How a case gets published
Every case in this library is somebody's patient. What follows is the whole route a case takes from the angio suite to a page you can read.
A verified contributor writes the case
Browsing DOTTIR is open to everyone. Contributing is not. Before an account can upload anything, we check it against professional registration and society membership — the GMC number, the BSIR or CIRSE membership, the institution. Until that check passes, the account simply cannot contribute.
A verified contributor then writes the case up in structure rather than prose: presentation, indication, planning, access and approach, technique, intra-procedural findings, complications, outcome, follow-up, discussion. The shape is deliberate. A case written as a paragraph is a story; a case written into those fields is something another interventionalist can compare against their own.
They anonymise the imaging before it leaves the hospital
Contributors are asked to anonymise imaging before uploading it. That is the correct order — the person who was in the room is the only one who knows what is identifying about the case, and de-identification done at the source is de-identification that never travels.
We do not rely on it. Everything below this point assumes the upload still carries patient data, because sooner or later one of them will.
It arrives in quarantine — not on the site
An upload does not go to the website. Scans, photographs and video all land in a holding area kept completely apart from the public site.
The point is not that we remember to keep these files out of sight. It is that, until a case has been reviewed, there is no way to reach them at all — no link exists, so there is nothing to share, guess at or stumble across.
We anonymise it again
DICOM has a standard for this, and we follow it: every study is anonymised to the DICOM PS3.15 basic confidentiality profile, which strips out the patient's name, identifiers, dates and institution, along with the long tail of hidden detail that travels inside a medical image.
Photographs and video have no equivalent standard, so we do the two things that always work. First, we check that a file really is what it claims to be, by opening it rather than trusting its name. Second, we rebuild it from the picture itself. Everything a camera or a workstation quietly attaches — where it was taken, when, on what — stays behind in the original, because it is never copied across.
A human looks at it
None of that touches text burnt into the image itself: a name along the top of a fluoroscopy run, an accession number in the corner of a screen grab, a date stamped into a video overlay. That detail is part of the picture. There is nothing to strip it out of.
So an editor looks. Not a glance at the first image and a quick scroll through the rest — every frame, because a name can appear for a single second in the middle of a run. The write-up is read with the same care: a referring clinician's name, or an unusually specific date, will identify a patient just as reliably as a face will.
This is the slow part of the process, and it is meant to be. A case waits until someone has properly looked at it. Nothing is hurried through because a contributor is keen to see their work go up.
Nothing reaches the public library without a person having looked at the actual images.
Only the reviewed version becomes the case
When a case is cleared, it is the reviewed version that becomes part of the library. The material it was built from does not sit quietly in the background afterwards. Once it has done its job, it is gone.
If a case is not cleared, nothing of it is kept at all. Either way the decision is final, which is exactly why it is made slowly.
The case can only be published once every item has passed
A case cannot be published while any part of it is still waiting on review, or has been turned down. Not should not — cannot. It is not a checklist someone works through and might one day skip a line of; it is a condition the case has to meet before publication is possible at all.
There is a second condition alongside it: the author must have set a name to be credited under. A published case carries a byline, so there has to be a byline to carry.
And then it is somebody else's teaching material
From there the case is in the library: searchable, filterable by procedure category, modality and difficulty, with its imaging openable in a full viewer rather than flattened into screenshots.
Found something we missed?
If you can identify a patient from anything on this site, that is a failure of ours and we want to hear about it immediately rather than politely. Email privacy@dottir.uk and the case will be taken down while we look, not after.