The curiosity of what lies within an object consumes everyone whether it is a child
taking apart a toy or a surgeon performing a diagnostic laparotomy. Radiology is the
branch of medicine that has been entrusted with taking pictures of deep structures
within the body and laying open what is inside. Plain X-rays and various radiographic
views provide a fascinating look at the skeleton and its diseases. These are easy
to interpret as they look like dry bones held in the hand. However, in the realm of
cross-sectional imaging, radiology spews out a profusion of peculiar pictures that
require to be stacked up mentally to imagine the whole organ or disease as we see
at dissection. Though radiologists are experts at such imagining of a whole structure
such as the common bile duct or ureter from cross sections, physicians and surgeons
find it very challenging to form such mental images. One surgeon friend mentioned
to me that he finds these as gibberishly challenging as a series of software code
([Fig. 1]).
Fig. 1 Cartoon showing a clinician perplexed by innumerable axial sections through the abdomen.
To help understand and analyze cross sections, three-dimensional (3D) reconstructions
of whole organs can now be performed by advanced computer software at workstations,
especially with computed tomographic (CT) scans. These allow us to visualize the human
body like never before. Sadly, these are often not done routinely, and the vast majority
of cross-sectional reports are dispatched with just the standard axial images. Volumetric
acquisition of scans that is key to these reconstructions is also not routinely performed.
The reason offered by most radiology departments for not providing detailed reconstructed
images is that it is time consuming and technically difficult. Some departments in
the West have 3D laboratories that have technologists generate and push these static
images to the PACS for the radiologist to read. This laboratory is an addition to
the radiology department cost causing concern. The static images are also vulnerable
to artifacts that are created due to technique and do not offer an opportunity for
real-time data manipulation and rotation of the image around the axis required. For
example, the common bile duct cancer and its relationship to the hepatic arteries
that are 2 to 3 mm in diameter, and sometimes smaller can be assessed by making an
oblique volumetric reconstruction along the plane of the duct and rotating it around
its axis to bring the artery in view as well as doing the same for the artery. These
complex manoeuvres are akin to an artist using the brush with dexterity bringing out
on the canvas the image that the mind has formed ([Fig. 2]).
Fig. 2 Panel A is an oblique coronal reconstruction in the plane of the common bile duct
showing its entire length and the cholangiocarcinoma (white arrow) involving the common
hepatic duct. Panel B is an artistic rendition of the same image as panel A.
Generation of 3D images requires an artistic disposition and eye. Today's radiologist
is neither trained in this nor understands the value of this. These are useful not
only for the referring clinician's use but also for the accuracy of image interpretation
and final diagnosis.
Goodman and Kelleher[1] in an interesting experiment connected radiology to art. They demonstrated that
after a focused session of interpretation, training at a local art gallery where art
experts taught novice radiology trainees how to thoroughly analyze a painting, first-year
residents’ ability to localize imaging abnormalities improved enormously.
There is significant emphasis on the way a radiology report should be written and
much current debate about the pros and cons of a structured versus unstructured report.
However, little literature exists on how cross-sectional radiology images should be
generated to demonstrate the pathology. Representative imaging views and spot filming
appear to be restricted to the two-dimensional X-ray and procedures such as barium
studies for the gastrointestinal tract and the pyelo and cystograms for the urinary
tract. As these studies become less and less utilized, the art of demonstration of
pathology by a specific view seems to be dying.
Generation of representative images to demonstrate the findings in a radiology report
is as important as the report itself. This is of great value especially in abdominal
radiology where subtle gray shade differentiation is all one can rely on to separate
disease from normal tissue. Images need to be in an interpretable format for a clinician
who may not have an in-depth knowledge of radiology or the time required to study
0.50-mm-thin sections from the dome of diaphragm to pubic symphysis in arterial, early
and late venous, and delayed phases that could yield a mind boggling 4,000 to 5,000
axial sections. The will and skill to produce one perfect masterpiece of a 3D picture
that would sum up several lines in a written report and take one look to comprehend
is what makes the science of radiology an art.
