Use of "Big Data" approaches using spatial dose objects
The classic tools used in radiation oncology to evaluate dose distributions do not quantify the spatial relationship of dose levels to structures. So if the response of the body depends not just on how big the dose is but also where it is, classic tools to not make that easy to detect. For example when the lung is irradiated, risk of pneumonitis may depend on how close the radiation is to the heart or in what part of the lung it is located; the classic tools will only tell us how much dose the lung got not where it was with respect to other structures. We demonstrated use of spatial-dose objects that let us measure these relationships. For this project we want to apply these tools to large patient sets and use Hadoop based analytics to identify patterns in response linked to the spatial dose objects.