Workflow integration

Commit a corrected remote study without retaining a duplicate

Send a study-level correction back to the archive and read back proof that it replaced rather than duplicated.

Desktop app only, 1.6.0 or newerdcmage pro

Problem

An archive that accepts a corrected study and keeps the old one has made the problem worse: the patient now has two studies, one right and one wrong, and downstream routing and review cannot tell which is which. A call that returned successfully is not, by itself, a study that was replaced.

Steps

  1. Correct the tags in the tag editor as you normally would. A study-level tag such as Study Description is written to the whole study. A series-level tag such as Series Description, Laterality or Body Part Examined is written to the series you edited it in, so a study with a left and a right series can carry a different Laterality in each.

  2. Open the change preview. Before anything is sent, it states the exact archive resource and Study Instance UID that would be targeted, every changed tag with the value the study holds today and the value you propose, which series each value is written to and how many instances that covers, and the identifier policy. By default the corrected study receives a new Study Instance UID, a new Series Instance UID, and a new SOP Instance UID for every instance.

    Tags that do not hold the same value across the study today are highlighted: committing overwrites every instance with the proposed value, including instances you never edited. A series given two different values is refused, because one series cannot hold both.

  3. Confirm the commit. dcmage asks the archive to correct the study itself without keeping the original, writes each series value to its own series, then reads the result back.

  4. Read the five conditions. Every one is reported, passed or failed:

    • the corrected study carries the approved value for every changed tag, in every series the value was written to
    • it holds the same number of instances the source held
    • the identifiers were regenerated at all three levels
    • the source resource is gone from the archive
    • the patient has no more studies than before; for a correction that files the study under another patient, the patient it left has one study fewer and the one it joined one more

    Only when all five pass does dcmage say the study was replaced. Anything less is reported as an attempt, with the readings that failed, so you can see what the archive is actually holding.

Keeping the study’s identifiers

Some studies are referenced by their UIDs elsewhere, in a report or a research dataset. If the Orthanc archive is configured to overwrite stored instances, the preview offers Keep the study’s identifiers. It is off until you tick it. Ticked, the archive writes your changes into the images it already holds, under the UIDs they carry now, and the commit is judged by two different conditions in place of the identifier and source checks: the study still carries its own UIDs, and it still resolves under the archive resource you opened.

Keeping the UIDs has a cost the preview states beside the option. A viewer, router or other archive that already received the study keeps its old values under the same UIDs, and nothing tells it the study changed. Regenerated UIDs make the corrected study a new study that those systems receive as new.

Correcting the patient

A change to the Patient ID, name, birth date or sex adds a warning, a table of the patient now, after the correction and in the archive’s record, and a confirmation of its own. See Correct the patient a remote study is filed under.

When the archive cannot promise a replacement

The preview refuses the commit and names the reason. An archive that can only accept new instances can still be searched and opened; it just cannot be corrected in place.

When the study’s pixels changed

A commit sends tags only. If you ran a pixel operation, replaced the pixel data, changed the compression or edited an RT structure set, a segmentation or a structured report, or drew a redaction box, the archive would keep its original pixels. The preview refuses the commit and says so in place of the button. Undo the change, or delete the box in Anon mode, to send the tags. To keep the changed pixels, export the file instead.

When the commit fails

Your local corrections are never discarded. The study stays open and edited exactly as it was. Check the archive and retry reads the archive first: only if the source study is still there and the patient’s study count is unchanged is the correction sent again. If the correction did reach the archive before the connection failed, dcmage tells you so and refuses to send it a second time, because a second attempt would leave a duplicate.

Result

The archive holds one corrected study and no source study, and you have the read-back evidence that says so, or you have a plain statement of what the archive is holding instead, with your local corrections intact and a recovery path that will not create a duplicate.