Development of an AI-Powered Decision Support System for Early Diagnosis and Treatment of Neurological Disorders

سال انتشار: 1402
نوع سند: مقاله کنفرانسی
زبان: انگلیسی
مشاهده: 107

نسخه کامل این مقاله ارائه نشده است و در دسترس نمی باشد

استخراج به نرم افزارهای پژوهشی:

لینک ثابت به این مقاله:

شناسه ملی سند علمی:

AIMS01_210

تاریخ نمایه سازی: 1 مرداد 1402

چکیده مقاله:

Introduction: Stroke, Alzheimer’s disease, and other neurodegenerative conditions afflict millionsof individuals every year and place a heavy financial strain on healthcare systems aroundthe globe. Better patient outcomes and lower healthcare expenditures are possible with earlierdiagnosis and treatment. This scoping review aims to investigate the present level of research intothe implementation of AI in the design of decision support systems for the prompt detection andtreatment of neurological illnesses.Methods: We used PubMed, Embase, and Scopus to perform a systematic review of the relevantliterature. Included were studies that looked into how AI can be used to improve decision supportsystems for the early detection and treatment of neurological illnesses. The primary results wereAI system type, system accuracy, and clinical applicability.Results: Our systematic review found ۳۸ studies that met the inclusion criteria. There were ۱۹ onstroke diagnosis, ۱۱ on Alzheimer’s disease, and ۸ on other neurodegenerative diseases. The majorityof studies (n = ۲۷) developed decision support systems using machine learning algorithms,with the remaining studies using deep learning (n = ۷) or hybrid models (n = ۴).The accuracy of AI-powered decision support systems ranged from ۷۲ to ۹۹.۲% across all studies,with an average accuracy of ۸۹.۷%. Deep learning algorithms produced the highest accuracy instudies. The clinical applicability of the systems was reported in ۱۹ studies, with ۱۶ demonstratingthe ability of AI-powered decision support systems to improve early diagnosis and treatmentof neurological disorders.Twelve of the studies on stroke diagnosis reported on the use of AI-powered decision supportsystems to predict patient outcomes. In all of these studies, the AI-powered system was able toaccurately predict patient outcomes such as stroke severity and functional disability. Furthermore,three studies reported on the use of artificial intelligence to develop personalized treatment plansfor stroke patients.Seven of the Alzheimer’s disease studies reported on the use of AI-powered decision supportsystems to aid in early diagnosis, with the remaining four focusing on disease progression prediction.AI-powered systems achieved accuracy ranging from ۷۸% to ۹۶.۸% in studies aimed atimproving early diagnosis. The accuracy of AI-powered systems in studies aimed at predictingdisease progression ranged from ۷۴% to ۸۹.۷%.Finally, AI-powered decision support systems were used in eight studies that focused on otherneurodegenerative conditions to aid in early diagnosis and treatment. The accuracy of these systemsranged from ۷۲% to ۸۴%.Conclusion: There is hope in the application of artificial intelligence (AI) to create decision-supportsystems for the early detection and treatment of neurological illnesses. Algorithms based onmachine learning can analyze massive volumes of data and spot patterns that would be invisibleto humans. By facilitating earlier diagnosis and individualized treatment regimens, these technologiesmay enhance patient outcomes. However, big datasets and cooperation between physicians,researchers, and data scientists are necessary for the creation and validation of these systems. Overall, our findings indicate that AI-powered decision support systems have the potential toimprove the early detection and treatment of neurological disorders. More research is needed,however, to determine the best AI algorithms and to validate these systems in larger, more diversepatient populations.

نویسندگان

Fariba Nakhaei

Faculty of Medicine, Birjand University of Medical Sciences, Birjand, Iran

Ali Rajabpour-Sanati

Faculty of Medicine, Birjand University of Medical Sciences, Birjand, Iran

Hamidreza Riasi

Faculty of Medicine, Birjand University of Medical Sciences, Birjand, Iran