DOI:10.2214/AJR.07.2426
AJR 2007; 189:528-534
© American Roentgen Ray Society
Volume CT: State-of-the-Art Reporting
Frank John Parrish1
1 Department of Radiology, MIA Victoria, 1355 High St., Malvern, Victoria 3144,
Australia.
Received December 21, 2006;
accepted after revision April 23, 2007.
Address correspondence to F. J. Parrish
(frank.parrish{at}gmail.com).
Abstract
OBJECTIVE. CT has undergone generational change that has led to true
volume imaging. Interpretation of volume images requires interaction between
the radiologist and the volume data sets. The aim of this review is to examine
postprocessing options and the evidence in the literature for changing the
process of reporting to digital volume reporting.
CONCLUSION. Diagnostic confidence and the accuracy of interpretation
of volume CT images have increased with improvements in postprocessing
techniques.
Keywords: CT • postprocessing • reporting • volume imaging
Introduction
CT became widely available in the mid 1970s with the introduction of
single-detector scanners. The first MDCT machine (CT-Twin, Els-cint) was
produced in 1992. It was not until the early 2000s, with the advent of
multiple detectors, that MDCT imaging began to flourish. The two slices per
rotation used initially expanded to 64 slices. The numbers are continuing to
increase: 128- and 256-MDCT machines are undergoing clinical evaluation. In
addition to the slice revolution, gantry speeds have increased from one
rotation every 2 seconds to three per second. The third major change is that
slice collimation results in thinner slices: the 10-mm collimation used
originally has decreased to sections as thin as 0.35 mm.
The term MDCT is no longer appropriate because the number of images per
detector has increased with the advent of multiple tubes, dual-energy tubes,
and flying focal spots. With the latest generation of machines, multiplanar
images can be reconstructed from the raw data rather than reconstructed axial
slices, so volume CT is a more accurate term. The effect of the changes is
that in 2 seconds, the latest generation of volume CT scanners can produce up
to 768 images at 0.6-mm slice collimation compared with the one image at 10-mm
collimation achievable in the mid 1970s.
Computing hardware and software have advanced the processing and
presentation of images. Reporting radiologists have a bewildering array of
image postprocessing options and thousands of potential images to review.
Thorough understanding of the capabilities of the technologic changes and
adaptation of reporting techniques are necessary so that patients can realize
the full benefit of CT technology. The processing steps can be broken down
into volume acquisition, volume image display, and volume reporting. The aim
of this review is to examine postprocessing options and the evidence in the
literature for changing the process of reporting to digital volume
reporting.
Thin-Slice Review
Decreasing the slice thickness decreases signal-to-noise ratio, but an
increased radiation dose is required to maintain image quality. For this
reason, many radiologists continued to use thick slices and axial images in
the early days of MDCT. With the advent of 64-MDCT, images can be acquired at
0.5- to 0.625-mm slice collimation. Several studies have shown that reducing
slice thickness increases detection of pathologic conditions.
Hong et al. [1] found that
decreasing slice thickness from 2.5 to 1.5 mm and then to 0.75 mm improved the
minimum amount of coronary artery calcium detected in a cardiac phantom.
Ketelslegers and Van Beers [2],
also using phantoms, found that detection and characterization of urinary
calculi improved with decreasing slice thickness. Furthermore, in renal stone
disease, thin-slice CT has been found helpful in differentiating small
phleboliths from urinary calculi
[3]. In a study of traumatic
injuries, Herzog et al. [4]
found improved depiction of fractures with thin-slice imaging compared with
thick-slice imaging and computerized radiography. In several studies
[5-7],
investigators have concluded that thin-slice image review is superior in the
diagnosis of pulmonary embolic disease. Schoepf et al.
[8] found that the average
yield of detected pulmonary emboli increased 40% when 3-mm slice thickness was
reduced to 1 mm. Those authors further concluded that the rate of
indeterminate reports decreased 70% and that intraobserver agreement improved.
Heuschmid et al. [9] concluded
that thin-slice collimation and thickness are mandatory for visualization of
segmental and subsegmental pulmonary emboli in patients with suspected
pulmonary embolus.
View larger version (109K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1A —78-year-old man with aortic aneurysm. CT scans show aortic
aneurysm (arrow) at 7-mm (A), 5-mm (B), 3-mm
(C), and 1-mm (D) slice thickness. All images obtained from same
data set acquired at 0.6-mm primary collimation.
|
|
View larger version (110K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1B —78-year-old man with aortic aneurysm. CT scans show aortic
aneurysm (arrow) at 7-mm (A), 5-mm (B), 3-mm
(C), and 1-mm (D) slice thickness. All images obtained from same
data set acquired at 0.6-mm primary collimation.
|
|
View larger version (114K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1C —78-year-old man with aortic aneurysm. CT scans show aortic
aneurysm (arrow) at 7-mm (A), 5-mm (B), 3-mm
(C), and 1-mm (D) slice thickness. All images obtained from same
data set acquired at 0.6-mm primary collimation.
|
|
View larger version (115K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1D —78-year-old man with aortic aneurysm. CT scans show aortic
aneurysm (arrow) at 7-mm (A), 5-mm (B), 3-mm
(C), and 1-mm (D) slice thickness. All images obtained from same
data set acquired at 0.6-mm primary collimation.
|
|
Weg et al. [10] examined
small liver lesions at differing slice thicknesses and found an increased rate
and confidence of detection of small liver lesions with thin slices. Similar
results were obtained by Fischbach et al.
[11] in the detection of
pulmonary nodules smaller than 5 mm. These findings are the result of the
partial volume effect that occurs with thick imaging of small structures,
especially those with lower inherent contrast. The small structures are
averaged out and not depicted. This principle applies as much to images
obtained at 0.5-mm collimation but reviewed only at 5 mm as to images obtained
at 5-mm primary collimation. In an example of an aortic dissection (Fig.
1A,
1B,
1C,
1D), images are shown at slice
thicknesses of 7, 5, 3, and 1 mm. The dissection is clearly visible only at 3
and 1 mm.
Because slice collimation is relatively fixed in the latest generation of
CT machines, the only issue is radiation dose. As the number of detectors
increases, beam geometry improves. As a consequence, there is no change in
radiation dose for the collimation options of 64-MDCT scanners
[12]. The images available for
clinician review are usually thick enough to reduce noise. The challenge is to
keep the radiation dose the lowest possible while ensuring the thin-slice
images contain sufficient information for diagnosis.
Multiplanar Reformatting
It is possible to use the isotropic voxels generated to display images in
any plane, including curved planes. This capability has long been an advantage
of MRI over CT. Volume CT now has the advantage of reproduction of any plane
with almost identical resolution after the examination has been completed and
the patient has left the radiology department.
Using reformats perpendicular to blood vessels compared with the
multiplanar reformation (MPR) of standard axial and coronal imaging, Brugel et
al. [13] found improved
sensitivity of prediction of vascular invasion of cancer of the pancreatic
head. These findings were confirmed by Fukushima et al.
[14] in the detection of
resectable pancreatic ductal adenocarcinoma. In studying intraductal papillary
mucinous tumor of the pancreas, Takada et al.
[15] found MPR imaging
significantly improved detection of communication of the tumors with the main
pancreatic duct. Furthermore, diagnostic performance increased with the
combination of axial and MPR imaging.
View larger version (161K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2 —42-year-old woman with bowel obstruction. Coronal multiplanar
reformation CT image shows small-bowel stricture (arrow) and
obstruction due to diverticular disease. Stricture lies in axial plane and was
not visualized on axial images.
|
|
Paulson et al. [16]
reported increased confidence in diagnosing acute appendicitis when using
coronal reformatting. There was, however, no statistical difference in
diagnostic accuracy with the addition of coronal reformatting. Jaffe et al.
[17] found a similar pattern
when looking at coronal reformatted images of small-bowel obstruction.
Figure 2 shows the increased
conspicuity of bowel obstruction on coronal images. The combination of MPR and
axial images significantly improves preoperative staging of colorectal
carcinoma [18]. In the
evaluation of gastric cancer, Shimizu et al.
[19] found MPR images a useful
guide for assessment of the z-axis extent of a tumor. The accuracy of
diagnosis of chest wall invasion by non-small cell lung carcinoma increases
with the use of thin MPR review
[20]. MPR images can be used
to show organs in their true longitudinal and short axes and hence show
pathologic findings more clearly (Fig.
3).
MPR imaging is necessary in cardiac imaging because the cardiac structures
do not lie in the standard planes, as shown in a case of aortic valve
incompetence (Fig. 4). Curved
MPR can be used to show vessels when the pathologic structure lies within the
vessel wall but not in the opacified lumen, as in a case of right coronary
artery aneurysm (Fig. 5). With
the use of oblique planes, the whole pathologic story can be shown on one
image. Figure 6 is an oblique
coronal image showing primary apical non-small cell lung cancer with nodal
disease eroding into the right main bronchus. Overall, MPR imaging improves
diagnostic accuracy and confidence. Improved depiction of the pathologic
features in the anatomic long-axis planes also is possible.
View larger version (106K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6 —62-year-old woman with non-small cell lung cancer. Oblique
coronal multiplanar reformatted CT image shows carcinoma (top left
arrow) with nodal metastatic lesion (bottom left arrow) and
erosion into right main bronchus (right arrow).
|
|
Maximum Intensity Projection
Maximum intensity projection (MIP) is 3D reconstruction entailing
projection of images of the highest-attenuation voxels within a slab of data
along imaginary rays [21].
Marten et al. [7] found MIP
imaging not suited for accurate diagnosis of pulmonary embolic disease. Those
investigators, however, used 5- and 10-mm MIP images, which would be
considered thick slices and have inherent reduced accuracy. MIP imaging has
been traditionally used to show vascular anatomy
(Fig. 7). Choi et al.
[22] found that thin (4 mm)
MIP images were comparable with 1-mm axial MPR images in the diagnosis of
hemodynamically significant stenosis of the coronary arteries.
Thin-slab MIP imaging has been found more accurate than MPR imaging for
assessment of the cervicocranial blood vessels
[23]. Use of a thick (10-20
mm) MIP slab with lung windows makes small pulmonary nodules conspicuous
(Fig. 8). Improved delineation
of the bronchovascular bundles allows differentiation of nodules, which are
similar in size to bronchovascular bundles, as separate structures. Valencia
et al. [24] found better
detection of small pulmonary nodules with 10-mm-thick coronal and axial MIP
images than with 5-mm MPR axial images. Care should be taken in the use of MIP
imaging for diagnosis because structures of low attenuation can be poorly
visualized (Fig. 9), and the
internal architecture of structures of high attenuation can be obscured (Fig.
10A,
10B).
View larger version (46K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 10A —63-year-old woman with peripheral vascular disease.
Maximum-intensity-projection CT image of leg artery shows lumen partially
obscured by two extraluminal clips (top arrow, A;
arrow, B) and by calcium (bottom arrow,
A).
|
|
View larger version (106K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 10B —63-year-old woman with peripheral vascular disease.
Maximum-intensity-projection CT image of leg artery shows lumen partially
obscured by two extraluminal clips (top arrow, A;
arrow, B) and by calcium (bottom arrow,
A).
|
|
Advanced Reconstructions
The number and scope of advanced reconstructions are rapidly developing.
Workstations have a bewildering array of reconstruction tools. Many of these
tools are system specific, such as colonic fly-through and dissection views
(Fig. 11A,
11B), brain perfusion imaging
(Fig. 12A,
12B,
12C), and vessel analysis
packages (Fig. 13). Virtual
endoscopy has been found to improve diagnostic tumor staging of malignant
gastrointestinal tumors [25,
26]. Three-dimensional volume
endoluminal measurements have been shown to be more accurate than standard 2D
displays in assessment of colonic polyps
[27]. Vessel analysis programs
allow automatic assessment of blood vessels with respect to true
cross-sectional and longitudinal length. These measurements have been found
reproducible and accurate [28,
29].
View larger version (42K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 13 —69-year-old man with carotid artery atherosclerosis with
endoluminal stent. On curved multiplanar reformation perpendicular to short
axis of vessel, vessel view shows carotid artery stent measurements.
|
|
Volume rendering is a 3D technique of assigning colors and opacities to
specific ranges of Hounsfield units (opacity transfer functions)
[21]. Johnson et al.
[30] found faster and more
accurate delineation of renal artery stenosis with volume rendering
(Fig. 14) than with MIP
imaging. In musculoskeletal (Fig.
15), vascular, and renal (Fig.
16) imaging, volume rendering provides a useful overview of the
pathologic features.
Computer-Aided Diagnosis
Computer-aided diagnosis is a rapidly evolving technique proving beneficial
in the detection of colonic polyps
[31-33]
and pulmonary nodules
[34-38],
especially in screening. In essence, the computer programs work by assessing
the whole volume of an organ for preset patterns and then highlighting the
abnormalities found. The techniques are in their infancy and have limited
commercial availability.
Reporting Techniques
CT findings can be reported with four methods or a combination approach.
These methods are film, PACS, thick client, and thin client. Film has been the
traditional technique for reporting but is rapidly being overtaken by digital
imaging. The approximately 1,000 thin images in a CT examination of the chest,
abdomen, and pelvis make interpretation of thin-slice images practically
impossible on film. In addition, there is no capacity for interaction with the
volume of data. Use of a PACS has replaced film in many institutions and will
continue to do so. With a PACS, thin images can be reviewed rapidly, and
windows can be used. As a stand-alone platform without a thin or thick client,
however, a PACS relies on the images produced by CT radiographers.
Many centers have been reluctant to transmit and store thin images because
the average data set for 1-mm images is approximately 1 GB of data per
examination. Meenan et al.
[39] proposed use of a
thin-section-only archive with 3D software access to the archive for long-term
storage. In their study, 1,869 CT examinations per month produced 588 GB of
data. It would be possible to reduce this figure 50% or more if only isotropic
coronal images were stored.
Thick client is a traditional workstation or 3D platform. It is
comprehensive and can incorporate all the specialized processes and
techniques. As a consequence, the system is complicated and usually operates
at a rate too slow for interpretation of all examinations at a busy center.
Thin clients are systems that reside on a separate server and are accessed
remotely via local area network, wide area network, or Internet. These systems
provide basic 3D processing functions but retain the high end or proprietary
processors and graphics cards of full workstations. The server rapidly
performs the required intensive processing and sends only processed images
over the network. A minimum network bandwidth of 1.5 MB is required for
real-time review and image manipulation. Thin clients are ideal for
integration with a PACS.
Conclusion
The evidence in the literature supports the reporting of volume CT data
from thin images with the use of techniques such as MPR, MIP, and volume
rendering as additional tools to increase diagnostic confidence and
sensitivity. Volume CT reporting allows radiologists to produce a few images
of the diagnosed pathologic condition in the best orientation and with the
most appropriate postprocessing method for referring clinicians.
It is not possible to perform volume reporting with film. One of the most
effective current methods for reporting volume CT data appears to be combining
a thin client with a PACS. The major CT manufacturers are developing and
releasing thin clients to work within the CT hardware of their volume CT
scanners.
The evidence for volume reporting includes many areas encountered in a
general CT interpretation session. All radiologists interpreting CT scans
should learn and perform volume reporting. The change in reporting techniques
from film to manipulation of CT volume data sets requires radiologists to have
access to volume reporting stations and the necessary training. This access
may be the rate-limiting step for improvement in the diagnostic accuracy of CT
by use of volume reporting and will only be overcome by the action of
radiologists.
Acknowledgments
I thank E. Lazarus for help proofreading the manuscript.
References
- Hong C, Pilgram TK, Zhu F, Joe BN, Towler DA, Bae KT. Improving
mass measurement of coronary artery calcification using threshold correction
and thin collimation in multidetector row computed tomography: in vitro
experiment. Acad Radiol 2003;10
: 969-977[CrossRef][Medline]
- Ketelslegers E, Van Beers BE. Urinary calculi: improved detection
and characterisation with thin-slice multidetector CT. Eur
Radiol 2006; 16:161
-165[CrossRef][Medline]
- Arac M, Celik H, Oner AY, Gultekin S, Gumus T, Kosar S.
Distinguishing pelvic phleboliths from distal ureteral calculi: thin slice CT
findings. Eur Radiol 2005;15
: 65-70[CrossRef][Medline]
- Herzog C, Ahle H, Mack MG, et al. Traumatic injuries of the pelvis
and thoracic and lumbar spine: does thin-slice multidetector-row CT increase
diagnostic accuracy? Eur Radiol 2004;14
: 1751-1760[Medline]
- Patel S, Kazerooni EA, Cascade PN. Pulmonary embolism: optimization
of small pulmonary artery visualization at multi-detector row CT.
Radiology 2003;227
: 455-460[Abstract/Free Full Text]
- Remy-Jardin M, Baghaie F, Bonnel F, Masson P Duhamel A, Remy J.
Thoracic helical CT: influence of subsecond scan time and thin collimation on
evaluation of peripheral pulmonary arteries. Eur
Radiol 2000; 10:1297
-1303[CrossRef][Medline]
- Marten K, Funke M, Obenauer S, Baum F, Grabbe E. The contribution
of different postprocessing methods for multislice spiral CT in acute
pulmonary embolism [in German]. Rofo2003; 175:635
-639[Medline]
- Schoepf UJ, Holzknecht N, Helmberger TK, et al. Subsegmental
pulmonary emboli: improved detection with thin-collimation multi-detector row
spiral CT. Radiology 2002;222
: 483-490[Abstract/Free Full Text]
- Heuschmid M, Mann C, Luz O, et al. Detection of pulmonary embolism
using 16-slice multidetector-row computed tomography: evaluation of different
image reconstruction parameters. J Comput Assist
Tomogr 2006; 30:77
-82[CrossRef][Medline]
- Weg N, Scheer MR, Gabor MP. Liver lesions: improved detection with
dual-detector-array CT and routine 2.5-mm thin collimation.
Radiology 1998;209
: 417-426[Abstract/Free Full Text]
- Fischbach F, Knollmann F, Griesshaber V, Freund T, Akkol E, Felix
R. Detection of pulmonary nodules by multislice computed tomography: improved
detection rate with reduced slice thickness. Eur
Radiol 2003; 13:2378
-2383[CrossRef][Medline]
- Dalrymple NC, Srinivasa RR, Fadi EM, Kedar CN. Price of isotropy in
multidetector CT. RadioGraphics 2007;22
: 49-62
- Brugel M, Link TM, Rummeny EJ, Lange P, Theisen J, Dobritz M.
Assessment of vascular invasion in pancreatic head cancer with multislice
spiral CT: value of multiplanar reconstructions. Eur
Radiol 2004; 14:1188
-1195[Medline]
- Fukushima H, Itoh S, Takada A, et al. Diagnostic value of curved
multiplanar images in multislice CT for the detection of resectable pancreatic
ductal adenocarcinoma. Eur Radiol 2006;16
: 1709-1718[CrossRef][Medline]
- Takada A, Itoh S, Suzuki K, et al. Branch type intraductal
papillary mucinous tumour: diagnostic value of multiplanar reformatted images
in multislice CT. Eur Radiol 2005;15
: 1888-1897[CrossRef][Medline]
- Paulson EK, Harris JP, Jaffe TA, Haugan PA, Nelson RC. Acute
appendicitis: added diagnostic value of coronal reformations from isotropic
voxels at multi-detector row CT. Radiology2005; 235:879
-885[Abstract/Free Full Text]
- Jaffe TA, Martin LC, Thomas J, Adamson AR, Delong DM, Paulson EK.
Small-bowel obstruction: coronal reformations from isotropic voxels at 16
slice multi-detector row CT. Radiology2006; 238:135
-142[CrossRef][Medline]
- Jin KN, Lee JM, Kim SH, et al. The diagnostic value of multiplanar
reconstruction on MDCT colonography for the preoperative staging of colorectal
cancer. Eur Radiol 2006;16
: 2284-2291[CrossRef][Medline]
- Shimizu K, Ito K, Matsunaga A, Shimizu A, Kawakami Y. Diagnosis of
gastric cancer with MDCT using water-filling method and multiplanar
reconstruction: CT-histologic correlation. AJR2005; 185:1152
-1158[Abstract/Free Full Text]
- Higashino T, Ohno Y, Takenaka D, et al. Thin section multiplanar
reformats from multidetector-row CT, utility for assessment of regional tumour
extent in non-small cell cancer. Eur J Radiol2005; 56:48
-55[CrossRef][Medline]
- Cody DD. AAPM/RNSA physics tutorial for residents: topics in
CT—image processing in CT. RadioGraphics2002; 22:1255
-1268[Abstract/Free Full Text]
- Choi JW, Seo JB, Do KH, et al. Comparison of transaxial source
images and 3-plane, thin slab maximal intensity projection images for the
diagnosis of coronary artery stenosis with using ECG-gated cardiac CT.
Korean J Radiol 2006;7
: 20-27[Medline]
- Ertl-Wagner BB, Bruening R, Blume J, et al. Relative value of
sliding-thin-slab maximum intensity projections as reformatting techniques in
multisection CT angiography of the cervicocranial vessels. Am J
Neuroradiol 2006; 27:107
-113[Abstract/Free Full Text]
- Valencia R, Denecke T, Lehmkuhl L, Fischbach F, Felix R, Knollmann
F. Value of axial and coronal maximum intensity projection (MIP) images in the
detection of pulmonary nodules by multislice spiral CT: comparison with axial
1-mm and 5-mm slices. Eur Radiol 2006;16
: 325-332[CrossRef][Medline]
- Panebianco V, Grazhdani H, Iafrate F, et al. 3D CT protocol in the
assessment of the oesophageal neoplastic lesions: can it improve TNM staging?
Eur Radiol 2006;16
: 414-421[CrossRef][Medline]
- Kim HJ, Kim AY, Oh ST, et al. Gastric cancer staging at
multi-detector row CT gastroscopy: comparison of transverse and volumetric CT
scanning. Radiology 2005;236
: 879-885[Abstract/Free Full Text]
- Pickhardt PJ, Lee AD, McFarland EG, Taylor AJ. Linear polyp
measurement at CT colonography: in vitro and in vivo comparison of
two-dimensional and three-dimensional displays.
Radiology 2005;236
: 872-878[Abstract/Free Full Text]
- Cury RC, Ferencik M, Achenbach S, et al. Accuracy of 16-slice
multi-detector CT to quantify the degree of coronary artery stenosis:
assessment of cross-sectional and longitudinal vessel reconstructions.
Eur J Radiol 2006;57
: 345-350[CrossRef][Medline]
- Zhang Z, Berg MH, Ikonen AE, Vanninen RL, Manninen HI. Carotid
artery stenosis: reproducibility of automated 3D CT angiography analysis
method. Eur Radiol 2004;14
: 665-672[CrossRef][Medline]
- Johnson PT, Halpern EJ, Kuszyk BS, et al. Renal artery stenosis: CT
angiography—comparison of real-time volume-rendering and maximum
intensity projection algorithms. Radiology1999; 211:337
-343[Abstract/Free Full Text]
- Yoshida H, Nappi J, MacEneaney P, Rubin DT, Dachman AH.
Computer-aided diagnosis scheme for detection of polyps in CT colonography.
RadioGraphics 2002;22
: 263-279
- Mani A, Napel S, Paik DS, et al. Computed tomography colonography:
feasibility of computer-aided polyp detection in a first reader paradigm.
J Comput Assist Tomogr 2004;28
: 318-326[CrossRef][Medline]
- Yoshida H, Matsutani Y, MacEneaney P, Rubin DT, Dachman AH.
Computerized detection of colonic polyps at CT colonography on the basis of
volumetric: pilot study. Radiology 2002;222
: 327-336[Abstract/Free Full Text]
- Awai K, Murao K, Ozawa A, et al. Pulmonary nodules: estimation of
malignancy at thin-section helical CT: effect of computer-aided diagnosis on
performance of radiologists. Radiology2006; 239:276
-284[Abstract/Free Full Text]
- Yuan R, Vos PM, Cooperberg PL. Computer-aided detection in
screening CT for pulmonary nodules. AJR2006; 186:1280
-1287[Abstract/Free Full Text]
- Li F, Li Q, Englemann R, et al. Improving radiologists'
recommendations with computer-aided diagnosis for management of small nodules
detected by CT. Acad Radiol 2006;13
: 943-950[CrossRef][Medline]
- Marten K, Engelke C. Computer-aided detection and automated CT
volumetry of pulmonary nodules. Eur Radiol2007; 17:888
-901[CrossRef][Medline]
- Martin K, Grillhosl A, Seyfarth T, Obenauer S, Rummeny EJ, Engelke
C. Computer-assisted detection of pulmonary nodules: evaluation of diagnostic
performance using an expert knowledge-based detection system with variable
reconstruction slice thickness settings. Eur Radiol2005; 15:203
-212[CrossRef][Medline]
- Meenan C, Daly B, Toland C, Nagy P. Use of thinsection archive and
enterprise 3D software for long-term storage of thin-slice CT data sets.
J Digit Imaging 2006;19
[suppl 1]:84
-88[CrossRef][Medline]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
