When used in conjunction with an experienced radiologist, the use of computer-aided detection (CAD) for lung nodule detection on chest radiograph with CT angiography (CTA) correlation presents very good sensitivity, specificity and accuracy, according to a study published online March 31 in the Journal of Digital Imaging.
William Moore, MD, and colleagues from the Stony Brook University Hospital in Stony Brook, N.Y., utilized a varied patient-population of those referred to CTA of the chest for the detection of lung nodules, to determine the sensitivity and specificity of a commercially available CAD system for detection of lung nodule on posterior-anterior (PA) chest radiograph. Noting that while chest radiographs usually identify lung nodules, there is a chance that lung cancer can be missed even though they may have been present on the initial study.
“Although there are several commercially available CAD platforms, the exact sensitivity and specificity of these devices have not been fully tested in a general population,” wrote the authors.
The researchers collected all CTA chest studies conducted from November 2007 to November 2008 for review and these studies were then cross referenced for patients who had a PA radiograph within a 24-hour time frame. After selection of 240 chest radiographs consisting of 86 male and 154 female patients ranging in age from 16 to 88 years, the studies were sent to an FDA-approved chest radiograph CAD system (IQQA-Chest, EDDA Technology) for review and all 240 cases presented a CAD result, said the authors.
Following review by CAD, a radiologist with five years of experience and 10 months of CAD experience reviewed all the CAD images and scored the findings on a five-point Likert scale. The CTA was then reviewed to determine if there were correlative nodules, explained the researchers, who noted that the presence of a correlative nodule between 0.5 cm and 1.5 cm was considered a positive result.
Moore and colleagues wrote that there were a total of 69 CTA-confirmed nodules in 49 patients from the CAD reading alone, two of which were determined to be malignant. There were a total of 165 CAD findings or regions of interest (ROIs) on the chest radiographs, of which 49 represented actual nodules seen by CTA.
In terms of the 165 CAD findings that were shown to the radiologist, 87 received a score of 1 or 2 by the radiologist and an average of 0.48 false-positive ROIs were observed per chest radiograph. Of the 240 cases, 121 had no CAD-suggested ROI and in the remaining 119 patients, 85 had false-positive ROIs, one case had four false-positive ROIs, six had three false-positive ROIs and 15 cases had two false-positive ROIs. The remaining 64 patients had one ROI.
“We found that this chest radiograph CAD system had a moderate sensitivity of 71 percent and a very low specificity at baseline. However, if the CAD was used as an interactive tool, we found a significant increase in the specificity to 78.1 percent, concluded the researchers. “Accuracy of 73.7 percent was achieved when the CAD device was used with an experienced radiologist.”