Computerized System for Early Detection of Lung Cancer

Identification and tracking are based on computer analysis of size, shape, density, location, and other nodule and body information contained on thin-section computed tomography (CT) scans.

The new system also automatically re-identifies nodules seen previously, so that changes in size can be monitored to assess a patient's response to treatment. Modern CT scans consist of 400 or more images, whereas the older exams produced only about 40. This has resulted in a ten-fold increase in the number of images for a radiologist to interpret and creates a potential for missing very small nodules. The computer-based system is specifically designed to detect micronodules as small as 3 mm. The system is being developed at the radiological sciences department of the School of Medicine at the University of California, Los Angeles (UCLA, USA).

The system supports the use of an interactive pen display, called Cintiq 18sx. With this, radiologists can contour nodules and make handwritten annotations on CT images using an electronic pressure-sensitive stylus instead of a mouse to draw directly on the screen. The pen operates without batteries or connecting cords by taking advantage of an electromagnetic signal sent from a sensor board under the display.

"We believe computer-assisted nodule detection will play a crucial role in lung cancer screening and follow-up in reducing user analysis time and assuring that micronodules are not missed among the many normal lung structures such as blood vessels and airways,” said Dr. Matthew S. Brown, assistant professor, radiological sciences, UCLA School of Medicine.