Revisit modality types from RDSR issue #445

Issue #580 resolved
Ed McDonagh created an issue

Refs #445

With kit that does fluoro and DX, this solution doesn't work. I need to be able to have a modality that is both RF and DX!

Issue ref #577 could help as the wall stand will always be DX. But the table detector is used for both fluoro and supine DX such as abdomen.

This either needs to be fixed for 0.8.0 or the exports for RF and DX need to be changed to allow exports. Prefer the former.

Comments (45)

  1. Ed McDonagh reporter

    ref #577 won't help with this as there is no detector ID in the RDSR, but I think I have a way to do this anyway. I think.

  2. Ed McDonagh reporter

    Ok, so we need to make some decisions @dplatten, @tcdewit, @LuukO

    If you have a single room that does fluoroscopy and general radiography, such as a Siemens Luminos dRF Max, then you will have a situation where you can have fluoroscopy events and radiography events in the same RDSR. And when you do your dose surveys, you'll want to consider the fluoro event with your other fluoro rooms, and the radiography event with your other radiography rooms.

    Currently, you can choose to consider RDSRs from such a room to be either DX or RF. But not both, or some studies one and other studies the other (which wouldn't work anyway if you have one study with both types of events in).

    Option 1: Do nothing

    • The only way I can analyse one of my rooms would be to reset modality type depending on what I am looking at. Things would probably break.

    Option 2: Propose that such rooms are considered to be fluoro

    • Probably need to improve the fluoro export to better cater for the radiography events
    • Would mean radiography events won't be considered in filtering and charts in the radiography section of the interface

    Option 3: Create a new modality type of dual or similar

    • Rewrite the filters to allow dual modalities to be listed in both radiography and fluoroscopy sections
    • Possibly change the detail pages to be more agnostic to RF or DX
    • Make sure the radiography and fluoroscopy exports can deal with both (or ignore events of the other type?)
    • Would mean that the same study would be listed twice on the home page
    • Would mean that a general listing of fluoro studies would include lots of radiography (for me, one room is maybe 98% radiography exams, so this is quite annoying if you want to review fluoroscopy!)

    Option 4: Magic! As 3, but with study by study modality selection

    • Not clear yet how to do this, but ideal I think would be to:
      • have studies with only radiography events in listed just in radiography
      • have studies with only flouro events in listed just in fluoroscopy
    • In the [unusual?] case of a fluoro exam that includes a radiography event, list in both?
    • Ideally fluoro 'acquisition' images, which could be single shot, wouldn't trigger being listed in radiography? Though I don't think there could be anything to distinguish them!

    What do you think?

    Thanks, Ed

  3. Ed McDonagh reporter

    @bearner and anyone else - I'd value any input on this! Doesn't need to be thought through to code level; I want to know how people would expect/like these studies to exist in the system.

  4. Ben Earner

    I'm probably not the best person to help decide but I think something like option 4 looks the best (although probably the hardest to implement!).

    Although for the mixed studies, would it be possible to split the study so that they're only shown in the appropriate modality section? Yes you wouldn't see all of the study, but at least you can compare apples with apples and for dose monitoring and isn't that what we want?

    For the single shot fluoro acquisition images, if they appear as DX in the RDSR and machine is behaving like it's doing a DX, then I wouldn't really mind it being in the DX list. (I don't really know enough about how these fluoro machines operate to know if this is true though!)

  5. David Platten

    I have a room that sort-of falls into this category. It's a Siemens room with a fluoroscopy and radiographic tube. The only things I receive from it are RDSRs of radiographic exposures that all currently sit in the RF category (I'm running 0.7.4 on my live system). At some point this room will be sending fluoroscopy studies too, then I'll be in the same boat as Ed.

    I think that magical option 4 sounds best. I would (naively?) imagine that there's a way of distinguishing a fluoroscopy acquisition (even single-shot) from a radiographic exposure.

    I'm in favour of trying to keep exams from a given modality together, without "pollution" from other modalities.

  6. Ed McDonagh reporter

    As far as I can tell, there would be no difference between a single shot cath lab style acquisition and a radiography event in an RDSR.

    I'm thinking we could set DX or RF on import, then have a button on the display name page to review all the existing studies to reset DX or RF subsequently.

    This would be easy for

    • DX only studies (total fluoro DAP is zero)
    • RF only studies (total fluoro DAP non-zero)
    • RF studies with cath lab style acquisition events along with fluoro (total fluoro DAP is non-zero)

    Not so clear what to do with a single study that has maybe a fluoro swallow and an abdomen x-ray in the same study. Just throw them into fluoro and assume they will be so infrequent that we don't need to be too concerned?

    I don't think listing some events in DX and other events from the same study in Fluoro would be very easy to implement.

  7. Ed McDonagh reporter

    @bearner, @dplatten, @tcdewit, @luukO - I have thrown a copy of this onto testing, and if you have a look at Display Names you can see some of the changes I have been making. If you go to set a modality type now, you have a third choice of dual. Once selected, that modality will appear in both tables on the display names page.

    You will also see that in the user-defined modality column, an admin user will see a link next to any dual modalities to 'reprocess'. This will then filter all the studies with that display name, and run through each one trying to determine if it is fluoroscopy or radiography, and resetting the modality_type accordingly. They will then appear in the correct table on the front page.

    Still outstanding:

    • Any independent verification of my implementation and wording for the links and messages etc
    • I do a count at the start and the end that gets displayed once the finished and the page is reloaded - the start count is always zero for each modality the first time I run it (0 are DX, 0 are CR, 0 are RF), but correctly populated at the end of the task and on future runs (453 are DX, 0 are CR, 6 are RF). I don't know why this is yet.
    • The page could do with some AJAX to make it less frustrating... my production database has 52 entries in the CT table, 28 in MG, 179 in DX, 14 in fluoro and 21 in other, and it takes ages to load!
  8. Ed McDonagh reporter

    So the reason the start count is zero for each modality is because they are set to dual... more work needed :-)

  9. Ed McDonagh reporter

    Initial draft of a better method - the resetting of the modality type takes place when assinging dual instead. More to do, including removing the previous method. Refs #580

    → <<cset acc5413e05f4>>

  10. Ed McDonagh reporter

    Attempting to review the broken studies in the OT table - it is looking like most of them are artifacts from studies that have been deleted... Refs #580

    → <<cset 4d9c76630689>>

  11. Luuk

    So, I'm a bit late...;-). But the chosen solution seems fine with me. I didn't have a look at the logic implemented for the decision on RF and DX, but I would suggest a quite simple approach: if total fluoro DAP is zero it is DX otherwise RF. Other option would be to decide on Performed Procedure Code Value (meaning). That would require to maintain a hospital specific list for each administrator (actually if regex or the like would be possible in this list, for my institution this would be quite easy as modality is part of the procedure code). Also in the HL7-implementation I'm using the modality type can be extracted and applied from the ORM message sent by the HIS/RIS/EMR (but in the current implementation, it won't if a user defined modality is set).

  12. Ed McDonagh reporter

    Each modality equipment list now populated using AJAX, each with study count and latest date. Need to work out why some are zero that I think shouldn't be! Refs #580

    → <<cset d3978f4be1a4>>

  13. Ed McDonagh reporter

    Moved fledgling study review code to AJAX to enable more of the logic to be done in Python and to keep page load times reasonable. Refs #580

    → <<cset a906419219cf>>

  14. Ed McDonagh reporter

    Reduced duplication by making ad-hoc reprocess use the initial function used when dual is selected. Also moved the review_summary_list function. Sort of refs #580, and is working towards ref #479

    → <<cset 33e000999f5c>>

  15. Ed McDonagh reporter
    • changed status to open

    New fluoro exams of a dual system don't seem to be going in as RF, but do if reprocessed... need to revisit

  16. Log in to comment