reconstruction of the abdominal and in-utero magnetic resonance imaging (MRI). The state-of-art methods are based on slice-to-volume registration (SVR) where multiple 2D image stacks are acquired in three orthogonal orientations. In this work, we present a novel reconstruction pipeline that only needs one orientation of 2D MRI scans and can reconstruct the full high-resolution image without masking or registration steps. The framework consists of two main stages: the respiratory motion estimation using a self-supervised recurrent neural network, which learns the respiratory signals that are naturally embedded in the asymmetry relationship of the neighborhood slices and cluster them according to a respiratory state. Then, we train a 3D deconvolutional network for super-resolution (SR) reconstruction of the sparsely selected 2D images using integrated reconstruction and total variation loss. We evaluate the classification accuracy on 5 simulated images and compare our results with the SVR method in adult abdominal and in-utero MRI scans. The results show that the proposed pipeline can accurately estimate the respiratory state and reconstruct 4D SR volumes with better or similar performance to the 3D SVR pipeline with less than 20\% sparsely selected slices. The method has great potential to transform the 4D abdominal and in-utero MRI in clinical practice., Comment: Accepted by MICCAI 2019 workshop on Machine Learning for Medical Image Reconstruction

In in-utero MRI, motion correction for fetal body and placenta poses a particular challenge due to the presence of local non-rigid transformations of organs caused by bending and stretching. The existing slice-to-volume registration (SVR) reconstruction methods are widely employed for motion correction of fetal brain that undergoes only rigid transformation. However, for reconstruction of fetal body and placenta, rigid registration cannot resolve the issue of misregistrations due to deformable motion, resulting in degradation of features in the reconstructed volume. We propose a Deformable SVR (DSVR), a novel approach for non-rigid motion correction of fetal MRI based on a hierarchical deformable SVR scheme to allow high resolution reconstruction of the fetal body and placenta. Additionally, a robust scheme for structure-based rejection of outliers minimises the impact of registration errors. The improved performance of DSVR in comparison to SVR and patch-to-volume registration (PVR) methods is quantitatively demonstrated in simulated experiments and 20 fetal MRI datasets from 28-31 weeks gestational age (GA) range with varying degree of motion corruption. In addition, we present qualitative evaluation of 100 fetal body cases from 20-34 weeks GA range.

In utero diffusion MRI provides unique opportunities to non-invasively study the microstructure of tissue during fetal development. A wide range of developmental processes, such as the growth of white matter tracts in the brain, the maturation of placental villous trees, or the fibres in the fetal heart remain to be studied and understood in detail. Advances in fetal interventions and surgery furthermore increase the need for ever more precise antenatal diagnosis from fetal MRI. However, the specific properties of the in utero environment, such as fetal and maternal motion, increased field-of-view, tissue interfaces and safety considerations, are significant challenges for most MRI techniques, and particularly for diffusion. Recent years have seen major improvements, driven by the development of bespoke techniques adapted to these specific challenges in both acquisition and processing. Fetal diffusion MRI, an emerging research tool, is now adding valuable novel information for both research and clinical questions. This paper will highlight specific challenges, outline strategies to target them, and discuss two main applications: fetal brain connectomics and placental maturation.

Diffusion MRI offers a unique probe into neural microstructure and connectivity in the developing brain. However, analysis of neonatal brain imaging data is complicated by inevitable subject motion, leading to a series of scattered slices that need to be aligned within and across diffusion-weighted contrasts. Here, we develop a reconstruction method for scattered slice multi-shell high angular resolution diffusion imaging (HARDI) data, jointly estimating an uncorrupted data representation and motion parameters at the slice or multiband excitation level. The reconstruction relies on data-driven representation of multi-shell HARDI data using a bespoke spherical harmonics and radial decomposition (SHARD), which avoids imposing model assumptions, thus facilitating to compare various microstructure imaging methods in the reconstructed output. Furthermore, the proposed framework integrates slice-level outlier rejection, distortion correction, and slice profile correction. We evaluate the method in the neonatal cohort of the developing Human Connectome Project (650 scans). Validation experiments demonstrate accurate slice-level motion correction across the age range and across the range of motion in the population. Results in the neonatal data show successful reconstruction even in severely motion-corrupted subjects. In addition, we illustrate how local tissue modelling can extract advanced microstructure features such as orientation distribution functions from the motion-corrected reconstructions., Comment: revision. 15 pages, 10 figures, 2 tables

Purpose: To develop a MRI acquisition and reconstruction framework for volumetric cine visualisation of the fetal heart and great vessels in the presence of maternal and fetal motion. Methods: Four-dimensional depiction was achieved using a highly-accelerated multi-planar real-time balanced steady state free precession acquisition combined with retrospective image-domain techniques for motion correction, cardiac synchronisation and outlier rejection. The framework was evaluated and optimised using a numerical phantom, and evaluated in a study of 20 mid- to late-gestational age human fetal subjects. Reconstructed cine volumes were evaluated by experienced cardiologists and compared with matched ultrasound. A preliminary assessment of flow-sensitive reconstruction using the velocity information encoded in the phase of dynamic images is included. Results: Reconstructed cine volumes could be visualised in any 2D plane without the need for highly-specific scan plane prescription prior to acquisition or for maternal breath hold to minimise motion. Reconstruction was fully automated aside from user-specified masks of the fetal heart and chest. The framework proved robust when applied to fetal data and simulations confirmed that spatial and temporal features could be reliably recovered. Expert evaluation suggested the reconstructed volumes can be used for comprehensive assessment of the fetal heart, either as an adjunct to ultrasound or in combination with other MRI techniques. Conclusion: The proposed methods show promise as a framework for motion-compensated 4D assessment of the fetal heart and great vessels.

Diffusion-weighted MRI measures the direction and scale of the local diffusion process in every voxel through its spectrum in q-space, typically acquired in one or more shells. Recent developments in microstructure imaging and multi-tissue decomposition have sparked renewed attention in the radial b-value dependence of the signal. Applications in motion correction and outlier rejection therefore require a compact linear signal representation that extends over the radial as well as angular domain. Here, we introduce SHARD, a data-driven representation of the q-space signal based on spherical harmonics and a radial decomposition into orthonormal components. This representation provides a complete, orthogonal signal basis, tailored to the spherical geometry of q-space and calibrated to the data at hand. We demonstrate that the rank-reduced decomposition outperforms model-based alternatives in human brain data, whilst faithfully capturing the micro- and meso-structural information in the signal. Furthermore, we validate the potential of joint radial-spherical as compared to single-shell representations. As such, SHARD is optimally suited for applications that require low-rank signal predictions, such as motion correction and outlier rejection. Finally, we illustrate its application for the latter using outlier robust regression., Comment: Accepted author manuscript, including supplementary materials