Computer Aided Diagnosis System for Lung Cancer Detection: a Survey

Please note! This essay has been submitted by a student.

Download PDF

Cancer is one of the most deadly diseases worldwide. Lung cancer is the major type of cancers which increases the death rate. According to The American Cancer Society about 234,030 peoples have been suffering from lung cancer. It can be cured if it is detected in the early stage which will increase the survival rate. Several methods have been used to detect the cancerous pulmonary nodules. Computer Aided Detection (CAD) is the computational diagnostic tool used to detect pulmonary nodules and there is an immense refinement in improving the performance of the CAD scheme. There are three main steps in detecting the cancerous nodules; they are segmentation, feature extraction and false positive reduction. In this paper, methods developed by the authors to detect the cancerous pulmonary nodules are exposed. This survey will help the researchers to use the existing methods as well as to develop new methods to improve the accuracy.

Essay due? We'll write it for you!

Any subject

Min. 3-hour delivery

Pay if satisfied

Get your price

Cancer is a disease of the cells in the body. There are different types of cell in the body, and different types of cancer which arise from different types of cell. Cancer affects when cells that are abnormal grow and spread very fast. Cancer cells usually group or clump together to form tumors. A growing tumor becomes a lump of cancer cells that can destroy the normal cells around the tumor and damage the body’s healthy tissues.

Among other types, lung cancer is the leading cause of death rates. In lung tumors can be benign or malignant. Cancer refers to the malignant tumors where benign tumors can be removed and it will not spread to other parts of the body. Lung cancer can be broadly classified into two types: small cell lung cancers (SCLC) and non-small lung cancer (NSCLC). This classification is based on the size of the cell in microscopic appearance. Mainly lung cancer contains four stages which are based on how far cancer has spread. This implies that early detection of cancer will improve the survival rate of the cancerous patients.

Initially, radiologist used chest x-ray images to diagnosis lung cancer but in x-rays, small lung nodules were missed. Those nodules may also be cancerous which will be dangerous. So that instead of x-rays Computed Tomography (CT) scans are used for diagnosis of lung cancer.

Computer Aided Diagnosis (CAD) is a technology designed to decrease observational oversights and thus the false negative rates of physicians interpreting medical images. It refers to software that analyses a radiographic finding to estimate the likelihood that the feature represents a specific disease process. Prospective clinical studies have demonstrated an increase in lung cancer detection with CAD assistance. Early detection of lung cancer is valuable thus CAD system improved the diagnostic performance of radiologists in a detection and diagnosis of lung nodules in CT images. A CAD system is highly depend on CAD’s performance thus a number of research works have been done on the improvement of CAD performance.

CT images can be collected from several medical image databases which may contain noise. Eliminating noise from the medical images will be a challenging task. Preprocessing of CT images will reduce the noise and enhance the image quality. It is also useful in removing unwanted parts in the background of the CT image. The methods used to remove the background part are known as segmentation.

In image processing, there are several segmentation techniques. Image segmentation is one of the important processes in image processing. It extracts the region of interest from an image. After segmentation feature extraction should be done. A CAD system is used in feature extraction and selection of suitable feature which will help the radiologist for further diagnosis. There are various types of features such as morphological, textural, fractal-based, topological, intensity-based features. Classification of suitable feature is the most important task to be done. Classifiers are used to improve the accuracy of the detection. There are several classifiers such as support vector machine, artificial neural network, k-nearest classifier, Bayes classifier.

Several studies show that the CAD system is very useful for the radiologist in the detection of cancer. CAD system mainly contains four steps they are data acquisition, preprocessing, segmentation, feature extraction, false positive reduction.

There are several medical image modalities available among that CT image is preferred for lung cancer detection. Publicly available databases were created to help the researcher among the world to utilize the large and diverse dataset to train and test their algorithms and models. Most commonly used databases are LIDC, LIDC-IDRI, and ANODE09.

Lung Image Database Consortium (LIDC) is a publicly available database which has been working since 2001 to develop the research resource for development, training, and evaluation of CAD. Seven academic centers and eight medical imaging companies collaborated to create this dataset which contains 1018 cases. These data are divided based on three categories (nodule>=3mm, nodule<3mm, non-nodule>=3mm). The combination of both CT and thoracic CT image will be known as LIDC-IDRI. An image repository of screening and diagnostic thoracic CT scans.

Associated metadata

It contains several images of different modalities which will help the researchers. [21] Automatic Nodule Detection 2009 (ANODE 09) and Lung Nodule Analysis 2016 are a database which contains CT images which help the researchers to validate, compare and evaluate the performance of their CAD system.

CT scans are difficult to diagnosis so that preprocessing is needed to improve the quality. Preprocessing techniques are used to remove the noise and to enhance the quality of the image. It mainly contains two steps, in the first step contrast of the image will be improved and in second step background noise will be reduced from the image. Preprocessing will help the radiologist to detect the nodules at an early stage.

Segmentation is the process of partitioning an image into a different region with similar properties such as gray level, texture, brightness, and color. In CT images it is essential to separate the suspicious areas from its surroundings. The separated area is known as Region Of Interest (ROI). Some of the techniques used for segmentation are discussed below.

Thresholding is the process of separates the pixel of an image into classes that are distinct based on a defined threshold. Define a threshold T and classify the output f(x,y) into a distinct group based on their relation to T.

Get quality help now


Verified writer

Proficient in: Impact of Technology, Medical Practice & Treatment, Oncology

4.8 (345 reviews)
“Writer-Justin was a very nice and great writer. He asked questioned as necessary to perform the job at the highest level. ”

+75 relevant experts are online

banner clock
Clock is ticking and inspiration doesn't come?
We`ll do boring work for you. No plagiarism guarantee. Deadline from 3 hours.

We use cookies to offer you the best experience. By continuing, we’ll assume you agree with our Cookies policy.