Posts by radtraveller

    Pacsdude:


    All the DCMSND studies are "CR"?


    ImportConverter0 = ifequal "%V0008,0060", "CR"; compression as
    Importconverter1 = ifequal "%V0008,0060", "US"; compression as


    though I suppose %m for modality works. I have always just preferred to use Dicom Tags.


    I always use acrenema.map to set compression by AE.



    "DecompressNon16BitsJpeg = 1" are your US images 16bit Jpegs coming in? If not, would not the Conquest Decompress then recompress them?



    Storage is so chesap these days, bleh 2TB SAS drives for ~$300 US each... Why compress at all on the archive? (Well, for smaller facilities anyway) and just use as exportconverter to remote destinations?


    Ip chain policy input / forward / output (ACCEPT) (internal network all firewalled )


    Disable apparmor.. No firewall between the two systems I am trying to move, (conquest linux to Conquest windows)
    acrnema is correct


    etc/hostname is correct.

    Grrrr... I physically moved the computer I have Conquest 1.4.16 on 64 bit Ubuntu. No change in IP or port configurations. I can search for studies, I can store studies to it, but every other system I try to move images to or use to try to retrieve images from this system, I get an error message of HOST 'name ' did not accept the connection. On the host I try to move images to, I cannot find any error messages. (Including a Conquest with debug to 4)



    HELP!?!?!?!?


    :-)


    Thanks

    dgate --modifyimage:*.*; ifequal "%V0008,0090[0,5]","000375"; set 0008, 0090 to "00375^LASTNAME^FIRSTNAME^(middle initial)";
    dgate --modifyimage:*.*; ifequal "%V0008,0090[0,5]","000470"; set 0008, 0090 to "00470^LASTNAME1^FIRSTNAME1^(middle initial)";
    dgate --modifyimage:*.*; ifequal "%V0008,0090[0,5]","000234"; set 0008, 0090 to "00234^LASTNAME2^FIRSTNAME2^(middle initial)";


    For every Physician ID?


    Or maybe just easier to install a second instance and run --grabimagesfromserver:AE,date


    from first instance and set multiple importconverters with ifequal running against each known physician ID?


    Unfortunately, it wasn't till I had the full database in Conquest that I could 'hook up" Crystal Reports to the SQL DB and start running some sample reports that I found the all the errors in physician namings.


    I also ran a total db query on patient demographics. Lots of DOBs incorrect, Misspelled names, patients with multiple MRN ids etc....


    Waiting for medical records to verify and give me corrections to clean up that data... Maybe I will wait and finish patient data fixes, then do another instance and use the importconverter statements for each physician ID to fix physician names. Would "newuids" be necessary (or a good idea) in the second instance dicom.ini? I 'thought' that newuids were only necessary when changing patient ID (MRN) shrug.


    I still have alot of work to do to clean up everything in addition to imaging patients in Xray and CT... sigh.


    I appreciate the help. I sometimes don't have time to do alot of my own research.
    Matt

    Hi all.


    So I finally have my interim database where I need to fix incorrect information. All the ID are normalized and Site information is corrected. These were all Global corrections so I could easily do an ImportConverter to do the fixes as I got them from the commercial PACS.


    Now I am at a stage where I need to fix hundreds of studies at a time on specific fields. The first I need to try is 0008 0090 for the Referring Physician. I have physician ID numbers in this field followed by ^LASTNAME^Firstname^(middle initial). The problem, of course, is not the Physician ID... The problem is that the names have variations in either spelling or some may have a ',' comma without a "^" seperator or they have both or just a "^" seperating name portions AND some just plain mispelling of the names.


    I am trying to figure out how to go in and basically do an in place batch converter for each Doctor (luckily I have less than 10 docs to worry about) where (for example Dr. with Physician ID of 000375):


    ifequal "%V0008,0090[0,5]","000375"; set 0008, 0090 to "00375^LASTNAME^FIRSTNAME^(middle intitial)";



    1. Is it possible to do this against a running DB?
    2. Which dgate command to use and does the above statement look correct?


    I have to do this muliple times in a 13k study database.


    Thanks
    Matt

    Hi


    What is the correct usage of --grabimagesfromserver?


    I tried running this against a running windows Conquest as "dgate.exe -grabimagesfromserver:RADIOLOGYBU,2010" in an effort to grab all images from RADIOLOGYBU that dont exist on this server.


    [PACSTESTMRN] Testing transfer: '1.2.840.10008.1.2' against list #0 = '1.2.840.10008.1.2'
    [PACSTESTMRN] 0000,0002 18 UI AffectedSOPClassUID "1.2.840.10008.1.1"
    [PACSTESTMRN] 0000,0100 2 US CommandField 48
    [PACSTESTMRN] 0000,0110 2 US MessageID 1
    [PACSTESTMRN] 0000,0800 2 US DataSetType 257
    [PACSTESTMRN] 0002,0010 17 UI TransferSyntaxUID "1.2.840.10008.1.2"
    [PACSTESTMRN] 9999,0300 42 UN "Server command sent using DGATE -- option"
    [PACSTESTMRN] 9999,0400 38 UN "grabimagesfromserver:RADIOLOGYBU,2010"
    [PACSTESTMRN] Server command sent using DGATE -- option
    [PACSTESTMRN] Grab - finished


    Is all I get.


    tried -gserver:RADIOLOGYBU,2010 also with no results.


    Basically I want to populate a new instance (B) with all studies/images from server (A).


    Advice?


    Thanks

    This really looks like a permissions or share misconfiguration.


    http://ubuntuforums.org/archive/index.php/t-288534.html


    Gives a really good tutorial for getting NAS/cifs up and mounted at boot.


    ah I thought you already had a conquest server with all of the studies on it. If you have a commercial pacs you are getting the data from that is QUITE different.


    If I understand you, all you want to do is keep an aggregate record of dose applied. You are not keeping dose records for individuals?




    One thing I am not is an expert in all the DICOM Tags. Is dose information asctually stored in all of those tags?
    Tag
    Name
    Keyword
    VR
    VM


    (0018,1411)
    Exposure Index
    ExposureIndex
    DS
    1


    (0018,1412)
    Target Exposure Index
    TargetExposureIndex
    DS
    1


    (0018,1413)
    Deviation Index
    DeviationIndex
    DS
    1


    ImportConverter0 = append "%V0018,1411" to doseaudit.csv
    ImportConverter1 = append "%V0018,1412" to doseaudit.csv
    ImportConverter2 = append "%V0018,1413" to doseaudit.csv
    Importconverter3 = destroy


    This should pull the data you are looking for from a commercial PACS that does not allow user queries on the actual sql db..


    I actually would need to research all the available fields where Dose information could be kept and that, I think, varies with manufacturer and their private tags.


    Once you have it in a csv for whatever time period you run the report (query) for, you can import that csv to a spreadsheet or to a database. (think excel or access or OO Base or Calc) then destroy the current csv or rename it to something and archive it somewhere (ie date_range_doseaudit.csv)

    Just leave out search on patient name... Go on MRN, Study then fields that hold the exposure info. Write that out to a csv, import into a spreadsheet then you can delete the column for MRN if you do not need it.


    That also gives you lots of data to play with and graph in many different ways

    Rainer,


    You need to create a second, NEW, instance of Conquest with compression set as jk, j1-j6 (I always set all three compressions to the same for the data base I am using (dropped, incoming, archived =jk) whatever. In the acrnema.map of the old server, make sure you have the compression for the NEW instance set as jk, j1-j6( After AETitl, IP and Port), then transmit all the images from the old server storing as .v2 to the new instance.


    OR


    Create a NEW instance of conquest, in that Dicom.ini set Dropppedfile, incoming and archiving compression as jk,j1-j6, whatever you want, then, open the new instance of conquest (or use kill and restart server if already running) to the user interface, then open the old conquest directory in explorer, go to the \data directory and highlight all the studies you want to transfer, then drag and drop them onto the new conquest interface. The dropped file argument should uncompress and recompress the images as they are loaded to the new instance when you drop them on the new server interface.


    Matt.

    ./configure make & make install jpeg-6c 1st
    then jasper:


    If jasper throws errors about something like 'incompatible jpeg' or some such do ./configure --disable-libjpeg (from /docs/jasper.pdf)
    then your make and make install


    That should get you through the ./configure make and make install for jasper, then assuming you have your mysql setup correctly, the maklinux_mysql should just work.


    PS I install conquest in /opt and then third partition (sda3) labeled "dicom" so MAG0 = /media/_dicom (675gb jk compression) on my test server


    Matt


    Don't forget to go to the /usr/local/cgi-bin and copy the acrnema.map and change the dicom.ini settings to match your configuration if you want web access.


    BTW, any info on an updated user/admin web interface? I remember reading about someone trying to build something some time ago.

    # Configuration of rules to modify, log or reject incoming DICOM slices
    ImportConverter0 = ifequal "%V0010,0020[0,3]","ABCD"; set 0010,0020 to "%V0010,0020";
    ImportConverter1 = ifnotequal "%V0010,0020[0,3]","ABCD"; set 0010,0020 to "ABCD%V0010,0020";


    So if you have an old database that has mixed number prefix/no prefix (ie 123456 and ABCD123456)


    These will import the old data, if it has ABCD prefix, it stores and does nothing else.
    If prefix "ABCD" is missing, it will prepend the number with "ABCD" and store


    Set this up in a new instance of conquest then did a find local missing patient. Small DB only 80gb. I did change tcp time out to 600.
    Suppose you could put newuids at the end if you wanted to.


    EDIT.. After going through a couple GBs of data and seeing a few errors regarding instance uids error, I appended newuids to my importconverter. Deleted my database and started over.. No errors now. Best to try stuff with a small subset and make sure everything is working as you intend before going all out. :-)


    I also then do a


    Exportconverters = 1
    ExportConverter0 = forward Series to RADIOLOGYBU (Backup device jk on a 2 TB Fantom USB Drive)


    Changed my export converter because old data needs to be fixed (double MRN for some patient, wrong spelling etc.)
    Since I know that all data after 6/1/2010 is good:


    ExportConverter0 = ifnumgreater "%V0008,0020","20100601" ; forward Series to RADIOLOGYBU

    I see a C-Find but no C-Move.


    2 things I would check:


    First is OSIRIX setup with the Conquest server information correctly (Caps counts!)
    Usually you can do a C-find against a server that does not know about you, but it will refuse get requests.


    Second, and I really can't say, maybe marcel can comment..


    Weren't older MACS Big Endian?


    From the Manual:


    LittleEndianImplicit 1.2.840.10008.1.2 transfer
    #LittleEndianExplicit 1.2.840.10008.1.2.1 transfer
    #BigEndianExplicit 1.2.840.10008.1.2.2 transfer

    Hi Marcel,


    Quick question...


    For a brand new linux installation:


    If you have a large amount of studies, can you copy them to the linux conquest MAG0 and then issue a regenerate DB command? Would this have any performance benefits vs transfer from another AE or the dgate --addimagefile:
    command?


    Thanks

    using alpha5.


    removed -m32 flag


    after ./configure, make and make install jasper and ./configure make and make install jpeg-6 stuff...


    test@test-desktop:~/dcm/conquest$ sudo sh maklinux_mysql
    /usr/local/lib/libjasper.a(jas_stream.o): In function `jas_stream_tmpfile':
    /home/test/dcm/conquest/jasper-1.900.1-6ct/src/libjasper/base/jas_stream.c:368: warning: the use of `tmpnam' is dangerous, better use `mkstemp'
    test@test-desktop:~/dcm/conquest$



    ??