Qasem Al-Tashi

Immune checkpoint inhibitors (ICIs), either as monotherapy (ICI-Mono) or combined with chemotherapy (ICI-Chemo), improves survival in advanced non-small cell lung cancer (NSCLC). However, prospective guidance for choosing between these options remains limited, and single-feature biomarkers like PD-L1 prove inadequate. We develop a machine learning model using clinicogenomic data from four cohorts (MD Anderson n = 750; Mayo Clinic n = 80; Dana-Farber n = 1077; Stand Up To Cancer n = 393) to predict individual benefit from adding chemotherapy. Benefit scores are calculated using five distinct functions derived from 28 genomic and 6 clinical features. Our integrated model, A-STEP (Attention-based Scoring for Treatment Effect Prediction), estimates heterogeneous treatment effects and achieves the largest reduction in 3-month progression risk, improving weighted risk reduction by 13–23% over stand-alone models. A-STEP recommends treatment changes for over 50% of patients, most often favoring ICI-Chemo. In simulation on external cohort, patients treated in accordance with A-STEP recommendations show improved 2-year progression-free survival (HR = 0.60 for ICI-Mono treatment arm; HR = 0.58 for ICI-Chemo treatment arm). Predictive features include FBXW7, APC, and PD-L1. In this study, we demonstrate how machine learning can fill critical gaps in immunotherapy selection for NSCLC, by modeling treatment heterogeneity with real-world clinicogenomic data, driving precision medicine beyond conventional biomarker boundaries.

The Arithmetic Optimization Algorithm (AOA) is a recently proposed metaheuristic algorithm that has been shown to perform well in several benchmark tests. The AOA is a metaheuristic that uses the main arithmetic operators’ distribution behavior, such as multiplication, division, subtraction, and addition. This paper proposes a binary version of the Arithmetic Optimization Algorithm (BAOA) to tackle the feature selection problem in classification. The algorithm’s search space is converted from a continuous to a binary one using the sigmoid transfer function to meet the nature of the feature selection task. The classifier uses a method known as the wrapper-based approach K-Nearest Neighbors (KNN), to find the best possible solutions. This study uses 18 benchmark datasets from the University of California, Irvine (UCI) repository to evaluate the suggested binary algorithm’s performance. The results demonstrate that BAOA outperformed the Binary Dragonfly Algorithm (BDF), Binary Particle Swarm Optimization (BPSO), Binary Genetic Algorithm (BGA), and Binary Cat Swarm Optimization (BCAT) when various performance metrics were used, including classification accuracy, selected features as well as the best and worst optimum fitness values.

Computed Tomography (CT) imaging has become a commonly used technique in healthcare to identify irregularities in the human body. However, CT scans involve exposure to electromagnetic radiation, which can pose health risks to patients. To address this, Low-Dose CT has been introduced, but it results in degraded image quality, including increased noise, artifacts, and loss of edge and feature contrast. This can limit the effectiveness of Computer-Aided Diagnosis systems. Denoising and preserving edge sharpness in Low-Dose CT images is a challenging task that conventional denoising techniques may not efficiently solve. Deep learning-based methods have emerged as a potential solution to this problem. This study proposes a new unsupervised LDCT image denoising algorithm called DEPnet (Denoise and Edge Preserve), which utilises a U-Net-based autoencoder with hybrid dilated convolution and batch normalization layers. The proposed algorithm has been evaluated on the KiTS19 Low-Dose CT Grand Challenge dataset and compared with other models such as Q-AE, Msaru-Net, and CT-ReCNN. The results demonstrate that DEPnet effectively reduces noise in LDCT images and preserves fine details, making it a promising solution for denoising Low-Dose CT images.

Data generated from sources such as wearable sensors, medical imaging, personal health records, pathology records, and public health organizations have resulted in a massive information increase in the medical sciences over the last decade. Advances in computational hardware, such as cloud computing, Graphical Processing Units (GPUs), and Tensor Processing Units (TPUs), provide the means to utilize these data. Consequently, many Artificial Intelligence (AI)-based methods have been developed to infer from large healthcare data. Here, we present an overview of recent progress in artificial intelligence and biosensors in medical and life sciences. We discuss the role of machine learning in medical imaging, precision medicine, and biosensors for the Internet of Things (IoT). We review the most recent advancements in wearable biosensing technologies that use AI to assist in monitoring bodily electro-physiological and electro-chemical signals and disease diagnosis, demonstrating the trend towards personalized medicine with highly effective, inexpensive, and precise point-of-care treatment. Furthermore, an overview of the advances in computing technologies, such as accelerated artificial intelligence, edge computing, and federated learning for medical data, are also documented. Finally, we investigate challenges in data-driven AI approaches, the potential issues that biosensors and IoT-based healthcare generate, and the distribution shifts that occur among different data modalities, concluding with an overview of future prospects.

Wind power’s increasing use and its integration into the utility grid have prompted scholars to focus more on the refinement of wind power harvesting systems, injecting stable power and providing ancillary services to the utility grid. This article introduces a novel hybrid approach termed as sine cosine algorithm and transient search optimization (HSCATSO) optimizer of better exploration and exploitation phases for optimal designing the power electronic converter control schemes (PECCS) of the grid-tied variable speed wind turbine (VSWT) system. The PECCS is simultaneously coordinated with a robust maximum power extraction algorithm (MPEA) to form enhanced control systems for achieving the best wind harnessing, improving the VSWT system performance, and supporting the utility grid stability. In this context, the HSCATSO optimally designs the PECCS parameters based on minimizing the summation of integral squared error (ISE) of multiple error signals in the developed control schemes in coordination with the MPEA. The superiority of the HSCATSO optimizer is validated using twenty benchmark functions and statistically analysed against four well-known optimization algorithms. Meanwhile, the effectiveness of the optimally designed PECCS using the HSCATSO is verified by the extensive simulation analysis in MATLAB/Simulink considering severe grid disturbance and real wind speed data taken from Kudat, Sabah, Malaysia to mimic realistic circumstances. The obtained results have been compared with that realized using the other algorithms-based design PECCS. The simulation outcomes affirmed that the PECCS designed by the HSCATSO and coordinated with MPEA resulted in higher power harvesting and enhanced the grid-tied VSWT system stability better than the competitive control schemes.

The identification of biomarkers plays a crucial role in personalized medicine, both in the clinical and research settings. However, the contrast between predictive and prognostic biomarkers can be challenging due to the overlap between the two. A prognostic biomarker predicts the future outcome of cancer, regardless of treatment, and a predictive biomarker predicts the effectiveness of a therapeutic intervention. Misclassifying a prognostic biomarker as predictive (or vice versa) can have serious financial and personal consequences for patients. To address this issue, various statistical and machine learning approaches have been developed. The aim of this study is to present an in-depth analysis of recent advancements, trends, challenges, and future prospects in biomarker identification. A systematic search was conducted using PubMed to identify relevant studies published between 2017 and 2023. The selected studies were analyzed to better understand the concept of biomarker identification, evaluate machine learning methods, assess the level of research activity, and highlight the application of these methods in cancer research and treatment. Furthermore, existing obstacles and concerns are discussed to identify prospective research areas. We believe that this review will serve as a valuable resource for researchers, providing insights into the methods and approaches used in biomarker discovery and identifying future research opportunities.

Minimizing energy costs while maintaining consumer satisfaction is a very challenging task in a smart home. The contradictory nature of these two objective functions (cost of energy and satisfaction level) requires a multi-objective problem formulation that can offer several trade-off solutions to the consumer. Previous works have individually considered the cost and satisfaction, but there is a lack of research that considers both these objectives simultaneously. Our work proposes an optimum home appliance scheduling method to obtain an optimum satisfaction level with a minimum cost of energy. To achieve this goal, first, an energy management system (EMS) is developed using a rule-based algorithm to reduce the cost of energy by efficient utilization of renewable energy resources and an energy storage system. The second part involves the development of an optimization algorithm for optimal appliance scheduling based on consumer satisfaction level, involving their time and device-based preferences. For that purpose, a multi-objective grey wolf accretive satisfaction algorithm (MGWASA) is developed, with the aim to provide trade-off solutions for optimal load patterns based on cost per unit satisfaction index (Cs_index) and percentage satisfaction (%S). The MGWASA is evaluated for a grid-connected smart home model with EMS. To ensure the accuracy of the numerical simulations, actual climatological data and consumer preferences are considered. The Cs_index is derived for six different cases by simulating (a) optimal load, (b) ideal load, and (c) base (random) load, with and without EMS. The results of MGWASA are benchmarked against other state-of-the-art optimization algorithms, namely, binary non-dominated sorting genetic algorithm-2 (NSGAII), multi-objective binary particle swarm optimization algorithm (MOBPSO), Multi-objective artificial bee colony (MOABC), and multi-objective evolutionary algorithm (MOEA). With the proposed appliance scheduling technique, a % reduction in annual energy cost is achieved. MGWASA yields Cs_index at 0.049$ with %S of 97%, in comparison to NSGAII, MOBPSO, MOABC, and MOEA, which yield %S of 95%, 90%, 92%, and 94% at 0.052$, 0.048$, 0.0485$, and 0.050$, respectively. Moreover, various related aspects, including energy balance, PV utilization, energy cost, net present cost, and cash payback period, are also analyzed. Lastly, sensitivity analysis is carried out to demonstrate the impact of any future uncertainties on the system inputs.

The process of selecting features or reducing dimensionality can be viewed as a multi-objective minimization problem in which both the number of features and error rate must be minimized. While it is a multi-objective problem, current methods tend to treat feature selection as a single-objective optimization task. This paper presents enhanced multi-objective grey wolf optimizer with Lévy flight and mutation phase (LMuMOGWO) for tackling feature selection problems. The proposed approach integrates two effective operators into the existing Multi-objective Grey Wolf optimizer (MOGWO): a Lévy flight and a mutation operator. The Lévy flight, a type of random walk with jump size determined by the Lévy distribution, enhances the global search capability of MOGWO, with the objective of maximizing classification accuracy while minimizing the number of selected features. The mutation operator is integrated to add more informative features that can assist in enhancing classification accuracy. As feature selection is a binary problem, the continuous search space is converted into a binary space using the sigmoid function. To evaluate the classification performance of the selected feature subset, the proposed approach employs a wrapper-based Artificial Neural Network (ANN). The effectiveness of the LMuMOGWO is validated on 12 conventional UCI benchmark datasets and compared with two existing variants of MOGWO, BMOGWO-S (based sigmoid), BMOGWO-V (based tanh) as well as Non-dominated Sorting Genetic Algorithm II (NSGA-II) and Multi-objective Particle Swarm Optimization (BMOPSO). The results demonstrate that the proposed LMuMOGWO approach is capable of successfully evolving and improving a set of randomly generated solutions for a given optimization problem. Moreover, the proposed approach outperforms existing approaches in most cases in terms of classification error rate, feature reduction, and computational cost.