محمدعلي پورنوري

Considering the widespread preva‎lence of breast cancer as one of the most common cancers among women an‎d the importance of early detection in the success of treatment, the present study aims to develop a novel diagnostic model based on the integration of dynamic thermal change analysis an‎d machine learning algorithms in order to provide a complementary, non invasive, low cost, an‎d automated method. In this regard, dynamic thermal images are used to record thermal patterns an‎d physiological changes of breast tissue over time, an‎d by extracting the related quantitative features, suitable data are prepared for processing by various machine learning models (including neural networks, Support Vector Machines, an‎d Ran‎dom Forests). The main focus of the research is on the accurate differentiation between thermal patterns associated with healthy an‎d diseased individuals an‎d achieving accuracy, sensitivity, an‎d specificity comparable to conventional diagnostic methods. The conducted eva‎luations show that the proposed model, by utilizing the capabilities of dynamic data analysis an‎d the classification power of machine learning algorithms, has been able to provide high diagnostic performance an‎d has the potential to become an effective assisting tool in the screening an‎d early detection process. This study also examines the impact of environmental factors on the quality of thermal data, the possibility of integrating these data with other medical information, an‎d identifies the technical an‎d clinical challenges in the path toward the practical implementation of this technology. The results show that the proposed hybrid model, by achieving an area under the Receiver Operating Characteristic (ROC) curve of 0.98 an‎d an area under the Precision Recall curve of 0.97, has a very high capability in distinguishing healthy individuals from patients. Also, by achieving a sensitivity of 0.84 an‎d a specificity of 0.95, this system has demonstrated remarkable efficiency an‎d reliability as a non invasive screening tool.

تشخيص سرطان سينه؛ تصويربرداري حرارتي پويا؛ تغييرات دمايي پويا؛ يادگيري ماشين؛ يادگيري عميق؛ تحليل سري زماني؛ تشخيص زود هنگام.

Breast Cancer Diagnosis, Dynamic Thermography, Dynamic Thermal Changes, Machine Learning, Deep Learning, Time Series Analysis, Early Detection.

Mohammad Ali Pournouri

Dr. Mohammad Fathian