Hiroshi Moriya MD, Presenter: Nothing to Disclose

Manabu Nakagawa, Abstract Co-Author: Nothing to Disclose

Koutarou Sakuma, Abstract Co-Author: Nothing to Disclose

To explain the clinical utility of a computer-assisted detection system (CAD) for pulmonary metastases of extra-thoracic-cancer patients. 1. To show the false-positive CT findings in CAD system. 2. To show the false-negative nodules in CAD system.  

[Materials and Methods] Consecutive 43 cases with post-operative status of digestive organ cancer or breast cancer. CT scanner:Aquilion ONE,plain chest scan, conventional dose,FC17/FC13. Computer-assisted lung nodule detection system:Xelis lung(effective diameter:1mm). Radiologist-detected-nodules and CAD-detected-nodules were compared. [Results] 995 nodules(diameter >3mm:536, 3mm>:459) of 43 cases were evaluated. When limited to the size of 5-10mm, sensitivity was 89%, and positive predictive value was 85%. [Conclusion] CAD detected a large number of nodules less than 3 mm. As a result, there was an increase in false positives. And, there were some large nodules in false negative of CAD, however, there was no oversight of the radiologists. Detection algorithm is quite different from the thinking patterns of radiologists, CAD can be used as a supportive system for pulmonary nodule screening.

http://abstract.rsna.org/uploads/2014/14010779/14010779_zr4l.pdf

Moriya, H, Nakagawa, M, Sakuma, K, Misleading Findings at a Chest CT Computer-assisted Detection System.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14010779.html