XRAYS ON THE NET - DIGITIZING IMAGES 1995
Date: Sun, 3 Dec 1995 18:54:20 -0500From: hwb@eideti.com (Bill Burman)On 29 Nov. 1995, a colleague (TDECOSTE@medusa.unm.edu) posted a messagedescribing a complex lower extremity orthopaedic injury. Quite shortlythere were two replies ("Alexander N. Chelnokov" alex@uniito.e-burg.su ; "Jeffrey Anglen" jeffrey_anglen@SURGERY.MISSOURI.EDU) in terms of case discussion. Thus, the stage is set for a fracture conference in cyberspace.Unfortunately, as noted by Dr. Chelnokov, there was a lack of image data(xrays, clinical photographs, etc.). This could inhibit case discussion ateven the local level - where the centerpiece of most clinical presentations is an image - rendered if need be on a chalk board if not a paper towel.For the past few years I have been involved with the digital presentationof orthopaedic cases with the AAOS CD-ROM project "Orthopaedic GrandRounds". I would like to offer to extend experience obtained in that exercise to the present problem of the text-bound orthopaedic trauma case presentations.I would caution at the outset that this experiment will not necessarily bean easy undertaking and all that participate should be steeled forfrustration and failure. On the other hand, it could present us all with anopportunity to gain hands-on experience with orthopaedic telemedicine.I would propose the following:Presenters could e-mail me the text and either e-mail or conventional-mailme the images associated with their cases. The address is :hwb@eideti.comRon Hoering and I would try to post the cases and the discussion in theOrthopaedic Trauma Forum of the HWB Foundation Home Page.http://www.hwbf.orgThere are some sample cases posted there already which may serve as abaseline in terms of suggesting a format for such discussions.The case text (sent preferably as e-mail so that I don't have to re-typeit) should indicate the position of the accompanying graphics.e.g. A 21 year old white male S/P MCA 8/8/95 is admitted with the followinginjuries : CHI Lt hemopneumothorax Rt comminuted femur shaft fx Rt intra-articular comminuted open proximal tibia fxFigures : AP Rt femur Lat Rt femur AP Rt tibia Lat Rt tibia Wound photograph Rt tibiaetc.If the images are in digital format (either PICT, BMP, Compuserve GIF orJPEG) they may be sent along as attachments to e-mail to:hwb@eideti.comBin-hexing and compression of the images prior to sending may be advisable.Some image formats may get chewed up with internet transmission so onewould have to be prepared for some adversity as we worked out the difficulties in the early phases of this process.If the images are not in digital format, there are several digitizing methods which may be considered. The digitized images could then be sent as attachments to e-mail.1) Radiology Xray scanner - Perhaps your radiology department is digitallyarchiving xrays at this time. The American College of Radiology isrecommending the DICOM 3 format for the image files. These are usuallylarge image files 2kx2k (8-10 Meg) and are not readily readable by PCcomputers without special software. The large size can make them a bulky,time-consuming transmission - impossibly slow over phone lines (manyhours). Also, we would quickly run out of disc space if we tried to storeimages of this size. It is possible to convert DICOM 3 images into themore common PC formats (such as PICT, BMP, Compuserve GIF or JPEG) whichcan more readily be compressed and transmitted. This would depend on anability to network access and open files in your radiology departmentarchive. This can be technically and politically tricky to do.Bottom line : A complex acquisition. Forget this option for now.2) Your own xray scanner - Common types would be the Vidar VXR-12, VidarVXR-8 , Lumisis, Howtek and Kodak film digitizers. Many of theseprofessional film SCSI digitizers can be driven with plug-ins to standardPC image manipulation software such as Adobe Photoshop. They can rapidlyscan standard 17"x24" films. The present price of such scanners is in the$12,000 to $20,000 range.Bottom line : Too expensive.3) Flatbed scanners - Flatbed scanners are becoming quite affordable ($400- 1000). They require backlighting to scan xrays. Generally, thebacklighting attachments transilluminate an area on the order of 6"x8"which is an inadequate size for most xrays. It is possible to make standardphotographic prints of xrays and soft tissue conditions, develop with a 1hour print service and then digitize the photograph. Xerox prints of xraysdo not appear to be of adequate resolution, however xerox camera shots ofwounds may be sufficient.Bottom line: Too small for xrays. Can be used to scan photographic prints.4) Slide scanners - Nikon Coolscan, Polaroid SprintScan run between $1400 -$2000. Standard 35 mm slides can be digitized with these devices. Apossibility exists to shoot high speed ektachrome film which can bedeveloped in an xray processor and then scanned with the slide scanner.Bottom line : OK if your images are first captured in 35 mm kodachrome format.5) Digital cameras - There are excellent digital cameras which take adigitized image which can be copied directly on to the hard drive of a PC.They cost upwards of $12,000. There are some low-end digital cameras (AppleQuicktake 150, Kodak DCS 40) for under $1000 which are mainly used forphotographing houses in the real estate market. The resolution of thesecameras are probably not of diagnostic quality but possibly, they are worthexperimenting with. The direct digital capture of xrays and other clinicalphotographs makes this method appealing. Another option for this so-called"camera on a stick", is an inexpensive NTSC digitizer attachment for acamcorder called "Snappy View".Bottom line : Simplest, most direct method of capturing images but theimages are probably inadequate when an inexpensive camera is used.At the Radiological Society of North America Annual meeting in Chicago thispast November, the jury is still out with respect to the minimum resolutionstandards of digitized xrays. Even with "optimal" images (2kx2k,recommended by ACR-NEMA for the DICOM 3 standard of diagnostic quality),musculoskeletal radiologists from Harborview missed subtle fractures whichthey were able to detect on conventional films (Radiology, August 1995,p565).If none of these options are available, the images as either photographicslides or prints or as xray duplicates could be sent via conventional mailto :Bill Burman, MDHWB Foundation604 Stage RdMonroe, NY 10950USAand we could undertake the digitization along with the internet posting ofthese materials.******************Date: Mon, 04 Dec 1995 09:28:56 -0600From: Jeffrey AnglenBill has made a very interesting proposal. The cases would be sent in bymembers of this discussion list, or others I assume, then assembled andposted to the web site, where everyone with a web browser couldview the images and read the details. The case could then be discussedon the list or perhaps by a designated OTA member, - in an interactivemanner with other list members; then the comments of the list discussioncould be added to the web page for the benefit of other docs, students,residents. This would be similar to the "fireside discussions" at the AOcourses where the participants bring cases and discuss them with agroup of faculty.The cases could be updated with followup data and Xrays by thesubmitters so that we could all actually follow the course of particularlyinstructive cases as they evolve, discuss and learn from each othersexperience. There are 30+ people presently on this list, and I expect it togrow; if we had one case a month each list member would have to puttogether a case every couple years. And of course we could ask nonlist members to contribute. Given the willingness of OTA members toserve as educators (witness the participation in the recent resident'scourse) I think it would be no problem with a little organization.This might even be a project for the OTA education committee?The web site would not be limited to static images, but could of coursepotentially have sound and video. Someday.What does everyone think?*******************Date: Mon, 4 Dec 95 21:22:45 -0600From: Andrew H. SchmidtI think that the proposals by Bill and Jeff sound great. I have also beenfollowing the Orthopedic Mailing List run by Wes Spears and Randale Sechrest,which has been fairly successfully utilizing images uploaded by the members ofthe list and archived for later retrieval by anyone interested.Another easy option for digitizing images that I didn't see in the list ofalternatives that Bill Burman presented is the Photo CD format, which is done atminimal cost and is readily available at most "professional" photo processinglabs. Once saved on a photoCD, I suspect the images could readily be convertedto one of the other formats (jpeg, etc) for transmittal. Anyone tried this?******************Date: Mon, 4 Dec 1995 19:45:11 -0500From: Bill BurmanThank you for mentioning this, Andy. It completely skipped my mind. Nearlyall the images used for the AAOS Orthopaedic Grand Rounds were digitizedvia Kodak Photo CD. It costs about $1.50 per image and the CD costs $10.You can store about 120 images on a single CD. The CD's are multisessionwhich means you can fill them up over time instead of having to save up 120images and do it all at once. Each image is stored on the CD in 5 differentresolutions:1) 192x128 - thumbnail2) 384x256 - quarter screen3) 768x512 - full screen4) 1536x1024 - full page ad NY Times5) 3072x2048 - good for highway billboardsThe advantages are that you don't have to fool with any scanning equipmentand techniques. You take your 35 mm film or slides to Kodak and they scanit for you with a $50,000 scanner and you get your images back on a CD-ROM.Needless to say, the quality is exceptional. You don't have to burn upspace on your hard disk(s) because it is archived on CD (for 100 years -good for long term follow-up - even in Iowa).The disadvantage is that unless you are near a busy service bureau,turn-around time can be slow. I have had to wait 10 business days. ThePhoto CD format usually requires conversion of the image to other formatsin order to permit image compression or import into other programs.Generally, this requires a program like Adobe Photoshop which can take sometime to get used to. Compression is probably necessary for this project asone full screen image, 768x512, could take up to an hour to transmit on aphone line.Bottom line : Great images, great storage, slow process, image file formatmanipulation usually required.*******************Date: Tue, 5 Dec 95 10:32:26 +0500From: Alexander N. ChelnokovHi> Presenters could e-mail me the text and either e-mail or conventional-mail> me the images associated with their cases. The address is :> hwb@eideti.com> Ron Hoering and I would try to post the cases and the discussion in the> Orthopaedic Trauma Forum of the HWB Foundation Home Page.> http://www.hwbf.org/hwb/There is also Randale Sechrest's web site"htp:/ww.cyberport.net/ortho/ortho.html"> 1) Radiology Xray scanner - Perhaps your radiology department is digitally[...]> Bottom line : A complex acquisition. Forget this option for now.Noted> 2) Your own xray scanner - Common types would be the Vidar VXR-12, Vidar> Bottom line : Too expensive.Yes...> 3) Flatbed scanners - Flatbed scanners are becoming quite affordable ($400> - 1000). They require backlighting to scan xrays. Generally, the> backlighting attachments transilluminate an area on the order of 6"x8"> Bottom line: Too small for xrays. Can be used to scan photographic prints.:-/ HP ScanJet II/III - is too small? It is 400 dpi, and with transparencyadaptor is ~$2000. I saw also Mustek Paragon Combo - it is A4 (too small forchest films but quite enough for any fracture), of 600x1200 dpi opticalresolution, with transparency adaptor and less than $1000 even in Russia.**********************Date: Tue, 5 Dec 1995 06:54:24 -0500From: Bill BurmanThe resolution provided by these scanners is more than adequate. When Isaid "small", I was referring to the physical size of the scanning bedwhich is generally 8"x12" - some going up to 11"x14". However, the backlittransparency adaptors usually transilluminate only a portion of the scanbed - generally 8"x10". This is probably OK for hand and foot films butxrays of the long bones, spine and pelvis generally are 14"x17".Presentation of these studies using conventional flatbed scanners couldlead to the serving up of the notorious "key-hole" views. In the absence ofalternatives, the argument can be made that a "key-hole" view is betterthan no view. But, of course, it can be misleading.******************Date: Wed, 6 Dec 95 10:21:39 +0500From: Alexander N. ChelnokovOk I realized what's a problem. Hmm...I can say that if somebody would pay me difference between A4 scanner priceand genuine Xray scanner i would scan for him large films partially and thenpaste the parts with Photo(Styler/Paint/Shop/etc.) or automatically (i supposethere is such software or it is easy to create).So my opinion is - if you need scanner for time-to-time job like to senda case to the list or prepare a presentation/slide/paper - cheap A4models with transparency adaptor would be quite enough (and you could use thisdevice for pictures and OCR). But if you need a systematic duty likeelectronic Xray archive of each patient - you need Xray scanner, better ifwith autofeeding.*********************Date: Wed, 6 Dec 1995 04:24:03 -0500From: Bill BurmanThis is possible - but time consuming and tricky. However, your suggestionof segmental scanning of xrays which are later stitched together brings tomind another idea.An xray could be placed on an xray view box laying horizontal on a table.They used to make hand held scanners which are inexpensive (ArcTec 800 dpi24 bit - $129.99), of reasonable resolution and used to come with"stitching" software which would assist the joining of the image segmentsresulting from the several passes that would be required to capture a largeimage. If somebody had access to this sort of equipment, it would becertainly worth a try.**********************From: Alexander N. ChelnokovDate: Wed, 6 Dec 95 23:04:14 +0500Hi againYes sure but AFAIK hand held scanners may cause problems with OCR.But it is another story.******************Date: Tue, 5 Dec 95 14:49:50 PSTFrom: CAPT A. Herbert Alexander, MC, USNIn regards to imaging of radiographs. . .I have a Microtek ScanMaker IIXE with a transparency adaptor. It limits thesize of my radiograph scans to about 8x10". I solved that problem by takinga 35mm slide of the radiograph I want and then scanning it. I don't know ifmy scanned image (.tif, .pcx. or etc.) can be transmitted intact via Internet.********************Date: Tue, 5 Dec 1995 17:17:45 -0500From: Bill BurmanI don't know either. You could try and send a small tif or bmp as an e-mailattachment to me at the above address and let's see if I can read it.*********************Date: Wed, 6 Dec 95 07:56:14 PSTFrom: CAPT A. Herbert Alexander, MC, USNX-Attachments:C:\PHOTOSHP\ACLXRAY1.JPG;C:\PHOTOSHP\ACLXRAY1.EPS;C:\PHOTOSHP\ACLXRAY1.TIF;Bill,Attached are files of the same x-ray in 3 formats: .jpg (7kb), .eps (10kb),.tif (48kb). Hopefully the smaller one, .jpg works. Let me know.I use Eudora and attached the files using the "Attach Document" command.Files transmitted as is without conversion or compression.********************Date: Wed, 6 Dec 95From: Bill BurmanWorks Fine ! I am on a Mac Powerbook using Eudora and I could receive andopen the images without a problem. I used a viewer called "JPEGVIEW". Ineeded Adobe Photoshop fot the eps file.********************Date: Thu, 07 Dec 1995 09:48:24 -0600From: Jeffrey AnglenI would be interested in seeing those images. I wonder if you attachedthe jpeg file to a message to the list, would ListProc copy the attachmentto everyone?********************Date: Thu, 7 Dec 1995 09:20:11 -0500From: Bill BurmanJeff,I will attach the jpg image to this letter which is being composed withEudora on a Mac. There is no special treatment of the file. Distributionwill be to you and the listserver. Perhaps Dr. Alexander can run the sameexperiment from his PC.*********************Date: Thu, 07 Dec 1995 16:29:37 -0600From: Jeffrey AnglenGot something, but I can't make it into an image. Apparently it can't betransferred back and forth through the email program.**********************Date: Thu, 7 Dec 1995 18:17:14 -0500From: Bill BurmanI was able to open and view the image sent by Dr. Alexander as well as theimage bounced off the listserver using a public domain program called JPEGView. The e-mail program here is Eudora.Look on your computer for the file "ACLXRAY1.JPG"What image viewing program are you using to open the image ?Does that program recognize the JPEG file format ? (Does it have an "openas" or "save as" jpg option ?)What is your e-mail program ?Was anybody else out there able to open and view the image ?************************Date: Fri, 8 Dec 95 12:10:34 +0500From: Alexander N. ChelnokovHiAll is fine - the image can be successfully decoded and viewed by anyjpeg viewer (i use QPEG (DOS) and ACDSee (Windoze))To unpack use MIME decoder instead of usual UUDECODE.**************************Date: Fri, 08 Dec 1995 08:57:46 -0600From: Jeffrey AnglenI got the data for the image not as a separate file but as a charactersequence appended to the message. I attempted to copy it as aseparate file and open it using a viewer called Vueprint whichrecognizes jpeg as well as gif images but it would not open. I don't havea decoder which is apparently the missing element. Anybody know of apublic domain or shareware decoder that I can ftp?My email program is Novell Groupwise 4.1.******************Date: Fri, 08 Dec 1995 15:46:30 -0700 (MST)From: Tom DeCosterI have posted xrays of the comminuted tibia shaft case which I previouslydiscussed on the OTA email group. It seemed best to me to post thoseradiographs on the orthopedic surgery mailing list ftp site on theinternet since our email didn't easily accept images. The files can beviewed by accessing the mailing list web site at:http://www.cyberport.net/ortho/ortho.htmldirections are given for gaining access to the ftp site and downloading imagesat the web site.***************************Date: Mon, 11 Dec 95 07:52:52 PSTFrom: CAPT A. Herbert Alexander, MC, USNX-Attachments: C:\PHOTOSHP\ACLXRAY1.JPG;The ACLXRAY1.JPG came back to me intact. I opened it with Adobe Photoshop.I did not post the original transmission to the bulletin board, not wishingto trouble everyone with the detail work. Here it is again for everyone.**********************Date: Mon, 11 Dec 1995 12:33:37 -0600From: Jeffrey AnglenThe jpeg image came through perfectly as an attached file! Veryexciting - apparently we can send graphic files through the ListProcprogram as attachments to list messages and they will be copied anddistributed intact.**********************Date: Mon, 11 Dec 1995 04:04:09 -0500From: Bill BurmanDr. Alexander,I am still working on the issue of trying to simplify the transmission ofgraphics for case presentation. Netscape (as well as Eudora) has theability to send mail with attachments.Could you please re-send the ACLXRAY1.JPG file to me by this time visitingthe HWB Page http://www.hwbf.org./hwb/ and click the "Send Mail"button near the top of the page. Type "Graphics Test" in the body of themessage and use the "Attach document" button to affix the ACLXRAY1.JPGfile. This method has worked on the Mac side and I would appreciate yourassistance in testing it on the PC side.Thanks for your help.*********************Date: Mon, 11 Dec 95 14:13:19 -0800From: A. Herbert Alexander, M.DAttachment converted: PB 520:x-ray on the net (JPEG/JVWR) (0000474A)Bill,Graphics Test. Let me know how it goes.********************Date: Mon, 11 Dec 1995 22:08:05 -0500From: Bill BurmanHerb,I was able to read the graphic you sent via Netscape loud and clear. Theonly problem that I am having is that the Netscape mailer renamed thegraphic to "xray on the net" - same as your message header.If the image file names are changed, then confusion in image placement canresult, especially if we hand the job of posting the www cases over toanyone who is not an orthopaedist. Eventually, everybody should be able tpwrite and arrange their own case web page (html source code) by convertersattached to standard word processing applications - but I think that is toocomplicated for right now.I am looking for the most straight-forward method of transmitting imagesand text for posting www case presentations. Perhaps, the best methodwould be to post mail with attached image files using Netscape. I do notknow what the results would be like with Mosaic. This assumes most peoplehave access to www browsers.It may be perfectly acceptable for case presentation to simply bouncee-mail with images off the reflector to each other or download the samefrom an ftp site. However, given the wide variety of e-mail programs, imageformats and time to separately find, download, open, organize and orientthe images, this path will probably be fraught with frustration. If youhave a Mac and you download a bunch of zipped (PC compressed) files - youare out of luck. Vice versa with Compact Pro.A lot of people who know a great deal about orthopaedics and perhaps not somuch about computers will be turned off and subsequently unavailable forthe discussion. While I have my own ideas about the management of the tibiafx presented, I would like to hear what Ken Johnson, Don Wiss, BobWinquist, Jim Kellam, Mike Bosse, Sig Hansen, Bob Brumback, MarcSwiontkowski, Jeff Mast, Joldas Kuldjanov, Dean Cole, Caudle and Stern andothers would have to say. While one does not need to be a physicist to readxrays in their own hospital, the ideal design of www case presentationshould not require them to be computer wizards to read or show xrays overthe internet.I would be grateful if you could help me with another experiment inshipping multiple graphics files (2 or more) as e-mail attachments to theNetscape mailer activated by clicking the "Send Mail" button athttp://www.hwbf.organd let's see if the names of these image files are properly retained.Thanks.****************Date: Wed, 13 Dec 95 11:44:30 PSTFrom: CAPT A. Herbert Alexander, MC, USNX-Attachments:C:\PHOTOSHP\DRUJ4P.JPG;C:\PHOTOSHP\ACLXRAY1.JPG;C:\PHOTOSHP\BBFX1.JPG;Bill,For some reason it has been difficult for me to connect to your Home Page.I don't know why, but either the message comes up that server is unknownor I get only the the list of the Board of Directors and cannot page downto the "send mail" icon.At any rate, when I did get through, I was allowed to list only 1 graphicsfile to attach. More than one listed resulted in an unrecognized filenotice. At any rate, I am sending you the files by Eudora to see if thatworks. The files are: our favorite aclxray1.jpg (which you already knowand love); bbfx1.jpg (x-ray of both bone forearm fracture); druj4p.jpg(intraop photo of a complex DRUJ dislocation).I think it is just a matter of time before everyone has easy access todigitized x-rays and can ZAP them anywhere they want. Some places alreadycan access x-rays on their monitors and are without the need to use hardcopy x-rays. I hope we are not trying to reinvent the wheel here.**********************Date: Wed, 13 Dec 1995 18:34:22 -0500From: Bill BurmanHerb>For some reason it has been difficult for me to connect to your>Home Page. I don't know why, but either the message comes up>that server is unknown or I get only the the list of the Board>of Directors and cannot page down to the "send mail" icon.I am sorry about that. I gave you a faulty address at first - with an extradot.. Did it clear up with "http://www.hwbf.org/hwb/" ? Are othersout there having trouble connecting to the HWB page.>when I did get through, I was allowed to list only 1 graphics file>to attach. More than one listed resulted in an unrecognized file notice.I had the same problem. That is a pity. If we use the Netscape mailer,people will have to bundle the graphics in a zip or compact pro file andsend that as the attachment. That is another step I was hoping to avoid.>At any rate, I am sending you the files by Eudora to see if that works. The>files are: our favorite aclxray1.jpg;bbfx1.jpg;druj4p.jpgGot them all and can see them just fine - nice DRUJ image.>I think it is just a matter of time before everyone has easy>access to digitized x-rays and can ZAP them anywhere they want.I hope you are right. I went to the RSNA meeting in Chicago 2 weeks ago tofind out what I could about capturing xray images. It is definitely coming.The question is when. The 2 big stumbling blocks now are the cost of filmdigitizing systems - starting at $15 - 20,000 and the time and know-howinvolved with manipulating, cropping, labelling and archiving the imagefiles. Teleradiology adds a few steps more to this already involvedprocess.> Some places already can access x-rays on their monitors and> are without the need to use hard copy x-rays.Usually these are well-endowed paces with direct digital radiography, fiberoptic backbones, optical jukeboxes, a vast array of high resolutionmonitors, technical support and right now, people yearning for the good olddays.> I hope we are not trying to reinvent the wheel here.While I have seen cases presented on the net, I haven't seen any examplesof real interactive presentation - where anybody, anywhere can pipe rightinto the discussion - grab down the xray as they would in a conference andput an arrow on the "critical" spot, throw it back up on the viewbox withappended comment or reach into their image archives to throw up a visualcounter-example or pop an abstract off their Medline CD ROM database. Idon't think this has been done yet - at least using home PC's.All the best.What do you think we ought to do with respect to presenting cases ?Remember not everybody is using Eudora.