Session Item

Legacy treatment data are a necessary evil for many radiotherapy researchers. Legacy data are images, structure contours, and dose distributions associated with treatment plans that were created, and are likely still stored in historical, retired treatment planning systems. Legacy data are commonly required by those looking to study the link between radiotherapy exposures and late effects with long latency periods of years to decades. Other researchers, however, may also find themselves considering the prospect of resurrecting legacy data to bolster cohort numbers when studying rare diseases and/or effects, or to improve data variety for advanced analytical techniques like data mining and deep learning. Despite their value, working with legacy data can be intimidating and onerous.

We at St. Jude Children’s Research Hospital have found it necessary to leverage treatment data created as many as 25 years ago for all of the above-mentioned reasons. Our journey to recover and employ legacy data involved inconsistent software versions, outdated data formats, questionable data veracity, database crashes, and many other hurdles. This presentation will draw on our experiences and focus on the challenges associated with retrieving and using legacy treatment data, as well as strategies to overcome them. The objective will be to provide attendees with: realistic expectations for working with legacy treatment data, appreciation for the value of legacy treatment data, and strategies for overcoming, or avoiding entirely, some potential obstacles.