If you all are looking for additional data sets to perform testing you might try checking out the CIP Survey of Biomedical Imaging Archives (https://wiki.nci.nih.gov/display/CIP/CIP...).  We have been trying to compile a list of what's out there for a couple years now and one of the criteria we evaluate is whether the archive has an explicit data sharing policy.

The Cancer Imaging Archive (http://cancerimagingarchive.net) is one in the list there which has a very large variety of DICOM data from over 25 different institutions so there are lots of different scanners/vendors/modalities.