CAD Systems Shown To Be Effective in Colon Cancer Detection

New findings have demonstrated that in two different symptomatic study populations, a computer-assisted detection (CAD) system designed for colon cancer screening is effective for the localization of the frequently difficult to detect morphologically flat early colonic cancerous lesions, as well as later stage colorectal cancers.

Previous data have supported CAD's role in highlighting areas of abnormality with computed tomography colonography (CTC) in asymptomatic patients, and this new data support the system's use in symptomatic patient populations. The system used in the study is called ColonCAD, which was developed by Medicsight (London, UK).

Dr. Stuart Taylor, a consultant radiologist at University College Hospital, London (UK), presented his study's findings in November 2007 for the first time at the annual Radiological Society of North America (RSNA) meeting held in Chicago, IL, USA. He said, "These results are interesting because CT colonography is well-established for the detection of adenomatous polyps in asymptomatic patients and the role of CAD is becoming increasingly established, but the potential for computer-aided cancer detection in symptomatic patients has been relatively neglected. The results show that ColonCAD effectively aids the detection of colon cancer in symptomatic patients, in particular bringing the difficult to detect, morphologically flat lesions to the attention of the radiologist. With the rising global incidence of colorectal cancer it is essential that technology [be] continually improved to enable the physician to detect abnormalities accurately and in the least invasive way for the patient. The ColonCAD technology used in the study is validated against one of the world's largest and most population diverse databases of verified CT scan data.”

One of the studies involved 30 symptomatic patients who had undergone a diagnostic colonoscopy and who were undergoing cancer staging. Their tumors were all endoscopically classified as having a flat morphology and were located and characterized by three independent radiologists in conjunction with the endoscopic reports and imaging. Medicsight's ColonCAD software was then applied at three different settings of sphericity. The analysis showed system detected 83.3% of the morphologically flat cancers present. This type of lesion has historically been difficult to detect with CT colonography.

The second larger study included 59 symptomatic patients with already proven cancers. ColonCAD was used at four different filter settings and results based on the sensitivity and specificity at the different settings showed the system to be effective for the detection of these later stage cancers. In addition, results showed that as in the detection of polyps, optimal results for CAD detection of these cancers require scans to be performed in both supine and prone CT acquisitions.

Colorectal cancer is the second leading cause of cancer-related deaths in the United States. With increasingly sophisticated radiological imaging hardware such as multi-detector CT scanners, radiologists are facing a growing challenge in the amount of detailed patient image data that they must review for each patient examination. Some CT scan examinations generate as many as 1,000 images per patient. Review of this data by the radiologist is not only time-consuming but also prone to error due to reader fatigue. CAD software can help the reviewing radiologist by analyzing the image data and automatically highlighting suspicious regions of interest for closer inspection. Without CAD software some potential abnormalities or areas of disease may be overlooked. This is critical for diagnosis and the management of patient outcomes as early detection of disease greatly increases the probability of successful treatment and a positive therapeutic outcome.

Medicsight's ColonCAD and LungCAD software use an advanced CAD algorithm to analyse CT scans of the colon and lung and automatically highlight suspicious areas that may be indicators of disease. CAD may highlight areas easily overlooked by the reviewing radiologist, such as small lesions or regions that are hidden from view behind folds in the colon or normal structures and surrounding tissue in the lung.

Both CAD systems integrate with the advanced 3D visualization platforms of industry's leading imaging equipment partners. The integrated systems provide sophisticated image viewing capabilities, including 3D reconstructed image data, with the added advantage of demonstrating automatic CAD findings to assist clinical end users in the detection and analysis of disease. This allows clinical end users to perform either a ‘second read', where CAD findings are displayed to the user after completion of an initial review of the CT scan data, or a ‘concurrent read' where CAD findings are displayed during the user's initial review of the original CT scan images.


Related Links:
Medicsight