What impact is multidetector CT having on workflow and data management?

Are you prepared for the information explosion headed your way? PACS experts from across Europe relate their experience with MDCT and impart some hard-earned wisdom on how to cope with ever-expanding data requirements.

Dr. Elmar Kotter
Head, imaging IT, Freibourg
University Hospital, Germany

Installation of four-slice spiral CT doubled the number of images we had previously generated on a single scanner. We started with about 10 GB per week in 1997 and approached 70 GB per week at the end of 2002.

This overall increase is not a serious problem for our PACS archive. We have simply extended our RAID system. Our main problem was that some of our older viewing stations were unable to handle the volume of data produced by each multidetector CT examination. We expect an even larger increase in data volume with 16-slice CT, which is due to be installed soon.

Be prepared for a sudden increase in data volume when installing MDCT. All components of your PACS must be ready, including the archive, network, and viewing stations.

Prof. Walter Hruby
Chair of radiology
Danube Hospital, Vienna

We do not have a MDCT scanner yet, mainly due to the problems we would likely face combining our current infrastructure with the large amount of image data generated. Archive space definitely has to be considered, but network capacity and workstation hardware are hidden key factors. It is one thing to store more images on your archive, but you also have to consider whether these images will reach their destinations in the hospital quickly enough and whether your workstation can cope with 10 times more slices per study.

We consequently decided not to introduce MDCT until our archive and our workstations have been upgraded. We will also wait for our network to be replaced. The continuous growth of image data is not going to stop, so we can only advise radiologists, industry representatives, and legislators to promote "nearly lossless" image compression rather than lossless compression.

Martin Peacock
Systems Development Officer
Mater Misericordiae University Hospital, Dublin

Archive space is really only a financial matter, and this is becoming considerably less of an issue as the cost of fast disk storage decreases. Any PACS should expand easily to match changing requirements. MDCT has highlighted the difference between retrieve-on-demand and preloaded (as distinct from prefetched) data. It may take two to three seconds to retrieve-on-demand a typical computed radiography image, which is well within most users' tolerance. A full CT study may take 30 to 40 seconds, which can result in frustration. Reporting stations typically work with preloaded images, but Web-based distribution outside of the radiology department should be viewed with caution.

Dr. Frits Barneveld Binkhuysen
Radiologist, Hospital Eemland
Amersfoort, the Netherlands

Our department installed four-slice CT six months ago and replaced it with a 16-slice CT last month. This generated much enthusiasm among radiologists, technicians, and referring physicians.

MDCT makes it possible to perform dozens of CT examinations every day. It is tough when you have to scroll through all of this data in different settings and also need some reconstructions. So we asked our technicians to do the image reconstructions using protocols. We are now considering splitting the daily CT workload in two shifts: one radiologist in the morning and one in the afternoon, to handle all procedures properly.

Dr. Raimund Vogl
Imaging information manager
Innsbruck Hospital, Austria

Our hospital has two 16-slice and three four-slice CT scanners. Data production per scanner has roughly doubled from single-slice to four-slice, and doubled again from four-slice to 16-slice. Workstation infrastructure had to be upgraded to cope with increased data load.

For PACS using an autorouting system, scanner-to-archive transfer times and time to autoroute data to workstations adds up to more than half an hour, which has a crucial impact on reporting. On-demand PACS, where workstations pull data directly from the central archive via the network, are also affected by the increased data volume. Network backbones need gigabit bandwidth, and server systems should be upgraded to guarantee acceptable loading times. Failure to do so can have a dramatic impact on productivity. We found that the exponential increase in images caused the central PACS databases to grow beyond expectations, leading to unacceptable performance reductions.

Prof. Dimiter Tscholakoff
Professor of radiology
Rudolfstiftung Hospital, Vienna

We expect MDCT to be installed in our department this year. Although this will place extra demands on our archive space, Vienna's hospital managers have decided that we are no longer obliged to store all CT data in DICOM format. After reporting and reconstructing all images we need for each case, we will compress a certain number for storage in JPEG file format.