2014 Volume 32 Issue 2 Pages 98-108
The performance of computer-assisted detection (CAD) software depends on the quality and quantity of the dataset used for supervised learning. If the data characteristics differ in development and practical use, the performance of CAD software will be degraded. Therefore, it is necessary to continuously collect data for supervised learning in practical use, and to improve CAD software by retraining with the collected data. In this paper, we describe the development, clinical use, and continuous performance improvement of CAD software based on multi-institutional collaboration in teleradiology environment. We developed additional functions of the web-based CAD software processing and evaluation platform (CIRCUS CS) for implementing into teleradiology environment, and a multi-institutional study has been started since September 2011. We investigated the performance improvement of CAD software for each institution based on retraining through a simulation-based study. According to the results, the performance of CAD software for each institution was improved by retraining. These results suggest that the multi-institutional data collection has potential to accelerate development and clinical use of CAD software for various institutions.