Synthetic Lung Tumor Data Sets for Comparison of Volumetric Algorithms.

We present a set of synthetic lung tumor data in which synthetic tumors of known volume are embedded in clinical lung computerized tomographic (CT) data in different background settings in the lung. Because the change in pulmonary nodules over time is an important indicator of lung tumor malignancy, it is important to be able to accurately measure changes in tumor size from measurements made at different times and possibly with different equipment. Standardized lung tumor data sets can be used to calibrate the differences between sets of scans as well as accurately compare volumetric measurement techniques. Our standard data sets combine the usefulness of phantom data with the clinical challenges of realistic CT scans.