Mesothelioma is a malignant tumor mostly seen in the membranes of heart and lungs. The exposure of these organs to substances such as asbestos and erionite causes mesothelioma disease. As a result of the deformation in these organs, shortness of breath, chest or back pain, cough and similar complaints occur. Because the symptoms of mesothelioma overlap with those of many other diseases, diagnosing the disease can be difficult and time-consuming. The goal of this study is to design an artificial intelligence-based early diagnosis system for mesothelioma disease. Two alternative neural network (NN) algorithms were utilized for this, and their results were analyzed. The performances of artificial neural network (ANN) and convolutional neural network (CNN) models were compared. F-measure rates for the designed ANN and CNN architectures were measured as 95% and 98%, respectively. The results showed that NN-based methods can be used in the early diagnosis of the disease. The software that will be built based on this model is expected to assist physicians in their decision-making processes.
|Subjects||Engineering, Basic Sciences, Medicine|
|Journal Section||Research Articles|
|Publication Date||April 30, 2022|
|Published in Issue||Year 2022, Volume 2, Issue 1|