The International Journal of Cardiovascular Imaging

, Volume 27, Issue 4, pp 579–586

Use of 100 kV versus 120 kV in cardiac dual source computed tomography: effect on radiation dose and image quality

Authors

    • Cardiac MR PET CT Program, Department of Radiology and Division of CardiologyMassachusetts General Hospital and Harvard Medical School
    • Non-invasive Cardiovascular Imaging Program, Department of Medicine (Cardiovascular Division) and RadiologyBrigham and Women’s Hospital
  • Michael A. Bolen
    • Cardiovascular SectionImaging Institute, Cleveland Clinic
  • Rodrigo Pale
    • Cardiac MR PET CT Program, Department of Radiology and Division of CardiologyMassachusetts General Hospital and Harvard Medical School
  • Meagan K. Murphy
    • Cardiac MR PET CT Program, Department of Radiology and Division of CardiologyMassachusetts General Hospital and Harvard Medical School
  • Amar B. Shah
    • Cardiac MR PET CT Program, Department of Radiology and Division of CardiologyMassachusetts General Hospital and Harvard Medical School
  • Hiram G. Bezerra
    • Cardiac MR PET CT Program, Department of Radiology and Division of CardiologyMassachusetts General Hospital and Harvard Medical School
  • Ammar Sarwar
    • Cardiac MR PET CT Program, Department of Radiology and Division of CardiologyMassachusetts General Hospital and Harvard Medical School
  • Ian S. Rogers
    • Cardiac MR PET CT Program, Department of Radiology and Division of CardiologyMassachusetts General Hospital and Harvard Medical School
  • Udo Hoffmann
    • Cardiac MR PET CT Program, Department of Radiology and Division of CardiologyMassachusetts General Hospital and Harvard Medical School
  • Suhny Abbara
    • Cardiac MR PET CT Program, Department of Radiology and Division of CardiologyMassachusetts General Hospital and Harvard Medical School
  • Ricardo C. Cury
    • Cardiac MR PET CT Program, Department of Radiology and Division of CardiologyMassachusetts General Hospital and Harvard Medical School
    • Baptist Cardiac and Vascular Institute
  • Thomas J. Brady
    • Cardiac MR PET CT Program, Department of Radiology and Division of CardiologyMassachusetts General Hospital and Harvard Medical School
Article

DOI: 10.1007/s10554-010-9683-3

Cite this article as:
Blankstein, R., Bolen, M.A., Pale, R. et al. Int J Cardiovasc Imaging (2011) 27: 579. doi:10.1007/s10554-010-9683-3
  • 170 Views

Abstract

To evaluate the effective radiation dose and image quality resulting from use of 100 vs. 120 kV among patients referred for cardiac dual source CT exam (DSCT). Prospective data was collected on 294 consecutive patients referred for DSCT. For each scan, a physician specializing in cardiac CT chose all parameters including tube current and voltage, axial versus helical acquisition, and use of tube current modulation. Lower tube voltage was selected for thinner patients or when lower radiation was desired for younger patients, particularly females. For each study, image quality (IQ) was rated on a subjective IQ score and contrast (CNR) and signal-to-noise (SNR) ratios were calculated. Tube voltage of 100 kV was used for 77 (26%) exams while 120 kV was used for 217 (74%) exams. Use of 100 kV was more common in thinner patients (weight 166lbs vs. 199lbs, P < .001). The effective radiation dose for the 100 and 120 kV scans was 8.5 and 15.4 mSv respectively. Among scans utilizing 100 and 120 kV, there was no difference in exam indication, use of beta blockers, heart rate, scan length and use of radiation saving techniques such as prospective ECG triggering and tube current modulation. The IQ score was significantly higher for 100 kV scans. While 100 kV scans were found to have higher image noise then those utilizing 120 kV, the contrast-to-noise and signal-to-noise were significantly higher (SNR: 9.4 vs. 8.3, P = .02; CNR: 6.9 vs. 6.0, P = .02). In selected non-obese patients, use of low kV results in a substantial reduction of radiation dose and may result in improved image quality. These results suggest that low kV should be used more frequently in non-obese patients.

Keywords

Cardiac CTRadiation doseTube voltage

Introduction

Recent technological advancements in cardiac computed tomography (CCT) have resulted in improved diagnostic accuracy and have contributed to the increased adoption of this non-invasive technique into clinical practice [1]. As potential applications of CCT continue to emerge, concerns exist in regards to patient radiation exposure [2]. Indeed, radiation dose remains a limitation of CCT, especially since other noninvasive tests which do not use any radiation (i.e. stress echo or cardiac MRI) are available. Due to such concerns, multiple recent mechanisms have been developed for lowering radiation dose with cardiac CT.

One such technique which allows for lower radiation dose during computed tomography is lowering the tube voltage from 120 to 100 kV [3, 4]. Such a decrease in tube voltage results in a reduction of both the peak and mean energy of emitted photons. Importantly, this results in a non-linear lowering of patient radiation dose such that even small changes in kV may result in substantial reduction of radiation dose. Heyer et al prospectively randomized 60 patients referred for pulmonary CT angiography for suspected pulmonary embolus to receive either 100 or 120 kV and found no significant difference in objective or subjective image quality [5]. While such a strategy of lowering kV with cardiac CT has been proposed as a technique to lower radiation dose [3, 4, 6, 7], the evidence supporting this technique is still limited [8]. Subsequently, a recent multi center observational study discovered that the use of 100 kV was only applied in 5% of scans. Moreover, for 2 out of the 4 vendors represented in this study, no scans were performed with 100 kV tube voltage [9].

The limited use of low kV may be explained by several factors, including a lack of awareness of this technique and a lack of confidence resulting from a paucity of data demonstrating the validity of this approach. Given that evidence on the use of low kV with the dual source CT (DSCT) is limited, the purpose of this study was to evaluate the effective radiation dose and image quality resulting from use of 100 vs. 120 kV among patients referred for a cardiac DSCT exam.

Materials and methods

Patient population

The institutional review board approved the study. Our study population included 294 consecutive patients referred for a cardiac DSCT exam at the Massachusetts General Hospital that were scanned with a tube voltage of either 100 or 120 kV. All data regarding image acquisition parameters and reason for the exam were collected prospectively.

Image acquisition

CCT was performed on the Definition dual-source 64-slice CT scanner (Siemens Medical Systems) with a gantry rotation time of 330 ms and standard detector collimation of 0.6 mm. A flying focus along the z-axis (z-sharp technology) was used to acquire 64 overlapping 0.6 mm slices using two 32-detector rows. The resulting temporal resolution was 83 ms (See example in Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs10554-010-9683-3/MediaObjects/10554_2010_9683_Fig1_HTML.jpg
Fig. 1

Example of coronary CT angiography acquired with tube voltage of 100 kV. 35 year old male with BMI of 28 kg/m2 who was referred for evaluation of atypical chest pain/suspected anomalous origin of the coronary arteries. For each vessel, the images show the curved multi-planar reformation on top and the average multi-planar reformation at the bottom. RCA right coronary artery; LAD left anterior descending

Prior to each scan a test bolus of 20 cc of contrast was administered at 5–6 cc/s and dynamic axial imaging was performed. The timing for image acquisition was determined by adding ~2 s to the time of peak contrast enhancement in the ascending aorta. Sublingual nitroglycerin (dose of 0.6 mg) was administered to all patients. Beta blockers were not routinely administered but were selectively given to patients with irregular or elevated (e.g >80 beats per minute) heart rates.

For the purposes of selecting kV and mAS, no pre-defined absolute cut off values based on either weight or BMI were utilized. Instead, one of 6 experienced physicians specializing in cardiac CT was present at each scan and selected the tube voltage (kV) and tube current (mAs) based on the following factors: (a) patient height, weight, and calculated BMI were provided for each patient. While physicians could choose whatever tube current and voltage they thought was most appropriate, it was recommended to consider the use lower voltage for patients with BMI < 30 kg/m2; (b) assessment of patient’s body habitus and chest wall attenuation—this was performed by both a physical inspection of the patient by the physician prior to the scan and by visualization of the coronal scout image and axial test-bolus images (obtained prior to the scan) to determine if excessive chest wall adiposity or soft tissue was present; (c) clinical indication—physicians were advised to consider lower tube voltage for indications that did not require the highest possible resolution (e.g. suspected anomalous origins of the coronary arteries).

Scans utilizing an axial acquisition using prospective triggering (Siemens: Sequential Scanning) were chosen when low and regular heart rate could be achieved in patients in whom radiation savings was thought to be most important (young patients ~age < 50, especially if female) . All such axial acquisitions were obtained at 65% of the R–R interval. For scans utilizing retrospective triggering, tube current modulation window width were determined by the physician based on the clinical indication, patient’s heart rate, and probability of CAD. For patients with elevated or irregular heart rate a window of 35–75% of the R–R interval was typically selected. On the other hand for patients with a lower heart rate, a window of 50–75% of the R–R interval was typically selected.

Image reconstruction and analysis

For retrospectively gated CCT, raw data from 5 to 95% of the cardiac cycle was used to reconstruct images at 10% intervals with 0.75 mm slice thickness and overlap of 0.4 mm using a medium smooth reconstruction kernel (B26f). For prospectively triggered scans, raw data (65% phase) was used to reconstruct a single data set using the same settings as above. Axial and double-oblique images viewed in thin-slab maximal intensity projections (MIP) and multi planar reformation (MPR) settings were used for image analysis.

Radiation exposure

CT dose index volume (CTDIvol) and dose-length product (DLP) were provided by the scanner console. Effective radiation dose was calculated by multiplying the dose-length product (DLP) of the cardiac scan (not including test bolus) by a constant (k = .017 mSv/mGy/cm).

Image quality

Image quality was determined based on three methods:
  • Per vessel analysis of non-evaluable vessels—For each patient who underwent CCT for coronary evaluation, clinical readings were used to identify any vessels which were deemed to have limited or non-evaluable segments. The advantage of this method is that it only identifies limitations that are thought to have an impact on the clinical interpretation.

  • Image quality score—subjective image quality (IQ) was determined retrospectively by use of subjective scale: 1 = poor; 2 = significantly reduced; 3 = mildly reduced; 4 = excellent. Each study was independently reviewed by two blinded readers (cardiologist and radiologist) who have each interpreted >500 coronary CT exams. The average of the two scores was used to represent the study IQ.

  • Contrast (CNR) and signal-to-noise (SNR) ratios—Image noise was derived from the standard deviation of the density values (in Hounsfield units) within a large region of interest in the left ventricle. The signal-to-noise ratio was defined as the ratio of the mean signal intensity divided by image noise. The contrast-to-noise ratio was defined as the difference between the mean density of the contrast-filled left ventricular chamber and the mean density of the left ventricular wall, which was divided by image noise. This method, which has been previously described [3], is relevant for a wide range of CCT studies, regardless of whether or not evaluation of the coronary arteries was performed.

Statistical analysis

Data analysis was performed using Stata IC version 10.0 (StataCorp LP, College Station, Texas). All continuous variables were expressed as mean ± standard deviation while categorical variables were expressed as percentage. Differences in continuous variables were assessed using student unpaired t-tests. Differences in dichotomous variables were assessed using the chi-square test, fisher’s exact test, or rank sum test, as appropriate. Kappa statistic was used to test for inter-observer agreement of image quality assessment. A P-value < 0.05 was considered statistically significant.

Results

Of the 294 DSCT exams included in the analysis, tube voltage of 100 kV was utilized for 77 (26%) exams while 120 kV was used for 217 (74%) exams.

Baseline characteristics

Among the 294 patients included in our study, the mean age was 60 ± 13 and 63% of patients were male. The average heart rate 68.6 ± 16 beats per minute. When considering the indication for exam, 60% were referred for coronary evaluation, while the remainder exams were referred for evaluation of other cardiac structures (e.g. pulmonary veins, aorta), often in addition to coronary CTA.

A comparison of patient characteristics of the 100 and 120 kV groups (Table 1) revealed no differences in exam indication, or patient’s age and gender. Use of 100 kV (vs. 120 kV) was more common in thinner patients (166lbs vs. 199lbs, P < .001) and among those with lower BMI (25.6 vs. 30.26 kg/m2, P < .001). When examining the CT acquisition parameters (Table 2), there were no differences in the use of beta blockers or heart rate during acquisition. There was also no difference in scan length or in the use of radiation saving techniques such as prospective triggering and tube current modulation. For those patients scanned with 100 kV, however, a lower tube current (mA or mAs) was selected, likely reflecting the fact that such patients were typically thinner.
Table 1

Patient characteristics

 

100 kV (N = 77)

120 kV (N = 217)

P-value

Indication

.259

 Coronary evaluation

45 (58%)

131 (60%)

 Pulmonary vein

20 (26%)

61 (28%)

 Other (aorta/congenital/mass)

12 (16%)

25 (12%)

Demographics

 Age (years)

55.4 ± 14.4

57.5 ± 13.0

.237

 Gender (females)

35 (45%)

73 (34%)

.065

 Height (inches)

67.1 ± 4.2

70.0 ± 4.6

.317

 Weight (pounds)

165.7 ± 35.1

198.6 ± 38.6

<.001

 BMI (kg/m2)

25.6 ± 4.0

30.2 ± 5.2

<.001

Table 2

Scan parameters and radiation dose

 

100 kV (N = 77)

120 kV (N = 217)

P-value

Scan parameters

Use of radiation dose lowering techniques:

 Prospective triggering

6 (8%)

30 (14%)

.17

 Tube current modulation

60 (78%)

148 (68%)

.11

 MinDosea

11 (14%)

34 (16%)

.77

Use of beta blockers

6 (8%)

21 (10%)

.62

Heart rate during acquisition (average)

69.5 ± 16

68.3 ± 15

.56

Tube currentb

 mA (for prospectively triggered scans)

159 ± 23

214 ± 38

.002

 mAs (for retrospective scans)

306 ± 57

338 ± 76

.002

 Pitch (average; helical scans only)

.29 ± .08

.29 ± .08

.72

 Scan length (cm)

19.2 ± 3.6

18.9 ± 5.3

.68

Radiation dose estimates

All scans

 Dose length product (DLP, mGyxcm)

498.6 ± 260

903.1 ± 572

<.0001

 CT dose index (CTDI) volume (mGy)

26.0 ± 12

47.0 ± 25

<.0001

 Effective radiation (mSv)

8.5 ± 4

15.4 ± 10

<.0001

Prospective triggering

 Dose length product (DLP, mGyxcm)

104

213.8

.0016

 CT Dose index (CTDI) volume (mGy)

5.3 ± 0.8

12.4 ± 2.2

<.0001

 Effective radiation (mSv)

1.8 ± 0.6

3.6 ± 1.3

.0016

Retrospective gating (Helical)

 Dose length product (DLP, mGyxcm)

532.0

1013.7

<.0001

 CT Dose index (CTDI) volume

27.8 ± 11.1

52.4 ± 22.8

<.0001

 Effective radiation (mSv)

9.0 ± 4.1

17.2 ± 9.2

<.0001

aMinDose (Siemens Medical Solutions, Forchheim Germany) is a method of tube current modulation reducing the tube current to 4% of the maximal value

bFor prospective triggering, tube current (mAs) is calculated as: [total mA × exposure time], while for retrospective triggering, tube current/rotation is calculated as: [total mA × gantry rotation time]

Radiation dose

The CTDIvol for the 100 and 120 kV scans was 26 and 47 mGy, yielding an estimated effective radiation doses of 8.5 and 15.4 mSv, respectively. (Table 2) Among scans that utilized prospective triggering, the CTDIvol was 5.3 and 12.4 mGy corresponding to an effective dose of 1.8 and 3.6 mSv, respectively for 100 and 120 kV scans. On the other hand, the CTDI vol was 27.8 and 52.4 mGy with a resulting calculated effective dose of 9.0 and 17.2 mSv, respectively, for scans that utilized retrospective gating (See Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs10554-010-9683-3/MediaObjects/10554_2010_9683_Fig2_HTML.gif
Fig. 2

Radiation Dose of Cardiac CT Utilizing 100 vs. 120 kV

Image quality

Table 3 summarizes the results of the image quality assessment between the 100 and 120 kV scans. The interobserver agreement for IQ scores between the two readers was excellent (kappa = 0.87). The mean IQ score was 3.2 for 100 kV scans and 3.0 for 120 kV scans (P = .03) suggesting that subjective image quality was slightly improved with use of low kV.
Table 3

Image quality assessment

Image quality

100 kV (N = 77)

120 kV (N = 217)

P-value

Image quality score (mean)

3.2

3.0

.0244

 1

2 (3%)

9 (4%)

P for trend = .028

 2

9 (12%)

36 (17%)

 3

33 (43%)

108 (50%)

 4

33 (43%)

63 (29%)

Non-evaluable vessels (e.g. ≥1 segment non-evaluable)

 Left main

3 (5%)

7 (4%)

.703

 Left anterior descending

11 (19%)

34 (19%)

.895

 Right coronary artery

8 (14%)

29 (17%)

.583

 Left circumflex

6 (10%)

28 (16%)

.272

Measurements

 Signal (hounsfield units)

463 ± 124

340 ± 68

<.001

 Noise (hounsfield units)

53.1 ± 16.7

45.7 ± 14.6

<.001

 Contrast-to-noise ratio

6.9 ± 2.8

6.0 ± 3.0

.0176

 Signal-to-noise ratio

9.4 ± 3.5

8.3 ± 3.7

.0244

While 100 kV scans were found to have higher image noise then those utilizing 120 kV, the contrast-to-noise and signal-to-noise were significantly higher for the low kV scans (SNR: 9.4 vs. 8.3, P = .02; CNR: 6.9 vs. 6.0, P = .02). Finally, there was no difference in the number or type (i.e. LAD, LCx, etc…) of limited or non-evaluable vessels between the 100 and 120 kV scans. (P-values of 0.7, 1.00, 0.39, 0.68 for left main, LAD, LCx, and RCA vessels, respectively).

Discussion

In this study evaluating the use of low tube voltage for cardiac CT we observed that lowering the tube voltage from 120 to 100 kV resulted in a substantial reduction of radiation dose of approximately 45%. Importantly, there was no compromise of image quality. In fact, both subjective image quality and objective measures such as contrast-to-noise and signal-to-noise were significantly higher with lower kV.

Similar findings have been reported by Hausleiter et al. [3] where changing from a 120 to 100 kV technique resulted in a 22 and 43% decrease in radiation dose with 16 and 64 slice scanners, respectively, and Leshcka et al. [6] where a decreased of effective dose of 25% was observed. While in both of these studies, measure of image quality between 120 and 100 kV were similar, in our study which employed an individualized selective use of lower tube voltage (and also has a larger number patient in which 100 kV was used) we were able to detect a significantly higher subjective and objective image quality associated with use of low kV.

The potential mechanism for similar image quality is that despite of more noise, use of low kV results in higher iodine X-ray absorption thus resulting in higher image signal. This is because a lower kV translates into lower effective photon energy (effective photon energy is approximately one half of the kV) and when the effective photon energy is closer to the K-edge of iodine (33.2 keV) CT attenuation increases. While the image noise increases with the lower kV, the attenuation of contrast material at 100 kV compared to 120 kV increases to a greater proportion, yielding an overall increase in the contrast to noise ratio (CNR) and signal to noise ratio (SNR).

Our study represents a non-randomized cohort in which use of kV was largely influenced by the patient’s body habitus. Therefore, the improved image quality which was seen with lower kV could also be attributed to more optimal patient size. Despite our non-randomized design, the two groups compared in our study appeared to be similar as there were no differences in age, gender, exam indication, use of retrospective ECG gating versus prospective ECG triggering, or scan length. The observational nature of our study also allowed us to identify actual practice patterns. For instance, we observed that 100 kV was selected for approximately one quarter of all cardiac CT exams. While this represents more frequent use of low kV in our center than the 5% which was observed in a recent multi center trial, [9] there is potential to use low kV techniques for even more patients.

A unique aspect of our study is that an experienced imaging physician was present at each exam and selected individualized scan parameters for each patient based on various metrics including body habitus—assessed by BMI, physical inspection, and visualization of scout and axial test-bolus images prior to the scan—and exam indication. Given the fact that it is not realistic to have an experienced physician at every scan and in light of the subjective nature of our approach, it is important to develop more objective future guidelines for use of low tube voltage. Such guidelines could promote wider adoption of radiation saving techniques and may translate into dramatic reduction in radiation dose for non-obese patients undergoing cardiac CT.

This concept is exemplified in the recent study by Alkadhi et al. which showed that a protocol of selective use of low kV and axial prospective triggering based on heart rate and BMI, results in a significant reduction in radiation while maintaining image quality [10]. In this European study, a tube voltage of 100 kV was used for all patients with a BMI < 25 kg/m2. Notably, over half of the patients randomized to receive the “tailored protocol” had a BMI < 25 and thus were imaged with low kV. In contrast, in the US the median BMI of patients undergoing cardiac CT is generally higher and establishing a higher cut-point, as is partially supported by our study, may allow more patients to benefit from radiation savings associated with low tube voltage.

There are several noteworthy considerations for selecting a single cut-point for use of lower tube voltage. While calculating BMI or measuring a patient’s weight are simple, both of these techniques are not perfect as due to differences in weight distribution, these measures do not always predict the amount of chest wall attenuation. Other potential strategies include integration of weight and BMI with visual inspection by a trained clinician or technologist (as was performed in our center). While this technique is, by nature, more subjective, it may be useful in situation when the BMI may be viewed as borderline (25–32 kg/m2). Perhaps, a “test-image” obtain prior to a patient’s scan would be a better, and more objective measure, for predicting how using different setting such as tube current and tube voltage could be most optimally lowered without compromising image quality. In addition to body habitus, any future recommendations should also consider factors which would identify patients who are more susceptible to the harmful effects of ionizing radiation (i.e. younger patients, particularly females). The use of 100 kV may be even more important in such “radiation vulnerable” populations. Finally, the use of CCT for indications which do not require the visualization of small low resolution structures such as plaques—namely, evaluation for suspected anomalous origins of the coronary arteries or evaluation of the pulmonary veins—should also be considered for use of low tube voltage.

While it is clear that selective use of low tube voltage (i.e. 100 kv) results in a substantial reduction of radiation dose without compromising image quality, whether this has an effect on the diagnostic accuracy of CCT is less apparent. Answering this question will require randomized blinded prospective randomized studies comparing high and low kV scans with invasive angiography. Given the already excellent diagnostic accuracy of CCT, the number of patients that would be required in order to detect the small, if any, differences in diagnostic accuracy would be very large, thus making such a trial unlikely to be performed.

Regardless of what acquisition parameters are used, the benefits of CCT should be weighed against potential hazards such as exposure to ionizing radiation and the ALARA principle, stating that radiation exposure to patients should be kept As Low As Reasonably Achievable, should be followed.

Acknowledgments

Drs. Blankstein and Rogers received support from NIH grant 1T32 HL076136-02.

Conflict of interest

None.

Copyright information

© Springer Science+Business Media, B.V. 2010