DICOM supplement 145 inquiry

[Mathieu Malaterre]

Hi Yves,

On Sat, Mar 16, 2013 at 9:16 AM, Yves Sucaet <sucaet at histogenex.com> wrote:
> I'm currently attending the 20th international DICOM conference in
> Bangalore, India (http://www.dicomconference.org). While most of the focus
> is on RX and topics like tele-radiology, there is at least a grassroot
> movement interested in pathology as well.

Great !

> I've talked to several people here about the DICOM sup. 145 standard for
> Digital Pathology. One problem is that there are currently no libraries or
> viewers that can properly view this file format. Some groups (Siemens in
> particular; eSiePath) has done some hypothetical work with the format,
> assuming that one day files would become available. However, they don't have
> an actual parser either.
>
> Part of the problem is that there are no publically DICOM sup. 145 sample
> files available. So my question here it two-fold:
>
> * Does anybody have DICOM sup. 145 encoded whole slide images available and
> would they be willing to contribute them to the OpenSlide reference
> repository of WSI formats?

Some people from wg26 at nema reported building some (including D.
Clunie). I know the DCMTK team has made some effort in that direction,
they have a converter NDPI -> DICOM/WSI. But they did not distribute
the datasets used AFAIK.

> * Is anybody else interested in seeing DICOM sup. 145 reading capabilities
> added to OpenSlide (or is working on that already)?

Short answer: I do :)
Long answer: DICOM/WSI is a portion of the DICOM standard. As such I
do not see the need for yet-another limited DICOM implementation for
DICOM/WSI. So anyone going in that direction would need to actually
implement a great portion of the DICOM standard. With my very biased
point of view, there are only two large DICOM open source
implementation: DCMTK and GDCM. I do not know enough of the internal
DCMTK details, but in GDCM we had to work on a streaming interface to
be able to deal with those datasets (or at least paved the way).
During GSoC 2011, we got pretty far with the experiment:

http://sourceforge.net/apps/mediawiki/gdcm/index.php?title=GSOC_Whole_Slide_Imaging_in_Pathology
http://gdcm.sourceforge.net/wiki/index.php/Gdcmstream

However there is still a long way to go to have a complete
implementation. I would think this is -at least- yet another GSoC to
wrap up the implementation. Finally both DCMTK and GDCM are c++ libs,
which would make OpenSlide even more complex to build.

Finally with all due respect, I believe the DICOM world will never use
the openslide API to deal with actual real world DICOM/WSI application
(think: meta data). IMHO  it would be a very much wisely spend time to
build up a converter Any-WSI => DICOM/WSI. That would help people move
forward from legacy (proprietary format) application to interoperable
DICOM apps.

> I'm willing to contribute resources to this effort, but I need to know if
> there is actual demand (and supply of sample files for that matter) for this
> first.

In that case quickly register to GSoC 2013 for OpenSlide and apply as
mentor on OpenSlide/DICOM/WSI !

2cts
-- 
Mathieu