هوش مصنوعی و پردازش تصاویر : مرز نوینی در پیش بینی عفونت محل جراحی
کد: G-1717
نویسندگان: Kimia Zeraatkar * ℗, طیبه بنی اسدی, اقلیما توکلی
زمان بندی: زمان بندی نشده!
برچسب: پردازش سیگنال های پزشکی
دانلود: دانلود پوستر
خلاصه مقاله:
خلاصه مقاله
Background and Aims: Surgical Site Infections (SSIs) are a significant source of postoperative complications, frequently diagnosed at advanced stages due to reliance on physical examination. Delayed detection leads to elevated morbidity, mortality, and prolonged hospital stays. Leveraging artificial intelligence (AI) and medical image processing presents a transformative solution for early SSI detection during the subclinical phase, enabling timely interventions. This study conducted a scoping review to evaluate the integration of image analysis and deep learning models (DLMs) for SSI prediction and prevention. Methods: A comprehensive literature search was performed in December 2024 across databases including PubMed, Scopus, and Google Scholar using keywords such as “medical image processing,” "Image Processing, Medical," “surgical site infection,” “surgical wound infection,” and “artificial intelligence.” Studies employing image processing methodologies for SSI prediction were included. Screening and data extraction were independently conducted by two reviewers. Thematic analysis synthesized insights, focusing on algorithmic performance, clinical outcomes, and implementation challenges. Results: NLP algorithms achieved 93% recall and 82% precision, outperforming manual chart reviews for detecting intra-abdominal SSIs. Deep learning models using preoperative CT imaging achieved 92% accuracy, with sensitivity and specificity of 76.7% and 53.3%, respectively, enhancing predictions of surgical complexity and SSI occurrence. Thermal imaging techniques demonstrated 92% accuracy and 100% specificity, effectively identifying abnormal wound healing. Smartphone-based applications using segmentation algorithms achieved sensitivity and specificity rates of 83% and 69%, offering scalable diagnostic potential for resource-limited environments. These advancements enabled early SSI detection, improving patient outcomes by reducing morbidity, mortality, and hospital stays. Preoperative imaging facilitated surgical planning, personalizing strategies and minimizing intraoperative complications. Remote applications expanded diagnostic accessibility, addressing healthcare disparities in underserved regions. Challenges include variability in image quality, algorithmic bias, and limited standardized datasets, underscoring the need for interdisciplinary collaboration and further research to optimize implementation. Conclusion: AI-driven image processing is revolutionizing SSI detection, offering unprecedented accuracy and accessibility, particularly in remote settings. Multimodal data frameworks, standardized datasets, and interdisciplinary approaches are essential for advancing integration into surgical care, reducing costs, and enhancing global healthcare equity.
کلمات کلیدی
Medical Image Processing, Surgical Site Infection
