is a Professor of Radiology, the Gerald D. Dodd, Jr.
Distinguished Chair in Diagnostic Imaging, and Director of
Academic Development; and
are Assistant Professors of Radiology at the University of Texas
M.D. Anderson Cancer Center, Houston, TX.
In the 1970s, while employing traditional step-and-shoot
computed tomographic (CT) technology, significant efforts were made
to prolong contrast administration so that it continued for the
duration of the scan.
Thus, examinations that required long scanning times (ie, excessive
coverage in the z-axis) necessitated the use of relatively large
volumes of contrast (150 to 175 mL) at slow rates. In order to
achieve adequate contrast enhancement of the liver, a bolus
followed by a drip infusion or multiple injection rates using a
power injector were used. Biphasic injections with an initial fixed
rate of 1 to 3 mL/sec for the early portion of the examination and
even slower rates for the latter portion of the study were common.
Concentrations of iodinated contrast rarely exceeded 300 mg I/mL.
Higher-concentration contrast materials were, in fact, detrimental,
since the volume needed would result in delivery of excessive grams
of iodine. The use of power injectors was found to provide a
greater level of consistency for contrast enhancement compared with
a manual injection technique, but the rate of contrast injection
was limited because of slow scanner technology and a general
inexperience by practicing radiologists.
Significant safety concerns regarding the adverse effects of
extravasation of commonly used ionic contrast media were
Helical CT contrast dynamics
The introduction of single-slice helical (spiral) CT scanning in
the 1980s allowed for the elimination of the standard interscan
delay of 6 to 7 seconds between slices. The continuous acquisition
of slices with a scan time of 1 second and breath-hold acquisitions
ranging from 15 to 30 seconds had a profound impact on allowing
imaging during optimal phases of contrast enchancement.
The organ for which this impact was most profound was the liver.
With the introduction of helical scanning, the entire liver could
be imaged during the portal venous phase (PVP). This was ideal for
detecting the vast majority of liver lesions, which are
hypovascular in nature and can be detected by virtue of their lower
density compared with the surrounding densely enhanced liver
parenchyma (Figure 1). Radiologists could perform examinations of
the entire liver, avoiding the equilibrium phase during which
lesions often become less visible or even isodense and invisible to
detection (Figures 2 and 3). Whole-organ scanning became a
With helical CT and subsecond helical scan times, optimal liver
imaging was combined with improved image quality. It also allowed
for extended examinations of the lower abdomen, pelvis, and chest
with decreased motion artifact. The standard contrast agents used
in the early years of CT were still primarily ionic, high-osmolar
contrast material (HOCM), which had the benefit of being relatively
However, with more widespread availability of newer and safer
nonionic, low-osmolar contrast material (LOCM) and a concomitant
decrease in their price, these new nonionic agents were used
selectively. With the widespread use of power injectors, increased
rates of contrast administration were employed to match faster
scanners. In order to avoid even mild adverse reactions, which
could ruin an entire helical acquisition, LOCM began to be used
more widely, almost universally.
Understanding the contrast dynamics in a structure as complex as
the liver became important with the new scanning flexibility
afforded by single-slice helical CT. Time-density curves of the
liver and lesions became more prevalent as a means to understanding
how protocols could be developed to better detect liver lesions
with this new technology (Figure 4).
The liver enhancement curve is characterized by a fairly steep
uprise, although not as dramatic an uprise as an aortic enhancement
curve. It reaches a plateau followed by a slow and prolonged
downward curve. When a curve is made of tumor enhancement for
hypovascular lesions, the vast majority of metastases generally
have a slower upswing with a longer time to peak enhancement, which
is shorter than the time to peak enhancement of the liver.
Hypovascular metastases then have a slow degradation of
enhancement. The point at which the liver and tumor enhancement
curves begin to parallel each other and decline is the equilibrium
Although it is not truly representative of equilibration of
contrast at the cellular level, this phase does illustrate an
important phenomenon. It is during this phase that it is most
dangerous to image the liver, since lesions are less easily
discriminated from the surrounding liver. This can result in
erroneous interpretation of disease becoming smaller for technical
reasons without a true improvement of the disease (ie, regression
or cure) (Figure 5).
Contrast-enhancement protocols for helical CT
Optimal characterization of lesions mandates careful technique.
It is imperative that routine hepatic scanning be performed during
the optimal PVP, which occurs approximately 70 seconds after the
initiation of injection and extends just to the point prior to the
beginning of the equilibrium phase. For optimal imaging, one would
like to have the maximal difference in enhancement between the
liver and the lesion during the time the scans of the liver are
performed. The faster the scanner, the more easily one can image
purely during this phase, eliminating the equilibrium phase and the
chance of liver lesions becoming less conspicuous.
Understanding contrast enhancement protocols, especially for
liver imaging, became more critical with helical scanning, since it
provided the flexibility of imaging during the optimal PVP to
detect hypovascular liver lesions.
Also, for the first time, exploitation of multiphasic imaging was
possible, allowing the behavior of lesions to be followed over
time. Characterization of benign lesions (such as cavernous
hemangioma, adenoma, and focal nodular hyperplasia) became a common
role for CT (Figure 6). With the use of multiphasic imaging,
scanning could be performed early in the bolus, the hepatic
arterial dominant phase (HADP), 15 to 20 seconds after contrast
administration, and then later at 65 to 70 seconds to catch the PVP
(Figures 7 through 10). Detection of hypervascular lesions by CT
could also be achieved utilizing early scans prior to peak liver
Malignant hypervascular lesions include primary tumors, such as
hepatocellular carcinoma, hemangioendothelioma, and angiosarcoma.
Hypervascular metastases include renal cell carcinoma, carcinoid,
thyroid carcinoma, neuroendocrine tumors, choriocarcinoma,
melanoma, and, in some cases, breast cancer. All of these
hypervascular tumors are best imaged when an early enhancement
phase, HADP, is included; frequently, these tumors can be imaged
only in this early phase.
Computer-automated scanning technology and bolus
Body-imaging protocols have been primarily constructed with
significant attention to optimizing imaging of the main target
organ, the liver. For liver imaging, the delay time prior to the
initiation of scanning has generally been fixed and has ranged in
different practices from 60 to 70 seconds between the onset of
injection of contrast and the initiation of diagnostic scanning.
Unfortunately, patients vary significantly in body mass, and there
can also be significant changes in cardiac output and circulation
time, depending on the clinical status of individual patients and
variations in an individual patientfrom scan to scan. Thus, one
size does not fit all (Table 1). A specific delay time for one
individual may not be optimal for the next individual. This becomes
most critical with the faster scanning provided with the evolution
of multislice CT (MSCT). Research on this problem has demonstrated
that a bolus-tracking computer-automated scanning technology (CAST)
can be highly effective in allowing scanning during optimal
enhancement for critical organs such as the liver.
This technology has been called various names by different
manufacturers, including SmartPrep (GE Medical Systems, Milwaukee,
WI) and Care (Siemens Medical Systems, Iselin, NJ). Employing CAST,
multiple low-radiation-dose scans can be performed at the mid-level
of the liver during the administration of contrast material using
the available software. Cursors denoting regions of interest (ROI)
can be placed on the liver and the aorta, and the enhancement of
the structures is tracked automatically on a graph (Figure 11).
Switching to routine diagnostic scanning when the enhancement of
the liver reaches 50 Houndsfield Units (HU) allows optimal imaging
of the liver by capturing the optimal level of enhancement. This
method allows for better and more reliable enhancement of the liver
and associated abdominal structures (Figures 12 and 13). It has
also been found that an equivalent degree of enhancement can even
be achieved using less contrast material, and that a higher degree
of enhancement can be achieved using the same amount of contrast
that would be normally used with a standard, fixed-delay time
period (Figure 14). This technology, when incorporated with MSCT,
can allow for precise timing to capture different phases of
contrast enhancement. Though most frequently used to time the
hepatic venous phase, which occurs for 65 to 75 seconds, it has
also been used in scanning the liver for hypervascular metastatic
disease by optimally timing for the arterial contrast phase.
Helical CT has greatly enhanced our understanding of the
pharmacokinetics of contrast administration for imaging of the
liver. The ability of helical scanning to scan more rapidly through
the liver has allowed better detection of hypovascular lesions. For
the first time, hypervascular lesions can be adequately assessed by
scanning through the liver during the hepatic arterial dominant
phase as well as during the portal venous phase. With the
technological advance of MSCT, truly practical multiphasic imaging
has become feasible.
Part II of this article will address the introduction of
multislice technology and will be published in the June 2003 issue