Going digital translates to going to meetings

Some years ago, we converted everything but mammography to PACS. It was the best move our department ever made. Recently, we finished the change by converting to digital mammography. I've had more fun at funerals.

The first big decision was selecting a vendor. Digital mammography equipment is expensive. We convinced our hospital that the only sensible way to do this conversion was all-or-none, which meant an outlay near

$2 million. Given the unwritten rule requiring that hospital administrators meet for two hours per dollar they spend, we had to go to a lot of meetings. Fortunately, my partner, our "chief mammographer," bore the brunt. I helped out a little, but only with a gun to my head.

One of the toughest hurdles was figuring out the workflow in our reading room. For efficiency, we have always had two reading stations about 10 feet apart. While the techs hang films on one, we read on the other, and shift sides as necessary. Because we wanted two old screen-film studies hung on every patient before we looked at her digital exam, we needed to keep this double-sided approach. For some reason, the powers that be balked at buying a second $100,000 workstation to sit 10 feet from the first one.

The solution was rather ingenious. A keyboard and mouse sit on each side of a single workstation and the monitor bank pivots. The shared computer is under the table. We can work with the monitors and a viewbox next to each other, while the techs load/unload on the other side. This solution only took 14,000 meetings to figure out, but then, the table did cost seven grand. We ended up going with Hologic, and we're all happy with the choice. The equipment is relatively user-friendly, and the images are exquisite-too much so, in fact. I started in the era of film. Then we went to Xerox. Then back to film with screens. Now digital.

With each change, the resolution improves, and the number of findings increases. This does not make our job faster or easier. All my mental thresholds have to be reset, again. How many calcs represent a "true suspicious cluster" with our new system? I'm sure my recall statistics and six-month follow-up rate have gone ballistic.

At last year's RSNA meeting, Hologic was showing its next technical advance for mammography: tomosynthesis. Using a variant of the old linear tomography, the company is able to obtain hundreds of tomographic slices through each breast with one exposure. Instead of one image in each projection, you get 50. When I whined that I could never read 200 images per exam, Hologic's development chief assured me the company's computer-aided detection system would make it easy.

Our current CAD points out about three to five benign findings per patient, so with 50 times as many images, I will have to check only 150 to 250 spots. How comforting.

Getting the requisite training hours was the easiest step of the process and also the least helpful. Ultimately, new techniques like this are learned by doing, not listening. The more I use our system, the more comfortable I am with it. Of course, I also become more aware of its flaws. The software is very good but not perfect. Like all PACS equipment, the fix is only an upgrade away. The next version of the workstation software is coming any day now. Uh huh.

If I sound skeptical, it may be because the PACS we bought five years ago still doesn't have all the features GE was showing on it at the RSNA meeting that year. Fortunately, my sales rep says they're in the next upgrade.

Dr. Tipler is a private-practice radiologist in Staunton, VA. He can be reached by fax at 540/332-4491 or by e-mail at btipler@medicaltees.com.