Dr. Siegel
is the Director of Imaging at the VA Maryland Healthcare System
and an Associate Professor at the University of Maryland School
of Medicine, Baltimore, MD. He is also a member of the Editorial
Board of this journal.
The long-anticipated transition from the innovators, to the
early adopters, and now to the early majority phase in the
transition from film-based to filmless imaging seems to have
finally arrived. Approximately 5% of hospitals or imaging
outpatient centers have implemented large-scale PACS or filmless
enterprise solutions, and more than 90% of radiologists are reading
at least a subset of their studies using computer workstations at
work or at home. The precipitous increase in the volume of
examinations, the emergence of technologies (such as multidetector
CT), the routine use of additional sequences (such as MR
angiography, diffusion, and perfusion imaging), and the acquisition
of other functional data are increasingly rendering film
ineffective as a tool for diagnostic interpretation. Recent
developments in other modalities (such as combined CT/PET, computed
radiography, or direct radiography using dual-energy subtraction or
tomosynthesis), and innovations in image processing and
enhancement, computer-aided detection (CAD) for mammography, and
lung-nodule detection result in the generation of data that
requires dynamic interactive image review, which also mandates the
use of soft-copy interpretation using a computer workstation.
Unfortunately, most radiologists in practice today do not have
the training or experience required to take advantage of all of the
potential benefits of soft-copy interpretation. Radiology resident
and fellowship training programs are not adequately preparing the
next generation of imaging specialists with background information
and basic strategies required to optimize the benefits of a
soft-copy reading environment. The American Board of Radiology has
not begun to emphasize the importance of focused training in the
use of computer workstations for primary diagnosis and the
requirements for quality control in a soft-copy reading
environment. For example, most programs do not provide formal
training in specific soft-copy navigation functions, such as the
use of "stack" rather than "frame" mode, in the interpretation of
complex cross-sectional imaging studies. A number of studies have
concluded that the use of stack mode results in substantial
improvements both in speed and diagnostic accuracy. The application
of disease- or indication-specific image processing, such as
unsharp masking (edge enhancement) or gray-scale inversion, to
general radiographic images has been shown to increase conspicuity
of life-support lines, pneumothoraces, and lung nodules, which
consequently improves interpretation times and accuracy. The use of
dual-energy subtraction with computed radiography or digital
radiography has been found to improve the radiologist's ability to
detect and distinguish malignant from benign lung nodules.
Unfortunately, these and many developments in image display and
navigation, post-processing, and acquisition have received little
attention in the radiology literature and have not yet been
incorporated into the basic curriculum of radiology training
programs. Other important topics that should be incorporated into
the basic diagnostic imaging curriculum include computer monitor
selection and optimization, DICOM and other imaging standards,
teleradiology, the use and abuse of image compression and "just in
time delivery" strategies, and speech recognition. Although topics
such as these are increasingly being discussed at meetings such as
SCAR (Society of Computer Applications in Radiology) or RSNA
(Radiological Society of North America), the majority of residents,
fellows, and practicing radiologists who do not demonstrate a
special interest or affinity for computer topics in medicine are
interpreting studies in this new soft-copy environment without
formal training or an "instruction manual." While a trial-and-error
approach to instruction and training works well for the newest
generation of video games, I fear that we cannot afford to practice
diagnostic image interpretation for our patients in this manner.
Based on the research that has been performed thus far in the area
of digital imaging, it seems clear that a detailed knowledge of the
issues related to soft-copy interpretation are necessary to be able
to provide optimal care to our patients and guidance to our
referring clinical colleagues.
We need to expand the opportunities for exposure of our
residents, fellows, and the general radiology community to the
unique issues, pitfalls, and challenges associated with digital
imaging and soft-copy interpretation. This should include a
thorough understanding of the benefits and trade-offs associated
with: acquisition devices (such as computed radiography and digital
radiography); the use of teleradiology, image and data display
strategies; basic informatics; clinical aspects of image
processing; and security. This digital imaging curriculum should be
incorporated into the basic core curriculum for radiology residents
and should be an important part of the written and oral radiology
boards as well. Leaving the digital imaging education of
radiologists and the imaging community to chance will undoubtedly
increase the likelihood that non-radiologists, or even non-clinical
professionals, will step in to take over the responsibility for and
associated benefits of digital imaging in this new millennium.