The Analytical Tool of Artificial I ntelligence, A Method For Early A nd Accurate Detection of Thyroid Cancer: A Meta- A nalysis

Thyroid cancer has seen an increase in both diagnosis rates and prevalence in recent years, making it the third most common malignancy among endocrine glands. This review investigates the potential of artificial intelligence (AI) in addressing the diagnostic challenges associated with thyroid cancer. Methods: This study examines the application of AI algorithms for cancer diagnosis by analyzing clinical data and imaging through a meta-analysis and comprehensive data evaluation. Results: Early and accurate diagnosis of thyroid cancer is crucial for improving clinical treatment outcomes. Conventional diagnostic methods, such as clinical examinations, imaging tests, and biopsies, often lack precision. The differentiation between benign and malignant nodules remains a significant challenge in clinical practice. Recent advancements in AI, including models such as ThyNet , Support Vector Machines (SVM), and Deep Neural Networks (DNN), are increasingly being applied as diagnostic tools for thyroid cancer. However, each of these models has limitations. For example, DNNs face challenges related to interpretability, while SVMs show lower accuracy in differential diagnosis. Consequently, there is a demand for models with higher accuracy and lower error rates. Our review suggests that ThyNet , a model based on convolutional neural networks (CNN), is a promising tool for thyroid cancer diagnosis. ThyNet enables precise analysis of medical imaging data and has been evaluated using metrics such as accuracy, sensitivity, specificity, and positive predictive value. Results indicate that the ThyNet AI model achieved an overall accuracy of ۹۲% in distinguishing between benign and malignant nodules. This accuracy is approximately ۸% higher than that of experienced radiologists, highlighting the model’s superior performance in nodule classification and its potential in reducing misdiagnoses. In a simulated trial, the number of fine-needle aspirations (FNAs) decreased from ۶۱?fore using the AI-assisted strategy to ۳۵?ter its implementation. This reduction suggests that AI can significantly reduce unnecessary FNAs when applied as a novel diagnostic approach. Conclusion: Based on a review of ۱۳۰ articles published from ۲۰۱۷ to the present, our study concludes that the use of AI-based models, such as ThyNet , leads to more accurate diagnoses and optimized treatment decisions for thyroid cancer patients. However, it is important to acknowledge that ThyNet is a relatively expensive model, which may limit its accessibility in certain clinical settings.

کلیدواژه ها:

Thyroid cancer ، Early diagnosis ، Artificial intelligence ، Depth learning ، ThyNet