Novel approach for detection of lung cancer in computed tomography images

Abstract

Cancer is a dangerous disease threatening people in all age groups newlineand the second major reason for death worldwide. Cancer cells can generate and spread to any part of our body. The cancer cells generating primarily in lungnodules are known as lung cancer and it is one of the leading causes of cancerdeath in both men and women. The early screening of the disease improves thesurvival rate of the patients and helps to decide the mode of treatment to beprovided. The advancement in the Computed Tomography (CT) scan helpstovisualize small lesions in the lung to detect and monitor the stages of lungcancer. People at high risk for lung cancer and mild symptoms should berecommended for annual screening of low dose computed tomography to easeearly detection and to improve their survival rate. The abnormal cancer cells inthe lung region are separated into two groups as Small Cell Lung Cancer(SCLC) and Non-Small Cell Lung Cancer (NSCLC). Small cell lung cancer ismostly found in frequent smokers and it addresses only about 20% of lungcancerdetected. Non-Small Cell Lung Cancer addresses several other categoriesof lung cancer; it includes adenocarcinoma (AC), squamous cell carcinoma newline(SCC), and large cell carcinoma (LCC). The Advancement of CAD in medicalimaging increases the accuracy of detection and helps to pinpoint the locationsof tumors. By improving the proficiency of the classifier employed in medicalimage processing, the accuracy of cancer detection can be improved. For theaccurate discrimination of lungs, a group of features is needed. While usingmore features more computational cost is required. To provide fast, accurate, andreliable lung cancer classification for clinical aided diagnosis and other potentialapplications, a new method is proposed. newline newline