Computer-aided detection (CAD) combined with multidetector CT (MDCT) improves radiologists’ ability to detect solid lung nodules early enough for them to be treated without increasing interpretation time, according to a recent study published in the October issue of the American Journal of Roentgenology.
“The comparison of a current examination with prior examinations is a time-consuming and tedious task,” said Philippe Grenier, MD, lead author of the study. “This study wanted to evaluate the potential of a computerized automated system to improve human efficiency in this way, and determine whether CAD systems improve the detection of actionable lung nodules,” he said.
The study, conducted by researchers at Hopital Pitie-Salpetriere in Paris, sought to “assess detection, tracking, and reading time of solid lung nodules [at least] 4 mm on pairs of MDCT chest screening examinations using a computer-aided detection system.”
Of 54 pairs of low-dose MDCT chest exams, two chest radiologists interpreted all paired examinations using the CAD workstation—first without and then with CAD input—to detect and track 52 lung nodules that were 4 mm or larger in 25 exams. A subset of 33 examination pairs was later read on a clinical workstation used in daily practice, and the results were compared for reading time with those on the CAD workstation.
According to the results, one cancer was initially missed by one radiologist but was correctly identified with CAD input. “On average, readers spent 4-5 minutes per case to read the paired exams on CAD and 6-8 seconds per CAD mark. The CAD system successfully matched 91.3 percent of nodules detected in both exams. The overall rate of available CAD growth assessment was 54.9 percent of all nodule pairs.”
The sensitivity of radiologists to detect lung nodules greater than 4 mm increased significantly with CAD input, according to the researchers.
Grenier said the results demonstrate “the added value of CAD systems as a second reader. CAD was sensitive allowing us the potential to assess more accurately the growth of indeterminate nodules, without compromising the reading time,” he said.