this is for holding javascript data
- If a financial claim is done, it must be substantiated. The costs of having N phantom subjects travel and be scanned at each site of a multi-centric study are certainly prohibitive, and it cannot be assumed that it is necessarily lower than harmonizing protocols and acquisitions, as well as scanning geometric phantoms at each site. The authors site ADNI as an example, and it is true that ADNI has gone through an extensive and expensive process of harmonizing sequences, building a phantom, and scanning the phantom before each scan. However, those are now sunk costs, which, thanks to ADNI, need not be borne by every new study : as a case in point, there is a large body of projects that now use the ADNI protocol as is, or with only slight variations, and thereofre needing a fraction of the cost to harmonize protocols. Further, ADNI itself has moved away from scanning geometric phantoms before each scan. It is an open question in the literature what the right frequency of phantom scan is to properly sample scanner drift, however the current rule of thumb falls between weekly and monthly; for a 15 minute scan, this is not an overly expensive proposition.
