International Journal of Clinical Case Reports, 2025, Vol.15, No.3, 98-109 http://medscipublisher.com/index.php/ijccr 100 2.3 The advantages and limitations of traditional technologies like CT and MRI Computed tomography (CT) and magnetic resonance imaging (MRI) are the cornerstone imaging modalities for the diagnosis of cerebrovascular accidents. Non-contrast CT (NCCT) is often the first-line imaging technique due to its widespread availability, speed, and ability to quickly rule out hemorrhage (Gilotra et al., 2023; Antón-Munárriz et al., 2023). However, its sensitivity for detecting early ischemic changes is limited, which can lead to diagnostic challenges in the acute setting (Tarnutzer et al., 2017; Antón-Munárriz et al., 2023). MRI, particularly diffusion-weighted imaging (DWI), offers superior sensitivity and specificity for detecting acute ischemic strokes and can provide detailed information about the extent and location of brain injury (Chennareddy et al., 2021; Gilotra et al., 2023). However, MRI is less accessible, more time-consuming, and may not be suitable for all patients, particularly those with contraindications such as implanted medical devices (Chennareddy et al., 2021; Deshpande et al., 2023). The integration of AI algorithms with these imaging modalities has shown potential in enhancing diagnostic accuracy and efficiency. AI-driven software can assist in the rapid detection of intracranial hemorrhages and ischemic changes, thereby expediting treatment decisions and improving patient outcomes (Chennareddy et al., 2021; Karataş et al., 2022; Deshpande et al., 2023; Gilotra et al., 2023). Despite these advancements, further studies are needed to validate the long-term benefits and feasibility of AI-assisted diagnostic systems in routine clinical practice (Chennareddy et al., 2021; Gilotra et al., 2023; Li et al., 2023). 3 Concept and Development of AI-Assisted Diagnostic Systems 3.1 Fundamentals of AI technology Artificial intelligence (AI) has become a pivotal technology in medical diagnostics, particularly through the use of machine learning (ML) and deep learning (DL) techniques. These technologies have shown significant promise in the field of medical imaging, where they are used to analyze complex imaging data and assist in the diagnosis of various conditions. For instance, deep learning, a subset of machine learning, has been widely applied in cardiovascular imaging to classify, segment, and detect abnormalities, demonstrating its potential to enhance diagnostic accuracy and efficiency (Wong et al., 2020; Wang et al., 2021). Similarly, AI algorithms, such as convolutional neural networks (CNNs), have been employed to detect large vessel occlusions (LVOs) in ischemic stroke patients, showing higher sensitivity compared to traditional methods (Murray et al., 2019). Despite these advancements, the practical application of DL in clinical settings remains limited due to several challenges. One major issue is the "black box" nature of DL models, which makes it difficult to understand their internal workings and ensure their reliability and interpretability (Wong et al., 2020). Additionally, the development of robust AI models requires large, high-quality datasets and rigorous validation to ensure their generalizability and effectiveness in real-world clinical scenarios (Hadjiiski et al., 2022). Addressing these challenges is crucial for the successful integration of AI technologies into routine medical practice. 3.2 System architecture and working principles AI-assisted diagnostic systems typically consist of several core components, including data acquisition, preprocessing, model training, and inference. The data acquisition phase involves collecting medical images from various modalities such as CT, MRI, and X-rays. These images are then preprocessed to enhance their quality and remove any artifacts that may interfere with the analysis (Zhu et al., 2019). The preprocessed images are used to train AI models, which learn to identify patterns and features associated with specific medical conditions. For example, CNNs are trained to detect LVOs in stroke patients by analyzing CT angiography images (Murray et al., 2019). Once trained, these models can be deployed in clinical settings to assist healthcare professionals in diagnosing and treating patients. The inference phase involves using the trained model to analyze new medical images and provide diagnostic recommendations. Some AI systems also include features for highlighting regions of interest in the images, making the diagnosis more transparent and interpretable for clinicians (Kermany et al., 2018). Additionally, AI-assisted systems can be integrated with hospital information systems to streamline workflows and improve the efficiency of emergency care (Murray et al., 2019).
RkJQdWJsaXNoZXIy MjQ4ODYzNA==