Netherlands' national breast cancer screening program goes digital

Going filmless for one radiology department brings many logistical and cost issues. Multiply those issues several times over, and you have what it takes to digitize an entire national breast cancer screening program. With a target population of about two million women, the Dutch breast cancer screening program aims to use digitization to increase the levels of efficiency and cost savings needed to provide the best diagnosis and clinical care.

In the Netherlands, all women aged 49 to 75 are invited every two years to participate in the breast cancer screening program. The country is divided into nine regions, and each has a number of mobile screening units, reducing travel distance for the participants. Digitized images of previous examinations will be put on CD-ROMs and transported to the units the day before the women are scheduled. The films are then read in reading centers where the films are archived. All images are read independently by two radiologists. In total there are 65 screening units and 25 reading units.

About 1% of the women are referred to a hospital for further examinations, and for about half of those women, cancer is detected.

DIGITIZING THE DATA

The Dutch program has devised an elaborate plan for going all-digital, realizing cost savings by eliminating film and equipment. In the mobile units, film-based mammograms will be replaced by a full-field digital mammograms, and film development equipment can be eliminated. Simple two-screen viewing stations with adequate resolution will allow the technician to view the newly acquired images and compare these with the previous images on CD-ROMs. This may lead to the acquisition of additional images. In the evening, that day's production will be transported on CD-ROM to a reading center.

In the reading center, advanced workstations will be available to view the images on high-resolution screens. CAD software available in the reading center will help in the analysis of the images and will give indications of suspect regions.

When the radiologists want to see images of older screenings they can request these from the digital archive. The reading process is suspended until the old images become available.

As for quality control, test images with known abnormalities can be inserted in the flow of images to be read.

ROUND-THE-CLOCK DIGITIZATION

We intend to switch to digital reading as soon as the digital acquisition apparatus is operational. Because the reading of digital images alongside traditional film is not recommended, we will digitize old film images that are still needed. This will be done on a routine basis for images from previous screenings and for older images on request. We foresee a digitization center where five to eight high-speed, high-resolution digitizers will work round-the-clock, at least for the first two years.

The digitized images will be stored in a central archive and duplicated for availability. We are making plans for special image management software to cope with the characteristic image usage pattern.

COST AND BENEFIT ANALYSIS

We have developed a cost model for this process. In steady state, the costs of digitized screening will not be higher than those of film-based screening. Savings come from the reduction of film, chemicals, development machines, and archive personnel. There were some extra costs, however, stemming from new workstations, viewing stations, digital mammograms, software, image archiving, and the development of the digitization center. During the transition, total extra costs will be about 15% of the total annual costs of the screening program.

In addition to cost savings, we expect to achieve better diagnosis due to a variety of factors: the addition of CAD software, test images, and truly independent reading. Going filmless will introduce less radiation exposure and reduce harmful environmental effects by eliminating chemicals. Because images will be available anywhere, there is the potential for lower logistical costs and the automatic creation of an accessible, large-scale image database for research purposes.

In the EU project "Screen trial," a pilot is being conducted with a digital mammogram in the nonmobile screening unit in Utrecht. We are still looking into further specification of the software needed for digitization of the full screening program. Although the money needed for migration to digital screening seems to be in the long-term financial planning of the Ministry of Health, we are waiting for financial support of the government to continue the preparation.

Ab Bakker, Ed Mos, and André van Peppen
e-mail:abakker@addabit.demon.nl; mos@met.tno.nl; avp@ikca.nl; avp@ikca.nl