To design and develop a World Wide Web-based education program that will allow trainees to interactively learn and improve the diagnostic capability of detecting pulmonary nodules on chest radiographs.

Chest radiographs with known diagnosis were retrieved and selected from our institutional clinical archives. A database was constructed by sorting radiographs into three groups: normal, nodule, and false positive (i.e., nodule-like focal opacity). Each nodule was assigned with the degree of detectability: easy, intermediate, difficult, and likely missed. Nodules were characterized by their morphology (well-defined, ill-defined, irregular, faint) and by other associated pathologies or potentially obscuring structures. The Web site was organized into four sections: study, test, record and information.

The Web site allowed a user interactively to undergo the training section appropriate to the user's diagnostic capability. The training was enhanced by means of clinical and other pertinent radiological findings included in the database. The outcome of the training was tested with clinical test radiographs that presented nodules or false positives with varying diagnostic difficulties.

A Word Wide Web-based education program is a promising technique that would allow trainees to interactively learn and improve the diagnostic capability of detecting and characterizing pulmonary nodules.