Transfer Learning for Alzheimer's Disease through Neuroimaging Biomarkers: A Systematic Review

A systematic literature review on the significance of deep learning and machine learning in predicting Alzheimer's disease

BACKGROUND

Alzheimer's disease (AD) is the most prevalent cause of dementia, characterized by a steady decline in mental, behavioral, and social abilities and impairs a person's capacity for independent functioning. It is a fatal neurodegenerative disease primarily affecting older adults.

OBJECTIVES

The purpose of this literature review is to investigate various AD detection techniques, datasets, input modalities, algorithms, libraries, and performance evaluation metrics used to determine which model or strategy may provide superior performance.

METHOD

The initial search yielded 807 papers, but only 100 research articles were chosen after applying the inclusion-exclusion criteria. This SLR analyzed research items published between January 2019 and December 2022. The ACM, Elsevier, IEEE Xplore Digital Library, PubMed, Springer and Taylor & Francis were systematically searched. The current study considers articles that used Magnetic Resonance Imaging (MRI), Positron Emission Tomography (PET), APOe4 genotype, Diffusion Tensor Imaging (DTI) and Cerebrospinal Fluid (CSF) biomarkers. The study was performed following Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines.

CONCLUSION

According to the literature survey, most studies (n = 76) used the DL strategy. The datasets used by studies were primarily derived from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. The majority of studies (n = 73) used single-modality neuroimaging data, while the remaining used multi-modal input data. In a multi-modality approach, the combination of MRI and PET scans is commonly preferred. Also, Regarding the algorithm used, Convolution Neural Network (CNN) showed the highest accuracy, 100 %, in classifying AD vs. CN subjects whereas the SVM was the most common ML algorithm, with a maximum accuracy of 99.82 %.

The Association of Ancillary Diagnostic Tests With Outcome in Dementia

OBJECTIVES

Dementia is a clinical diagnosis without curative treatment. It is uncertain whether ancillary testing is beneficial for patients. This study investigates the association between use of diagnostic tests and time to poor outcome and health care costs.

DESIGN

Nationwide register-based cohort study using health care reimbursement data in the Netherlands.

SETTING AND PARTICIPANTS

All Dutch hospitals, including 13,312 patients diagnosed with dementia in 2018.

METHODS

Diagnostic testing included computed tomography or magnetic resonance imaging (CT/MRI), neuropsychological examination (NPE), nuclear imaging (PET/SPECT), electroencephalography (EEG), and cerebrospinal fluid (CSF) testing. We compared time to poor outcome (institutionalization or death) and costs per month from 2018 to 2021 between those who underwent a specific diagnostic test in previous years to controls, propensity score matched for age, sex, type of hospital, and comorbidity.

RESULTS

Time to poor outcome in those who underwent CT/MRI, EEG, or CSF testing was similar to those who did not, but was longer for those who underwent NPE. Time to poor outcome was shorter in patients who underwent PET/SPECT. Patients who underwent CSF testing or PET/SPECT had higher mean total health care costs as compared to controls (CSF €248, 95% CI 64-433; PET/SPECT: €315, 95% CI 179-451). NPE during the diagnostic trajectory was associated with lower total health care cost (-€127, 95% CI -62, -193).

CONCLUSION AND IMPLICATIONS

NPE was associated with longer time to poor outcome and lower health care costs, potentially due to confounding by indication. Patients who underwent neuroimaging (CT, MRI, SPECT/PET), CSF testing, or EEG for dementia diagnostics did not experience a longer time to poor outcome or lower health care costs. This emphasizes the importance of clinical examination as anchor for the diagnosis of dementia.

Performance Evaluation of Deep, Shallow and Ensemble Machine Learning Methods for the Automated Classification of Alzheimer's Disease

Artificial intelligence (AI)-based approaches are crucial in computer-aided diagnosis (CAD) for various medical applications. Their ability to quickly and accurately learn from complex data is remarkable. Deep learning (DL) models have shown promising results in accurately classifying Alzheimer's disease (AD) and its related cognitive states, Early Mild Cognitive Impairment (EMCI) and Late Mild Cognitive Impairment (LMCI), along with the healthy conditions known as Cognitively Normal (CN). This offers valuable insights into disease progression and diagnosis. However, certain traditional machine learning (ML) classifiers perform equally well or even better than DL models, requiring less training data. This is particularly valuable in CAD in situations with limited labeled datasets. In this paper, we propose an ensemble classifier based on ML models for magnetic resonance imaging (MRI) data, which achieved an impressive accuracy of 96.52%. This represents a 3-5% improvement over the best individual classifier. We evaluated popular ML classifiers for AD classification under both data-scarce and data-rich conditions using the Alzheimer's Disease Neuroimaging Initiative and Open Access Series of Imaging Studies datasets. By comparing the results to state-of-the-art CNN-centric DL algorithms, we gain insights into the strengths and weaknesses of each approach. This work will help users to select the most suitable algorithm for AD classification based on data availability.

Optimising brain age estimation through transfer learning: A suite of pre-trained foundation models for improved performance and generalisability in a clinical setting

Estimated age from brain MRI data has emerged as a promising biomarker of neurological health. However, the absence of large, diverse, and clinically representative training datasets, along with the complexity of managing heterogeneous MRI data, presents significant barriers to the development of accurate and generalisable models appropriate for clinical use. Here, we present a deep learning framework trained on routine clinical data (N up to 18,890, age range 18-96 years). We trained five separate models for accurate brain age prediction (all with mean absolute error ≤4.0 years, R2  ≥ .86) across five different MRI sequences (T2 -weighted, T2 -FLAIR, T1 -weighted, diffusion-weighted, and gradient-recalled echo T2 *-weighted). Our trained models offer dual functionality. First, they have the potential to be directly employed on clinical data. Second, they can be used as foundation models for further refinement to accommodate a range of other MRI sequences (and therefore a range of clinical scenarios which employ such sequences). This adaptation process, enabled by transfer learning, proved effective in our study across a range of MRI sequences and scan orientations, including those which differed considerably from the original training datasets. Crucially, our findings suggest that this approach remains viable even with limited data availability (as low as N = 25 for fine-tuning), thus broadening the application of brain age estimation to more diverse clinical contexts and patient populations. By making these models publicly available, we aim to provide the scientific community with a versatile toolkit, promoting further research in brain age prediction and related areas.

Medical image identification methods: A review

The identification of medical images is an essential task in computer-aided diagnosis, medical image retrieval and mining. Medical image data mainly include electronic health record data and gene information data, etc. Although intelligent imaging provided a good scheme for medical image analysis over traditional methods that rely on the handcrafted features, it remains challenging due to the diversity of imaging modalities and clinical pathologies. Many medical image identification methods provide a good scheme for medical image analysis. The concepts pertinent of methods, such as the machine learning, deep learning, convolutional neural networks, transfer learning, and other image processing technologies for medical image are analyzed and summarized in this paper. We reviewed these recent studies to provide a comprehensive overview of applying these methods in various medical image analysis tasks, such as object detection, image classification, image registration, segmentation, and other tasks. Especially, we emphasized the latest progress and contributions of different methods in medical image analysis, which are summarized base on different application scenarios, including classification, segmentation, detection, and image registration. In addition, the applications of different methods are summarized in different application area, such as pulmonary, brain, digital pathology, brain, skin, lung, renal, breast, neuromyelitis, vertebrae, and musculoskeletal, etc. Critical discussion of open challenges and directions for future research are finally summarized. Especially, excellent algorithms in computer vision, natural language processing, and unmanned driving will be applied to medical image recognition in the future.

Machine Learning-Derived MRI-Based Neurodegeneration Biomarker for Alzheimer's Disease: A Multi-Database Validation Study

BACKGROUND

Pilot study showed that Alzheimer's disease resemblance atrophy index (AD-RAI), a machine learning-derived MRI-based neurodegeneration biomarker of AD, achieved excellent diagnostic performance in diagnosing AD with moderate to severe dementia.

OBJECTIVE

The primary objective was to validate and compare the performance of AD-RAI with conventional volumetric hippocampal measures in diagnosing AD with mild dementia. The secondary objectives were 1) to investigate the association between imaging biomarkers with age and gender among cognitively unimpaired (CU) participants; 2) to analyze whether the performance of differentiating AD with mild dementia from CU will improve after adjustment for age/gender.

METHODS

AD with mild dementia (n = 218) and CU (n = 1,060) participants from 4 databases were included. We investigated the area under curve (AUC), sensitivity, specificity, and balanced accuracy of AD-RAI, hippocampal volume (HV), and hippocampal fraction (HF) in differentiating between AD and CU participants. Among amyloid-negative CU participants, we further analyzed correlation between the biomarkers with age/gender. We also investigated whether adjustment for age/gender will affect performance.

RESULTS

The AUC of AD-RAI (0.93) was significantly higher than that of HV (0.89) and HF (0.89). Subgroup analysis among A + AD and A- CU showed that AUC of AD-RAI (0.97) was also higher than HV (0.94) and HF (0.93). Diagnostic performance of AD-RAI and HF was not affected by age/gender while that of HV improved after age adjustment.

CONCLUSIONS

AD-RAI achieves excellent clinical validity and outperforms conventional volumetric hippocampal measures in aiding the diagnosis of AD mild dementia without the need for age adjustment.

A robust class decomposition-based approach for detecting Alzheimer's progression

Computer-aided diagnosis of Alzheimer's disease (AD) is a rapidly growing field with the possibility to be utilized in practice. Deep learning has received much attention in detecting AD from structural magnetic resonance imaging (sMRI). However, training a convolutional neural network from scratch is problematic because it requires a lot of annotated data and additional computational time. Transfer learning can offer a promising and practical solution by transferring information learned from other image recognition tasks to medical image classification. Another issue is the dataset distribution's irregularities. A common classification issue in datasets is a class imbalance, where the distribution of samples among the classes is biased. For example, a dataset may contain more instances of some classes than others. Class imbalance is challenging because most machine learning algorithms assume that each class should have an equal number of samples. Models consequently perform poorly in prediction. Class decomposition can address this problem by making learning a dataset's class boundaries easier. Motivated by these approaches, we propose a class decomposition transfer learning (CDTL) approach that employs VGG19, AlexNet, and an entropy-based technique to detect AD from sMRI. This study aims to assess the robustness of the CDTL approach in detecting the cognitive decline of AD using data from various ADNI cohorts to determine whether comparable classification accuracy for the two or more cohorts would be obtained. Furthermore, the proposed model achieved state-of-the-art performance in predicting mild cognitive impairment (MCI)-to-AD conversion with an accuracy of 91.45%.

Diagnostic Efficacy and Clinical Relevance of Artificial Intelligence in Detecting Cognitive Decline

Cognitive impairment is an age-associated disorder of increasing prevalence as the aging population continues to grow. Classified based on the level of cognitive decline, memory, function, and capacity to conduct activities of daily living, cognitive impairment ranges from mild cognitive impairment to dementia. When considering the insidious nature of the etiologies responsible for varying degrees of cognitive impairment, early diagnosis may provide a clinical benefit through the facilitation of early treatment. Typical diagnosis relies heavily on evaluation in a primary care setting. However, there is evidence that other diagnostic tools may aid in an earlier diagnosis of the different underlying pathologies responsible for cognitive impairment. Artificial intelligence represents a new intersecting field with healthcare that may aid in the early detection of neurodegenerative disorders. When assessing the role of AI in detecting cognitive decline, it is important to consider both the diagnostic efficacy of AI algorithms and the clinical relevance and impact of early interventions as a result of early detection. Thus, this review highlights promising investigations and developments in the space of artificial intelligence and healthcare and their potential to impact patient outcomes.

Improving Alzheimer's Disease Classification in Brain MRI Images Using a Neural Network Model Enhanced with PCA and SWLDA

The examination of Alzheimer's disease (AD) using adaptive machine learning algorithms has unveiled promising findings. However, achieving substantial credibility in medical contexts necessitates a combination of notable accuracy, minimal processing time, and universality across diverse populations. Therefore, we have formulated a hybrid methodology in this study to classify AD by employing a brain MRI image dataset. We incorporated an averaging filter during preprocessing in the initial stage to reduce extraneous details. Subsequently, a combined strategy was utilized, involving principal component analysis (PCA) in conjunction with stepwise linear discriminant analysis (SWLDA), followed by an artificial neural network (ANN). SWLDA employs a combination of forward and backward recursion methods to choose a restricted set of features. The forward recursion identifies the most interconnected features based on partial Z-test values. Conversely, the backward recursion method eliminates the least correlated features from the same feature space. After the extraction and selection of features, an optimized artificial neural network (ANN) was utilized to differentiate the various classes of AD. To demonstrate the significance of this hybrid approach, we utilized publicly available brain MRI datasets using a 10-fold cross-validation strategy. The proposed method excelled over existing state-of-the-art systems, attaining weighted average recognition rates of 99.35% and 96.66%, respectively, across all the datasets.

Navigating Early Alzheimer's Diagnosis: A Comprehensive Review of Diagnostic Innovations

The hunt for early Alzheimer's disease detection has created cutting-edge diagnostic instruments with enormous promise. This article examines the many facets of these developments, focusing on how they have revolutionised diagnosis and patient outcomes. These tools make it possible to detect tiny brain changes even before they give birth to clinical symptoms by combining cutting-edge biomarkers, neuroimaging methods, and machine-learning algorithms. A significant opportunity for therapies that can slow the course of the disease exists during this early detection stage. Additionally, these cutting-edge techniques improve diagnostic precision, objectivity, and accessibility. Liquid biopsies and blood-based biomarkers provide non-invasive alternatives, filling accessibility gaps in diagnosis. While issues with standardisation, ethics, and data integration continue, collaboration within research, clinical practice, and policy realms fuels positive developments. As technology advances, the way towards better Alzheimer's diagnosis becomes more evident, giving patients and families dealing with this difficult illness fresh hope. The synergy between scientific advancement and compassionate treatment is crucial for improving Alzheimer's disease diagnosis, as this paper emphasises.

The convergence of traditional and digital biomarkers through AI-assisted biosensing: A new era in translational diagnostics?

Advances in consumer electronics, alongside the fields of microfluidics and nanotechnology have brought to the fore low-cost wearable/portable smart devices. Although numerous smart devices that track digital biomarkers have been successfully translated from bench-to-bedside, only a few follow the same fate when it comes to track traditional biomarkers. Current practices still involve laboratory-based tests, followed by blood collection, conducted in a clinical setting as they require trained personnel and specialized equipment. In fact, real-time, passive/active and robust sensing of physiological and behavioural data from patients that can feed artificial intelligence (AI)-based models can significantly improve decision-making, diagnosis and treatment at the point-of-procedure, by circumventing conventional methods of sampling, and in person investigation by expert pathologists, who are scarce in developing countries. This review brings together conventional and digital biomarker sensing through portable and autonomous miniaturized devices. We first summarise the technological advances in each field vs the current clinical practices and we conclude by merging the two worlds of traditional and digital biomarkers through AI/ML technologies to improve patient diagnosis and treatment. The fundamental role, limitations and prospects of AI in realizing this potential and enhancing the existing technologies to facilitate the development and clinical translation of "point-of-care" (POC) diagnostics is finally showcased.

Repurposing drugs against Alzheimer's disease: can the anti-multiple sclerosis drug fingolimod (FTY720) effectively tackle inflammation processes in AD?

Therapeutic approaches providing effective medication for Alzheimer's disease (AD) patients after disease onset are urgently needed. Previous studies in AD mouse models and in humans suggested that physical exercise or changed lifestyle can delay AD-related synaptic and memory dysfunctions when treatment started in juvenile animals or in elderly humans before onset of disease symptoms. However, a pharmacological treatment that can reverse memory deficits in AD patients was thus far not identified. Importantly, AD disease-related dysfunctions have increasingly been associated with neuro-inflammatory mechanisms and searching for anti-inflammatory medication to treat AD seems promising. Like for other diseases, repurposing of FDA-approved drugs for treatment of AD is an ideally suited strategy to reduce the time to bring such medication into clinical practice. Of note, the sphingosine-1-phosphate analogue fingolimod (FTY720) was FDA-approved in 2010 for treatment of multiple sclerosis patients. It binds to the five different isoforms of Sphingosine-1-phosphate receptors (S1PRs) that are widely distributed across human organs. Interestingly, recent studies in five different mouse models of AD suggest that FTY720 treatment, even when starting after onset of AD symptoms, can reverse synaptic deficits and memory dysfunction in these AD mouse models. Furthermore, a very recent multi-omics study identified mutations in the sphingosine/ceramide pathway as a risk factor for sporadic AD, suggesting S1PRs as promising drug target in AD patients. Therefore, progressing with FDA-approved S1PR modulators into human clinical trials might pave the way for these potential disease modifying anti-AD drugs.

Development and validation of a deep learning signature for predicting lymph node metastasis in lung adenocarcinoma: comparison with radiomics signature and clinical-semantic model

OBJECTIVE

To develop and validate a deep learning (DL) signature for predicting lymph node (LN) metastasis in patients with lung adenocarcinoma.

METHODS

A total of 612 patients with pathologically-confirmed lung adenocarcinoma were retrospectively enrolled and were randomly divided into training cohort (n = 489) and internal validation cohort (n = 123). Besides, 108 patients were enrolled and constituted an independent test cohort (n = 108). Patients' clinical characteristics and CT semantic features were collected. The radiomics features were derived from contrast-enhanced CT images. The clinical-semantic model and radiomics signature were built to predict LN metastasis. Furthermore, Swin Transformer was adopted to develop a DL signature predictive of LN metastasis. Model performance was evaluated by area under the receiver operating characteristic curve (AUC), sensitivity, specificity, calibration curve, and decision curve analysis. The comparisons of AUC were conducted by the DeLong test.

RESULTS

The proposed DL signature yielded an AUC of 0.948-0.961 across all three cohorts, significantly superior to both clinical-semantic model and radiomics signature (all p < 0.05). The calibration curves show that DL signature predicted probabilities fit well the actual observed probabilities of LN metastasis. DL signature gained a higher net benefit than both clinical-semantic model and radiomics signature. The incorporation of radiomics signature or clinical-semantic risk predictors failed to reveal an incremental value over the DL signature.

CONCLUSIONS

The proposed DL signature based on Swin Transformer achieved a promising performance in predicting LN metastasis and could confer important information in noninvasive mediastinal LN staging and individualized therapeutic options.

KEY POINTS

• Accurate prediction for lymph node metastasis is crucial to formulate individualized therapeutic options for patients with lung adenocarcinoma. • The deep learning signature yielded an AUC of 0.948-0.961 across all three cohorts in predicting lymph node metastasis, superior to both radiomics signature and clinical-semantic model. • The incorporation of radiomics signature or clinical-semantic risk predictors into deep learning signature failed to reveal an incremental value over deep learning signature.

A review of the application of three-dimensional convolutional neural networks for the diagnosis of Alzheimer's disease using neuroimaging

Alzheimer's disease (AD) is a degenerative disorder that leads to progressive, irreversible cognitive decline. To obtain an accurate and timely diagnosis and detect AD at an early stage, numerous approaches based on convolutional neural networks (CNNs) using neuroimaging data have been proposed. Because 3D CNNs can extract more spatial discrimination information than 2D CNNs, they have emerged as a promising research direction in the diagnosis of AD. The aim of this article is to present the current state of the art in the diagnosis of AD using 3D CNN models and neuroimaging modalities, focusing on the 3D CNN architectures and classification methods used, and to highlight potential future research topics. To give the reader a better overview of the content mentioned in this review, we briefly introduce the commonly used imaging datasets and the fundamentals of CNN architectures. Then we carefully analyzed the existing studies on AD diagnosis, which are divided into two levels according to their inputs: 3D subject-level CNNs and 3D patch-level CNNs, highlighting their contributions and significance in the field. In addition, this review discusses the key findings and challenges from the studies and highlights the lessons learned as a roadmap for future research. Finally, we summarize the paper by presenting some major findings, identifying open research challenges, and pointing out future research directions.

A review on Alzheimer's disease classification from normal controls and mild cognitive impairment using structural MR images

Alzheimer's disease (AD) is an irreversible neurodegenerative brain disorder that degrades the memory and cognitive ability in elderly people. The main reason for memory loss and reduction in cognitive ability is the structural changes in the brain that occur due to neuronal loss. These structural changes are most conspicuous in the hippocampus, cortex, and grey matter and can be assessed by using neuroimaging techniques viz. Positron Emission Tomography (PET), structural Magnetic Resonance Imaging (MRI) and functional MRI (fMRI), etc. Out of these neuroimaging techniques, structural MRI has evolved as the best technique as it indicates the best soft tissue contrast and high spatial resolution which is important for AD detection. Currently, the focus of researchers is on predicting the conversion of Mild Cognitive Impairment (MCI) into AD. MCI represents the transition state between expected cognitive changes with normal aging and Alzheimer's disease. Not every MCI patient progresses into Alzheimer's disease. MCI can develop into stable MCI (sMCI, patients are called non-converters) or into progressive MCI (pMCI, patients are diagnosed as MCI converters). This paper discusses the prognosis of MCI to AD conversion and presents a review of structural MRI-based studies for AD detection. AD detection framework includes feature extraction, feature selection, and classification process. This paper reviews the studies for AD detection based on different feature extraction methods and machine learning algorithms for classification. The performance of various feature extraction methods has been compared and it has been observed that the wavelet transform-based feature extraction method would give promising results for AD classification. The present study indicates that researchers are successful in classifying AD from Normal Controls (N_rmC) but, it still requires a lot of work to be done for MCI/ N_rmC and MCI/AD, which would help in detecting AD at its early stage.

Artificial intelligence in dementia

PURPOSE OF REVIEW

Artificial intelligence tools are being rapidly integrated into clinical environments and may soon be incorporated into dementia diagnostic paradigms. A comprehensive review of emerging trends will allow physicians and other healthcare providers to better anticipate and understand these powerful tools.

RECENT FINDINGS

Machine learning models that utilize cerebral biomarkers are demonstrably effective for dementia identification and prediction; however, cerebral biomarkers are relatively expensive and not widely available. As eye images harbor several ophthalmic biomarkers that mirror the state of the brain and can be clinically observed with routine imaging, eye-based machine learning models are an emerging area, with efficacy comparable with cerebral-based machine learning models. Emerging machine learning architectures like recurrent, convolutional, and partially pretrained neural networks have proven to be promising frontiers for feature extraction and classification with ocular biomarkers.

SUMMARY

Machine learning models that can accurately distinguish those with symptomatic Alzheimer's dementia from those with mild cognitive impairment and normal cognition as well as predict progressive disease using relatively inexpensive and accessible ocular imaging inputs are impactful tools for the diagnosis and risk stratification of Alzheimer's dementia continuum. If these machine learning models can be incorporated into clinical care, they may simplify diagnostic efforts. Recent advancements in ocular-based machine learning efforts are promising steps forward.

Artificial Intelligence Models in the Diagnosis of Adult-Onset Dementia Disorders: A Review

Background: The progressive aging of populations, primarily in the industrialized western world, is accompanied by the increased incidence of several non-transmittable diseases, including neurodegenerative diseases and adult-onset dementia disorders. To stimulate adequate interventions, including treatment and preventive measures, an early, accurate diagnosis is necessary. Conventional magnetic resonance imaging (MRI) represents a technique quite common for the diagnosis of neurological disorders. Increasing evidence indicates that the association of artificial intelligence (AI) approaches with MRI is particularly useful for improving the diagnostic accuracy of different dementia types. Objectives: In this work, we have systematically reviewed the characteristics of AI algorithms in the early detection of adult-onset dementia disorders, and also discussed its performance metrics. Methods: A document search was conducted with three databases, namely PubMed (Medline), Web of Science, and Scopus. The search was limited to the articles published after 2006 and in English only. The screening of the articles was performed using quality criteria based on the Newcastle–Ottawa Scale (NOS) rating. Only papers with an NOS score 7 were considered for further review. Results: The document search produced a count of 1876 articles and, because of duplication, 1195 papers were not considered. Multiple screenings were performed to assess quality criteria, which yielded 29 studies. All the selected articles were further grouped based on different attributes, including study type, type of AI model used in the identification of dementia, performance metrics, and data type. Conclusions: The most common adult-onset dementia disorders occurring were Alzheimer’s disease and vascular dementia. AI techniques associated with MRI resulted in increased diagnostic accuracy ranging from 73.3% to 99%. These findings suggest that AI should be associated with conventional MRI techniques to obtain a precise and early diagnosis of dementia disorders occurring in old age.

End-to-End Deep Learning Architectures Using 3D Neuroimaging Biomarkers for Early Alzheimer’s Diagnosis

This study uses magnetic resonance imaging (MRI) data to propose end-to-end learning implementing volumetric convolutional neural network (CNN) models for two binary classification tasks: Alzheimer’s disease (AD) vs. cognitively normal (CN) and stable mild cognitive impairment (sMCI) vs. AD. The baseline MP-RAGE T1 MR images of 245 AD patients and 229 with sMCI were obtained from the ADNI dataset, whereas 245 T1 MR images of CN people were obtained from the IXI dataset. All of the images were preprocessed in four steps: N4 bias field correction, denoising, brain extraction, and registration. End-to-end-learning-based deep CNNs were used to discern between different phases of AD. Eight CNN-based architectures were implemented and assessed. The DenseNet264 excelled in both types of classification, with 82.5% accuracy and 87.63% AUC for training and 81.03% accuracy for testing relating to the sMCI vs. AD and 100% accuracy and 100% AUC for training and 99.56% accuracy for testing relating to the AD vs. CN. Deep learning approaches based on CNN and end-to-end learning offer a strong tool for examining minute but complex properties in MR images which could aid in the early detection and prediction of Alzheimer’s disease in clinical settings.

Transfer learning for medical image classification: a literature review

BACKGROUND

Transfer learning (TL) with convolutional neural networks aims to improve performances on a new task by leveraging the knowledge of similar tasks learned in advance. It has made a major contribution to medical image analysis as it overcomes the data scarcity problem as well as it saves time and hardware resources. However, transfer learning has been arbitrarily configured in the majority of studies. This review paper attempts to provide guidance for selecting a model and TL approaches for the medical image classification task.

METHODS

425 peer-reviewed articles were retrieved from two databases, PubMed and Web of Science, published in English, up until December 31, 2020. Articles were assessed by two independent reviewers, with the aid of a third reviewer in the case of discrepancies. We followed the PRISMA guidelines for the paper selection and 121 studies were regarded as eligible for the scope of this review. We investigated articles focused on selecting backbone models and TL approaches including feature extractor, feature extractor hybrid, fine-tuning and fine-tuning from scratch.

RESULTS

The majority of studies (n = 57) empirically evaluated multiple models followed by deep models (n = 33) and shallow (n = 24) models. Inception, one of the deep models, was the most employed in literature (n = 26). With respect to the TL, the majority of studies (n = 46) empirically benchmarked multiple approaches to identify the optimal configuration. The rest of the studies applied only a single approach for which feature extractor (n = 38) and fine-tuning from scratch (n = 27) were the two most favored approaches. Only a few studies applied feature extractor hybrid (n = 7) and fine-tuning (n = 3) with pretrained models.

CONCLUSION

The investigated studies demonstrated the efficacy of transfer learning despite the data scarcity. We encourage data scientists and practitioners to use deep models (e.g. ResNet or Inception) as feature extractors, which can save computational costs and time without degrading the predictive power.

Transfer Learning Approaches for Neuroimaging Analysis: A Scoping Review

Deep learning algorithms have been moderately successful in diagnoses of diseases by analyzing medical images especially through neuroimaging that is rich in annotated data. Transfer learning methods have demonstrated strong performance in tackling annotated data. It utilizes and transfers knowledge learned from a source domain to target domain even when the dataset is small. There are multiple approaches to transfer learning that result in a range of performance estimates in diagnosis, detection, and classification of clinical problems. Therefore, in this paper, we reviewed transfer learning approaches, their design attributes, and their applications to neuroimaging problems. We reviewed two main literature databases and included the most relevant studies using predefined inclusion criteria. Among 50 reviewed studies, more than half of them are on transfer learning for Alzheimer's disease. Brain mapping and brain tumor detection were second and third most discussed research problems, respectively. The most common source dataset for transfer learning was ImageNet, which is not a neuroimaging dataset. This suggests that the majority of studies preferred pre-trained models instead of training their own model on a neuroimaging dataset. Although, about one third of studies designed their own architecture, most studies used existing Convolutional Neural Network architectures. Magnetic Resonance Imaging was the most common imaging modality. In almost all studies, transfer learning contributed to better performance in diagnosis, classification, segmentation of different neuroimaging diseases and problems, than methods without transfer learning. Among different transfer learning approaches, fine-tuning all convolutional and fully-connected layers approach and freezing convolutional layers and fine-tuning fully-connected layers approach demonstrated superior performance in terms of accuracy. These recent transfer learning approaches not only show great performance but also require less computational resources and time.