Abstract:
Aims: Assess AI methodologies in early disease detection, identify application areas like imaging and prognosis, determine key findings and clinical impacts, explore AI's prospects for accuracy and personalization, and analyze implementation challenges.Methodology:This narrative review examines AI's role in medical diagnostics by analyzing peer-reviewed articles from 2019-2024 from PubMed. A comprehensive search strategy identified 338 publications, which were filtered for relevance, resulting in 10 key studies. Focus areas include AI techniques like machine learning and their applications and challenges in disease diagnosis.Results:Results showed that in 2024, notable AI studies included a large experimental study with 27,558 samples, a comparative study with 1,653 samples, and a cross-sectional study with 20 samples. Previous years saw a retrospective cross-sectional study with 90,000 samples, an observational study with 1,052 samples, and a retrospective study with 219 samples. AI techniques featured EfficientNet-B2, CNNs, VGG-16, and ResNet variants, with transfer learning models like VER-Net and methods such as COVID-DSNet. Advancements highlighted EfficientNet-B2's 97.57% accuracy in malaria detection, VER-Net's superior lung cancer detection, and AI’s effectiveness in diagnosing retinal diseases, heart conditions, diabetic nephropathy, and COVID-19. Challenges included computational demands and dataset needs, with recommendations for model optimization and clinical integration.Scientific Novelty:This review is about the integration of advanced AI techniques in disease diagnostics, showing new algorithms and machine learning models that improve accuracy, speed, and personalized treatment strategies in medical practice.Conclusion:This study has shown that there has been a significant progress in AI-based disease diagnostics, with examples of high performing models such as EfficientNet-B2 and VER-Net. Despite challenges like computational requirement and interpretability of the model, AI has the potential to revolutionize diagnosis.
