In Vivo Comparison of Radiation Exposure of Dual-Energy CT Versus Low-Dose CT Versus Standard CT for Imaging Urinary Calculi

Abstract

Purpose: Dual-energy computed tomography (DECT) is an emerging imaging modality with the unique capability of determining urinary stone composition. This study compares radiation exposure of DECT, standard single-energy CT (SECT), and low-dose renal stone protocol single-energy CT (LDSECT) for the evaluation of nephrolithiasis in a single in vivo patient cohort.

Materials and Methods: Following institutional review board (IRB) approval, we retrospectively reviewed 200 consecutive DECT examinations performed on patients with suspected urolithiasis over a 6-month period. Of these, 35 patients had undergone examination with our LDSECT protocol, and 30 patients had undergone examination of the abdomen and pelvis with our SECT imaging protocol within 2 years of the DECT examination. The CT dose index volume (CTDI_vol) was used to compare radiation exposure between scans. Image quality was objectively evaluated by comparing image noise. Statistical evaluation was performed using a Student's t-test.

Results: DECT performed at 80/140 kVp and 100/140 kVp did not produce a significant difference in radiation exposure compared with LDSECT (p=0.09 and 0.18, respectively). DECT performed at 80/140 kVp and 100/140 kVp produced an average 40% and 31%, respectively, reduction in radiation exposure compared with SECT (p<0.001). For patients imaged with the 100/140 kVp protocol, average values for images noise were higher in the LDSECT images compared with DECT images (p<0.001) and there was no significant difference in image noise between DECT and SECT images in the same patient (p=0.88). Patients imaged with the 80/140 kVp protocol had equivocal image noise compared with LDSECT images (p=0.44), however, DECT images had greater noise compared with SECT images in the same patient (p<0.001). Of the 75 patients included in the study, stone material was available for 16; DECT analysis correctly predicted stone composition in 15/16 patients (93%).

Conclusion: DECT provides knowledge of stone composition in addition to the anatomic information provided by LDSECT/SECT without increasing patient radiation exposure and with minimal impact on image noise.

Introduction

Computed tomography (CT) scanning is the preferred initial diagnostic imaging technique in patients with suspected urinary calculi,^1,2 due to its high sensitivity and specificity.^3,4 Aside from the initial diagnostic CT imaging, patients are also exposed to radiation during treatment procedures (fluoroscopic imaging) and while monitoring for treatment response (additional CTs or radiographs). Since radiation exposure is intimately associated with the diagnosis and management of renal stone disease, it is important for urologists to have an understanding of the available CT imaging techniques and their associated radiation exposure. With the recent introduction of dual-energy CT (DECT) for the evaluation of renal calculi, radiation exposure continues to be a concern for patients with urolithiasis.

DECT is an emerging imaging modality with the unique ability to accurately determine urinary stone composition as either uric acid (UA) or non-UA.^5–10 Both material-specific and diagnostic images are created from a single acquisition (Fig. 1). In the future, DECT will likely be able to distinguish other stone types.^11,12 Other than analysis of collected stone material, dual-energy evaluation is the only currently available noninvasive method for determining stone composition. Conventional stone imaging characteristics (>4 mm, HU <500) in combination with urine pH (<5.5) have been suggested to predict UA composition calculi with a positive predictive value of 90%.¹³ However, DECT is able to provide the information of stone composition at the same time of diagnosis without requiring further testing. Therefore, DECT is able to offer a benefit during preoperative treatment planning. A common misconception of DECT imaging is that dual energy implies a doubling of the radiation exposure.^14,15 Before the introduction of dual-source (DS), rapid switching single-source, and dual-layer detector CT scanners, dual-energy examinations were performed by scanning the patient twice at two different kVp, thus doubling the radiation exposure. However, more advanced, commercially available DECT scanners do not appear to significantly increase radiation exposure when compared with a conventional single-energy CT (SECT) scanner, while still preserving image quality.^15,16

FIG. 1.

Axial (A) dual-energy computed tomography (DECT) material-specific (B) diagnostic and (C) conventional single-energy low-dose renal stone protocol images through the left kidney in the same patient demonstrate a small renal calculus. The calculus is purple on the material-specific dual-energy images indicating a nonuric acid stone. The arrow in each image is pointing to the stone.

Awareness of radiation exposure from repetitive imaging has prompted most institutions to use a low-dose renal stone protocol single-energy CT (LDSECT) for evaluation of urinary calculi. The purpose of this study is to compare in vivo radiation exposure of DS DECT imaging for renal stones versus conventional SECT of the abdomen and pelvis and LDSECT within the same patient cohort.

Materials and Methods

Radiation dose reporting

Radiation dose is often reported as CT dose index volume (CTDI_vol), dose length product (DLP), or the effective dose (ED). Both CTDI_vol and DLP are automated calculations reported for every CT scan. The DLP is determined by multiplying the CTDI_vol by the scan length. The ED is determined by multiplying the DLP by a constant, which varies dependent upon the part of the body being scanned. When interpreting radiation exposure, it is important to realize that CTDI_vol is a standardized parameter (provided by the manufacturer) to measure the scanner radiation output and is not the actual individualized patient dose.¹⁷ The relationship between CTDI_vol and patient dose is dependent upon many factors, including patient size and body composition. Since DLP and ED are derived from the CTDI_vol, these values are also not true measures of patient dose and may increase the magnitude of error in the estimate of the patient dose.

In addition, many CT scanners now have automatic exposure control (AEC) software to adjust the technique parameters (and as a result the CTDI_vol) to achieve specified image quality across a range of patient sizes. In larger patients, the radiation exposure is increased by increasing the tube current (mA) to preserve image quality. It should be noted that the dose to internal organs does not increase linearly with an increase in tube current settings, since most of the additional X-ray dose is absorbed by excess adipose tissue.^17–20 Thus, dose calculations based on CTDI_vol will be slightly overestimated in larger patients and underestimated in smaller patients. Furthermore, dose estimations derived from the CTDI_vol, namely, DLP and ED, can further magnify this error in dose estimation. For these reasons, CTDI_vol has been proposed as a way of comparing radiation exposure from CT scanning and is used in this article.^17–20

Patients and study protocol

Following institutional review board (IRB) approval, we retrospectively reviewed 200 consecutive DECT examinations performed on patients with suspected urolithiasis between May 2012 and December 2012. Of these, 35 patients had undergone an examination with our standard LDSECT and 30 patients had undergone an examination of the abdomen and pelvis with our standard SECT imaging protocol within 2 years of the DECT examination for a total of 65 patients included in the study. Of these 65 patients imaged with DECT, 25 (38%) were females and 40 (62%) were males; the average patient age was 59.6 years (range 31–86); the average body–mass index (BMI) was 28.2 kg/m² (range 18.5–48.8 kg/m²).

DECT protocol

All patients were imaged according to our standard DECT renal stone protocol with a (DS DECT) Siemens Definition Flash Scanner (Siemens Healthcare, Forchheim, Germany) in dual-energy mode. Continuous noncontrast images were acquired from just above the diaphragm through the pubic symphysis. In patients with a cross-sectional diameter (measured at the level of the kidneys) of 35 cm and below, the tube voltages and reference effective tube current–time products were set to 80 kVp/419 mAs and 140 kVp/162 mAs with quality reference CTDI_vol=16 mGy. In patients with a cross-sectional diameter greater than 35 cm, the tube voltages and reference effective mAs were set to 100 kVp/210 mAs and 140 kVp/162 mAs with quality reference CTDI_vol=17 mGy. Other scan parameters were constant regardless of the cross-sectional diameter: collimation of 32×0.6 mm, pitch of 0.7. Attenuation-based tube current modulation (CareDose4D; Siemens Healthcare) was utilized. Axial and coronal images were reconstructed using a mixed (low and high kVp) dataset with 5 mm slice thickness/2.5 mm slice interval (axial) and 3 mm slice thickness/2.5 mm slice interval (coronal) with a standard soft tissue (B30f) convolution kernel. Syngo post-processing software (VE36A; Siemens Healthcare) was utilized to create material-specific image reconstructions in the axial and coronal planes with 1.0 mm slice thickness, 0.8 mm slice interval, and D30f convolution kernel.

LDSECT protocol

The 35 patients imaged with the LDSECT protocol had the following parameters. Continuous noncontrast images were acquired from just above the diaphragm through the pubic symphysis. The single source tube voltage was set to 120 kV with reference effective tube current–time product of 160 mAs and CTDI_vol of 11 mGy. Other scan parameters included collimation of 24×1.2 mm and pitch of 1. Attenuation-based tube current modulation (CareDose4D; Siemens Healthcare) was utilized. Axial (5 mm slice thickness and slice interval) and coronal (3 mm slice thickness and 2.5 mm slice interval) reconstructions with a B30f convolution kernel were performed.

SECT protocol

All patients imaged with routine abdomen and pelvis protocol had the following parameters. Continuous images were acquired from just above the diaphragm through the pubic symphysis. The single-source tube voltage was set to 120 kV with reference effective tube current–time product of 240 mAs and CTDI_vol=16 mGy. Other scan parameters included collimation of 24×1.2 mm and pitch of 1. Attenuation-based tube current modulation (CareDose4D; Siemens Healthcare) was utilized. Axial reconstructions were performed (5 mm slice thickness and slice interval, B30f convolution kernel).

Image quality

Image quality was assessed quantitatively by measuring the attenuation and standard deviation (SD) of circular regions of interest on the axial reconstructions in the aorta at the level of the renal pelvis and in the bladder at the level of the femoral heads (Fig. 2).²¹ Regions of interest were sized appropriately for each individual patient. DECT and non-DECT images were measured at equivalent levels between the two scans. Image noise was determined by averaging the SD at both levels. For consistency, a single author (M.A.J.) performed all of the measurements.

FIG. 2.

Axial (A) DECT diagnostic image and (B) conventional single-energy image through the abdomen in the same patient. A region of interest is drawn in the aorta for comparison of image noise.

Statistical analysis

Reported CTDI_vol values were acquired from the automated calculations produced at the end of each patient examination. Statistical evaluation was performed using a Student's t-test; a p-value of 0.05 was considered to indicate a statistically significant difference.

Results

Results for all patients in the study group imaged with DECT are summarized in Table 1. Overall, DECT had an average CTDI_vol of 12.4 mGy. When divided into subgroups based upon the waist circumference, DECT performed at the lower setting (80/140 kVp) had an average CTDI_vol of 9.7 mGy, and DECT performed at the higher setting (100/140 kVp) had an average CTDI_vol of 15.6 mGy.

Table 1.

Radiation Exposure and Noise Summary for All DECT Examinations

	DECT
Protocol (no. of patients)	Mean CTDIvol (mGy)	Mean noise (SD)
All (65)	12.4	17
80/140 kVp (35)	9.7	14.6
100/140 kVp (30)	15.6	15.8

Summary of mean CTDI_vol and noise for all 65 patients imaged with DECT. Mean CTDI_vol and image noise were lower in patients imaged with the 80/140 kVp protocol compared with patients imaged with the 100/140 kVp protocol. SD is average HU standard deviation for ROIs drawn in the bladder and the aorta at the level of the renal pelvis.

CTDI_vol=computed tomography dose index volume; DECT=dual-energy computed tomography; ROI=region of interest.

Data from the patients imaged with DECT and SECT are summarized in Table 2. Patients imaged with the low kVp (80/140) DECT protocol had a 40% reduction in radiation exposure (p<0.0001); image noise measurements in this group were slightly higher for DECT images (mean SD=13.7) compared with SECT images (mean SD=11.9), (p=0.001). The high kVp (100/140) DECT protocol had a 31% reduction in radiation exposure (p<0.0001); however, there was no difference in the image quality based on image noise measurements between DECT (SD=17.2) and SECT (SD=17.3) images (p=0.88).

Table 2.

Radiation Exposure and Noise Summary for DECT Versus SECT

	Mean CTDIvol (mGy)			Mean noise (SD)
Protocol (no. of patients)	DECT	SECT	% reduction in radiation exposure	DECT	SECT	p-Value
80/140 kVp (15)	9.3	15.7	40	14	11.6	<0.001
100/140 kVp (15)	16	24.2	31	17.2	17.3	0.88
All (30)	12.7	20	36	15.6	14.5	0.03

Summary of mean CTDI_vol and noise for all 30 patients imaged with DECT and SECT. DECT examinations had a lower mean CTDI_vol compared with SECT. Image noise was equivalent in the 100/140 kVp protocol; however, patients imaged with the 80/140 kVp protocol had greater DECT image noise compared to SECT images. p-Values were calculated using a Student's t-test. % reduction in radiation exposure=(SECT_CTDIvol−DECT_CTDIvol)/SECT_CTDIvol. SD is average HU standard deviation for ROIs drawn in the bladder and the aorta at the level of the renal pelvis.

SECT=single-energy computed tomography.

The patients imaged with both DECT and LDSECT are summarized in Table 3. Radiation exposure from DECT performed at either setting was not significantly different (p=0.18 for 100/140 kVp; p=0.09 for 80/140 kVp) compared to LDSECT. Image noise in the 100/140 kVp DECT images (SD=17.2) was less than the LDSECT images (SD=21.4), p<0.001, whereas image noise in the 80/140 kVp DECT images was not significantly different compared with the LDSECT images (15.1 vs. 15.6; p=0.44).

Table 3.

Radiation Exposure and Noise Summary for DECT Versus LDSECT

	Mean CTDIvol (mGy)			Mean noise (SD)
Protocol (no. of patients)	DECT	LDSECT	% reduction in radiation exposure	DECT	LDSECT	p-Value
80/140 kVp (20)	10	10.9	—	15.1	15.6	0.44
100/140 kVp (15)	14.7	14	—	17.2	21.4	<0.001
All (35)	12.2	12.2	—	16	18	0.001

Summary of mean CTDI_vol and noise for all 35 patients imaged with DECT and LDSECT. DECT and LDSECT are dose neutral regardless of DECT protocol performed. The 80/140 kVp protocol resulted in equivalent image noise, whereas the 100/140 kVp protocol resulted in less image noise in the DECT group. p-Values were calculated using a Student's t-test. SD is average HU standard deviation for ROIs drawn in the bladder and the aorta at the level of the renal pelvis. % reduction in radiation exposure=(SECT_CTDIvol−DECT_CTDIvol)/SECT_CTDIvol.LDSECT

LDSECT=low-dose renal stone protocol single-energy computed tomography.

Of the 75 patients included in the study, stone material was available for 16; DECT analysis correctly predicted stone composition in 15/16 patients (93%). The one mischaracterized stone was less than 3 mm.

Discussion

The results of our study indicate that the implemented DECT protocol does not increase patient radiation exposure and maintains image quality compared to SECT or LDSECT. Therefore, radiation exposure concern should not deter utilization of DECT compared to standard or low-dose standard energy CT for evaluation of renal calculi. In addition to the information attained in a SECT, DECT further aids in the determination of stone composition, which may affect patient management decisions and thereby may result in less radiation exposure due to repeat studies. Furthermore, DECT has been reported to create a unique color contrast between ureteral stents and residual stone material, which may aid in improving detection and characterization of retained stone fragments, thereby also improving patient management decisions and indirectly reducing radiation exposure.^22,23

DS DECT utilizes two X-ray tubes (and detector arrays) mounted at an ∼90° angle, which allows for simultaneous acquisition of both low and high kVp data as well as independent tube currents to each X-ray source. This is important for minimizing radiation exposure, while maintaining optimal spectral separation (important for determination of stone composition) and image quality (noise reduction). These DECT settings have been reliably tested with a reported accuracy in the literature of nearly 100% for distinguishing stone composition as either UA or non-UA in stones 3 mm or greater.^5–11

Smaller patients had a greater reduction in radiation exposure (40% less than SECT) than larger patients (31% less than SECT) when imaged with the DECT protocol; this variation in radiation exposure is expected to be based upon the higher kVp DECT protocol used for larger patients. This study utilized a second-generation DS DECT Siemens Definition Flash scanner, which has a tin filter (allowing for greater spectral separation) and uses 100 kVp (rather than 80 kVp) as the low-energy beam for patients with a cross-sectional diameter greater than 35 cm. The 100 kVp setting allows for better image quality, but typically results in slightly increased CTDI_vol for larger patients compared to those of average size. Furthermore, we utilized AEC software (CareDose4D) to preserve image quality, which also typically results in increased CTDI_vol in larger patients and decreased CTDI_vol in smaller patients when compared with patients of standard size.

One advantage of our study is that by comparing scans within the same patient cohort, we internally control for patient gender, size, and body composition for the comparison of radiation exposure between imaging protocols. Since most institutions use AEC when acquiring all types of CT scans, there is a general increase in CTDI_vol for larger patients regardless of the scanning protocol. Thus, when evaluating the radiation exposure of any particular scanning protocol, it is important for the ordering physician to be aware of the relative exposure difference between protocols in addition to the reported average radiation exposure, the latter of which may be based upon an average size patient/phantom or an in vivo population of varying sizes.

Reported average radiation exposure in our study may be slightly overestimated since the average BMI of our patient population was in the overweight range (with nearly half the patients being in the obese range, BMI >30 kg/m²). The absolute values of radiation exposure in this study are less important than the finding that DS DECT is at least dose neutral with preserved image quality compared with our LDSECT scanning protocol in the same patient.

Interestingly, DECT image noise for smaller patients was slightly more noisy compared to SECT images but equivalent to LDSECT images, whereas DECT image noise in larger patients was equivalent to SECT images and slightly less noisy than LDSECT images. These findings are likely related to the kVp used for DECT image acquisition in larger patients, iterative reconstruction postprocessing used to reduce noise in DECT images, and the inherent higher noise in all imaging performed upon larger patients. With continued advances in iterative reconstruction software and filters to create spectral separation, DECT has a greater potential than LDSECT or SECT for reducing noise through postprocessing, allowing for lower X-ray tube settings and thereby reducing patient radiation exposure.¹⁵

Conclusion

This in vivo, internally controlled study demonstrates that DECT is able to preserve image quality without increasing patient radiation exposure compared to LDSECT or SECT. Concerns about radiation exposure should not be a deterrent to ordering DECT to diagnose or manage urolithiasis.

Abbreviations Used

AEC

automatic exposure control

BMI

body–mass index

CT

computed tomography

CTDI_vol

Computed tomography dose index volume

DECT

dual-energy computed tomography

DLP

dose length product

DS

dual source

ED

effective dose

IRB

institutional review board

LDSECT

low-dose renal stone protocol single-energy computed tomography

mA

tube current

mAs

current–time product

ROI

region of interest

SD

standard deviation

SECT

single-energy computed tomography

UA

uric acid

Acknowledgments

This project was supported by a grant from the National Institutes of Health (Mayo Clinic Urology Research Center grant DK 83007) and the Mayo Clinic. A portion of the data was presented at the 2014 Southeast Section of Urology Annual Meeting: March 20, 2014, Hollywood, Florida.

Disclosure Statement

No competing financial interests exist.