DOI:10.2214/AJR.07.3116
AJR 2008; 190:691-695
© American Roentgen Ray Society
Real-Time Temporal Maximum-Intensity-Projection Imaging of Hepatic Lesions with Contrast-Enhanced Sonography
Stephanie R. Wilson1,2,
Hyun-Jung Jang1,
Tae Kyoung Kim1,
Hiroko Iijima1,3,
Naohisa Kamiyama4 and
Peter N. Burns5
1 Department of Medical Imaging, Toronto General Hospital, University of
Toronto, Toronto, Ontario, Canada.
2 Present address: Diagnostic Imaging, Foothills Medical Centre, 1403 29 St. NW,
Calgary, Alberta T2R 1M5, Canada.
3 Present address: Department of Medicine, Hyogo University, Hyogo, Japan.
4 Toshiba Medical Systems, Tokyo, Japan.
5 Departments of Medical Biophysics and Medical Imaging, University of Toronto,
and Imaging Research, Sunnybrook Health Sciences Centre, Toronto, Ontario,
Canada.
Received January 26, 2007;
accepted after revision September 28, 2007.
Address correspondence to S. R. Wilson.
Supported by the Terry Fox Programme of the National Cancer Institute of
Canada and the Canadian Institutes of Health Research. S. R. Wilson is the
recipient of a research grant from Bristol-Myers Squibb and is an advisor to
Siemens Medical Solutions and Philips Medical Systems. N. Kamiyama is an
employee of Toshiba Medical Systems.
Abstract
OBJECTIVE. We sought to perform a preliminary evaluation of temporal
maximum intensity projection (MIP) of focal hepatic masses in selected
patients. The technique processes real-time contrast-enhanced sonography
images by integrating the path of moving bubbles to depict vascular
morphology. Following a high-intensity ultrasound pulse that disrupts bubbles
within the scan plane, MIP images the trajectories of fresh bubbles
replenishing the plane and revealing their course.
CONCLUSION. Vascular morphology is depicted at a level or detail not
seen before with sonography. High-frame-rate sequences of less than one second
uniquely show arterial structure in liver lesions.
Keywords: contrast agents • liver tumors • maximum intensity projection • microflow imaging • sonography
Introduction
Sonography with microbubble contrast agents and nonlinear techniques such
as pulse inversion imaging [1]
is capable of depicting flow at both the vascular and organ perfusion levels.
One unusual feature of micro-bubble contrast agents is the exceedingly low
doses administered for a clinical study. A typical adult dose, 0.1 mL, of
5-µm-diameter perflutren microspheres corresponds to 200–2,000
microbubbles per milliliter of blood. In malignant tumors, in which the
relative vascular volume is often less than 10%, only 20–200 bubbles are
present in a cubic centimeter of tissue. Although state-of-the-art nonlinear
imaging methods are capable of depicting individual bubbles, the sparseness of
microbubble scatterers in tissue presents a challenge to imaging of the
microcirculation: At any one time, owing to the position of the microbubbles
in space, not the blood vessels but specific locations within the vessels are
depicted. This problem results in a random speckle pattern typical of
contrast-enhanced tissue on microbubble sonograms.
We describe a technique in which the limitation of microbubble technology
is exploited to produce a new depiction of vascular structure. The approach is
simple and takes advantage of two unique aspects of microbubble contrast
perfusion imaging: first, that microbubbles can be imaged in real time with
sonography and, second, that a series of relatively high-amplitude ultrasonic
pulses can disrupt bubbles within the scan plane. Immediately after
disruption, fresh bubbles flow into the scan plane, replenishing the vascular
space [2]. In this study, we
evaluated use of this technique on focal hepatic lesions in selected
patients.
Subjects and Methods
Temporal Maximum Projection Imaging
In essence the method is analogous to photographing a moving light at night
with a long exposure (Fig. 1A).
Figure 1A depicts sparks
randomly and sparsely distributed as they fall toward earth. By integrating
the exposure for an extended period (e.g., by keeping the camera shutter
open), the photographer can produce an image of the route taken by the sparks,
revealing their trajectories. The image displays the aggregate exposure at
each point, known as a maximum intensity projection (MIP). In this study MIP
was implemented on a pulse-inversion contrast-enhanced sonographic image with
relatively simple processing. The method was first described for tracing the
vascular morphologic pattern in malignant tumors of the breast in humans
[3]. In the context of imaging
of lesions of the liver, the effects of tissue motion result in blurring, so
MIP imaging is initiated during a breath-hold after a sono-graphic flash frame
disrupts bubbles in the field of view (Fig.
1B). MIP records the paths taken by new bubbles entering the scan
plane. In a final component of the processing, the strength of echoes is given
an intensity weight that decreases with the elapsed time from the first
appearance of echoes on the image, so that the tracks of echoes gradually fade
from the real-time image, as does the trail of a comet observed in the night
sky.
View larger version (140K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1A —Principles of real-time temporal maximum-intensity-projection
imaging technique. Open-shutter photograph of fireworks in night sky is
comparable with temporal maximum-intensity-projection image. Shutter of camera
is held open for sufficient time to trace path of bright, moving object, such
as sparks of fireworks. Method can be applied to echoes of individual bubbles
of contrast agent detected with nonlinear sonography. (Courtesy of Ben
Burns)
|
|
View larger version (10K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1B —Principles of real-time temporal maximum-intensity-projection
imaging technique. Schematic shows two maximum-intensity-projection (MIP)
imaging sequences. MIP imaging is initiated as bubbles arrive in field of
view. Signal intensifies as bubble paths are tracked. Second sequence is
initiated by high-mechanical-index frames that cause bubble disruption.
Low-mechanical-index imaging then depicts new bubbles as blood flow carries
them into scan plane. MIP processing produces image of track of echoes,
revealing vascular structure.
|
|
Patient Selection
Successfully obtained MIP images were selected from nonconsecutive
contrast-enhanced sono-graphic studies of 65 patients with liver tumors.
Patients with deeply positioned masses, fatty liver, or inability to hold
their breath were excluded. The 37 female and 28 male patients had a mean age
of 49 years (range, 14–92 years). The diagnoses were 19 cases of
hepatocellular carcinoma, 26 of focal nodular hyperplasia, 13 of hemangioma,
three of metastasis, and four of adenoma. The mean diameter of the lesions was
5 cm (range, 1.1–19 cm). The study was approved by our institutional
review board; patients gave informed consent.
Technique
Fifty-five of the patients were examined with maximum intensity pulse
inversion imaging on a prototype system (Toshiba Aplio 80, Toshiba Medical
Systems). Ten patients were examined with comparable technology subsequently
developed (Acuson Sequoia, Siemens Medical Solutions). The contrast agent
perflutren (Definity, DMP115, Bristol-Myers Squibb) was administered IV in
successive boluses of 0.1–0.2 mL, to a maximum dose of 10 µL/kg. From
the first two boluses, digital still and cine loop sequences were recorded in
the arterial and portal phases according to a standard protocol for
contrast-enhanced sonography of the liver
[4]. The mechanical index was
less than 0.2, and the imaging frame rate was 15–20 Hz, depending on
depth. Additional boluses were given for the MIP study according to two
methods. In the first, MIP was initiated at first wash-in of the contrast
agent to the liver. In the second method, a single high-mechanical-index
(mechanical index > 1.0) disruptive flash frame was triggered manually at
the peak of enhancement, and MIP was initiated. With suspended maximum
inspiration and a fixed transducer position, MIP imaging showed the bubbles
either filling or replenishing the scan plane, depending on the method.
Data Analysis
Qualitative assessment of the scans selected for the MIP method was
performed by a single experienced reviewer who was otherwise un-involved in
acquisition and interpretation of the scans. For each patient, depiction of
vascular structure, filling direction, and pattern of enhance ment on MIP
images was compared with that on conventional contrast-enhanced sono-graphic
scans and was graded as worse, equal, or better. Results were tabulated for
each lesion diagnosis. Because the reading was unblinded (MIP images were
readily identifiable), patient selection biased, and the sample size small, no
statistical analysis was performed in this preliminary study.
Results
The reviewer responses are shown in
Table 1. Tumors with linear
vascularity and arter ial hypervascularity benefited most from the MIP method
(Fig. 2A,
2B). (See
www.ajronline.org
for Figs. S2C and S2D.) All four adenomas and approximately two thirds of
focal nodular hyperplastic lesions, hepatocellular carcinomas, and metastatic
lesions had improved delineation of the morphologic features of blood vessels
on MIP images (Fig. 3A,
3B). (See
www.ajronline.org
for Fig. S3C.) Some hypovascular metastatic lesions had dysmorphic arteries
not seen with conventional techniques (Fig.
4A,
4B). (See
www.ajronline.org
for Figs. S4C–S4D.) Assessment of the filling direction was better with
MIP than with conventional contrast imaging of such lesions. Hemangioma,
characterized by peripheral puddles rather than linear vascularity, exhibited
unconvincing improve ment in the parameters assessed. Overall, MIP was found
comparable with conventional imaging of this lesion. Hemangioma was the only
lesion with a grade of worse for all three features; this finding was reported
in approximately one third of cases. Nonetheless, MIP imaging of some rapidly
filling hemangiomas well depicted the vascular structure
(Fig. 5). (See
www.ajronline.org
for Fig. S5.)
View this table:
[in this window]
[in a new window]
|
TABLE 1: Results of Single-Reviewer Assessment of 65 Lesions Comparing
Conventional with Maximum-Intensity-Projection Imaging in Arterial Phase of
Contrast-Enhanced Sonography
|
|
View larger version (160K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2A —91-year-old man with hepatocellular carcinoma. Advantage of
maximum-intensity-projection imaging of highly vascularized lesion. See also
Figures S2C and S2D, cine loops, in supplemental data online. Conventional
contrast-enhanced sonographic image shows only heterogeneous bright ball of
enhancement with no vessel detail as contrast agent rapidly fills entire
vascular bed.
|
|
View larger version (171K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2B —91-year-old man with hepatocellular carcinoma. Advantage of
maximum-intensity-projection imaging of highly vascularized lesion. See also
Figures S2C and S2D, cine loops, in supplemental data online.
Maximum-intensity-projection image obtained 0.5 second after flash shows
morphologic features of individual tumor vessels.
|
|
View larger version (152K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3A —24-year-old asymptomatic woman with incidentally discovered
liver mass (focal nodular hyperplasia). Maximum-intensity-projection imaging
shows vascular morphologic features and direction of lesional filling in
highly vascularized lesion. See also Figure S3C, cine loop, in supplemental
data online. Conventional sonographic image obtained 9 seconds after the end
of saline flush shows homogeneous enhancement of mass. Lesional vessels are
not visible because of very rapid homogeneous filling of vessels.
|
|
View larger version (149K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3B —24-year-old asymptomatic woman with incidentally discovered
liver mass (focal nodular hyperplasia). Maximum-intensity-projection imaging
shows vascular morphologic features and direction of lesional filling in
highly vascularized lesion. See also Figure S3C, cine loop, in supplemental
data online. Maximum-intensity-projection image obtained 0.4 second after
flash shows stellate vessels and centrifugal filling pattern.
|
|
View larger version (183K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4A —45-year-old man with jaundice due to biopsy-proven, poorly
differentiated adenocarcinoma. Images show advantage of
maximum-intensity-projection imaging of poorly vascularized lesion. See also
Figures S4C and S4D, cine loops, in supplemental data online. Conventional
sonographic image obtained during wash-in of contrast agent shows isolated
bubbles within tumor without vessel detail.
|
|
View larger version (163K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4B —45-year-old man with jaundice due to biopsy-proven, poorly
differentiated adenocarcinoma. Images show advantage of
maximum-intensity-projection imaging of poorly vascularized lesion. See also
Figures S4C and S4D, cine loops, in supplemental data online.
Maximum-intensity-projection image obtained 5.8 seconds after flash shows
vessel detail within hypoperfused lesion.
|
|
View larger version (147K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5 —35-year-old man with hemangioma. See also Figure S5, cine
loop, in supplemental data online. Maximum-intensity-projection (MIP) image
shows rapidly perfused lesion. MIP image obtained 9.4 seconds after onset of
arterial phase enhancement shows puddles of contrast material around periphery
of lesion. Fine-vessel detail is evident in surrounding normal liver. If
hemangioma is extremely slowly perfused, MIP technique may not depict
vascularization within single breath-hold.
|
|
Discussion
Noninvasive diagnosis of focal hepatic masses with CT and MRI is based on
contrast enhancement patterns. Most studies of characterization of liver
masses with contrast-enhanced sonography have focused on the same enhancement
features. Vessel structure as a parameter for characterization of liver masses
is rarely mentioned. The direction of lesional filling has not been described,
to our knowledge, undoubtedly because of the very rapid enhancement of the
microvasculature as the contrast agent enters the field of view.
MIP is a reproducible method for showing both vascular structure and the
direction of lesional filling. Our preliminary experience suggests that the
technique is not suitable for all patients because of the requirement for a
motionless field of view. Inability to maintain suspended respiration was the
most common reason for failed MIP imaging. However, our findings and those of
Linden et al. [5] suggest that
when successful the MIP technique has considerable potential benefit. The
greatest advantage of the technique appears to be assessment of rapidly
filling hypervascular lesions. With conventional contrast-enhanced imaging of
such masses, filling of the entire vascular bed, including both feeding
vessels and capillaries, is often so rapid that the lesion is instantly seen
as a bright echogenic ball
[6–8].
Before the availability of MIP, we used other maneuvers to show filling
direction and vessel structure, including selection of a slightly higher
mechanical index to reduce filling of the vascular bed and emphasize contrast
enhancement in larger vessels with faster-flowing blood.
Our study showed consistent improvement in imaging of hypervascular masses
such as adenoma, focal nodular hyperplasia, and hepatocellular carcinoma.
Experience thus far suggests that imaging of focal nodular hyperplasia and
adenoma will benefit most from the MIP technique because filling direction and
vascular structure are the main components in differentiation of these
lesions. Imaging of hemangioma, on the other hand, which may not show uniform
distribution of peripheral nodularity and for which movement of the transducer
is used to localize peripheral puddles of enhancement, did not have such a
benefit. Furthermore, slow filling of many hemangiomas makes evaluation of
their vascularity within the time frame of a breath-hold impossible. We rarely
use MIP in imaging of patients with suspected hemangioma. The findings in the
few cases of metastasis studied suggest that MIP may improve characterization
of lesional vascularity and direction of filling. Identification of dysmorphic
arteries, seen with MIP of malignant liver masses, also has potential benefit
for diagnosis. For example, identification of dysmorphic linear vessels in a
hypovascular mass may be convincing evidence of the presence of metastasis,
whereas identification of peripheral pooling would suggest the presence of
hemangioma.
Selection of patients for this study was clearly biased to favor the
technique; even so, we encountered frequent technical failures due to motion.
However, at imaging of many patients with good breath-holding skills and a
relatively superficial hypervascular mass, excellent images were obtained with
unprecedented depiction of lesional and liver vasculature
(Fig. 6). (See
www.ajronline.org
for Fig. S6.) We use MIP in approximately 10% of contrast-enhanced sonographic
examinations, particularly of patients with a well-depicted superficial liver
mass, especially if thought to be focal nodular hyperplasia or adenoma. The
adverse effect of motion is important: A failed contrast injection cannot be
salvaged because the arterial phase is missed and the use of flash technique
makes further interpretation of subsequent portal venous phase images
unreliable. Further development with tissue correlation methods of motion
correction may improve the technique. Furthermore, MIP imaging is qualitative.
Integration results in an image that cannot be relied on as an indicator of
microbubble concentration or perfusion.
View larger version (161K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6 —54-year-old man with inflammatory bowel disease and
incidentally detected liver mass. See also Figure S6, cine loop, in
supplemental data online. Maximum-intensity-projection image of normal liver
vasculature shows accumulated enhancement in 11 seconds after contrast
material arrives in liver. Unprecedented depiction of vessel structure to
fifth order branching is evident. Focal unenhanced region is slowly perfusing
hemangioma, which does not have contrast accumulation, making diagnosis
impossible.
|
|
Contrast-enhanced sonographic MIP imaging shows structural vascular details
previously unavailable on conventional cross-sectional images. The benefit of
this technique for diagnosis is yet to be tested in a prospective trial. The
adverse effect of motion is the principal factor limiting general use of the
technique.
References
- Burns PN, Wilson SR, Hope Simpson D. Pulse inversion imaging of
liver blood flow: an improved method for characterization of focal masses with
microbubble contrast. Invest Radiol 2000;35
: 58–71[CrossRef][Medline]
- Wei K, Jayaweera AR, Firoozan S. Quantification of myocardial blood
flow with ultrasound-induced destruction of microbubbles administered as a
constant venous infusion. Circulation1998; 97:473
–483[Abstract/Free Full Text]
- Muradali D, Kulkarni S, Villet C, Burns PN, Kalman P, Wilson SR.
Angiogenesis mapping with pulse inversion: a new and sensitive method to image
vascularity in breast carcinomas. (abstr) Radiology2002; 225(P):367[Abstract/Free Full Text]
- Wilson SR, Burns PN. An algorithm for the diagnosis of focal liver
masses using microbubble contrast-enhanced pulse-inversion sonography.
AJR 2006; 186:1401
–1412[Abstract/Free Full Text]
- Linden RA, Gittens PR, Forsberg F, Trabulsi EJ, Gomella LG, Halpern
EJ. Directed prostate biopsies utilizing microflow imaging during contrast
enhanced ultrasound. (abstr 352) 2007 Prostate Cancer
Symposium. Alexandria, VA: American Society of Clinical Oncology,2007
- Dietrich CF, Schuessler G, Trojan J, Fellbaum C, Ignee A.
Differentiation of focal nodular hyperplasia and hepatocellular adenoma by
contrast-enhanced ultrasound. Br J Radiol2005; 78:704
–707[Abstract/Free Full Text]
- Albrecht T, Blomley M, Bolondi L, et al. Guidelines for the use of
contrast agents in ultrasound. Ultraschall Med2004; 25:249
–256[CrossRef][Medline]
- Burns PN, Wilson SR. Focal liver masses: enhancement patterns on
contrast-enhanced images—concordance of US scans with CT scans and MR
Images. Radiology 2007;242
: 162–174[CrossRef][Medline]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
|
|
 
S. R. Wilson, L. D. Greenbaum, and B. B. Goldberg
Contrast-Enhanced Ultrasound: What Is the Evidence and What Are the Obstacles?
Am. J. Roentgenol.,
July 1, 2009;
193(1):
55 - 60.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
