| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Diagnostic Imaging, University of Calgary Foothills Medical
Centre, 1403 29 St., NW, Calgary, AB T2N 2T9, Canada.
2 Department of Community Health Services, University of Calgary, Calgary, AB,
Canada.
3 Department of Medical Imaging, Toronto General Hospital, University of
Toronto, Toronto, ON, Canada.
Received December 18, 2008;
accepted after revision February 26, 2009.
S. R. Wilson is on the Advisory Board for Ultrasound, Philips Healthcare,
and received a research grant from Lantheus Medical Imaging.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Mechanical volumetric acquisitions of each abdominal organ, including the liver, gallbladder, pancreas, kidneys, spleen, bowel, and aorta, were performed in 200 consecutive patients.
RESULTS. One thousand four hundred fifty-four volume data sets were graded for feasibility of performance and technical adequacy from I (impossible, incomplete) to V (excellent, complete). The most successfully imaged organ was the right kidney (grades IV and V, 95.0%) and the least successfully imaged, the spleen (grades IV and V, 69.0%). Very good to excellent grades (IV and V) were obtained in 1,215 (83.6%) of the 1,454 volumes. One hundred twelve (7.7%) of the 1,454 volumes were failures (grades I and II). The three organs with the highest success compared with the right kidney were the left kidney, gallbladder, and liver. The data sets of all the other organs showed a statistically significant difference in the feasibility of performance from the right kidney. Liver acquisition failures were associated with end-stage liver cirrhosis (n = 6), fatty liver (n = 3), and obesity (n = 3). Other acquisition failures, similar to conventional sonography, were associated with bowel gas interference and poor acoustic window. The technical limitations include poor resolution in the B and C planes and a limited range of frequencies; these limitations can be overcome in the future with matrix transducers and introduction of the technology to a broader frequency range.
CONCLUSION. Volumetric acquisition in the abdomen performed using defined guidelines is feasible with recognized limitations. Technology advances will improve this imaging technique in the future.
Keywords: 3D volume imaging • abdomen • abdominal imaging • bowel • imaging guidelines • ultrasound technique
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Ultrasound has many advantages for diagnostic imaging including excellent spatial and contrast resolution, a relatively low cost, and lack of ionizing radiation. However, there are two additional important aspects of ultrasound acquisition that maintain ultrasound as a major player in the assessment of abdominal pathology: its real-time and multiplanar capabilities, which allow imaging in the ideal plane for the best interpretation of pathology within a region of interest. Both of these advantages are included in the most recently introduced technique for ultrasound imaging known as "volumetric image acquisition."
Today, the acquisition of a volume of ultrasound data is performed freehand or by a mechanically driven multielement array transducer that can acquire a data set as the array sweeps through a predetermined angle of acquisition during a breath-hold. The volumetric data set results from the combination of information through each plane of the sweep. The data acquired then allow electronic presentation of images in the acquisition A plane; the perpendicular B plane; and the unachievable plane in real-time sonography, the C plane. These data may then be reviewed either online or off-line as cine files in the three planes or as series of images, analogous to the stacking of images used on CT and MRI. Multiplanar reconstruction (MPR) allows the creation of image planes that are at unique and often unattainable angles. The surface-rendered image, familiar in obstetric volume imaging [1], plays a small role in imaging the abdomen, where fluid surrounding the surface of organs is not the norm.
In spite of identified problems with resolution and image display, volumetric imaging techniques are increasingly popular and accepted in obstetric [1-6] and cardiac [7, 8] ultrasound. It is thought intuitively that abdominal ultrasound might not lend itself well to volumetric imaging for several reasons including the multiplicity of organs, each requiring its own volume data set; the large size and variable shapes of organs; and the presence of bowel gas, which may interfere with visualization.
In 2007, the capability to perform volumetric evaluation of the abdominal organs was introduced to our ultrasound facility. In this article, we present the feasibility of performing volumetric acquisitions of data of the abdominal organs of 200 consecutive patients using our performance guidelines developed in a preliminary experience at our institution.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The study group included 97 men and 103 women who ranged in age from 18 to 85 years (mean age, 51.5 years). Our referral base included three dominant patient populations: those with risk factors for chronic liver disease sent for hepatoma surveillance (n = 61) or hepatoma evaluation (n = 15), those with known or suspected inflammatory bowel disease (IBD) referred to determine the extent and activity of bowel disease (n = 99), and those referred from our hepatobiliary and pancreatic surgical oncology team (n = 13). Other indications (n = 12) included various medical problems such as elevated liver function test results, elevated creatinine level, and renal mass characterization.
All volumetric scans were obtained on an ultrasound system (iU22 with a V6-2 transducer, Philips Healthcare). The V6-2 volumetric probe is a mechanical volumetric data acquisition probe. The piezoelectric crystals are moved mechanically in an arc with equal movement to each side of the selected center point. The extent of each sweep is determined by selection of an angle of acquisition, between 50° and 75°. The hand holding the transducer is maintained in a fixed position for the duration of volume capture.
Volumetric data were analyzed on two dedicated off-line software systems (View Forum 6.1 and Q Laboratory, Philips Healthcare).
Ultrasound Volumetric Data Acquisition Technique
All of the volumetric acquisitions were performed according to guidelines
formulated based on a succession of introductory scans obtained in our
ultrasound department from August 2007 until March 2008. These guidelines are
summarized in Table 1 and
include four steps: a prescan, the volume acquisition, a volume review, and an
additional scan to include any missed components in the volume data set before
storage and completion of volumetric scanning. The prescan includes a
real-time evaluation of the organ in question, performed with several
important objectives. First, the prescan guides transducer
placement—subcostal or intercostal, for example—and the optimal
plane and size of the angle for the acquisition volume. Further, it determines
the optimal patient position, either supine or lateral decubitus, and the
optimal phase of respiration for the breath-hold. Last and most important, the
prescan allows the recognition of abnormalities within the organ.
|
The volume acquisition was performed with a breath-hold to minimize motion artifact in the data set, using the optimal plane and transducer placement. A broad focal zone, set at the center of the acquisition plane, is ideal to encompass as much of the region of interest as possible. The data set was reviewed live and stored if acceptable or was repeated if unacceptable. Ultimately, before the completion of the examination, determination that all observed abnormalities had been included within the volume was essential. If the acquired volume was insufficient, it was repeated or additional acquisitions were performed as necessary.
In the following sections, we discuss organ-specific guidelines for volumetric acquisition.
The kidneys—The ideal volume is acquired in the long axis of the kidney in either a sagittal or a coronal plane (Fig. 1). An angle of sweep of approximately 50° is generally adequate, but the angle should be enlarged to include focal abnormalities or enlargement of the kidney.
|
|
|
|
|
The spleen—The technique for splenic volume acquisition is the most variable for the abdominal organs and reflects the difficult visualization of the spleen related to its high position in the left upper quadrant deep to the inferior costal margin. Nonetheless, we recommend a supine patient position and an intercostal transducer placement to show the maximal span of the spleen in a single image, suspension of respiration, and acquisition of the volume. Alternatively, if the spleen is better seen with subcostal transducer placement, obtaining the volume in the long axis of the spleen is best. This is true especially if the spleen is enlarged.
The bowel—The best acquisition plane of the bowel is the long axis so that both the bowel and the perienteric soft tissues are included. The focal zone should be relatively narrow and should be centered on the abnormal bowel loop. The angle of acquisition is set at the default, about 50°.
The aorta—The volume is acquired in the long axis of the aorta in two segments: the first to include the upper abdominal aorta and the second, the aortic bifurcation.
Supine patient positioning and a narrow acquisition angle are most often optimal.
Data Analysis
For each organ examined, the volumetric acquisitions were graded for
technical success or feasibility of performance from I (total failure [i.e.,
0%]) to V (total success [i.e., 100%]) at the time of their performance by the
sonographer or physician performing the scanning. Initially the volume sweep
was reviewed on the iU22 ultrasound system. If the radiologist or the
technologist was satisfied with the quality of the sweep, it was stored on the
hard drive of the iU22 ultrasound system. The stored sweeps were then
transferred to the workstations, Q Laboratory and View Forum. The data were
transferred from the ultrasound system to the dedicated off-line workstations
at the same time or at a later date. The data were not altered before review
by the radiologist. The radiologist reconstructed the data in all three planes
using the Q Laboratory or View Forum workstation. Depending on the complexity
of the findings in a particular case, the reconstruction and review time
varied from 10 to 20 minutes for all the volumes in a single patient.
A grade of V was associated with a technically successful acquisition with complete inclusion of the organ and its abnormality within the volume. A grade of I occurred when an acoustic window allowing the acquisition of a volume of data could not be obtained. Grades of II, III, and IV describe poor (25%), satisfactory (50%), and good (75%) acquisition results, respectively. When a grade of I (failure) or II (poor) was assigned to an organ, two additional assessments were made: the technical success of routine single-frame imaging and the reason for the failed or poor volumetric scan.
Display
Before a grade was assigned, the data were reviewed with multiple-frame
cine files in the acquisition A, perpendicular B, and C planes. The acquired
volume data can be displayed in various formats. Cine files, which allow
real-time evaluation of the entire region of interest, can be viewed not only
in the acquisition plane but also in the B and C planes with the use of MPR.
From these files, single images, similar to those obtained with standard
technique, are also possible in any plane. Unique to the technique is the
ability to produce a multislice image that can have variable slice thickness.
We generally prefer a nine-on-one format with the center point set on the
pathology or center point of the organ. This point is also generally the
fulcrum of the acquisition. The number of images in a multislice image file
can be chosen according to the need and size of the organ, ranging from three
images to as many as the sonologist wants. The thickness of the slices can
also be selected according to the need from 1 mm, which produces the best
resolution, to thicker slices, which progressively lose resolution and also
have a more substantial skip area that may result in omission of information
from the file.
In grading the feasibility of the performance of volumetric acquisitions, the loss of resolution in the B and C planes associated with the mechanical method of acquisition was ignored. However, our equipment provided only a single transducer, a V6-2, with a mean operating frequency suitable for scanning patients and organs to a standard depth of up to 15 cm. This transducer did not allow successful acquisitions of superficial organs, especially in thin patients, and we graded the acquisition as poor if the area of interest could not be adequately resolved for a proper evaluation.
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The results for feasibility of performance of volumetric acquisition are summarized graphically as a horizontal bar chart in Figure 6. The most successfully imaged organ, the right kidney (grades IV and V, 95.0%), is displayed at the top and the least successfully imaged organ, the spleen (grades IV and V, 69.0%), is shown at the bottom.
Table 2 shows the percentage of volumetric acquisitions graded as IV and V (75-100% success) with CIs by organ, as compared with the most successfully studied organ, the right kidney, and a p value for the difference between the successes. The three organs with the highest success as compared with the reference standard were the left kidney, gallbladder, and liver. The remaining organs—the pancreas, aorta, bowel, and spleen—showed statistically significant differences in the percentage of acquisitions graded as good or as a total success as compared with the reference organ.
|
Volumetric acquisitions of the liver, unlike those of the right kidney, however, were graded I or II in 12 patients (6%); this poor performance is attributed to end-stage cirrhosis (n = 6), gross fatty liver (n = 3), or morbid obesity (n = 3). Less common problems causing acquisition failure in two additional patients were a gigantic exophytic liver mass measuring greater than 20 cm in diameter and an enormous caudate lobe in a cirrhotic liver, both of which prevented placement of the transducer in a satisfactory position to allow acquisition of the complete volume of data.
Volumetric acquisitions of the pancreas and the aorta showed identical and more varied results. Both showed excellent (grades IV and V, 76.5%) and also very poor (grades I and II, 9%) results. Acquisition failures, similar to those encountered in routine pancreatic scanning, were related to interference by bowel gas in the epigastrium preventing visualization of all or of a portion of the gland. Poor visualization of the aorta was associated with both interference by bowel gas and morbid obesity.
Many of the splenic volumes included a band of interference as a rib shadow that was frequently incorporated in the volume acquisition. Nonetheless, this shadow did not, in most cases, make the data uninterpretable, and splenic acquisitions were graded IV or V in 69.0% of the cases and I or II in 11%. The grades for splenic acquisitions showed a direct relationship with splenic size: All acquisitions of spleens greater than 13 cm in length were graded IV or V.
In 83 of the 99 patients with possible or known IBD, abnormal bowel was shown on the baseline scan. Volumetric acquisitions were obtained in these patients, with 75.9% of the acquisitions graded as IV or V. Because many of the patients with abnormal bowel were very thin, the single transducer available, a V6-2, with an operating frequency in the range of 2-6 MHz, was too low to resolve the bowel and perienteric soft tissue positioned in the near field, especially in thin women. This accounts for a substantial number of bowel cases (18%) graded as I or II.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
The choice of the optimal plane and orientation for the acquisition will also allow a complete examination with the fewest volume data sets. This is desirable for two reasons: first, less demand on the storage system and, second, reduced time requirements for the interpretation because multiple unnecessary and redundant volumes require a lengthy interpretation time and are tedious and cumbersome. However, with complex pathology or large organ size, more than a single volume per organ may be required.
Similar to routine single-image sonography of the abdomen, volume acquisition has variable success depending on the available acoustic window. However, in our study, very good to excellent (grades IV and V) volumetric acquisitions were obtained in 1,215 (83.6%) of the 1,454 acquisitions. A few acquisitions (n = 127, 8.7%) were assigned a grade of III, consistent with satisfactory-quality data. Grade III volumetric data, similar to a satisfactory scan obtained with single-frame images, allow the reader to make reasonably accurate observations and interpret the primary pathology; however, very subtle observations might not be depicted as optimally.
On an organ-by-organ basis, organs with a single long and two short dimensions, such as the kidneys, gallbladder, aorta, and even the pancreas, lend themselves optimally to volume acquisition. Further, organs with a good acoustic window will obviously produce the best result. Changes in patient position and respiratory techniques may improve the acoustic window, as with conventional scanning.
Analysis of failed or poor examinations in our study showed that the absence of a good acoustic window was the most consistent explanation for acquisition failure and could affect any organ. Failure of volumetric acquisition of the liver is highly associated with severe end-stage cirrhosis, which makes a good acoustic window from a subcostal approach impossible, and with severe fatty liver, which limits penetration of the ultrasound beam. If the liver is visible only from an intercostal approach, a volume of data is unlikely to include the entire organ; in this case, either multiple volumes of a single organ or additional single-frame images should augment the obtained volumetric data.
Similar to traditional single-image data acquisition for ultrasound, volume imaging of normal bowel is rarely informative. In the patient with pathology of the bowel, however, volumetric acquisition may provide valuable information and show relationships with a capability beyond standard 2D ultrasound. Volumetric analysis of the abnormal bowel is easily performed as long as it is quiet and does not show hyperperistalsis, which may introduce significant motion artifact. Furthermore, our experience has shown us that the volumetric technique must be introduced on a wide range of transducer frequencies to allow evaluation of superficial structures, such as the bowel, with an appropriate high-frequency transducer. We believe that this single advancement, although most important for bowel scanning, would have decreased the total number of examinations of all organs graded as I or II (112/1,454 [7.7%]) in this study.
Before the introduction of volumetric imaging of the abdomen in our facility, we wondered if this technique might make the performance of abdominal ultrasound easier and reduce operator dependency. We wondered further about reduction of scanning times and improved efficiencies. Our initial efforts, however, have shown us that we needed guidelines, as we have described, to allow a high likelihood of success with the fewest data volumes for analysis. A high skill level with routine scanning seemingly also translates into improved volume data acquisition, therefore not necessarily making the performance of the scan easier. The benefit of volume data acquisition in terms of time and efficiencies has been discussed by many authors in the literature and is beyond the scope of this article. Further, we found that acquisitions performed without a plan did not afford good MPR and also increased the number of acquisitions that required review on completion of the examination. In agreement with our commitment to performing volume acquisitions according to established protocols, Nelson et al. [9] previously advocated a rigorous standardized protocol for both acquisition and review to preserve quality and diagnostic accuracy.
We now are using volumetric data acquisition as an adjunct to rather than as a replacement of the conventional real-time multiplanar technique that is so important in abdominal sonography. The development and advancement of volumetric imaging for ultrasound have been hampered by slow technologic advancement. With mechanical acquisition of data, resolution in the acquisition A plane is excellent and is analogous to any other cine sweep through a region of interest. There is, however, a loss of resolution in the B plane and an even greater loss of resolution in the C plane. In the future, the technique will improve with the introduction of matrix transducers, which will allow the electronic acquisition of data in the volume with the use of a multielement array, which scans through the entire volume electronically without actual movement of the transducer array itself. The loss of resolution should be corrected with this electronic data acquisition. Further, the introduction of a DICOM standard for the transmission of volumetric data and more universally available software platforms for manipulation of acquired data sets will enhance this technique in the future.
The major weakness of our study is the inclusion of the routine single-image scan before volumetric data acquisition. However, because volumetric acquisition is an unproven technique in imaging the abdomen, it would not be acceptable to perform the volume acquisition alone. Further, we were not using the volumetric data for interpretation of pathology. Rather, the data sets were evaluated only for technical feasibility in terms of inclusion of the entire organ and also inclusion of abnormalities.
Although experiences in both obstetric and cardiac applications cite huge efficiencies in terms of the time required to obtain the ultrasound scans [1], the efficiency impact of volumetric acquisition of data in abdominal ultrasound is yet to be determined [3, 4]. Further, a volume may allow visualization of a structure relative to its surroundings in a manner that was not previously possible. The impact this capability might have on diagnostic interpretation is also unknown.
In summary, our study has shown good success for the performance of volumetric data acquisitions of the abdominal organs when guideline protocols for performance are closely followed. Use of the guidelines we describe should facilitate the acceptance and dissemination of this exciting technique not only for better evaluation of abdominal pathology, but also for equally important reassurance of the absence of pathology in a particular organ. The ability to view cine files in all planes and single-frame images in any plane from one volume acquisition is the highlight of the technique. We predict that with the imminent release of new transducer technology for electronic volumetric data acquisition and the introduction of this technique over a broader range of transducer frequencies, the potential for this technique will be realized. Future investigations will evaluate the diagnostic capability of volumetric acquisition as compared with conventional techniques for the detection, diagnosis, and surveillance of abdominal pathology and also the efficiency impact.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
