Topic

Image storage in Radiotherapy

Susan Chadwick
Susan Chadwick • 11 February 2025
10 comments

Hi all. In radiotherapy we are required to keep patients radiotherapy records for 30years and I am concerned that we store more imaging data than is necessary, having implications on the carbon footprint for radiotherapy. 

It is not unusal for a RT patient to have a repeat planning CT scan following laxatives, may have scans with full and empty bladder, or scans may be reprocessed for metal artefact reduction, extended FoV etc. We may also store copies of diagnostic scans that have been used to assist with tumour delineation.  In addition daily on treatment imaging is stored.  All these images are left on the planning system and presumably stored for 30+years.

I would have thought it is really only necessary to store the scan that was used to plan the treatment on. Does anyone know if there has been any research into the carbon implications of long-term image  storage, or guidance on what is strictly necessary to be stored for 30 years

Comments (10)

Gerry Lowe
Gerry Lowe • 2 months ago

Hi Susan

I don't have anything to add, but I'm interested in this topic - have briefly discussed with Rob - could you keep me informed, please? Thanks ... Gerry

Huw Walters
Huw Walters • 2 months ago

I've been looking at the financial and carbon costs of storing data over the past few months in our radiology department, and have found it quite challenging due to a lack of good data on how much energy is required to store a set volume of data. We're hoping to publish our findings soon, but as a brief summary:

Estimates vary quite a lot on the energy required to store 1 GB data, ranging from 0.1 kWh/GB to 7 kWh/GB (the higher figure is often quoted but actually quite poorly sourced and I can't find evidence to back it up). I think it's reasonable to assume the energy required is at the lower end, based on how much companies like google charge to store data in the cloud. If the figure was as high as 7 kWh/GB they would be spending vast amounts of money on electricity and it most likely wouldn't be cost effective as a business model.

From our rough calculations it would be approximately 30g of CO2 from storing a CT CAP for a year. Not a huge amount, but obviously these are high volume studies so this adds up quite quickly. The longer we store them, the higher the energy cost.

Ultimately the data is fairly lacking but that's a rough ballpark estimation. This doesn't include the impacts of data transfer from data centres etc and assumes power is coming from the national grid.

Hope that helps :)

Gerry Lowe
Gerry Lowe • 3 weeks ago

Really interesting, yes would really like to see your paper when published. What time period is assumed for the 0.1kWh/GB storage figure? Many thanks - Gerry

Susan Chadwick
Susan Chadwick • 2 months ago

Thanks Huw. Do please post a link when this is published.
I think part of the problem for radiotherapy is the legal requirement to keep RT data for a minimum of 30 years, coupled with the fact that, in my department a least, more imaging data than is necessary is being stored,. Also RT planning scans are becoming ever more complex, eg 4d scans are common whereas 20 years ago ant and lat radiographs might be all that were used.

Huw Walters
Huw Walters • 2 months ago

Do you know if that 30 year rule applies to general radiology as well? We're looking at reducing what we save without necessarily losing data (e.g. getting rid of the MPRs which can always be reconstituted from the thin slices).

Susan Chadwick
Susan Chadwick • 2 months ago

Don't quote me, but I think it is specific to radiotherapy because we need to be able to access details of previous radiotherapy dose distribution should the patient require further radiotherapy down the line

Please log in or sign up to comment.