Diagnosing Pulmonary Embolism Using Artificial Neutral Network

Abstract

Pulmonary Embolism (PE) is a blockage of an artery in the lungs usually from tumors that have invaded the circulatory system or from other sources such as amniotic fluid, air, fat, bone marrow, and foreign substances. This kind of disease is rare but life-threatening. A patient with PE needs immediate medical attention that is why prompt diagnosis and therapy are necessary. The standard way of diagnosing PE is through the use of Lung Scan. The lung scans are analyzed by radiologists or nuclear medicine physician. However, there are only a few of them and other hospitals and clinics still need to send the images to a radiologists and nuclear medicine physician. The delivery of the image may consume some time and the condition of the patient who actually has PE might become critical. A possible solution to this problem is to automate the diagnosis even without the presence of the specialist through the use of Artificial Neural Network (ANN). ANN was trained to diagnose PE. The ANN was evaluated and the result shows that the system can diagnose lung scan of patients with suspected PE with an accuracy rate of 60 percent.