Contrast-enhanced to non-contrast-enhanced image translation to exploit a clinical data warehouse of T1-weighted brain MRI.

BACKGROUND: Clinical data warehouses provide access to massive amounts of medical images, but these images are often heterogeneous. They can for instance include images acquired both with or without the injection of a gadolinium-based contrast agent. Harmonizing such data sets is thus fundamental to guarantee unbiased results, for example when performing differential diagnosis. Furthermore, classical neuroimaging software tools for feature extraction are typically applied only to images without gadolinium. The objective of this work is to evaluate how image translation can be useful to exploit a highly heterogeneous data set containing both contrast-enhanced and non-contrast-enhanced images from a clinical data warehouse. METHODS: We propose and compare different 3D U-Net and conditional GAN models to convert contrast-enhanced T1-weighted (T1ce) into non-contrast-enhanced (T1nce) brain MRI. These models were trained using 230 image pairs and tested on 77 image pairs from the clinical data warehouse of the Greater Paris area. RESULTS: Validation using standard image similarity measures demonstrated that the similarity between real and synthetic T1nce images was higher than between real T1nce and T1ce images for all the models compared. The best performing models were further validated on a segmentation task. We showed that tissue volumes extracted from synthetic T1nce images were closer to those of real T1nce images than volumes extracted from T1ce images. CONCLUSION: We showed that deep learning models initially developed with research quality data could synthesize T1nce from T1ce images of clinical quality and that reliable features could be extracted from the synthetic images, thus demonstrating the ability of such methods to help exploit a data set coming from a clinical data warehouse.

Automatic motion artefact detection in brain T1-weighted magnetic resonance images from a clinical data warehouse using synthetic data.

Containing the medical data of millions of patients, clinical data warehouses (CDWs) represent a great opportunity to develop computational tools. Magnetic resonance images (MRIs) are particularly sensitive to patient movements during image acquisition, which will result in artefacts (blurring, ghosting and ringing) in the reconstructed image. As a result, a significant number of MRIs in CDWs are corrupted by these artefacts and may be unusable. Since their manual detection is impossible due to the large number of scans, it is necessary to develop tools to automatically exclude (or at least identify) images with motion in order to fully exploit CDWs. In this paper, we propose a novel transfer learning method from research to clinical data for the automatic detection of motion in 3D T1-weighted brain MRI. The method consists of two steps: a pre-training on research data using synthetic motion, followed by a fine-tuning step to generalise our pre-trained model to clinical data, relying on the labelling of 4045 images. The objectives were both (1) to be able to exclude images with severe motion, (2) to detect mild motion artefacts. Our approach achieved excellent accuracy for the first objective with a balanced accuracy nearly similar to that of the annotators (balanced accuracy>80 %). However, for the second objective, the performance was weaker and substantially lower than that of human raters. Overall, our framework will be useful to take advantage of CDWs in medical imaging and highlight the importance of a clinical validation of models trained on research data.

Evaluation of MRI-based machine learning approaches for computer-aided diagnosis of dementia in a clinical data warehouse.

A variety of algorithms have been proposed for computer-aided diagnosis of dementia from anatomical brain MRI. These approaches achieve high accuracy when applied to research data sets but their performance on real-life clinical routine data has not been evaluated yet. The aim of this work was to study the performance of such approaches on clinical routine data, based on a hospital data warehouse, and to compare the results to those obtained on a research data set. The clinical data set was extracted from the hospital data warehouse of the Greater Paris area, which includes 39 different hospitals. The research set was composed of data from the Alzheimer's Disease Neuroimaging Initiative data set. In the clinical set, the population of interest was identified by exploiting the diagnostic codes from the 10th revision of the International Classification of Diseases that are assigned to each patient. We studied how the imbalance of the training sets, in terms of contrast agent injection and image quality, may bias the results. We demonstrated that computer-aided diagnosis performance was strongly biased upwards (over 17 percent points of balanced accuracy) by the confounders of image quality and contrast agent injection, a phenomenon known as the Clever Hans effect or shortcut learning. When these biases were removed, the performance was very poor. In any case, the performance was considerably lower than on the research data set. Our study highlights that there are still considerable challenges for translating dementia computer-aided diagnosis systems to clinical routine.

Automatic quality control of brain T1-weighted magnetic resonance images for a clinical data warehouse.

Many studies on machine learning (ML) for computer-aided diagnosis have so far been mostly restricted to high-quality research data. Clinical data warehouses, gathering routine examinations from hospitals, offer great promises for training and validation of ML models in a realistic setting. However, the use of such clinical data warehouses requires quality control (QC) tools. Visual QC by experts is time-consuming and does not scale to large datasets. In this paper, we propose a convolutional neural network (CNN) for the automatic QC of 3D T1-weighted brain MRI for a large heterogeneous clinical data warehouse. To that purpose, we used the data warehouse of the hospitals of the Greater Paris area (Assistance Publique-Hôpitaux de Paris [AP-HP]). Specifically, the objectives were: 1) to identify images which are not proper T1-weighted brain MRIs; 2) to identify acquisitions for which gadolinium was injected; 3) to rate the overall image quality. We used 5000 images for training and validation and a separate set of 500 images for testing. In order to train/validate the CNN, the data were annotated by two trained raters according to a visual QC protocol that we specifically designed for application in the setting of a data warehouse. For objectives 1 and 2, our approach achieved excellent accuracy (balanced accuracy and F1-score >90%), similar to the human raters. For objective 3, the performance was good but substantially lower than that of human raters. Nevertheless, the automatic approach accurately identified (balanced accuracy and F1-score >80%) low quality images, which would typically need to be excluded. Overall, our approach shall be useful for exploiting hospital data warehouses in medical image computing.

Primary progressive aphasias associated with C9orf72 expansions: Another side of the story.

C9orf72 repeat expansions are rarely associated with primary progressive aphasias (PPA). In-depth characterization of the linguistic deficits, and the underlying patterns of grey-matter atrophy in PPA associated with the C9orf72 expansions (PPA-C9orf72) are currently lacking. In this study, we comprehensively analyzed a unique series of 16 patients affected by PPA-C9orf72. Eleven patients were issued from two independent French and Finnish cohorts, and five were identified by means of literature review. Voxel-based morphometry (VBM) studies were performed on three of them. This study depicts the spectrum of C9orf72-related aphasic phenotypes, and illustrates their linguistic presentation. The non-fluent/agrammatic variant was the most frequent phenotype in our series (9/16 patients, 56%), with apraxia of speech being the main defining feature. Left frontal lobe atrophy was present in these subjects, peaking in inferior frontal gyrus. Three patients (19%) showed the semantic variant, with progression of atrophy in temporo-polar regions, later involving orbitofrontal cortex. Anterior temporal lobe dysfunction was also particularly relevant in two patients (12.5%) with mixed forms of PPA. Lastly, two patients (12.5%) had unclassifiable PPA with predominating word-finding difficulties. No PPA-C9orf72 patients in our series fulfilled the criteria of the logopenic variant. Importantly, this study underlines the role of C9orf72 mutation in the disruption of the most anterior parts of the language network, including prefrontal and temporo-polar areas. It provides guidelines for C9orf72 testing in PPA patients, with important clinical impact as gene-specific therapies are upcoming.

Clinica: An Open-Source Software Platform for Reproducible Clinical Neuroscience Studies.

We present Clinica (www.clinica.run), an open-source software platform designed to make clinical neuroscience studies easier and more reproducible. Clinica aims for researchers to (i) spend less time on data management and processing, (ii) perform reproducible evaluations of their methods, and (iii) easily share data and results within their institution and with external collaborators. The core of Clinica is a set of automatic pipelines for processing and analysis of multimodal neuroimaging data (currently, T1-weighted MRI, diffusion MRI, and PET data), as well as tools for statistics, machine learning, and deep learning. It relies on the brain imaging data structure (BIDS) for the organization of raw neuroimaging datasets and on established tools written by the community to build its pipelines. It also provides converters of public neuroimaging datasets to BIDS (currently ADNI, AIBL, OASIS, and NIFD). Processed data include image-valued scalar fields (e.g., tissue probability maps), meshes, surface-based scalar fields (e.g., cortical thickness maps), or scalar outputs (e.g., regional averages). These data follow the ClinicA Processed Structure (CAPS) format which shares the same philosophy as BIDS. Consistent organization of raw and processed neuroimaging files facilitates the execution of single pipelines and of sequences of pipelines, as well as the integration of processed data into statistics or machine learning frameworks. The target audience of Clinica is neuroscientists or clinicians conducting clinical neuroscience studies involving multimodal imaging, and researchers developing advanced machine learning algorithms applied to neuroimaging data.

Primary Progressive Aphasia Associated With GRN Mutations: New Insights Into the Nonamyloid Logopenic Variant.

OBJECTIVE: To determine relative frequencies and linguistic profiles of primary progressive aphasia (PPA) variants associated with GRN (progranulin) mutations and to study their neuroanatomic correlates. METHODS: Patients with PPA carrying GRN mutations (PPA-GRN) were selected among a national prospective research cohort of 1,696 patients with frontotemporal dementia, including 235 patients with PPA. All patients with amyloid-positive CSF biomarkers were excluded. In this cross-sectional study, speech/language and cognitive profiles were characterized with standardized evaluations, and gray matter (GM) atrophy patterns using voxel-based morphometry. Comparisons were performed with controls and patients with sporadic PPA. RESULTS: Among the 235 patients with PPA, 45 (19%) carried GRN mutations, and we studied 32 of these. We showed that logopenic PPA (lvPPA) was the most frequent linguistic variant (n = 13, 41%), followed by nonfluent/agrammatic (nfvPPA; n = 9, 28%) and mixed forms (n = 8, 25%). Semantic variant was rather rare (n = 2, 6%). Patients with lvPPA, qualified as nonamyloid lvPPA, presented canonical logopenic deficit. Seven of 13 had a pure form; 6 showed subtle additional linguistic deficits not fitting criteria for mixed PPA and hence were labeled as logopenic-spectrum variant. GM atrophy involved primarily left posterior temporal gyrus, mirroring neuroanatomic changes of amyloid-positive-lvPPA. Patients with nfvPPA presented agrammatism (89%) rather than apraxia of speech (11%). CONCLUSIONS: This study shows that the most frequent PPA variant associated with GRN mutations is nonamyloid lvPPA, preceding nfvPPA and mixed forms, and illustrates that the language network may be affected at different levels. GRN testing is indicated for patients with PPA, whether familial or sporadic. This finding is important for upcoming GRN gene-specific therapies.

AD Course Map charts Alzheimer's disease progression.

Alzheimer's disease (AD) is characterized by the progressive alterations seen in brain images which give rise to the onset of various sets of symptoms. The variability in the dynamics of changes in both brain images and cognitive impairments remains poorly understood. This paper introduces AD Course Map a spatiotemporal atlas of Alzheimer's disease progression. It summarizes the variability in the progression of a series of neuropsychological assessments, the propagation of hypometabolism and cortical thinning across brain regions and the deformation of the shape of the hippocampus. The analysis of these variations highlights strong genetic determinants for the progression, like possible compensatory mechanisms at play during disease progression. AD Course Map also predicts the patient's cognitive decline with a better accuracy than the 56 methods benchmarked in the open challenge TADPOLE. Finally, AD Course Map is used to simulate cohorts of virtual patients developing Alzheimer's disease. AD Course Map offers therefore new tools for exploring the progression of AD and personalizing patients care.

Deep learning for brain disorders: from data processing to disease treatment.

In order to reach precision medicine and improve patients' quality of life, machine learning is increasingly used in medicine. Brain disorders are often complex and heterogeneous, and several modalities such as demographic, clinical, imaging, genetics and environmental data have been studied to improve their understanding. Deep learning, a subpart of machine learning, provides complex algorithms that can learn from such various data. It has become state of the art in numerous fields, including computer vision and natural language processing, and is also growingly applied in medicine. In this article, we review the use of deep learning for brain disorders. More specifically, we identify the main applications, the concerned disorders and the types of architectures and data used. Finally, we provide guidelines to bridge the gap between research studies and clinical routine.

Reproducible Evaluation of Diffusion MRI Features for Automatic Classification of Patients with Alzheimer's Disease.

Diffusion MRI is the modality of choice to study alterations of white matter. In past years, various works have used diffusion MRI for automatic classification of Alzheimer's disease. However, classification performance obtained with different approaches is difficult to compare because of variations in components such as input data, participant selection, image preprocessing, feature extraction, feature rescaling (FR), feature selection (FS) and cross-validation (CV) procedures. Moreover, these studies are also difficult to reproduce because these different components are not readily available. In a previous work (Samper-González et al. 2018), we propose an open-source framework for the reproducible evaluation of AD classification from T1-weighted (T1w) MRI and PET data. In the present paper, we first extend this framework to diffusion MRI data. Specifically, we add: conversion of diffusion MRI ADNI data into the BIDS standard and pipelines for diffusion MRI preprocessing and feature extraction. We then apply the framework to compare different components. First, FS has a positive impact on classification results: highest balanced accuracy (BA) improved from 0.76 to 0.82 for task CN vs AD. Secondly, voxel-wise features generally gives better performance than regional features. Fractional anisotropy (FA) and mean diffusivity (MD) provided comparable results for voxel-wise features. Moreover, we observe that the poor performance obtained in tasks involving MCI were potentially caused by the small data samples, rather than by the data imbalance. Furthermore, no extensive classification difference exists for different degree of smoothing and registration methods. Besides, we demonstrate that using non-nested validation of FS leads to unreliable and over-optimistic results: 5% up to 40% relative increase in BA. Lastly, with proper FR and FS, the performance of diffusion MRI features is comparable to that of T1w MRI. All the code of the framework and the experiments are publicly available: general-purpose tools have been integrated into the Clinica software package ( www.clinica.run ) and the paper-specific code is available at: https://github.com/aramis-lab/AD-ML .

Gaussian Graphical Model Exploration and Selection in High Dimension Low Sample Size Setting.

Gaussian graphical models (GGM) are often used to describe the conditional correlations between the components of a random vector. In this article, we compare two families of GGM inference methods: the nodewise approach and the penalised likelihood maximisation. We demonstrate on synthetic data that, when the sample size is small, the two methods produce graphs with either too few or too many edges when compared to the real one. As a result, we propose a composite procedure that explores a family of graphs with a nodewise numerical scheme and selects a candidate among them with an overall likelihood criterion. We demonstrate that, when the number of observations is small, this selection method yields graphs closer to the truth and corresponding to distributions with better KL divergence with regards to the real distribution than the other two. Finally, we show the interest of our algorithm on two concrete cases: first on brain imaging data, then on biological nephrology data. In both cases our results are more in line with current knowledge in each field.

Predicting the progression of mild cognitive impairment using machine learning: A systematic, quantitative and critical review.

We performed a systematic review of studies focusing on the automatic prediction of the progression of mild cognitive impairment to Alzheimer's disease (AD) dementia, and a quantitative analysis of the methodological choices impacting performance. This review included 172 articles, from which 234 experiments were extracted. For each of them, we reported the used data set, the feature types, the algorithm type, performance and potential methodological issues. The impact of these characteristics on the performance was evaluated using a multivariate mixed effect linear regressions. We found that using cognitive, fluorodeoxyglucose-positron emission tomography or potentially electroencephalography and magnetoencephalography variables significantly improved predictive performance compared to not including them, whereas including other modalities, in particular T1 magnetic resonance imaging, did not show a significant effect. The good performance of cognitive assessments questions the wide use of imaging for predicting the progression to AD and advocates for exploring further fine domain-specific cognitive assessments. We also identified several methodological issues, including the absence of a test set, or its use for feature selection or parameter tuning in nearly a fourth of the papers. Other issues, found in 15% of the studies, cast doubts on the relevance of the method to clinical practice. We also highlight that short-term predictions are likely not to be better than predicting that subjects stay stable over time. These issues highlight the importance of adhering to good practices for the use of machine learning as a decision support system for the clinical practice.

Convolutional neural networks for classification of Alzheimer's disease: Overview and reproducible evaluation.

Numerous machine learning (ML) approaches have been proposed for automatic classification of Alzheimer's disease (AD) from brain imaging data. In particular, over 30 papers have proposed to use convolutional neural networks (CNN) for AD classification from anatomical MRI. However, the classification performance is difficult to compare across studies due to variations in components such as participant selection, image preprocessing or validation procedure. Moreover, these studies are hardly reproducible because their frameworks are not publicly accessible and because implementation details are lacking. Lastly, some of these papers may report a biased performance due to inadequate or unclear validation or model selection procedures. In the present work, we aim to address these limitations through three main contributions. First, we performed a systematic literature review. We identified four main types of approaches: i) 2D slice-level, ii) 3D patch-level, iii) ROI-based and iv) 3D subject-level CNN. Moreover, we found that more than half of the surveyed papers may have suffered from data leakage and thus reported biased performance. Our second contribution is the extension of our open-source framework for classification of AD using CNN and T1-weighted MRI. The framework comprises previously developed tools to automatically convert ADNI, AIBL and OASIS data into the BIDS standard, and a modular set of image preprocessing procedures, classification architectures and evaluation procedures dedicated to deep learning. Finally, we used this framework to rigorously compare different CNN architectures. The data was split into training/validation/test sets at the very beginning and only the training/validation sets were used for model selection. To avoid any overfitting, the test sets were left untouched until the end of the peer-review process. Overall, the different 3D approaches (3D-subject, 3D-ROI, 3D-patch) achieved similar performances while that of the 2D slice approach was lower. Of note, the different CNN approaches did not perform better than a SVM with voxel-based features. The different approaches generalized well to similar populations but not to datasets with different inclusion criteria or demographical characteristics. All the code of the framework and the experiments is publicly available: general-purpose tools have been integrated into the Clinica software (www.clinica.run) and the paper-specific code is available at: https://github.com/aramis-lab/AD-DL.

Reproducible evaluation of classification methods in Alzheimer's disease: Framework and application to MRI and PET data.

A large number of papers have introduced novel machine learning and feature extraction methods for automatic classification of Alzheimer's disease (AD). However, while the vast majority of these works use the public dataset ADNI for evaluation, they are difficult to reproduce because different key components of the validation are often not readily available. These components include selected participants and input data, image preprocessing and cross-validation procedures. The performance of the different approaches is also difficult to compare objectively. In particular, it is often difficult to assess which part of the method (e.g. preprocessing, feature extraction or classification algorithms) provides a real improvement, if any. In the present paper, we propose a framework for reproducible and objective classification experiments in AD using three publicly available datasets (ADNI, AIBL and OASIS). The framework comprises: i) automatic conversion of the three datasets into a standard format (BIDS); ii) a modular set of preprocessing pipelines, feature extraction and classification methods, together with an evaluation framework, that provide a baseline for benchmarking the different components. We demonstrate the use of the framework for a large-scale evaluation on 1960 participants using T1 MRI and FDG PET data. In this evaluation, we assess the influence of different modalities, preprocessing, feature types (regional or voxel-based features), classifiers, training set sizes and datasets. Performances were in line with the state-of-the-art. FDG PET outperformed T1 MRI for all classification tasks. No difference in performance was found for the use of different atlases, image smoothing, partial volume correction of FDG PET images, or feature type. Linear SVM and L2-logistic regression resulted in similar performance and both outperformed random forests. The classification performance increased along with the number of subjects used for training. Classifiers trained on ADNI generalized well to AIBL and OASIS. All the code of the framework and the experiments is publicly available: general-purpose tools have been integrated into the Clinica software (www.clinica.run) and the paper-specific code is available at: https://gitlab.icm-institute.org/aramislab/AD-ML.

An Automated Pipeline for the Analysis of PET Data on the Cortical Surface.

We present a fully automatic pipeline for the analysis of PET data on the cortical surface. Our pipeline combines tools from FreeSurfer and PETPVC, and consists of (i) co-registration of PET and T1-w MRI (T1) images, (ii) intensity normalization, (iii) partial volume correction, (iv) robust projection of the PET signal onto the subject's cortical surface, (v) spatial normalization to a template, and (vi) atlas statistics. We evaluated the performance of the proposed workflow by performing group comparisons and showed that the approach was able to identify the areas of hypometabolism characteristic of different dementia syndromes: Alzheimer's disease (AD) and both the semantic and logopenic variants of primary progressive aphasia. We also showed that these results were comparable to those obtained with a standard volume-based approach. We then performed individual classifications and showed that vertices can be used as features to differentiate cognitively normal and AD subjects. This pipeline is integrated into Clinica, an open-source software platform for neuroscience studies available at www.clinica.run.