Ahmed Sharafeldeen

Accurate prediction of breast cancer (BC) response to neoadjuvant chemotherapy (NAC) is critical for tailoring treatment strategies and improving patient outcomes. This study introduces a novel deep learning-based framework that integrates multi-parametric magnetic resonance imaging (MRI) (i.e., T1, T2, STIR, and DWI), along with clinical and molecular subtype markers, to classify tumor response into pathological complete response (pCR), partial response (PR), and stable disease (SD). First, tumor regions are delineated across MRI modalities and then modeled using a translation-invariant Markov-Gibbs random field (MGRF) with analytical parameter estimation to capture modality-specific spatial appearance patterns correlated with NAC response. Subsequently, diffusion-weighted MRI is processed to generate apparent diffusion coefficient (ADC) maps, offering quantitative assessment of intratumoral water diffusion and cellularity. Afterward, an adaptive rescaling module (ARM) is proposed to adjust spatial resolution and project volumetric inputs into 2D, enabling compatibility with pretrained convolutional networks. Finally, a customized SEResNet architecture, augmented with Squeeze-and-Excitation (SE) blocks, is introduced to extract modality-specific features which are then fused with clinical and molecular subtypes descriptors for final classification. Evaluated on a cohort of 109 BC patients using leave-one-subject-out (LOSO) cross-validation method, the system achieved an accuracy of 96.33%, a precision of 96.51%, a recall of 96.33%, an F1-score of 96.23%, and a Cohen’s kappa of 94.08%, outperforming its individual components, various pretrained deep learning models, and a state-of-the-art method. These results underscore the value of integrating the appearance model, functional (i.e., ADC) model, adaptive rescaling module, SE blocks, and clinical and molecular subtype markers for the precise prediction of NAC outcomes.

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder that significantly impairs cognitive functions, leading to memory loss and other behavioral changes. It is the seventh leading cause of death worldwide, with millions of people affected. Early and accurate detection of AD is critical for improving patient outcomes and slowing disease progression. Recent advancements in machine learning (ML) and deep learning (DL) models have demonstrated significant potential for detecting AD using patient’s speech signals, as subtle changes in speech patterns, such as reduced fluency, pronunciation difficulties, and cognitive decline, can serve as early indicators of the disease, offering a non-invasive and cost-effective method for early diagnosis. This survey paper provides a comprehensive review of the current literature on the application of ML and DL techniques for AD detection through the analysis of a patient’s speech signal, utilizing various acoustic and textual features. Moreover, it offers an overview of the changes in the brain caused by the disease, associated risk factors, publicly available datasets, and future directions for leveraging ML and DL in the detection of AD.

The primary objective of this paper is to develop a machine learning-based approach capable of predicting the treatment response of neoadjuvant chemotherapy (NAC) to enhance breast cancer treatment management. The proposed system aims to predict NAC outcomes across three categories: pathological complete response (CR), partial response (PR), and stable disease (SD), by analyzing multimodal magnetic resonance images with clinical and molecular subtype markers. To ensure the comprehensiveness of our system design, texture radiomics were extracted from T1, T2, and STIR MRI modalities, along with functional radiomics from diffusion-weighted MRI at various b-values. The main rationale behind employing multiple b-values in collecting DW-MRI is to effectively capture the complexities of blood diffusion within the tumor microstructure. The proposed system comprises several key steps: i) extracting texture and functional radiomics from T1, T2, STIR MRI, and DW-MRI data; ii) identifying the most significant radiomics correlated with NAC treatment using a genetic algorithm; iii) initially predicting the PR from alternative treatment responses utilizing the extracted textures and functional radiomics; and iv) subsequently integrating clinical and molecular subtype markers with imaging radiomics to differentiate between CR and SD. Our proposed system is trained and validated through the utilization of a leave-one-subject-out (LOSO) cross-validation approach on various MRI scans from 109 subjects, of whom 27 had complete responses, 54 had partial responses, and 28 had no responses. The performance of the proposed system was assessed through the utilization of Cohen’s Kappa and accuracy metrics, achieving 81.31% and 88.07%, respectively. Our various experiments showed that integrating clinical and molecular subtype markers with radiomics highlights the proposed system’s efficiency in evaluating the tumor’s response to NAC efficiently, outperforming predictions based solely on individual radiomics.

We propose an automated, explainable artificial intelligence (xAI) system for age-related macular degeneration (AMD) diagnosis. Mimicking the physician's perceptions, the proposed xAI system is capable of deriving clinically meaningful features from optical coherence tomography (OCT) B-scan images to differentiate between a normal retina, different grades of AMD (early, intermediate, geographic atrophy (GA), inactive wet or active neovascular disease [exudative or wet AMD]), and non-AMD diseases. Particularly, we extract retinal OCT-based clinical imaging markers that are correlated with the progression of AMD, which include: ( i ) subretinal tissue, sub-retinal pigment epithelial tissue, intraretinal fluid, subretinal fluid, and choroidal hypertransmission detection using a DeepLabV3+ network; ( ii ) detection of merged retina layers using a novel convolutional neural network model; ( iii ) drusen detection based on 2D curvature analysis; ( iv ) estimation of retinal layers' thickness, and first-order and higher-order reflectivity features. Those clinical features are used to grade a retinal OCT in a hierarchical decision tree process. The first step looks for severe disruption of retinal layers' indicative of advanced AMD. These cases are analyzed further to diagnose GA, inactive wet AMD, active wet AMD, and non-AMD diseases. Less severe cases are analyzed using a different pipeline to identify OCT with AMD-specific pathology, which is graded as intermediate-stage or early-stage AMD. The remainder is classified as either being a normal retina or having other non-AMD pathology. The proposed system in the multi-way classification task, evaluated on 1285 OCT images, achieved 90.82% accuracy. These promising results demonstrated the capability to automatically distinguish between normal eyes and all AMD grades in addition to non-AMD diseases.

Breast cancer stands out as the most frequently identified malignancy, ranking as the fifth leading cause of global cancer-related deaths. The American College of Radiology (ACR) introduced the Breast Imaging Reporting and Data System (BI-RADS) as a standard terminology facilitating communication between radiologists and clinicians; however, an update is now imperative to encompass the latest imaging modalities developed subsequent to the 5th edition of BI-RADS. Within this review article, we provide a concise history of BI-RADS, delve into advanced mammography techniques, ultrasonography (US), magnetic resonance imaging (MRI), PET/CT images, and microwave breast imaging, and subsequently furnish comprehensive, updated insights into Molecular Breast Imaging (MBI), diagnostic imaging biomarkers, and the assessment of treatment responses. This endeavor aims to enhance radiologists’ proficiency in catering to the personalized needs of breast cancer patients. Lastly, we explore the augmented benefits of artificial intelligence (AI), machine learning (ML), and deep learning (DL) applications in segmenting, detecting, and diagnosing breast cancer, as well as the early prediction of the response of tumors to neoadjuvant chemotherapy (NAC). By assimilating state-of-the-art computer algorithms capable of deciphering intricate imaging data and aiding radiologists in rendering precise and effective diagnoses, AI has profoundly revolutionized the landscape of breast cancer radiology. Its vast potential holds the promise of bolstering radiologists’ capabilities and ameliorating patient outcomes in the realm of breast cancer management.

A novel unsupervised-based segmentation method is introduced to accurately delineate the lung region in 3D CT images based on appearance and geometric models. First, a probabilistic model that utilized a linear combination of Gaussian (LCG) tuned by a modified expectation maximization (EM) algorithm, is employed to model the density distribution of 3D CT chest volume. Subsequently, the initial labeling of the 3D CT chest volume is mapped to a probability distribution based on a 3D Markov Gibbs random field (MGRF) for refining. Finally, a geometric model is employed to refine the proposed segmentation by interpolating/connecting two points on its boundary with high curvature. The effectiveness of the proposed approach on 3D computed tomography (CT) chest scans of 26 patients diagnosed with different severity of coronavirus disease 2019 (COVID-19) is evaluated using four different metrics: overlap coefficient, Dice similarity coefficient (DSC), absolute lung volume difference (ALVD), and 95 ^th -percentile bidirectional Hausdorff distance (95 ^th HD). The proposed method achieved 94.89% ±2.39% , 97.36% ±1.27% , 1.79 ±1.89 , and 4.75 ±2.3 , respectively. Compared to three state-of-the-art methods based on deep learning approaches, the proposed method achieved superior performance in segmenting pathological lung tissues, demonstrating the promising of the proposed segmentation system.

If untreated, diabetic retinopathy (DR) can result in a severe health complication, leading to visual loss. This study focuses on developing a computer-assisted diagnostic (CAD) system that utilizes 3D optical coherence tomography (OCT) images for detecting DR. To begin with, the 3D OCT images are subjected to a process where the retinal layers are isolated from the input. Following this, from each individual retinal layer, two key 3D characteristics, namely thickness and first-order reflectivity, are computed. Eventually, classification is carried out using backpropagation neural networks. Utilizing 10-folds cross-validation on 188 cases, experiments validate the benefits of the developed system over competing approaches, with an accuracy of 94.74% ± 5.55%. These results demonstrate the method’s potential for DR detection utilizing OCT images.