This paper presents the design of a temporal image database system and its application in thoracic imaging. The design of this information system is based on the client/server architecture. The system consists of a chest imaging database server, a library of image processing modules, a link to the picture archiving and communication system (PACS) archive, and a low end client workstation with motif-based graphic user interface (GUI). The database system can be used to aid the radiologists in quantitating solitary or multiple long nodules and in assessing effectiveness of therapeutic procedures for these lung cancers. The GUI allows a user to retrieve any patient study from PACS. After a nodule is visually identified, it will be segmented automatically to obtain relevant features, such as the center of mass, volume, and surface area. Such 3D nodule information, together with the patient textual information, is subsequently organized in the chest imaging database to facilitate outcome analysis.