| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Medical Physics, S-171 76 Karolinska University Hospital,
Stockholm, Sweden.
2 Department of Diagnostic Radiology, Karolinska University Hospital, Stockholm,
Sweden.
Received May 11, 2004;
accepted after revision July 14, 2004.
Address correspondence to A. Özgün
(vahit123{at}hotmail.com).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Twenty-four latex phantoms that simulate the large bowel and contain artificial polyps of different sizes and shapes were constructed. The polyps were divided into three size groups (diameter, 0-2, 2-5, and 5-10 mm) and were classified into four shape groups: pedunculated; broad-based; ulcerated or depressed; and sessile or flat. The colon phantoms were submerged in a water tank and scanned on a 4-MDCT scanner using 12 protocols with various settings of slice thickness, pitch, and tube current. The images were independently evaluated by three radiologists using axial 2D multiplanar reconstruction images and a 3D surface-rendering technique (fly-through).
RESULTS. At a constant dose (i.e., dose-length product [DLP]), the polyp detection rate increased with increasing axial spatial resolution. For the standard protocol (2.50-mm slice thickness, 1.5 pitch), the detection rate for all polyp sizes decreased from approximately 70% at 100 mA to 55% at 40 mA. Between a 60- and 100-mA tube current, the detection rate for the largest polyps (> 5 mm) was almost constant, close to 90%.
CONCLUSION. The detection of polyps in the large bowel using a standard protocol can be improved without dose penalty by increasing the axial spatial resolution of the image acquisition and adjusting the tube current setting. If the analysis can be restricted to polyps larger than 5 mm, the dose can be substantially reduced without compromising the detection rate.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Recently, the use of MDCT scanners has allowed acquisition of larger data sets with breath-hold techniques. This contributes to the improved spatial resolution and faster image acquisition that is possible with MDCT scanners compared with single-detector CT.
A concern with CT procedures in general is the radiation dose to the patient. For a given scanner configuration, the radiation dose depends on the tube voltage, tube current, pitch factor, and collimated beam size. In the design of low-dose CT sequences, there is always a trade-off between the radiation dose and spatial resolution properties of the image data. Low-dose protocols typically involve high tube voltage, low tube current settings (calling for thicker slices), and a high pitch. Recent patient studies using a single or two different low-dose milliampere-second levels in CT colonography [12] or simulating low doses by postexamination modification of imaging raw data indicate that despite reduction of image quality below the level of 30 mAs, polyp detection in CT colonography remains acceptable [4].
The aim of this study was to systematically evaluate the impact of radiation dose and axial spatial resolution on detection of colonic polyps using 4-MDCT. In the study design, emphasis was put on simulating as closely as possible the clinical conditions under which CT colonography is performed.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Colon Phantom
Twenty-four phantoms with a length of 50 cm were prepared using a
latex-based hose (attenuation value, -97 H) with a thickness of 2.0 ±
0.1 (SD) mm. This material is flexible and can easily be molded into different
shapes. The colon phantoms were sealed and filled with air. At every 3-5 cm, a
fold yielding various degrees of distention was simulated to create inner
diameters varying from 1 to 7 cm.
A total of 240 synthetic polyps were made from the latex material. The
polyps were divided into three size groups (X, Y, and Z), with 0 < X 
2
mm, 2 < Y 5 mm, and 5< Z 10 mm). The polyps were evenly
distributed among the groups, resulting in 80 polyps of each size interval.
The shapes of polyps found in patients can be characterized as pedunculated;
broad-based; ulcerated or depressed with elevated margins; or sessile or flat.
Flat polyps were unique in that they were all rectangular with a size ratio
width-height equal to or larger than 1. To simulate the in vivo situation as
closely as possible, we manually manufactured "polyps" of these
four shapes. Their size and shape were visually confirmed with the help of
optical magnification instruments and precision rulers. There was an even
distribution of polyps among the different shapes, resulting in 20 polyps of
each shape in each size interval. Finally, the polyps were glued to the inside
wall of the phantoms using a cyanoacrylate-based glue. Axial and endoluminal
views of the different types of polyps on CT images are displayed in Figures
1 and
2, respectively.
|
|
Experimental Protocol
The experimental protocol included a series of 12 CT scanning sequences. An
outline of the protocol is given in Table
1. The tube potential was kept constant (120 kV) in all sequences.
All measurements were 2D. Estimates of the dose-length product (DLP) were
calculated from measured normalized CT dose index-weighted
(nCTDIw) values for the different beam collimations
used—that is, 5 (or 4 x 1.25), 10 (4 x 2.5), and 15 (4
x 3.75) mm, respectively (Appendix
1). The standard CT sequence used for colonography in the clinic
today is based on the following settings: 100 mA, 1.5 pitch, 2.5-mm slice
thickness (CT8 in Table
1). The DLP of the standard sequence, assuming a scanning length
of 30 cm, was defined as the 100% dose level. The ratio of DLP for any given
sequence (30-cm scan) and DLP of the standard sequence was calculated and is
referred to as the "relative DLP" in
Table 1. In this way, a DLP of
200% corresponds to an effective dose to the patient that is two times that of
the standard sequence, while a DLP of 60% corresponds to an effective dose of
60% that of the standard protocol and so on. The relative DLP values in the
range of 40-140% of the standard sequence were chosen, and the corresponding
tube current values were calculated. In addition, the sequence with the
highest axial spatial resolution (CT4) was used in combination with
the highest possible tube current setting (200 mA, 486% relative DLP). Because
of restrictions in the scanner software, small deviations in the tube current
settings from those calculated had to be accepted for some sequences. The
impact on relative DLP was at most 4% for protocols in the DLP range of
40-140%. For the purpose of this study, such deviations were disregarded, and
the DLP level initially chosen for each sequence was used during data
analysis. The corresponding effective dose values were calculated and are
reported in Table 1.
|
|
To simulate the circumstances under which a clinical CT colonography examination is performed, it was deemed essential that the observer be blinded to the number and shape of the polyps in the colon phantoms. In addition, the same observer should evaluate a specific phantom only one time. Finally, the total number and size of the polyps to be evaluated for each CT sequence should be kept constant. On the basis of these requirements, three criteria were defined and applied to each of the 12 CT sequences. The first criterion was that evaluations should be performed independently by three observers with similar experience with CT colonography evaluations. The second criterion was that six different colon phantoms should be included (two for each radiologist). The third criterion was that a total of 60 polyps (20 in each size group) should be distributed among the six phantoms.
Three radiologists with at least 2 years of experience in CT colonography were chosen to participate as observers in the study. Their performance in evaluating CT colonography was regarded as equal. A total of 24 colon phantoms (12 CT scanning sequences x 2 colon phantoms per radiologist and sequence) were constructed. The colon phantoms were submerged in a tank constructed from polymethylmethacrylate and filled with water. The dimensions of the tank are representative of the object size for adult bowel CT examinations (length, 40 cm; width, 25 cm; height, 19.5 cm). The scanning procedure included two colon phantoms in each acquisition. A typical setup with colon phantoms, inflated with air and submerged in the water tank, is displayed in Figure 3. Before acquisition of the image data, a standard scout view was acquired to evaluate the degree of distention and to check that the phantoms were completely immersed in water.
|
The 24 phantoms were distributed in four groups, each with six phantoms. Each phantom group was then used to acquire data for three CT scanning sequences. The rotational scheme of the phantoms in each group is shown in Figure 4. After completion of the study, each radiologist had examined all 24 phantoms.
|
The impact of residual fluid on polyp detection in the colon lumen was evaluated by rescanning phantoms in group 4 (i.e., sequences 1, 7, and 10) after the addition of some water. The water filled approximately 20-30% of the lumen volume. The submerged polyps were not visible with the viewing settings used in the evaluations.
To complete the experimental protocol, we performed a total of 45 data acquisitions: 36 (12 x 3) with phantoms having no water inside and nine (3 x 3) with phantoms partially filled with water. The study design included a total of 720 polyps (12 x 60) to be evaluated in colon phantoms without water and an additional 180 polyps (3 x 60) in phantoms with water.
Evaluation Protocol
The radiologists assessed the CT scans on a dedicated workstation
(Advantage Windows 3.1, GE Healthcare) using a software tool (Navigator
[version 2], GE Healthcare). All evaluations were performed with the same
viewing settings of the workstation monitor as those used during clinical CT
colonography examinations (window width, 1,600 H; window level, 400 H). The
radiologists were informed of the different polyp shapes and size intervals
prior to data evaluation.
Two different evaluation methods were used: axial images were displayed in three orthogonal directions (referred to as "2D multiplanar reconstruction [MPR] technique") and virtual colonoscopy images, yielded by a 3D surface-rendering technique (referred to as "fly-through technique"), were shown. With the 2D MPR technique, the number of polyps in each size group (X, Y, Z) was defined for each colon phantom. With the fly-through technique, the total number of polyps of each shape, independent of size, was extracted using forward and backward fly-through of the phantom. No polyp matching was done.
The image noise of the different CT sequences was calculated from the SD of attenuation values within regions of interest (ROIs) in water (STDwater). For this purpose, circular ROIs of 2,088 mm2 were positioned in the center of the water tank. For each CT sequence, ROI data were obtained from four consecutive images from the central part of the phantom and the mean of the STDwater in these images was used as a measure of image noise.
Shape Analysis
The ability to define the polyp shape was evaluated using the fly-through
data. The number of registered polyps of each of the four shapes (groups 1-4)
including all three polyp size intervals was obtained from the data. If the
number of polyps registered for a given shape exceeded the correct number,
those in excess were regarded as misclassified and all others were regarded as
correctly classified.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
The results of splitting the 2D MPR evaluations of Figure 5 into the different polyp size groups are displayed in Figure 6. At all DLP levels, the detection rate increased for larger polyps. For instance, at 60% DLP (relative to the standard protocol), approximately 20% of the smallest polyps were detected, with the detection rate increasing to almost 80% for medium-sized polyps and reaching 90% for the largest polyps.
|
Effect of Axial Spatial Resolution on Polyp Detection Rate
The axial spatial resolution—that is, the slice thickness and the
pitch factor—significantly affected the detection rate of polyps. When
the radiation dose (DLP) was kept constant and equal to that of the standard
sequence (2.5-mm slice thickness, 100 mA, 1.5 pitch), the highest detection
rate of polyps larger than 2 mm was obtained using the sequence with 1.25-mm
section thickness and 0.75 pitch, followed by the sequence with 1.25-mm slice
thickness and 1.5 pitch (2D MPR technique). This is despite the significantly
reduced tube current settings used with the narrower slice sequences to
balance DLP (40 and 80 mA, respectively). The results are shown in
Figure 7. Using the
high-resolution protocol with the maximum tube current setting (CT4
in Table 1; DLP = 486%), the
detection of the smaller polyps (2 mm) improved compared with the
corresponding data at a DLP of 100%, and there was not a significant
difference in the detection of polyps larger than 2 mm.
|
Image Noise
The noise variance in CT images (i.e., STD2water) is
inversely proportional to the number of detected photons
[13]. This means that the
noise varies linearly with the product of slice thickness and tube current.
The measured STDwater values followed closely the predicted
behavior, although they were slightly higher than those anticipated from a
purely quantum mottle noise. For example, the measured image noise
(STDwater) of the protocol using 1.25-mm slice thickness, 80 mA,
and 1.5 pitch was a factor 1.8 higher than that with the standard protocol
(2.5 mm, 100 mA, 1.5 pitch). Theoretic predictions, taking into account the
slice thickness and tube current setting, yield a value of 1.6.
Effect of Residual Fluid and Feces on Polyp Detection
A common problem during a CT examination of the colon is the presence of
residual liquid and stool in the colon. As we previously mentioned, the effect
of residual content on polyp detection was simulated by the addition of water
in phantom group 4. The results are shown in
Figure 8.
|
Shape Analysis
The evaluation of polyp shape yielded small differences between the CT
sequences with regard to misclassified polyps, and the number of such polyps
was always three or fewer per CT sequence ( 5% of all polyps) for
sequences with no added water. For sequences with water added to the phantom,
the number of misclassified polyps increased substantially and resulted in
between seven and 11 misclassified polyps (up to 18%). The misclassified
polyps in phantoms containing water were always registered in polyp shape
group 2 (Figs. 1 and
2), indicating that those were
most likely water drops attached to the phantom wall.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The detection rate of the smallest polyps (< 2 mm) was relatively low in comparison with the detection rates of medium-sized and large polyps for all CT sequences. In addition, the limited tube current settings of the high-resolution data acquisitions (1.25-mm slice thickness) reduced the detection rate of the smallest polyps to that obtained using the standard sequence (0.75 pitch) or to an even lower rate (1.5 pitch). This is most likely related to the increased noise level in the 1.25-mm data. However, polyps smaller than 5 mm are usually of less clinical importance [14, 15].
The results of this study indicate that imaging for polyps larger than approximately 5 mm can be performed at significantly reduced dose levels without compromising the detection rate. For instance, with the standard protocol the detection rate was almost constant (almost 90%) within the 60-100% range of DLP. Because radiation dose, contrast and detail resolution, and stochastic noise are closely related in X-ray imaging, the possibility of dose reductions will reflect the specific requirements on image quality of each diagnostic task. CT colonography involves the definition of objects (polyps) having a high contrast to the surroundings (air in the colon). The detection of objects in such high-contrast images is, in general, less susceptible to an increased level of stochastic noise and therefore also to a decrease in the tube current (and dose) of the CT sequences. The results of this study show that the dose—that is, the tube current setting—using a typical clinical CT colonography protocol can be substantially reduced without compromising the detection rate of polyps larger than 5 mm. The use of special low-dose protocols in CT colonography has been successfully reported by other groups [2, 4, 12, 16]. Although there is evidence that polyps of clinically relevant sizes (5-10 mm) can be detected with reduced milliampere-second settings in CT colonography [12], little comparative clinical data that prove the relative benefit of using such high-resolution protocols in combination with reduced milliampere-second settings have been reported to date. This must be evaluated further using the latest available MDCT technology.
The results of this study were obtained using a water-filled phantom with a cross-section of approximately 25 x 19.5 cm2. This was regarded as representative of a normal-sized adult. The adjustment of tube current to the actual size of the patient is common practice at many CT installations. This adjustment is needed to counteract the decrease in image quality otherwise obtained for thicker patients [4] and to balance the dose, especially to thin patients. The tube current settings reported in this study should be adjusted in a similar fashion at the time of performing patient examinations.
A common criticism about the use of phantom measurements in the optimization of clinical imaging protocols is the difficulty in simulating the complex structures and tissue components encountered in a patient. In this study, both the phantom and the polyps were constructed from latex, yielding a constant density of the polyps and the colon wall, together with a very smooth inner surface of the wall. This differs from the clinical situation in which so-called polyp edge-definition problems occur, reflecting a nonhomogeneous density across a polyp. The result is a more error-prone definition of the size of the polyp in patients. It can also be argued that the relatively rough surface of the colon wall in a patient could influence the detection rate of polyps, making smaller polyps more difficult to detect. The simulated haustral folds in the colon phantoms could potentially result in false-positive findings of polyps, but this could occur also in the clinical situation. Finally, with this phantom there are no objects surrounding the colon that potentially could influence the diagnostic outcome, as in the case of a patient.
Although there are differences between these phantoms and patients, the radiologists regarded them as relevant descriptors of the clinical situation. Important characteristics of the phantoms in our study that contribute to their relevance in the clinical setting are that they were not rigid, could be inflated, and contained simulated haustral folds. It should be mentioned that when residual fluid or feces are present inside the colon during the examination, computer-based removal of such high-density material has been tested [17]. Such procedures were not included in this study because it is, to date, not commonly used in the clinical setting.
The study design did not allow an evaluation of false-positive findings of polyps. Such findings could have occurred and potentially affected the results of the analysis. We believe that false-positive findings in our study would appear mainly in the regions close to the simulated colon folds in the phantoms and in the phantoms with water. Because the number of folds and their "severity" were similar for each phantom, we think that the number of false-positive findings would be equally distributed among the phantoms and CT sequences (phantoms without water). This means that the presence of false-positive findings would yield a decreased detection rate for all protocols but would not affect the overall results and conclusions of the study. Another argument in support of this assumption is that the shape analysis indicated a similar number of misclassified polyps with regard to shape for the different CT sequences. The presence of water increased the detection rate using a 2D MPR evaluation, whereas the results of the flythrough technique tended toward the opposite.
We conclude that misclassification of water drops attached to the wall in the water-filled phantoms is always a risk and that such false polyps will influence the detection rate. An explanation of the discrepancies between the results of the two evaluation techniques has not been found. However, our experience gained from this study indicates that the capacity to distinguish different polyp shapes is improved with the fly-through technique compared with the 2D MPR technique. This is believed to have influenced the results presented in Figure 8, reducing the number of water drops classified as polyps with the fly-through compared with the 2D MPR technique.
At dose levels used in diagnostic X-ray procedures, the associated risk to the patient relates to the stochastic induction of cell damage that could cause long-term effects, such as cancer. The current belief is that stochastic effects can be induced regardless of the radiation dose, with a higher incidence at higher dose levels. The radiation burden to the public from CT examinations is significant. A relatively recent study from the United Kingdom showed that although CT accounted for only about 4% of all X-ray examinations, it contributed 40% of the dose to the population [18]. Similar results have been presented by groups in other countries [19, 20].
The question of screening patients using techniques based on ionizing radiation is controversial, and it has not been the intention of this work to discuss the topic. If the CT colonography technique could perform significantly better than other screening techniques in detecting colon cancer and if diagnosis of the disease using screening could clearly influence clinical outcome, there might be arguments for introducing CT colonography as a screening for parts of the population. CT colonography indeed offers a number of advantages over current screening techniques for colorectal cancer.
In conclusion, the results of this study indicate that axial spatial resolution plays a dominant role for the detection of polyps in the colon using CT colonography. CT colonography conducted at a given dose level and spatial resolution could therefore possibly perform better at an increased spatial resolution (low pitch, thin slices), keeping the dose constant by reducing the tube current setting. The major limitation with an increased axial spatial resolution is the prolonged examination time, restricting the use of breath-hold techniques. The results also show that a substantial increase in image noise (from reducing the tube current setting) can be accepted when the analysis is restricted to polyps larger than approximately 5 mm. Hence, for a given CT colonography protocol, the radiation dose can potentially be reduced when only the larger polyps are of interest. This study indicated that a dose reduction by approximately 40% in a "normal"-sized patient compared with the present standard protocol could be used in such situations.
Acknowledgments
The assistance from Madeleine Leidner and Yvonne Eriksson Alm
(radiographers) during CT scanning, and Björn Hellberg, Håkan
Eriksson, Jan Westin, and Kjell Leanderson (engineers) for their help in
construction of the phantoms is acknowledged.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  A. H. de Vries, H. W. Venema, J. Florie, C. Y. Nio, and J. Stoker Influence of Tagged Fecal Material on Detectability of Colorectal Polyps at CT: Phantom Study Am. J. Roentgenol., October 1, 2008; 191(4): W181 - W189. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
= polyps > 5 mm; 
= 2- to 5-mm polyps; 
= 0- to 2-mm polyps.
