Computer aided diagnosis for an earlier and more accurate diagnosis of lung cancer

The substantial workload of today’s physicians has led to an increase in false and late diagnosis. Computer aided diagnosis (CAD) has the potential to solve this problem. CAD checks digital images for suspicious areas and offers input to narrow the physician’s focus. This improves accuracy of diagnosis, assists in early detection of diseases, and reduces exam evaluation time. We will analyze the technical features and the ethical implications associated with the use of CAD for lung cancer. CAD is an interdisciplinary technology that utilizes artificial intelligence developed by computer engineers and bioengineering image processing. We will focus on the bioengineering contribution. CAD of lung cancer follows four main steps: segmentation of lung cancer, detection of nodules inside the lung fields, segmentation of detected nodules, and diagnosis of nodules as malignant or benign. Ethical concerns regarding software engineers’ medical obligation to patients pose a potential problem. There is also doubt about CAD’s sustainability as a technology because it needs to be continuously updated and it is not extremely cost effective. Although these dilemmas must be addressed, the benefits of CAD are worth the potential drawbacks. For example, the detection rate of lung cancer using CAD is 2.610 times greater than when using traditional practices. This technology has the potential to significantly benefit cancer patients because it increases the accuracy of early diagnosis, an important factor in cancer treatment. Key words-Lung Cancer Diagnosis, Nodule Segmentation, Bioimaging, Computer Aided Diagnosis, Lung Tumor Detection WHY COMPUTER AIDED DIAGNOSIS? When a patient is diagnosed with late stage cancer, they begin their fight with a much lower likelihood of survival. The current early stage diagnosis rate of lung cancer is only 16%, and the survival rate for patients with tumors that have spread to organs other than the lungs is only 4% [2]. Early diagnosis is often responsible for more effective treatment and survival therefore, oncologists need new technology that will increase their early diagnosis. The concept of automated diagnosis (CAD) originated early in the 1980’s and has continued to develop since then [1]. Today it provides a potential solution to the oncology field’s problem of late diagnosis. CAD allows physicians to diagnose their patients during earlier stages of their disease, when treatments are more effective. Considering lung cancer specifically, CAD segments each individual nodule of each lung and determines which sections are likely to be malignant. The physician can use this information to quickly find the affected sections of the lungs and eliminate time previously wasted examining healthy sections. In addition to allowing for earlier diagnosis, the use of CAD also decreases the number of misdiagnoses by using complex algorithms to determine the probability of a detected structure (nodule) being malignant or benign. Figure 1 shows some subtle missed cancer lesions (indicated by the circles), which were detected correctly by the CAD system.