| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Report |
1 Division of Abdominal Imaging and Intervention, Department of Radiology,
Massachusetts General Hospital and Harvard Medical School, White 270-E, 55
Fruit St., Boston, MA 02114.
2 GE Healthcare, Waukesha, WI.
Received March 26, 2004;
accepted after revision May 18, 2004.
Address correspondence to M. K. Kalra.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CONCLUSION. Z-modulation is associated with a 34.1% increase in the mean tube current–time product and no change in the extent of streak artifacts in patients with a metallic prosthesis, compared with patients without a prosthesis. However, compared with the fixed-tube-current technique, z-modulation is associated with a 28.9% decrease in the mean tube current–time product.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Although recent studies [2, 5] have reported advantages to the use of angular and z-modulation techniques for minimizing the radiation dose associated with CT of the abdomen and pelvis, the effects of metallic prostheses on these techniques remain undocumented. Tube current has been observed to increase in the presence of metallic prostheses when CT is used with z-modulation technique. The purpose of our study was to assess the effect of orthopedic metallic prostheses on the radiation dose associated with z-axis automatic modulation of tube current in a phantom and patients.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Automatic tube-current-modulation techniques are based on the principle that objective image noise in CT is based on the quantum noise in the X-ray projections. Thus, automatic tube-current-modulation techniques function by adjusting the tube current to follow changing regional anatomy to maintain a constant quantum noise in the image and to improve radiation dose efficiency. The z-modulation technique used in the present study automatically adjusts the tube current to maintain a constant user-specified quantum noise level in the reconstructed images. The technique estimates the attenuation and shape of the anatomic region of interest from a single localizer radiograph of the patient and uses this information to determine the extent of tube-current modulation for each image in the scanning direction (z-axis). The system calculates the tube current using the patient's localizer radiograph data and a set of empirically determined noise prediction coefficients for the reference technique, which includes an arbitrary 2.5-mm slice thickness at the selected peak kilovoltage and 200 mAs. The noise index is approximately equal to the image noise in the central region of the image obtained from scanning a uniform water phantom. For this technique, the technologist selects a noise index depending on desired image quality and a minimum and maximum milliamperage to define the acceptable range or limits of tube-current modulation. We use a noise index of 15 H and a minimum and maximum tube current of 75 and 380 mA, respectively, for routine abdominal–pelvic CT [5]. In general, every 5% decrease in the noise index (Hounsfield units; H) increases radiation exposure by 10%. Conversely, with every 5% increase in noise index, radiation exposure decreases by approximately 10%, with a concomitant decrease in image quality.
With the exception of tube current, all other scanning parameters were kept identical for scanning of the phantom using z-modulation technique. As per our departmental protocol for routine abdominal–pelvic CT, a noise index of 15 H and a minimum and maximum milliamperage of 75 and 380, respectively, were selected for scanning the phantom with z-modulation technique. Next, a cobalt–chrome metallic hip endoprosthesis (Omnifit EON, Stryker Howmedica Osteonics) with a titanium acetabular cup (Trilogy acetabular cup, Zimmer), used for total hip arthroplasty, was placed in the phantom to determine the effect of metallic density on z-modulation technique. To avoid movement of the prosthesis during scanning, the prosthesis was fixed with adhesive tape to the dependent aspect of the inner side of the phantom. The phantom was placed in the same position in the gantry isocenter and rescanned using identical fixed-tube-current and z-modulation techniques. For each series of images acquired with fixed-tube-current and z-modulation techniques, standard CT radiation dose descriptors—the CT dose index volume and dose length product—were recorded from the scanner user interface. The tube current used by the scanner for acquiring images with z-modulation technique was recorded for all slice positions in series with and without the prosthesis.
Patient Study
The Human Research Committee of the institutional review board approved the
study protocol with waiver of informed consent. We retrospectively reviewed
650 consecutive enhanced abdominal–pelvic studies obtained with 16-MDCT
scanners using z-modulation technique to identify patients with
metallic prostheses in the lumbar spine or hips. Of the 28 patients with
metallic prostheses in the spine or hips, eight patients were excluded because
they had not previously undergone CT with fixed-tube-current technique. Thus,
20 patients (eight men and 12 women; age range, 44–88 years; mean, 71
years) composed the study cohort. Fifteen patients had total hip arthroplasty
prostheses (10 with a right hip prosthesis, three with a left hip prosthesis,
and two with bilateral prostheses), and five patients had a spinal prosthesis.
The weight range of patients in the study cohort was 54.4–107.5 kg
(mean, 79.7 kg). The standard departmental protocol for z-modulation
was used for scanning the patients and included 140 kVp, a 0.5-sec rotation
time, a 0.938:1 beam pitch, a 1.25 x 16 detector configuration, and an
18.75-mm table feed per gantry rotation. Images were reconstructed to obtain a
5-mm slice thickness at 5-mm intervals using a standard reconstruction
algorithm and full reconstruction mode. All patients were scanned at a 15-H
noise index, with minimum and maximum tube currents of 75 and 380 mA,
respectively.
For each patient with metallic prostheses, z-modulation examinations were compared with previous abdominal–pelvic CT studies obtained with fixed-tube-current technique and the same metallic prostheses. The mean interval between examinations performed with z-modulation and fixed-tube-current techniques was 6 months (range, 2–11 months). Except for tube current (180–300 mA), the scanning parameters for examinations performed with fixed-tube-current technique were identical to those used in z-modulation studies.
Two subspecialty radiologists (one with 5 years' experience and one with 3 years' experience) evaluated images acquired with z-modulation and fixed-tube-current techniques for the presence of streak artifacts from metallic prostheses. The images were displayed on the diagnostic workstation of a digital picture-archiving system (Impax RS 3000 1K review station, AGFA Technical Imaging Systems). All images were assessed at the same window level and width level: 40 and 400 H, respectively. In addition, the severity or extent of streak artifacts from the metallic prostheses was compared between the two techniques using a 3-point scale (1, worse; 2, equal; 3, better). The tube current for each image in examinations performed with z-modulation and fixed-tube-current techniques was recorded.
To determine whether z-modulation technique increases tube current in patients with metallic prostheses, compared with patients without metallic prostheses, we identified 20 patients (eight men and 12 women; age range, 42–84 years; mean, 70 years) who had no metallic prostheses and who underwent abdominal–pelvic CT using identical z-modulation scanning parameters. The patients without metallic prostheses (control cohort) were matched for age and sex to the patients who had metallic prostheses. The weight range of patients without metallic prosthesis was 54.4–104.3 kg (mean, 76.0 kg). The tube current for each image from examinations performed using z-modulation was also recorded.
Statistical Analysis
The tube current–time products (milliampere-seconds) used for
scanning the phantom with and without the metallic prosthesis using
z-modulation technique were compared with Student's t test
(Excel, Microsoft). The statistical significance of differences in weight
between the study patients and the control cohort was determined using
Student's t test. Qualitative scores for the extent or severity of
metallic streak artifacts in images acquired with z-modulation and
fixed-tube-current techniques were compared using Wilcoxon's signed rank test
(MedCalc software, MedCalc). The degree of interobserver agreement was
determined using the kappa test (MedCalc).
Mean milliampere-seconds and sample error of mean for examinations performed with the two techniques were calculated for the study and control cohorts. The tube current–time product used for examinations performed with z-modulation and fixed-tube-current techniques were compared using the Student's t test. For each technique, percentage differences in mean milliampere-seconds in patients with and without metallic prostheses were also estimated. A p value of less than 0.05 was considered a statistically significant difference.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
Patient Study
No statistically significant difference in weight was found between
patients with and patients without metallic prostheses (p = 0.548).
Streak artifacts due to metallic prostheses were seen in all patients in the
study cohort for examinations performed with z-modulation and
fixed-tube-current techniques. No significant difference was noted between
qualitative scores for the extent or severity of streak artifacts on images
acquired with z-modulation and fixed-tube-current scanning
(p = 0.8). Strong interobserver agreement was noted between the two
radiologists for grading of the streak artifacts from metallic prostheses
(
= 0.8; p < 0.01).
In patients with metallic prostheses, the mean numbers of milliampere-seconds with fixed-tube-current scanning were 177.1 (sample of mean, 5.9) and 125.9 (sample of mean, 9.2) with z-modulation technique (p = 0.0002). In patients without metallic prostheses who underwent abdominal–pelvic CT using z-modulation technique, the mean numbers of milliampere-seconds and sample error of mean were 82.9 and 9.9, respectively. In patients with metallic prostheses, z-modulation resulted in a 28.9% (125.9/177.1 mAs) reduction in the mean number of milliampere-seconds, compared with fixed-tube-current scanning. However, z-modulation technique was associated with a 34.1% (82.9/125.9 mAs) lower number of milliampere-seconds in the control cohort than in the study cohort, and the tube current–time product was 53.2% (82.9/177.1 mAs) lower for z-modulation technique than for fixed-tube-current scanning in the study cohort. Thus, CT was associated with a significantly lower tube current–time product for patients in the control cohort than for patients in the study cohort whose examinations were performed with z-modulation (p = 0.003) or fixed-tube-current (p = 0.000001) technique (Fig. 3A, 3B, 3C).
|
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Metallic orthopedic prostheses generate artifacts that can substantially degrade CT image quality [8]. These artifacts are typically seen as starbursts or streaks that occur when projection data are missing from reconstructions because the incident X-ray beams are blocked by the metallic prostheses [9]. These artifacts, which limit evaluation of prostheses and surrounding structures before revision surgery, are more prominent in helical CT than in conventional axial CT [10]. The severity or extent of streak artifacts from metallic prostheses depends on the composition of the prostheses. Studies have shown that titanium prostheses cause fewer artifacts with CT than do cobalt–chrome alloy or stainless-steel prostheses [11, 12]. Review of images at an extended CT scale (maximum window width, 40,000 H) rather than the standard window (maximum window width, 4,000 H) has been reported to improve imaging of metallic prostheses [10]. In addition, multiplanar reconstruction of data for 3D images has been reported to reduce the artifacts [9]. A recent study also has reported the use of a newly developed algorithm for reduction of metal artifacts on CT scans of patients with total hip endoprostheses [12].
On the other hand, interpolation of missing projection data from an extended 420° scan arc has not been found useful in reducing the artifacts [9]. Similarly, contrary to the general conception, studies have shown that tube current and tube potential (peak kilovoltage) do not affect streak artifacts from metallic prostheses [8]. These findings are consistent with the results of our phantom experiment, which showed no improvement in streak artifacts resulting from the increase in tube current with z-axis automatic tube-current modulation. Furthermore, the patient study showed no difference in the occurrence and extent of streak artifacts resulting from the increase in tube current with z-modulation technique.
To the best of our knowledge, no peer-reviewed study has reported the effect of metallic prostheses on CT with automatic tube-current-modulation techniques, which are now available on most state-of-the-art MDCT scanners. In this regard, given the prevalence of metallic prostheses in patients undergoing CT, the findings of our study highlight an important limitation of z-modulation technique. Our study showed that CT with z-modulation technique in patients with metallic prostheses results in a substantial increase in tube current, compared with the use of z-modulation in patients without metallic prostheses, without any change in the streak artifacts. Because all other scanning parameters, including scan length, peak kilovoltage, beam pitch, table feed, gantry rotation time, and region of interest, were identical in age-, sex-, and weight-matched patients with and without metallic prostheses, an increase in tube current translates into an increase in radiation dose in patients with metallic prostheses. Despite the increase in tube current with z-modulation technique in patients with metallic prostheses, the average tube current–time product with z-modulation technique was still substantially lower than the tube current used with fixed-tube-current scanning in our study. However, if a lower noise index is used as an initial level (e.g., compared with a 15-H noise index, an 11- to 12-H noise index implies a 30–40% higher radiation dose, if all other scanning parameters are constant), tube current with z-modulation potentially could exceed fixed-tube-current scanning [5].
Although z-modulation attempts to maintain the X-ray noise data by increasing the tube current in response to highly attenuating prostheses, streak artifacts from such prosthetic devices represent a current limitation of CT, and X-ray noise is not a significant contributing factor to these artifacts. Because z-modulation technique cannot respond to the presence of metallic prostheses and select the tube current by excluding the contribution of these devices, technologists, radiologists, and manufacturers must ensure that the technique is appropriately modified in the presence of metallic devices. Technologists must note the presence of any metallic prostheses in the localizer radiographs. Radiologists and technologists must determine whether z-modulation technique can be used or whether it should be modified to ensure that the tube current does not increase. Previous studies have shown that z-modulation technique with a noise index of 15 H at 75–380 mA can reduce the mean tube current–time product for abdominal–pelvic CT by 55%, compared with fixed-tube-current technique [13]. In addition, because our study showed a substantially reduced tube current in patients with metallic prosthesis with z-modulation, compared with fixed-tube-current scanning, z-modulation can be used in these patients with some modifications to ensure that tube current does not increase. A higher noise index can be selected in patients with metallic prostheses if the clinical indication of the study permits greater overall image noise. Alternatively, in patients with metallic prostheses, the maximum milliamperage with z-modulation technique can be set at the level that the technologist would have chosen for fixed-tube-current scanning. Paradoxically, this will be associated with subjective selection of tube current and can limit the range of tube-current modulation necessary to obtain the desired image quality. Manufacturers must focus on disseminating this information to the users, devising special techniques to enable prospective recognition of metallic prostheses in the localizer radiograph data, and eliminating the potential contribution to the overall regional anatomic density used for determining tube-current modulation.
The present study had some limitations. Our phantom was a simplified model and did not have surrounding structures such as vertebrae, retroperitoneal fat, and viscera to simulate cross-sectional anatomy. However, the phantom study provided a straightforward proof of concept for the increase in tube current with z-modulation technique in the presence of metallic prostheses. The noise index value of 15 H used in our study represented our departmental protocol and was not based on the vendor's recommendation (11- to 12-H noise index for abdominal–pelvic CT) [5]. It is possible that at the lower noise index recommended by the vendor, the tube current with the use of z-modulation in patients with metallic prostheses might have exceeded the tube current associated with fixed-tube-current scanning. Although we did not assess the effect of a lower noise index, our study design permitted us to make a vital conclusion:in the regions with metallic prosthesis, z-modulation technique in patients with metallic prostheses results in a 34.1% increase in the tube current–time product, compared with examinations that are performed with z-modulation in patients without metallic prostheses. Another limitation of our study was that we did not perform a power analysis to determine the size of the study cohort and the control cohort, because no published studies have assessed the effect of metallic prostheses on the determination of body attenuation using z-modulation technique. In addition, we did not study the effect of the composition of metallic prostheses (stainless steel vs titanium vs cobalt chrome) on z-modulation technique.
In summary, in the presence of metallic prostheses, z-modulation is associated with a 34.1% increase in the mean tube current–time product for abdominal–pelvic CT, with no change in the extent of streak artifacts, compared with scanning of patients without metallic prostheses using z-modulation. However, a substantial reduction (28.9%) in the mean tube current–time product was noted with z-modulation for the entire abdominal–pelvic CT scan, compared with fixed-tube-current scanning, in patients with metallic prostheses. Therefore, radiologists and technologists must pay special attention to the presence of metallic prostheses in the anatomic region of interest on the localizer radiograph so that an increase in radiation dose with z-modulation technique can be avoided by selection of a lower maximum milliamperage or use of a higher noise index.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  T. Dalal, M. K. Kalra, S. M. R. Rizzo, B. Schmidt, C. Suess, T. Flohr, M. A. Blake, and S. Saini Metallic Prosthesis: Technique to Avoid Increase in CT Radiation Dose with Automatic Tube Current Modulation in a Phantom and Patients Radiology, August 1, 2005; 236(2): 671 - 675. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
[open box], respectively), compared with fixed-tube-current
scanning (x). Tube current is increased when phantom with prosthesis is
scanned using z-modulation technique.
