A review of forty research papers that discuss lung cancer detection technologies highlights the gaps in the various approaches to the diagnosis of this potentially lethal disease and reveals how research might be targeted to improve detection and thus prognosis. Writing in the International Journal of Bioinformatics Research and Applications, Malayil Shanid and A. Anitha of the Information & Communication Engineering department at Noorul Islam Centre for Higher Education, in Kanyakumari District, Tamil Nadu, India explain the context of their review and its implications.
Lung cancer is one of the biggest killers of the modern age. Cancer is the second leading cause of death globally and is responsible for an estimated 10 million or so deaths annually, which amounts to 1 in 6 deaths. Of that approximately 10 million cancer deaths, about 1 in 5 is due to lung cancer. As with most cancers, early detection can greatly improve the prognosis of the disease, assuming appropriate treatment is available and undertaken. It also allows less invasive treatments to be employed, particularly reducing the level of surgery required, for instance.
Image processing coupled with machine learning has led to many improvements in the identification of malignant tissue in scan images for a wide range of diseases including lung cancer. The various techniques commonly look to distinguish between benign and malignant lesions seen in the scan. Computerised tomography is the tool of choice for detecting pulmonary nodules that might sit in either camp. A benign nodule can be treated relatively easily in contrast to a malignant one, which may develop rapidly and metastasize if not treated quickly.
Shanid, M. and Anitha, A. (2020) ‘An exhaustive study on the lung cancer risk models’, Int. J. Bioinformatics Research and Applications, Vol. 16, No. 2, pp.151–172.