Computer Aided Detection

Computer-aided detection (CAD) is a technique used by radiologists to help interpret medical images (mainly mammograms). In some cases, CAD stands for computer-aided diagnosis, which is misleading because CAD is not used for sole diagnostic purposes, only as a detection tool to assist radiologists in reading the images(1). CAD has been used in the detection of a variety of cancers such as breast cancer, prostate cancer, and lung lesions. The technique specifically is a computer system that scans an image (e.g., mammogram, CT-scan, MRI) and shows suspicious areas that could be the appearance of the medical condition.

Mainly radiologists interpret medical images, but radiologists are limited by three factors; “the nonsystematic search patterns of humans, the presence of structure noise in the image, and the presentation of complex disease states requiring the integration of vast amounts of image data and clinical information”(2). These factors can easily become inconsequential with the involvement of a computer-assisted system, thereby theoretically improving the performance of radiologists. This is the main reason why CAD systems have been developed and even though the true efficacy of these systems remains unknown, many researchers are conducting studies to determine whether or not CAD systems really improve radiologist performance.

Brief History

The concept of computer-aided detection was first reported by Winsberg et. al. in 1967. Winsberg and his team examined the use of computers to analyze the detection of abnormalities on mammograms(1). In 1972, Ackerman and Gose expanded the concept of CAD by classifying four properties a computer could use to detect lesions on mammograms(1): calcification, spiculation, roughness and shape(3). Many other features have been defined since this initial classification(1). The ImageChecker CAD system, produced by R2 Technology, Inc, was approved by the FDA for radiographic mammography in June 1998 and for digital mammography in 2001(4). In 2001, approximately 130 units were in use, mainly at academic centers and in 2005, it was reported that more than 1600 CAD units were in active use at a multitude of clinical practices. CAD units identify lesions by placing a mark or symbol on the suspicious area. This does not indicate that the area is cancerous, but rather indicates an area of suspicion to which a radiologist should review(1). A variation of CAD software that has developed over recent years is GridCAD. GridCAD allows the user to compare the image to an image database, to view images from the database, and to run more than one CAD algorithm to the set of images(5).

References:

1. Ulissey MJ. Mammography - Computer - Aided Detection. 2005; Available at: http://www.emedicine.com/radio/topic879.htm. Accessed Nov/19, 2007.
2. Giger ML, Karssemeijer N, Armato SG,3rd. Computer-aided diagnosis in medical imaging. IEEE Trans Med Imaging 2001 Dec;20(12):1205-1208.
3. Houssami N, Given-Wilson R. Incorporating new technologies into clinical practice without evidence of effectiveness in prospective studies: computer-aided detection (CAD) in breast screening reinforces the need for better initial evaluation. Breast 2007 Jun;16(3):219-221.
4. Morris T. R2 Unveils First System to Optimize CAD Display for Digital Mammography. 2007; Available at: http://news.thomasnet.com/companystory/477381. Accessed Nov/19, 2007.
5. Pan TC, Gurcan MN, Langella SA, Oster SW, Hastings SL, Sharma A, et al. Informatics in radiology: GridCAD: grid-based computer-aided detection system. Radiographics 2007 May-Jun;27(3):889-897.