- Email: [email protected]
Incidental adrenal lesions detected on enhanced abdominal dual-energy CT: Can the diagnostic workup be shortened by the implementation of virtual unenhanced images?
European Journal of Radiology 83 (2014) 1746–1751
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Incidental adrenal lesions detected on enhanced abdominal dual-energy CT: Can the diagnostic workup be shortened by the implementation of virtual unenhanced images? Diomidis Botsikas a,∗ , Frederic Triponez b,1 , Sana Boudabbous a,2 , Catrina Hansen a,3 , Christoph D. Becker a,4 , Xavier Montet a,5 a
Geneva University Hospital, Department of Imaging and Medical Information Sciences, Division of Radiology, Rue Gabrielle-Perret-Gentil 4, 1211 Genève 4, Switzerland b Geneva University Hospital, Department of Surgery, Division of Thoracic Surgery, Rue Gabrielle-Perret-Gentil 4, 1211 Genève 4, Switzerland
a r t i c l e
i n f o
Article history: Received 27 February 2014 Received in revised form 11 June 2014 Accepted 14 June 2014 Keywords: Adrenal incidentalomas Adrenal lesion characterization Dual-energy CT Virtual unenhanced images
a b s t r a c t Objective: To determine whether post-processing of the data from portal-phase enhanced dual-energy CT (DECT), with or without the addition of a late enhanced phase acquisition, may enable characterization of adrenal lesions without the need for acquisition of pre-contrast images. Materials and methods: Twenty-two patients with 24 adrenal lesions underwent unenhanced, venous and delayed phase DECT. Of these lesions, 20 were found to be adrenal adenomas, on the basis of histopathology, unenhanced attenuation values between 0 and −10 HU, or stability over at least 6 months. For all 24 lesions, true and virtual unenhanced attenuation values were measured based on the data of the portal (VNCp) and the delayed (VNCd) DECT acquisition. The absolute washout values based on the true non-contrast (TNC) and the VNCp and VNCd image series were also measured. The washout was also calculated based on the iodine concentration measured from both contrast-enhanced acquisitions. Results: Mean virtual unenhanced attenuation values of all lesions calculated from the portal phase images was 12.6 HU, and was 4.02 HU higher than the values based on true unenhanced images (p = 0.020). Washout values calculated from virtual unenhanced attenuation based on the VNCp were also significantly different (p = 0.0304) while those calculated from VNCd and from iodine concentration correlated with the corresponding values based on the true unenhanced values (p > 0.999). Conclusions: Our data indicate that attenuation values of adrenal adenomas based on virtual unenhanced images are significantly higher than those obtained with true unenhanced images. An incidental adrenal lesion with a virtual unenhanced attenuation lower than 10 HU can thus be safely characterized as an adenoma. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Incidental adrenal masses occur in up to 5% of CT examinations [1–3]. Most of these lesions are benign, adrenal adenomas being the
∗ Corresponding author. Tel.: +41 79 56 30 778; fax: +41 22 37 27 072. E-mail addresses: [email protected] (D. Botsikas), [email protected] (F. Triponez), [email protected] (S. Boudabbous), [email protected] (C. Hansen), [email protected] (C.D. Becker), [email protected] (X. Montet). 1 Tel.: +41 79 55 32 152. 2 Tel.: +41 79 55 32 444; fax: +41 22 37 27 072. 3 Tel.: +41 79 55 32 490; fax: +41 22 37 27 072. 4 Tel.: +41 22 37 27 001; fax: +41 22 37 27 072. 5 Tel.: +41 79 55 32 485; fax: +41 22 37 27 072. http://dx.doi.org/10.1016/j.ejrad.2014.06.017 0720-048X/© 2014 Elsevier Ireland Ltd. All rights reserved.
most common diagnosis, even in oncologic patients [2]. Combined enhanced and unenhanced computed tomography is established as the method of choice for characterization of adrenal lesions [2,4]. The currently accepted CT protocol for adrenal mass characterization consists of a three phase acquisition, including an unenhanced a venous and a delayed phase [4]. The values examined for this purpose are the unenhanced attenuation and the absolute washout values of adrenal lesions. Unenhanced CT attenuations allow characterizing lipid rich adenomas. The ideal thresholds for optimal specificity for the detection of adenomas were reported to be 10 HU on unenhanced images. A lesion with attenuation values below this threshold can be considered as a lipid rich adenoma (sensitivity 71%; specificity 98%) [5,6] The calculation of absolute washout values allows characterizing all adenomas (lipid rich and lipid poor). The absolute
D. Botsikas et al. / European Journal of Radiology 83 (2014) 1746–1751
washout value is calculated with the mathematical type “Absolute Washout = [(E − D)/(E − U)] × 100”, where E is the enhanced attenuation value measured from venous phase image series, D is the delayed enhanced attenuation value measured from the 15 min delayed phase image series, and U is the unenhanced value. A threshold of 60% (sensitivity 86–88%; specificity 92–96%) [2,5,7] is reported to be the ideal compromise for this purpose. If unenhanced CT is not available the diagnosis is based on the delayed enhanced phase, with calculation of relative washout values from the mathematical type “Relative Washout = (E − D)/E × 100” where E is the enhanced attenuation value measured from venous phase image series, D is the delayed enhanced attenuation value measured from the 15 min delayed phase image series. A lesion that demonstrates a relative washout more than 50% based on a 9 min delayed phase (sensitivity 98%; specificity 100%) [8] or 40% based on a 15 min delayed phase (sensitivity 96%;specificity 100%) [2] being safely characterized as an adenoma. DECT is a relatively recent application in the domain of CT. Dualsource DECT disposes two different X-ray tubes that are functioning at the same time in different kilovoltage settings. On the basis of reconstruction of the two datasets, different kind of information can be obtained. Tissue material composition analysis is feasible, based on the unique CT attenuation properties of different materials at different tube voltage settings. Thus, tissue iodine can be identified and quantified based on the data of the enhanced acquisition. Moreover, based on the same data, a virtual non-contrast (VNC) image series can be reconstructed [9,10]. VNC attenuation values of adrenal lesions could potentially be used to establish the diagnosis of adenoma, thus replacing true unenhanced series allowing a significant radiation dose reduction. There is no unanimity on this topic based on the results of the studies published up to now. According to Gnannt et al. [11] there is no statistically significant difference between attenuation values of incidental adrenal masses between true and virtual unenhanced series. Kim et al. reported that mean attenuation values of the lipid-rich adenomas on early virtual unenhanced CT images were significantly greater than those on true unenhanced [12]. It was the aim of our current study to add supplementary information on this domain. In a different approach, tissue lipid can be identified on DECT, as it is associated with an important reduction of HU at the low kVp series compared with the high kVp series of the DE acquisition. Gupta et al. have published that based on this fact, dual-energy CT can be used to help differentiate some lipid-poor adrenal adenomas from metastatic lesions [13]. In the same direction according to a recently published study, using the reduction in attenuation values between the 140 and 80 kVp series allows to differentiate between adenomas and metastases with areas under the ROC curve reported to be at 0.964 [14]. The objective of this study was to determine whether postprocessing of the data from portal-phase enhanced dual-energy CT (DECT), with or without the addition of a late enhanced phase acquisition, may enable characterization of adrenal lesions without the need for acquisition of pre-contrast images.
2. Materials and methods 2.1. Patients This retrospective study was approved by the institutional review board (IRB) of our hospital. Patient informed consent was waived by the IRB. The examinations of 22 consecutive patients that underwent dual-energy CT for characterization of adrenal lesions from September 2010 through March 2012 were located and retrieved by using the information system of our hospital.
1747
Indications for CT for the 22 patients were: clinically suspected adrenal lesion (n = 4), adrenal lesion found on ultrasound (n = 5), adrenal lesion found on a non dedicated CT that was not diagnostic (n = 13). One patient was excluded due to important motion artifacts on the venous phase acquisition. The remaining 21 patients formed the final population of the study. 2.2. Imaging protocol All patients were scanned on a dual source dual-energy CT (Flash Definition, Siemens Medical Solutions, Forchheim, Germany). At our institution the diagnostic protocol for characterization of adrenal masses consists of an unenhanced series covering the superior abdomen and two enhanced series, 60 s and a 15 min after i.v. injection of 1.5 mL/kg of iohexol 350 mg/mL (Accupaque 350, GE Healthcare, little Chafford, UK) at a rate of 3.5 mL/s, covering the abdomen and pelvis and the superior abdomen respectively. This protocol had been used for the CT examinations of our study. All three acquisitions were performed on a dual-energy mode. The imaging parameters were as follows: gantry rotation time 0.5 s, pitch 0.7 for the unenhanced and 0.6 for the 2 enhanced acquisitions, detector configuration 32 mm × 2 mm × 0.6 mm, reconstruction slice thickness 2 mm, reference tube current 250 mAs for the tube functioning at 100 kVp and automatically chosen for the tube functioning at 140 mm kVp, with 4D dose modulation (4D care dose, Siemens medical solutions, Forchheim, Germany). Radiation dose was calculated for each acquisition separately and was expressed in mSv. These values were estimated by multiplying the total dose-length product provided by the CT console for each acquisition, by a normalizing coefficient of 0.015. 2.3. Image analysis The image analysis was performed on a separate workstation (Syngo, Siemens Medical Solutions, Erlangen, Germany) by two board certified radiologists with 11 and 5 years experience in abdominal radiology respectively, in consensus. An adrenal lesion was defined as a focal enlargement of the adrenal gland with a diameter of more than 1 cm on the shortest transverse axis. The lesion’s size was measured as its greatest diameter on the axial plane. The mean attenuation values of the adrenal lesions on the 100 kVp, the 140 kVp and the “120 kVp weighted average” image series for all three acquisitions were measured. The mean attenuation of the spleen and the extra-abdominal fat were also determined for all patients in the 100 and 140 kVp series of the unenhanced acquisition. The lesions’ mean attenuation values were also measured on the VNC series calculated from the data of the portal (VNCp) and of the delayed (VNCd) acquisitions, using the “liver VNC” software of the Syngo workstation. Finally mean iodine concentration for each lesion was determined using the same software. Two round or oval regions of interest (ROI) were placed in each adrenal lesion and the mean of the 2 measurements was defined as the mean attenuation or the mean iodine concentration respectively. The ROIs were the largest possible, so as to remain in the contours of the lesion, and to avoid any calcification, blood vessel or evident necrosis, ideally covering more than two-thirds of the lesion (ROI range 0.4–2.9 cm2 ). In the same manner in the spleen and fat, ROI were placed in a way to include the maximum area coverage avoiding vessels, calcifications or areas of fat infiltration. The true absolute washout (washout TNC) was calculated by using the equation: washout TNC = [(E − D)/(E − UTNC )] × 100, where E is the enhanced attenuation value measured from the weighted average 120 kVp image series of the venous phase, D
1748
D. Botsikas et al. / European Journal of Radiology 83 (2014) 1746–1751
is the delayed enhanced value measured from the corresponding weighted average 120 kVp image series, and UTNC is the unenhanced value as calculated from the corresponding 120 kVp weighted average series. This equation issues by adapting the known equation described in the literature [5]. The virtual absolute enhancement was calculated based on the same mathematical equation, in 2 different ways: Absolute washout from VNCp (washout VNCp) = [(E − D)/(E − UVNCp )] × 100, where UVNCp is the attenuation calculated from the VNC series issuing from the portal phase acquisition; absolute washout from VNCd (washout VNCd) = [(E − D)/(E − UVNCd )] × 100 where UVNCd is the attenuation calculated from the VNC series issuing from the delayed phase acquisition. Finally the washout based on the iodine concentration was calculated based on the equation; washout CIodine = [(CIodine p − CIodine d)/(CIodine p − CIodine u)] × 100, where CIodine p and “CIodine d” is the iodine concentration calculated from the portal and delayed acquisitions, respectively while “CIodine u” is the iodine concentration from the unenhanced series that is always equal to 0.
Fig. 1. Attenuation values for all lesions measured from true unenhanced (TNC) virtual unenhanced calculated from portal (VNCp) and from delayed (VNCd) phase acquisition.
2.4. Reference standard Adrenal lesions were considered as adenomas if at least one of the following criteria were fulfilled: mean CT attenuation values of 0 HU or less in true unenhanced series; no change for at least 6 months; diagnosis confirmed by histological analysis. Adenomas with unenhanced attenuation values less than +10 HU were characterized as lipid-rich adenomas and those with attenuation values of more than +10 HU as lipid-poor adenomas. 2.5. Statistical analysis An a priori power analysis based on T-test statistics with a power of 80% estimated that a total number of 21 patients should be included for the purpose of the study. All values were tested for Gaussian distribution with Kolmogorov–Smirnov analysis. Normally distributed values were compared using one-way ANOVA with Dunnett’s multiple comparisons test. Non-normally distributed datasets were compared with Friedman test with Dunn’s multiple comparison test post hoc. For statistical analysis we used the software Graphpad Prism 6® (Graphpad, CA, USA). A value of p < 0.05 was considered as statistically significant. 3. Results The final diagnosis was made through pathological analysis for four non-adenomatous lesions and for one lipid poor adenoma. Fourteen lesions were considered as adenomas as they were stable in imaging over a period for at least 6 months (18.5 ± 12.7 months). In 9 of these lesions the diagnosis of adenoma was further supported by true unenhanced densities between 10 and −10. In 5 lesions the diagnosis was made based on a true unenhanced density between 0 and −10 HU. Of all 21 enrolled patients 19(12 men, 7 women; mean weight 78.51 ± 3.53, range 63/103 kg) had one or more adrenal tumors. Two of them had 4 non-adenomatous lesions (1 patient with 2 pheocromocytomas and one patient with 1 adrenal metastasis on one side and adrenal hyperplasia on the other side). 17 patients had a total of 20 adrenal adenomas. Three patients had 2 adenomas (three on the right and 3 on the left) and 14 had one adenoma (8 on the right and 6 on the left). Six of these adenomas were lipid poor and 14 were lipid rich. The mean transverse diameter of all lesions was 2.65 ± 1.59 cm. The mean transverse diameter of the
20 adenomas was 2.42 ± 0.72 cm. (2.4 ± 0.9 cm for the lipid poor and 2.43 ± 0.69 cm for the lipid rich). The mean attenuation values of all lesions at TNC image series was 8.58 ± 15.2 HU at VNCp series 12.6 ± 14.9 HU and at VNCd series 8.51 ± 13.6 HU. The mean attenuation values of all lesions, all adenomas, lipid-poor and lipid-rich adenomas are shown on Table 1. For all but one lesions, VNCp values were higher than TNC ones as shown in Fig. 1. There was a statistically significant difference between attenuation values at VNCp and TNC (p = 0.020) but not between VNCd and TNC (p = 0.9974). The mean value of “washout TNC” for all lesions was 67.5 ± 21.5% (73.8 ± 17.5% for adenomas and 44.2 ± 22.2 for the non-adenomatous lesions). The mean values of “washout VNCp”, “washout VNCd” washout “CIodine ” for all lesions were, 75.4 ± 27.3%, 67.7 ± 22.7% and 69.3 ± 24.6%, respectively. The mean values of “washout VNCp”, “washout VNCd” washout “CIodine ” for adenomas were, 83.3 ± 23.8%, 73.8 ± 20.1% and 76 ± 21.8%, respectively. The mean value of “washout VNCp” for all lesions was statistically different than “washout TNC” (p = 0.0304) There was no significant difference between “washout VNCd” and washout “CIodine values compared to “washout TNC” (p > 0.999) (Fig. 2). The attenuation of adenomas decreased as the tube voltage settings changed from 140 to 100 kVp with a mean value of change of 3.89 ± 8.43 HU for all adenomas, of 1.32 ± 6.16 HU for lipid poor and 5.3 ± 9.41 HU for lipid rich ones. The mean change of the attenuation of the spleen was 4.89 ± 4.16 HU and of fat 22.87 ± 3.36 HU. The attenuation change of the 4 non-adenomatous lesions was −3.88 ± 3.97. The detailed attenuation values of adenomas, of the spleen and of the fat is shown on Table 2. The average radiation dose for the non-enhanced phase was 5.63 ± 2.03 mSv, for the portal phase 7.89 ± 3.94 mSv and for the delayed enhanced phase 4.79 ± 2.38 mSv. 4. Discussion Incidentally discovered adrenal lesions are a frequent finding on enhanced abdominal CT performed for another clinical indication [2,3]. The established protocol for adrenal lesion characterization includes an unenhanced acquisition, and a lesion with a density lower than 10 HU can be safely characterized as an adrenal adenoma. Though, when the adrenal lesion is incidentally found on an enhanced CT, and no unenhanced acquisition is available, the patient has to be recalled to perform a dedicated CT examination.
D. Botsikas et al. / European Journal of Radiology 83 (2014) 1746–1751
1749
Table 1 Attenuation values of all lesions, all adenomas, lipid-rich and lipid poor adenomas and non-adenomatous lesions calculated on TNC, VNCp and VNCd. Attenuation values of all lesions calculated from TNC, VNCp and VNCd Attenuation (HU)
All lesions
All adenomas
Lipid-rich adenomas
Lipid-poor adenomas
Nonadenomatous lesions
TNC VNCp VNCd
8.58 ± 15.2 12.6 ± 14.9 8.51 ± 13.6
6.8 ± 15.2 11.2 ± 14.8 6.07 ± 12.7
−2.5 ± 5.6 2.23 ± 7.8 −1.25 ± 6.24
25.9 ± 6.8 28 ± 5.26 21.28 ± 5.88
21.4 ± 12.4 24.75 ± 12.42 23.5 ± 10.89
Table 2 Attenuation values for all adenomas, lipid poor adenomas, lipid rich adenomas, non-adenomatous lesions, spleen and fat measured from the 140 and 100 kVp series and the respective attenuation change between these series. Attenuation values of all lesions and different tissues at 140 and l00 kVp All adenomas
Lipid-rich adenomas
Lipid-poor adenomas
Non adenomatous Lesions
Spleen
Fat
140 kVp 1OO kVp Attenuation Change
11.01 ± 13.33 7.12 ± 16.81 3.89 ± 8.43
2.39 ± 4.86 −2.91 ± 11.07 5.3 ± 9.41
26.83 ± 7.62 25.51 ± 5.71 1.32 ± 6.16
24.17 ± 9.35 28.05 ± 11.27 −3.88 ± 3.97
54.11 ± 3.03 49.22 ± 4.14 4.89 ± 4.16
−88.3 ± 6.79 −111.17 ± 7.80 22.87 ± 3.36
p Values All adenomas Lipid-rich adenomas Lipid-poor adenomas Spleen Fat
N/A >0.9999 >0.9999 >0.9999 <0.0001
>0.9999 N/A >0.9999 >0.9999 0.0051
>0.9999 >0.9999 N/A >0.9999 0.0010
>0.9999 0.7288 >0.9999 0.7850 <0.0001
>0.9999 >0.9999 >0.9999 N/A 0.0003
<0.0001 0.0007 0.0002 0.0003 N/A
Since its introduction into clinical radiology in 2006 [15] DECT has been used for a continuously growing spectrum of clinical applications in different domains of radiology, mainly thoracic, abdominal, vascular, and, especially, genitourinary radiology [16–19]. It allows differentiating materials, as for example, urinary tract calculi according to their composition. Detection and characterization of calculi even through opacified urine is also possible with DECT-urography [20–22] Another application is evaluation of enhancement of hyperdense renal lesions and thus differentiation of dense cystic lesions from solid tumors on the basis of an enhanced acquisition, without the need of an unenhanced series [23–25].
Fig. 2. Washout values for all lesions measured from true unenhanced (washout TNC) virtual unenhanced calculated from portal (washout VNCp), from delayed (washout VNCd) phase acquisition and from Iodine concentration values (washout CIodine ).
In incidental adrenal masses, the possibility of creating virtual unenhanced image series based on the data of an enhanced acquisition potentially offers the opportunity to characterize a lesion without the need to recall the patient for additional imaging. In a recent study, Gnannt et al. [11] in an attempt to address this question, found that there was no significant difference in incidental adrenal mass attenuation between unenhanced and virtual unenhanced images calculated from portal phase. According to this study the mean values of adrenal lesions attenuation on the unenhanced images was 5.9 ± 21.0 HU and on virtual unenhanced images was 7.0 ± 20.6 HU, with a mean change in attenuation of 1.1 HU. Kim et al. found that mean attenuation values of the lipidrich adenomas on early virtual unenhanced (corresponding to VNCp in the present study) CT images (11.7 HU ± 9.5) were significantly greater than those on unenhanced CT images (0.7 HU ± 7.2) (p = 0.001) and delayed virtual unenhanced (corresponding to VNCd in the present study) CT images (6.6 HU ± 8.4) (p = 0.01) [12]. Ho et al. found that there was no significant difference between densities measured from virtual and true unenhanced images, but still in this series the mean density measured from virtual unenhanced was higher than the one from true unenhanced series (14.7 ± 15.1 and 12.9 ± 13.4 HU, respectively) (p = 0.2) [26]. Virtual unenhanced images acquired with DECT with rapid kilovoltage switching technique (RSDE) has also been investigated. The difference in this case is that in contrast to dual-source DECT that provides a HU measurement for tissues in the VNC image series, RSDE CT calculates tissue densities in terms of mg/ml of the material being investigated. According to Morgan et al., based on this technique, many of the incidentally found adrenal lesions on a RSDE CT for other clinical indications can be accurately characterized using post contrast material decomposition, aiming to lipid quantification [27]. The results of our study indicate that mean attenuation of all adrenal lesions measured on the virtual unenhanced images calculated from the dual energy data of the portal image series is 12.6 ± 14.9 HU and thus significantly different from the attenuation measured on the true unenhanced images (8.58 ± 15.2 HU), with a mean difference of 4.02 HU between the TNC and the VNCp series. This difference is not only statistically significant (p = 0.020) but
1750
D. Botsikas et al. / European Journal of Radiology 83 (2014) 1746–1751
Fig. 3. A 64-year-old male patient with a lipid rich left adrenal adenoma on DECT. True unenhanced (a), portal phase enhanced (b), virtual unenhanced (c) and iodine map overlay (d) images show the adenoma (white arrow in a, b, c, d). Note on (d) the ROI with the information provided from the processing of the data of the DE acquisition, as iodine concentration and the VNC attenuation. Pathologic specimen (e) shows the adenoma after resection (white arrow).
can influence the characterization of an adrenal nodule, as the cutoff of 10 HU that is established for the characterization of a lesion as lipid rich adenoma can be easily overpassed, thus changing the final diagnosis. On the other hand, the mean difference of the lesions’ attenuation between the TNC and VNCd series in our study was 0.07 HU. This difference is not statistically significant, so it can be considered safe enough to use the data from this series in order to characterize an adrenal nodule based on its unenhanced attenuation. The data from our study and the observations of other investigators [11,12,26] show a slight tendency to overestimate the attenuation values on the VNC images calculated from the portal phase. Particularly in our study, VNCp attenuation values of all but one lesions were higher than TNC ones (Fig. 1). The only lesion for which VNCp value (−15 HU) was lower than TNC value (−11 HU) was a lipid rich adenoma, with no change of the diagnosis in this case. So, on everyday clinical practice an incidental adrenal lesion detected on one DE portal acquisition only, can be evaluated on the VNC series calculated from this enhanced acquisition. If its attenuation value on the VNC series is less than 10 HU it could safely be characterized as a lipid rich adenoma, thus avoiding further workup. In our study, 10 patients had 13 lesions that had VNCp attenuation values lower than 10 HU. These 10 patients would have avoided further work up. This means that there would be no need to repeat a late phase acquisition at 15 min. This would allow to spare time and radiation dose (4.79 ± 2.38 mSv.) for the patient. In the same manner, there would be no need to recall any of these patients for a dedicated CT for adrenal characterization. This would allow sparing between 5.63 and 18.31 mSv, depending on the need of a single unenhanced series or repetition of a 3 phases enhanced CT. On the contrary, if a lesion has a VNC attenuation greater than 10 HU, a further workup with a delayed phase acquisition is considered necessary. This latter will allow not only the calculation of its VNCd attenuation which can be considered practically identical to the TNC attenuation value, but also the calculation of its absolute washout values. In current practice, in a CT protocol tailored to characterize adrenal lesions, when an adrenal lesion has a density greater than
10 HU on unenhanced images, it can be either a lipid poor adenoma, or another type of lesion, usually malignant. In this case, the workout will continue with a 15 min delayed acquisition, in order to calculate the absolute washout percentage [4]. If the CT was performed with another indication, in most cases an unenhanced series is not available. The characterization of the adrenal lesion is still feasible in this case, by calculating the relative washout [2]. According to the results of our study, the absolute washout values calculated from the data of the VNCp are significantly higher than those calculated from the true unenhanced series There was no statistically significant difference between “washout VNCd” and “washout TNC”. In the same manner, washout calculated from the Iodine concentration is not statistically different from washout calculated from true unenhanced series and could also be implemented on the characterization of adrenal nodules (Fig. 3). In the everyday clinical practice, after having treated the data of our first patients, we decided not to perform a true unenhanced series for the purpose of adrenal lesions characterization for all patients that were scanned with the dual-source DECT scanner of our department. A resident radiologist was assigned the task to treat DECT data with the dedicated software and calculate VNCp attenuation values for any adrenal lesion detected. If this value is higher than 10 HU, he repeats a 15 min late phase acquisition. If the VNCp attenuation is lower than 10 HU the lesion is classified as an adenoma and no further workout is needed. From another perspective, the utility of DECT in the characterization of adrenal adenomas has been examined by Gupta et al. [13] and Shi et al. [14] by studying attenuation change of adenomas between the series of different kilovoltages. Gupta et al. found that the mean attenuation change of all adenomas was 0.4 ± 7.1 HU and that of metastatic lesions 9.2 ± 4.3 HU while the change for the spleen and fat was found to be 7.4 ± 3.1 and −26.6 ± 2.7 HU. Based on their data, Gupta et al. concluded that DECT can be used to help differentiate some lipid-poor adrenal adenomas from metastatic lesions [13]. Our data show a mean attenuation change of all adenomas of 3.89 ± 8.43 HU between the series of high and low kilovoltage. The mean attenuation change of non-adenomatous lesions
D. Botsikas et al. / European Journal of Radiology 83 (2014) 1746–1751
was −3.88 ± 3.97 HU. In our opinion we cannot compare attenuation change of adenomas to that of non-adenomatous lesions as we found only 4 non-adenomatous lesions, two of which were pheochromocytomas that are known to be as “great mimickers” in imaging. For this reason we used the attenuation change of the spleen as an approximate equivalent of a, lesion, with minimal fat content. The spleen was used for this reason, because its fat content is minimal, about 3%, and splenic lipidosis is a very rare condition that has been described post-mortem in 12 out of 1500 autopsies and only in patients that had received intravenous lipid therapy [28] The mean attenuation change for the spleen in our study was 4.89 ± 4.16 and was not significantly different from adrenal adenomas (p > 0.9999). These results are not concordant with Gupta’s. These differences can be explained by the small number of subjects in both studies. The fact that the two studies were performed with different technical equipment (dual-source versus single source DECT) and different voltage parameters (we used 100 and 140 kVp respectively for the two tubes, while Gupta et al. performed the acquisition with the single tube alternating between 80 and 140 kVp) can also offer a possible explanation to this discrepancy. On the same topic, Shi et al. [14] found that the change of densities of all adenomas (including lipid-poor and lipid-rich) were significantly higher than those of metastases (p < 0.001). In our opinion research should be continued in this direction in order to prove the utility of measuring attenuation changes of adrenal lesions to their characterization. Most importantly, the acquisition parameters and technique should be strictly standardized, in order to set cut-off values of density changes for characterization purposes and provide reproducible measurements. Another concern about tissue characterization in general and subsequently adrenal lesion characterization with CT is that two independent studies have reported that different single detector and MDCT helical scanners produce slightly different but statistically significant unenhanced attenuation levels, which could lead to potentially erroneous categorization. This could potentially be considered as a potential pitfall with DECT as well [29,30]. Limitations of this study are its retrospective nature and the relatively small patient population, including 19 patients instead of 21 as indicated by the a priori statistical analysis. The consecutive patients selection minimizes selection bias related to the retrospective nature of the study. Another limiting factor is the fact that there were only 4 non-adenomatous lesions, 2 of which were pheochromocytomas. Thus it was not possible to compare adenomas with those lesions. For the purpose of comparison between attenuation values and washout values measured at TNC, VNCp and VNCd series, all lesions were included. In conclusion, according to the results of this study, incidental adrenal lesions can be characterized as adenomas, if their attenuation values based on the virtual unenhanced images calculated from the portal phase DECT acquisition are lower than 10 HU. If this value is ≥10 HU, it is recommended to add a 15 min delayed phase. This will allow calculation of virtual unenhanced attenuation from the data of the delayed acquisition which correlates better with the true unenhanced attenuation, as well as calculation of washout values (“washout VNCd” and washout “CIodine ”) that are found to correlate well with the washout calculated from a true unenhanced series.
Conflict of interest All authors declare no conflict of interest.
1751
References [1] Bovio S, Cataldi A, Reimondo G, Sperone P, Novello S, Berruti A, et al. Prevalence of adrenal incidentaloma in a contemporary computerized tomography series. J Endocrinol Invest 2006;29(4):298–302. [2] Caoili EM, Korobkin M, Francis IR, Cohan RH, Platt JF, Dunnick NR, et al. Adrenal masses: characterization with combined unenhanced and delayed enhanced CT. Radiology 2002;222(3):629–33. [3] Musella M, Conzo G, Milone MM, Corcione F, Belli G, De Palma M, et al. Preoperative workup in the assessment of adrenal incidentalomas: outcome from 282 consecutive laparoscopic adrenalectomies. BMC Surg 2013;13(1):57. [4] Dunnick NR, Korobkin MM. Imaging of adrenal incidentalomas: current status. AJR Am J Roentgenol 2002;179(3):559–68. [5] Blake MA, Cronin CG, Boland GW. Adrenal imaging. AJR Am J Roentgenol 2010;194(6):1450–60. [6] Boland GW, Lee MJ, Gazelle GS, Halpern EF, McNicholas MM, Mueller PR. Characterization of adrenal masses using unenhanced CT: an analysis of the CT literature. AJR Am J Roentgenol 1998;171(1):201–4. [7] Korobkin M, Brodeur FJ, Francis IR, Quint LE, Dunnick NR, Londy F. CT timeattenuation washout curves of adrenal adenomas and nonadenomas. AJR Am J Roentgenol 1998;170(3):747–52. [8] PeÒa CS, Boland GW, Hahn PF, Lee MJ, Mueller PR. Characterization of indeterminate (lipid-poor) adrenal masses: use of washout characteristics at contrast-enhanced CT. Radiology 2000;217(3):798–802. [9] Graser A, Johnson TRRC, Chandarana H, Macari M. Dual energy CT: preliminary observations and potential clinical applications in the abdomen. Eur Radiol 2009;19(1):13–23. [10] Coursey CA, Nelson RC, Boll DT, Paulson EK, Ho LM, Neville AM, et al. Dualenergy multidetector CT: how does it work, what can it tell us, and when can we use it in abdominopelvic imaging? Radiographics 2010;30(4):1037–55. [11] Gnannt R, Fischer M, Goetti R, Karlo C, Leschka S, Alkadhi H. Dual-energy CT for characterization of the incidental adrenal mass: preliminary observations. AJR Am J Roentgenol 2012;198(1):138–44. [12] Kim YK, Park BK, Kim CK, Park SY. Adenoma characterization: adrenal protocol with dual-energy CT. Radiology 2013;267(1):155–63. [13] Gupta RT, Ho LM, Marin D, Boll DT, Barnhart HX, Nelson RC. Dual-energy CT for characterization of adrenal nodules: initial experience. AJR Am J Roentgenol 2010;194(6):1479–83. [14] Shi JWJ, Dai HZ, Shen LH, Xu DFD. Dual-energy CT: clinical application in differentiating an adrenal adenoma from a metastasis. Acta Radiol 2013;55(4):505–12. [15] Johnson TRC, Krauss B, Sedlmair M, Grasruck M, Bruder H, Morhard D, et al. Material differentiation by dual energy CT: initial experience. Eur Radiol 2007;17(6):1510–7. [16] Remy Jardin M, Faivre JBJB, Pontana FJB, Remy JJB. Dual energy CT for thoracic applications. Rev Mal Respir 2012;29(10):1268–71. [17] Morgan DE. Dual-energy CT of the abdomen. Abdom Imaging 2014;39(1):108–34. [18] Vlahos I, Chung R, Nair A, Morgan R. Dual-energy CT: vascular applications. AJR Am J Roentgenol 2012;199(5 Suppl.):S87–97. [19] Kaza RK, Platt JF, Cohan RH, Caoili EM, Al Hawary M, Wasnik A. Dual-energy CT with single- and dual-source scanners: current applications in evaluating the genitourinary tract. Radiographics 2012;32(2):353–69. [20] Botsikas D, Hansen C, Stefanelli S, Becker CD, Montet X. Urinary stone detection and characterisation with dual-energy CT urography after furosemide intravenous injection: preliminary results. Eur Radiol 2014;24(3):709–14. [21] Primak AN, Fletcher JG, Vrtiska TJ, Dzyubak OP, Lieske JC, Jackson ME, et al. Noninvasive differentiation of uric acid versus non-uric acid kidney stones using dual-energy CT. Acad Radiol 2007;14(12):1441–7. [22] Stolzmann P, Kozomara M, Chuck N, Müntener M, Leschka S, Scheffel H, et al. In vivo identification of uric acid stones with dual-energy CT: diagnostic performance evaluation in patients. Abdom Imaging 2010;35(5):629–35. [23] Leschka S, Stolzmann P, Baumüller S, Scheffel H, Desbiolles L, Schmid B, et al. Performance of dual-energy CT with tin filter technology for the discrimination of renal cysts and enhancing masses. Acad Radiol 2010;17(4):526–34. [24] Neville AM, Gupta RT, Miller CM, Merkle EM, Paulson EK, Boll DT. Detection of renal lesion enhancement with dual-energy multidetector CT. Radiology 2011;259(1):173–83. [25] Arndt N, Staehler M, Siegert S, Reiser MF, Graser A. Dual energy CT in patients with polycystic kidney disease. Eur Radiol 2012;22(10):2125–9. [26] Ho LM, Marin D, Neville AM, Barnhart HX, Gupta RT, Paulson EK, et al. Characterization of adrenal nodules with dual-energy CT: can virtual unenhanced attenuation values replace true unenhanced attenuation values? AJR Am J Roentgenol 2012;198(4):840–5. [27] Morgan DE, Weber AC, Lockhart ME, Weber TM, Fineberg NS, Berland LL. Differentiation of high lipid content from low lipid content adrenal lesions using single-source rapid kilovolt (peak)-switching dual-energy multidetector CT. J Comput Assist Tomogr 2013;37(6):937–43. [28] Forbes GB. Splenic lipidosis after administration of intravenous fat emulsions. J Clin Pathol 1978;31(8):765–71. [29] Hahn PF, Blake MA, Boland GW. Adrenal lesions: attenuation measurement differences between CT scanners. Radiology 2006;240(2):458–63. [30] Stadler A, Schima W, Prager G, Homolka P, Heinz G, Saini S, et al. CT density measurements for characterization of adrenal tumors ex vivo: variability among three CT scanners. AJR Am J Roentgenol 2004;182(3):671–5.
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Incidental adrenal lesions detected on enhanced abdominal dual-energy CT: Can the diagnostic workup be shortened by the implementation of virtual unenhanced images? Diomidis Botsikas a,∗ , Frederic Triponez b,1 , Sana Boudabbous a,2 , Catrina Hansen a,3 , Christoph D. Becker a,4 , Xavier Montet a,5 a
Geneva University Hospital, Department of Imaging and Medical Information Sciences, Division of Radiology, Rue Gabrielle-Perret-Gentil 4, 1211 Genève 4, Switzerland b Geneva University Hospital, Department of Surgery, Division of Thoracic Surgery, Rue Gabrielle-Perret-Gentil 4, 1211 Genève 4, Switzerland
a r t i c l e
i n f o
Article history: Received 27 February 2014 Received in revised form 11 June 2014 Accepted 14 June 2014 Keywords: Adrenal incidentalomas Adrenal lesion characterization Dual-energy CT Virtual unenhanced images
a b s t r a c t Objective: To determine whether post-processing of the data from portal-phase enhanced dual-energy CT (DECT), with or without the addition of a late enhanced phase acquisition, may enable characterization of adrenal lesions without the need for acquisition of pre-contrast images. Materials and methods: Twenty-two patients with 24 adrenal lesions underwent unenhanced, venous and delayed phase DECT. Of these lesions, 20 were found to be adrenal adenomas, on the basis of histopathology, unenhanced attenuation values between 0 and −10 HU, or stability over at least 6 months. For all 24 lesions, true and virtual unenhanced attenuation values were measured based on the data of the portal (VNCp) and the delayed (VNCd) DECT acquisition. The absolute washout values based on the true non-contrast (TNC) and the VNCp and VNCd image series were also measured. The washout was also calculated based on the iodine concentration measured from both contrast-enhanced acquisitions. Results: Mean virtual unenhanced attenuation values of all lesions calculated from the portal phase images was 12.6 HU, and was 4.02 HU higher than the values based on true unenhanced images (p = 0.020). Washout values calculated from virtual unenhanced attenuation based on the VNCp were also significantly different (p = 0.0304) while those calculated from VNCd and from iodine concentration correlated with the corresponding values based on the true unenhanced values (p > 0.999). Conclusions: Our data indicate that attenuation values of adrenal adenomas based on virtual unenhanced images are significantly higher than those obtained with true unenhanced images. An incidental adrenal lesion with a virtual unenhanced attenuation lower than 10 HU can thus be safely characterized as an adenoma. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Incidental adrenal masses occur in up to 5% of CT examinations [1–3]. Most of these lesions are benign, adrenal adenomas being the
∗ Corresponding author. Tel.: +41 79 56 30 778; fax: +41 22 37 27 072. E-mail addresses: [email protected] (D. Botsikas), [email protected] (F. Triponez), [email protected] (S. Boudabbous), [email protected] (C. Hansen), [email protected] (C.D. Becker), [email protected] (X. Montet). 1 Tel.: +41 79 55 32 152. 2 Tel.: +41 79 55 32 444; fax: +41 22 37 27 072. 3 Tel.: +41 79 55 32 490; fax: +41 22 37 27 072. 4 Tel.: +41 22 37 27 001; fax: +41 22 37 27 072. 5 Tel.: +41 79 55 32 485; fax: +41 22 37 27 072. http://dx.doi.org/10.1016/j.ejrad.2014.06.017 0720-048X/© 2014 Elsevier Ireland Ltd. All rights reserved.
most common diagnosis, even in oncologic patients [2]. Combined enhanced and unenhanced computed tomography is established as the method of choice for characterization of adrenal lesions [2,4]. The currently accepted CT protocol for adrenal mass characterization consists of a three phase acquisition, including an unenhanced a venous and a delayed phase [4]. The values examined for this purpose are the unenhanced attenuation and the absolute washout values of adrenal lesions. Unenhanced CT attenuations allow characterizing lipid rich adenomas. The ideal thresholds for optimal specificity for the detection of adenomas were reported to be 10 HU on unenhanced images. A lesion with attenuation values below this threshold can be considered as a lipid rich adenoma (sensitivity 71%; specificity 98%) [5,6] The calculation of absolute washout values allows characterizing all adenomas (lipid rich and lipid poor). The absolute
D. Botsikas et al. / European Journal of Radiology 83 (2014) 1746–1751
washout value is calculated with the mathematical type “Absolute Washout = [(E − D)/(E − U)] × 100”, where E is the enhanced attenuation value measured from venous phase image series, D is the delayed enhanced attenuation value measured from the 15 min delayed phase image series, and U is the unenhanced value. A threshold of 60% (sensitivity 86–88%; specificity 92–96%) [2,5,7] is reported to be the ideal compromise for this purpose. If unenhanced CT is not available the diagnosis is based on the delayed enhanced phase, with calculation of relative washout values from the mathematical type “Relative Washout = (E − D)/E × 100” where E is the enhanced attenuation value measured from venous phase image series, D is the delayed enhanced attenuation value measured from the 15 min delayed phase image series. A lesion that demonstrates a relative washout more than 50% based on a 9 min delayed phase (sensitivity 98%; specificity 100%) [8] or 40% based on a 15 min delayed phase (sensitivity 96%;specificity 100%) [2] being safely characterized as an adenoma. DECT is a relatively recent application in the domain of CT. Dualsource DECT disposes two different X-ray tubes that are functioning at the same time in different kilovoltage settings. On the basis of reconstruction of the two datasets, different kind of information can be obtained. Tissue material composition analysis is feasible, based on the unique CT attenuation properties of different materials at different tube voltage settings. Thus, tissue iodine can be identified and quantified based on the data of the enhanced acquisition. Moreover, based on the same data, a virtual non-contrast (VNC) image series can be reconstructed [9,10]. VNC attenuation values of adrenal lesions could potentially be used to establish the diagnosis of adenoma, thus replacing true unenhanced series allowing a significant radiation dose reduction. There is no unanimity on this topic based on the results of the studies published up to now. According to Gnannt et al. [11] there is no statistically significant difference between attenuation values of incidental adrenal masses between true and virtual unenhanced series. Kim et al. reported that mean attenuation values of the lipid-rich adenomas on early virtual unenhanced CT images were significantly greater than those on true unenhanced [12]. It was the aim of our current study to add supplementary information on this domain. In a different approach, tissue lipid can be identified on DECT, as it is associated with an important reduction of HU at the low kVp series compared with the high kVp series of the DE acquisition. Gupta et al. have published that based on this fact, dual-energy CT can be used to help differentiate some lipid-poor adrenal adenomas from metastatic lesions [13]. In the same direction according to a recently published study, using the reduction in attenuation values between the 140 and 80 kVp series allows to differentiate between adenomas and metastases with areas under the ROC curve reported to be at 0.964 [14]. The objective of this study was to determine whether postprocessing of the data from portal-phase enhanced dual-energy CT (DECT), with or without the addition of a late enhanced phase acquisition, may enable characterization of adrenal lesions without the need for acquisition of pre-contrast images.
2. Materials and methods 2.1. Patients This retrospective study was approved by the institutional review board (IRB) of our hospital. Patient informed consent was waived by the IRB. The examinations of 22 consecutive patients that underwent dual-energy CT for characterization of adrenal lesions from September 2010 through March 2012 were located and retrieved by using the information system of our hospital.
1747
Indications for CT for the 22 patients were: clinically suspected adrenal lesion (n = 4), adrenal lesion found on ultrasound (n = 5), adrenal lesion found on a non dedicated CT that was not diagnostic (n = 13). One patient was excluded due to important motion artifacts on the venous phase acquisition. The remaining 21 patients formed the final population of the study. 2.2. Imaging protocol All patients were scanned on a dual source dual-energy CT (Flash Definition, Siemens Medical Solutions, Forchheim, Germany). At our institution the diagnostic protocol for characterization of adrenal masses consists of an unenhanced series covering the superior abdomen and two enhanced series, 60 s and a 15 min after i.v. injection of 1.5 mL/kg of iohexol 350 mg/mL (Accupaque 350, GE Healthcare, little Chafford, UK) at a rate of 3.5 mL/s, covering the abdomen and pelvis and the superior abdomen respectively. This protocol had been used for the CT examinations of our study. All three acquisitions were performed on a dual-energy mode. The imaging parameters were as follows: gantry rotation time 0.5 s, pitch 0.7 for the unenhanced and 0.6 for the 2 enhanced acquisitions, detector configuration 32 mm × 2 mm × 0.6 mm, reconstruction slice thickness 2 mm, reference tube current 250 mAs for the tube functioning at 100 kVp and automatically chosen for the tube functioning at 140 mm kVp, with 4D dose modulation (4D care dose, Siemens medical solutions, Forchheim, Germany). Radiation dose was calculated for each acquisition separately and was expressed in mSv. These values were estimated by multiplying the total dose-length product provided by the CT console for each acquisition, by a normalizing coefficient of 0.015. 2.3. Image analysis The image analysis was performed on a separate workstation (Syngo, Siemens Medical Solutions, Erlangen, Germany) by two board certified radiologists with 11 and 5 years experience in abdominal radiology respectively, in consensus. An adrenal lesion was defined as a focal enlargement of the adrenal gland with a diameter of more than 1 cm on the shortest transverse axis. The lesion’s size was measured as its greatest diameter on the axial plane. The mean attenuation values of the adrenal lesions on the 100 kVp, the 140 kVp and the “120 kVp weighted average” image series for all three acquisitions were measured. The mean attenuation of the spleen and the extra-abdominal fat were also determined for all patients in the 100 and 140 kVp series of the unenhanced acquisition. The lesions’ mean attenuation values were also measured on the VNC series calculated from the data of the portal (VNCp) and of the delayed (VNCd) acquisitions, using the “liver VNC” software of the Syngo workstation. Finally mean iodine concentration for each lesion was determined using the same software. Two round or oval regions of interest (ROI) were placed in each adrenal lesion and the mean of the 2 measurements was defined as the mean attenuation or the mean iodine concentration respectively. The ROIs were the largest possible, so as to remain in the contours of the lesion, and to avoid any calcification, blood vessel or evident necrosis, ideally covering more than two-thirds of the lesion (ROI range 0.4–2.9 cm2 ). In the same manner in the spleen and fat, ROI were placed in a way to include the maximum area coverage avoiding vessels, calcifications or areas of fat infiltration. The true absolute washout (washout TNC) was calculated by using the equation: washout TNC = [(E − D)/(E − UTNC )] × 100, where E is the enhanced attenuation value measured from the weighted average 120 kVp image series of the venous phase, D
1748
D. Botsikas et al. / European Journal of Radiology 83 (2014) 1746–1751
is the delayed enhanced value measured from the corresponding weighted average 120 kVp image series, and UTNC is the unenhanced value as calculated from the corresponding 120 kVp weighted average series. This equation issues by adapting the known equation described in the literature [5]. The virtual absolute enhancement was calculated based on the same mathematical equation, in 2 different ways: Absolute washout from VNCp (washout VNCp) = [(E − D)/(E − UVNCp )] × 100, where UVNCp is the attenuation calculated from the VNC series issuing from the portal phase acquisition; absolute washout from VNCd (washout VNCd) = [(E − D)/(E − UVNCd )] × 100 where UVNCd is the attenuation calculated from the VNC series issuing from the delayed phase acquisition. Finally the washout based on the iodine concentration was calculated based on the equation; washout CIodine = [(CIodine p − CIodine d)/(CIodine p − CIodine u)] × 100, where CIodine p and “CIodine d” is the iodine concentration calculated from the portal and delayed acquisitions, respectively while “CIodine u” is the iodine concentration from the unenhanced series that is always equal to 0.
Fig. 1. Attenuation values for all lesions measured from true unenhanced (TNC) virtual unenhanced calculated from portal (VNCp) and from delayed (VNCd) phase acquisition.
2.4. Reference standard Adrenal lesions were considered as adenomas if at least one of the following criteria were fulfilled: mean CT attenuation values of 0 HU or less in true unenhanced series; no change for at least 6 months; diagnosis confirmed by histological analysis. Adenomas with unenhanced attenuation values less than +10 HU were characterized as lipid-rich adenomas and those with attenuation values of more than +10 HU as lipid-poor adenomas. 2.5. Statistical analysis An a priori power analysis based on T-test statistics with a power of 80% estimated that a total number of 21 patients should be included for the purpose of the study. All values were tested for Gaussian distribution with Kolmogorov–Smirnov analysis. Normally distributed values were compared using one-way ANOVA with Dunnett’s multiple comparisons test. Non-normally distributed datasets were compared with Friedman test with Dunn’s multiple comparison test post hoc. For statistical analysis we used the software Graphpad Prism 6® (Graphpad, CA, USA). A value of p < 0.05 was considered as statistically significant. 3. Results The final diagnosis was made through pathological analysis for four non-adenomatous lesions and for one lipid poor adenoma. Fourteen lesions were considered as adenomas as they were stable in imaging over a period for at least 6 months (18.5 ± 12.7 months). In 9 of these lesions the diagnosis of adenoma was further supported by true unenhanced densities between 10 and −10. In 5 lesions the diagnosis was made based on a true unenhanced density between 0 and −10 HU. Of all 21 enrolled patients 19(12 men, 7 women; mean weight 78.51 ± 3.53, range 63/103 kg) had one or more adrenal tumors. Two of them had 4 non-adenomatous lesions (1 patient with 2 pheocromocytomas and one patient with 1 adrenal metastasis on one side and adrenal hyperplasia on the other side). 17 patients had a total of 20 adrenal adenomas. Three patients had 2 adenomas (three on the right and 3 on the left) and 14 had one adenoma (8 on the right and 6 on the left). Six of these adenomas were lipid poor and 14 were lipid rich. The mean transverse diameter of all lesions was 2.65 ± 1.59 cm. The mean transverse diameter of the
20 adenomas was 2.42 ± 0.72 cm. (2.4 ± 0.9 cm for the lipid poor and 2.43 ± 0.69 cm for the lipid rich). The mean attenuation values of all lesions at TNC image series was 8.58 ± 15.2 HU at VNCp series 12.6 ± 14.9 HU and at VNCd series 8.51 ± 13.6 HU. The mean attenuation values of all lesions, all adenomas, lipid-poor and lipid-rich adenomas are shown on Table 1. For all but one lesions, VNCp values were higher than TNC ones as shown in Fig. 1. There was a statistically significant difference between attenuation values at VNCp and TNC (p = 0.020) but not between VNCd and TNC (p = 0.9974). The mean value of “washout TNC” for all lesions was 67.5 ± 21.5% (73.8 ± 17.5% for adenomas and 44.2 ± 22.2 for the non-adenomatous lesions). The mean values of “washout VNCp”, “washout VNCd” washout “CIodine ” for all lesions were, 75.4 ± 27.3%, 67.7 ± 22.7% and 69.3 ± 24.6%, respectively. The mean values of “washout VNCp”, “washout VNCd” washout “CIodine ” for adenomas were, 83.3 ± 23.8%, 73.8 ± 20.1% and 76 ± 21.8%, respectively. The mean value of “washout VNCp” for all lesions was statistically different than “washout TNC” (p = 0.0304) There was no significant difference between “washout VNCd” and washout “CIodine values compared to “washout TNC” (p > 0.999) (Fig. 2). The attenuation of adenomas decreased as the tube voltage settings changed from 140 to 100 kVp with a mean value of change of 3.89 ± 8.43 HU for all adenomas, of 1.32 ± 6.16 HU for lipid poor and 5.3 ± 9.41 HU for lipid rich ones. The mean change of the attenuation of the spleen was 4.89 ± 4.16 HU and of fat 22.87 ± 3.36 HU. The attenuation change of the 4 non-adenomatous lesions was −3.88 ± 3.97. The detailed attenuation values of adenomas, of the spleen and of the fat is shown on Table 2. The average radiation dose for the non-enhanced phase was 5.63 ± 2.03 mSv, for the portal phase 7.89 ± 3.94 mSv and for the delayed enhanced phase 4.79 ± 2.38 mSv. 4. Discussion Incidentally discovered adrenal lesions are a frequent finding on enhanced abdominal CT performed for another clinical indication [2,3]. The established protocol for adrenal lesion characterization includes an unenhanced acquisition, and a lesion with a density lower than 10 HU can be safely characterized as an adrenal adenoma. Though, when the adrenal lesion is incidentally found on an enhanced CT, and no unenhanced acquisition is available, the patient has to be recalled to perform a dedicated CT examination.
D. Botsikas et al. / European Journal of Radiology 83 (2014) 1746–1751
1749
Table 1 Attenuation values of all lesions, all adenomas, lipid-rich and lipid poor adenomas and non-adenomatous lesions calculated on TNC, VNCp and VNCd. Attenuation values of all lesions calculated from TNC, VNCp and VNCd Attenuation (HU)
All lesions
All adenomas
Lipid-rich adenomas
Lipid-poor adenomas
Nonadenomatous lesions
TNC VNCp VNCd
8.58 ± 15.2 12.6 ± 14.9 8.51 ± 13.6
6.8 ± 15.2 11.2 ± 14.8 6.07 ± 12.7
−2.5 ± 5.6 2.23 ± 7.8 −1.25 ± 6.24
25.9 ± 6.8 28 ± 5.26 21.28 ± 5.88
21.4 ± 12.4 24.75 ± 12.42 23.5 ± 10.89
Table 2 Attenuation values for all adenomas, lipid poor adenomas, lipid rich adenomas, non-adenomatous lesions, spleen and fat measured from the 140 and 100 kVp series and the respective attenuation change between these series. Attenuation values of all lesions and different tissues at 140 and l00 kVp All adenomas
Lipid-rich adenomas
Lipid-poor adenomas
Non adenomatous Lesions
Spleen
Fat
140 kVp 1OO kVp Attenuation Change
11.01 ± 13.33 7.12 ± 16.81 3.89 ± 8.43
2.39 ± 4.86 −2.91 ± 11.07 5.3 ± 9.41
26.83 ± 7.62 25.51 ± 5.71 1.32 ± 6.16
24.17 ± 9.35 28.05 ± 11.27 −3.88 ± 3.97
54.11 ± 3.03 49.22 ± 4.14 4.89 ± 4.16
−88.3 ± 6.79 −111.17 ± 7.80 22.87 ± 3.36
p Values All adenomas Lipid-rich adenomas Lipid-poor adenomas Spleen Fat
N/A >0.9999 >0.9999 >0.9999 <0.0001
>0.9999 N/A >0.9999 >0.9999 0.0051
>0.9999 >0.9999 N/A >0.9999 0.0010
>0.9999 0.7288 >0.9999 0.7850 <0.0001
>0.9999 >0.9999 >0.9999 N/A 0.0003
<0.0001 0.0007 0.0002 0.0003 N/A
Since its introduction into clinical radiology in 2006 [15] DECT has been used for a continuously growing spectrum of clinical applications in different domains of radiology, mainly thoracic, abdominal, vascular, and, especially, genitourinary radiology [16–19]. It allows differentiating materials, as for example, urinary tract calculi according to their composition. Detection and characterization of calculi even through opacified urine is also possible with DECT-urography [20–22] Another application is evaluation of enhancement of hyperdense renal lesions and thus differentiation of dense cystic lesions from solid tumors on the basis of an enhanced acquisition, without the need of an unenhanced series [23–25].
Fig. 2. Washout values for all lesions measured from true unenhanced (washout TNC) virtual unenhanced calculated from portal (washout VNCp), from delayed (washout VNCd) phase acquisition and from Iodine concentration values (washout CIodine ).
In incidental adrenal masses, the possibility of creating virtual unenhanced image series based on the data of an enhanced acquisition potentially offers the opportunity to characterize a lesion without the need to recall the patient for additional imaging. In a recent study, Gnannt et al. [11] in an attempt to address this question, found that there was no significant difference in incidental adrenal mass attenuation between unenhanced and virtual unenhanced images calculated from portal phase. According to this study the mean values of adrenal lesions attenuation on the unenhanced images was 5.9 ± 21.0 HU and on virtual unenhanced images was 7.0 ± 20.6 HU, with a mean change in attenuation of 1.1 HU. Kim et al. found that mean attenuation values of the lipidrich adenomas on early virtual unenhanced (corresponding to VNCp in the present study) CT images (11.7 HU ± 9.5) were significantly greater than those on unenhanced CT images (0.7 HU ± 7.2) (p = 0.001) and delayed virtual unenhanced (corresponding to VNCd in the present study) CT images (6.6 HU ± 8.4) (p = 0.01) [12]. Ho et al. found that there was no significant difference between densities measured from virtual and true unenhanced images, but still in this series the mean density measured from virtual unenhanced was higher than the one from true unenhanced series (14.7 ± 15.1 and 12.9 ± 13.4 HU, respectively) (p = 0.2) [26]. Virtual unenhanced images acquired with DECT with rapid kilovoltage switching technique (RSDE) has also been investigated. The difference in this case is that in contrast to dual-source DECT that provides a HU measurement for tissues in the VNC image series, RSDE CT calculates tissue densities in terms of mg/ml of the material being investigated. According to Morgan et al., based on this technique, many of the incidentally found adrenal lesions on a RSDE CT for other clinical indications can be accurately characterized using post contrast material decomposition, aiming to lipid quantification [27]. The results of our study indicate that mean attenuation of all adrenal lesions measured on the virtual unenhanced images calculated from the dual energy data of the portal image series is 12.6 ± 14.9 HU and thus significantly different from the attenuation measured on the true unenhanced images (8.58 ± 15.2 HU), with a mean difference of 4.02 HU between the TNC and the VNCp series. This difference is not only statistically significant (p = 0.020) but
1750
D. Botsikas et al. / European Journal of Radiology 83 (2014) 1746–1751
Fig. 3. A 64-year-old male patient with a lipid rich left adrenal adenoma on DECT. True unenhanced (a), portal phase enhanced (b), virtual unenhanced (c) and iodine map overlay (d) images show the adenoma (white arrow in a, b, c, d). Note on (d) the ROI with the information provided from the processing of the data of the DE acquisition, as iodine concentration and the VNC attenuation. Pathologic specimen (e) shows the adenoma after resection (white arrow).
can influence the characterization of an adrenal nodule, as the cutoff of 10 HU that is established for the characterization of a lesion as lipid rich adenoma can be easily overpassed, thus changing the final diagnosis. On the other hand, the mean difference of the lesions’ attenuation between the TNC and VNCd series in our study was 0.07 HU. This difference is not statistically significant, so it can be considered safe enough to use the data from this series in order to characterize an adrenal nodule based on its unenhanced attenuation. The data from our study and the observations of other investigators [11,12,26] show a slight tendency to overestimate the attenuation values on the VNC images calculated from the portal phase. Particularly in our study, VNCp attenuation values of all but one lesions were higher than TNC ones (Fig. 1). The only lesion for which VNCp value (−15 HU) was lower than TNC value (−11 HU) was a lipid rich adenoma, with no change of the diagnosis in this case. So, on everyday clinical practice an incidental adrenal lesion detected on one DE portal acquisition only, can be evaluated on the VNC series calculated from this enhanced acquisition. If its attenuation value on the VNC series is less than 10 HU it could safely be characterized as a lipid rich adenoma, thus avoiding further workup. In our study, 10 patients had 13 lesions that had VNCp attenuation values lower than 10 HU. These 10 patients would have avoided further work up. This means that there would be no need to repeat a late phase acquisition at 15 min. This would allow to spare time and radiation dose (4.79 ± 2.38 mSv.) for the patient. In the same manner, there would be no need to recall any of these patients for a dedicated CT for adrenal characterization. This would allow sparing between 5.63 and 18.31 mSv, depending on the need of a single unenhanced series or repetition of a 3 phases enhanced CT. On the contrary, if a lesion has a VNC attenuation greater than 10 HU, a further workup with a delayed phase acquisition is considered necessary. This latter will allow not only the calculation of its VNCd attenuation which can be considered practically identical to the TNC attenuation value, but also the calculation of its absolute washout values. In current practice, in a CT protocol tailored to characterize adrenal lesions, when an adrenal lesion has a density greater than
10 HU on unenhanced images, it can be either a lipid poor adenoma, or another type of lesion, usually malignant. In this case, the workout will continue with a 15 min delayed acquisition, in order to calculate the absolute washout percentage [4]. If the CT was performed with another indication, in most cases an unenhanced series is not available. The characterization of the adrenal lesion is still feasible in this case, by calculating the relative washout [2]. According to the results of our study, the absolute washout values calculated from the data of the VNCp are significantly higher than those calculated from the true unenhanced series There was no statistically significant difference between “washout VNCd” and “washout TNC”. In the same manner, washout calculated from the Iodine concentration is not statistically different from washout calculated from true unenhanced series and could also be implemented on the characterization of adrenal nodules (Fig. 3). In the everyday clinical practice, after having treated the data of our first patients, we decided not to perform a true unenhanced series for the purpose of adrenal lesions characterization for all patients that were scanned with the dual-source DECT scanner of our department. A resident radiologist was assigned the task to treat DECT data with the dedicated software and calculate VNCp attenuation values for any adrenal lesion detected. If this value is higher than 10 HU, he repeats a 15 min late phase acquisition. If the VNCp attenuation is lower than 10 HU the lesion is classified as an adenoma and no further workout is needed. From another perspective, the utility of DECT in the characterization of adrenal adenomas has been examined by Gupta et al. [13] and Shi et al. [14] by studying attenuation change of adenomas between the series of different kilovoltages. Gupta et al. found that the mean attenuation change of all adenomas was 0.4 ± 7.1 HU and that of metastatic lesions 9.2 ± 4.3 HU while the change for the spleen and fat was found to be 7.4 ± 3.1 and −26.6 ± 2.7 HU. Based on their data, Gupta et al. concluded that DECT can be used to help differentiate some lipid-poor adrenal adenomas from metastatic lesions [13]. Our data show a mean attenuation change of all adenomas of 3.89 ± 8.43 HU between the series of high and low kilovoltage. The mean attenuation change of non-adenomatous lesions
D. Botsikas et al. / European Journal of Radiology 83 (2014) 1746–1751
was −3.88 ± 3.97 HU. In our opinion we cannot compare attenuation change of adenomas to that of non-adenomatous lesions as we found only 4 non-adenomatous lesions, two of which were pheochromocytomas that are known to be as “great mimickers” in imaging. For this reason we used the attenuation change of the spleen as an approximate equivalent of a, lesion, with minimal fat content. The spleen was used for this reason, because its fat content is minimal, about 3%, and splenic lipidosis is a very rare condition that has been described post-mortem in 12 out of 1500 autopsies and only in patients that had received intravenous lipid therapy [28] The mean attenuation change for the spleen in our study was 4.89 ± 4.16 and was not significantly different from adrenal adenomas (p > 0.9999). These results are not concordant with Gupta’s. These differences can be explained by the small number of subjects in both studies. The fact that the two studies were performed with different technical equipment (dual-source versus single source DECT) and different voltage parameters (we used 100 and 140 kVp respectively for the two tubes, while Gupta et al. performed the acquisition with the single tube alternating between 80 and 140 kVp) can also offer a possible explanation to this discrepancy. On the same topic, Shi et al. [14] found that the change of densities of all adenomas (including lipid-poor and lipid-rich) were significantly higher than those of metastases (p < 0.001). In our opinion research should be continued in this direction in order to prove the utility of measuring attenuation changes of adrenal lesions to their characterization. Most importantly, the acquisition parameters and technique should be strictly standardized, in order to set cut-off values of density changes for characterization purposes and provide reproducible measurements. Another concern about tissue characterization in general and subsequently adrenal lesion characterization with CT is that two independent studies have reported that different single detector and MDCT helical scanners produce slightly different but statistically significant unenhanced attenuation levels, which could lead to potentially erroneous categorization. This could potentially be considered as a potential pitfall with DECT as well [29,30]. Limitations of this study are its retrospective nature and the relatively small patient population, including 19 patients instead of 21 as indicated by the a priori statistical analysis. The consecutive patients selection minimizes selection bias related to the retrospective nature of the study. Another limiting factor is the fact that there were only 4 non-adenomatous lesions, 2 of which were pheochromocytomas. Thus it was not possible to compare adenomas with those lesions. For the purpose of comparison between attenuation values and washout values measured at TNC, VNCp and VNCd series, all lesions were included. In conclusion, according to the results of this study, incidental adrenal lesions can be characterized as adenomas, if their attenuation values based on the virtual unenhanced images calculated from the portal phase DECT acquisition are lower than 10 HU. If this value is ≥10 HU, it is recommended to add a 15 min delayed phase. This will allow calculation of virtual unenhanced attenuation from the data of the delayed acquisition which correlates better with the true unenhanced attenuation, as well as calculation of washout values (“washout VNCd” and washout “CIodine ”) that are found to correlate well with the washout calculated from a true unenhanced series.
Conflict of interest All authors declare no conflict of interest.
1751
References [1] Bovio S, Cataldi A, Reimondo G, Sperone P, Novello S, Berruti A, et al. Prevalence of adrenal incidentaloma in a contemporary computerized tomography series. J Endocrinol Invest 2006;29(4):298–302. [2] Caoili EM, Korobkin M, Francis IR, Cohan RH, Platt JF, Dunnick NR, et al. Adrenal masses: characterization with combined unenhanced and delayed enhanced CT. Radiology 2002;222(3):629–33. [3] Musella M, Conzo G, Milone MM, Corcione F, Belli G, De Palma M, et al. Preoperative workup in the assessment of adrenal incidentalomas: outcome from 282 consecutive laparoscopic adrenalectomies. BMC Surg 2013;13(1):57. [4] Dunnick NR, Korobkin MM. Imaging of adrenal incidentalomas: current status. AJR Am J Roentgenol 2002;179(3):559–68. [5] Blake MA, Cronin CG, Boland GW. Adrenal imaging. AJR Am J Roentgenol 2010;194(6):1450–60. [6] Boland GW, Lee MJ, Gazelle GS, Halpern EF, McNicholas MM, Mueller PR. Characterization of adrenal masses using unenhanced CT: an analysis of the CT literature. AJR Am J Roentgenol 1998;171(1):201–4. [7] Korobkin M, Brodeur FJ, Francis IR, Quint LE, Dunnick NR, Londy F. CT timeattenuation washout curves of adrenal adenomas and nonadenomas. AJR Am J Roentgenol 1998;170(3):747–52. [8] PeÒa CS, Boland GW, Hahn PF, Lee MJ, Mueller PR. Characterization of indeterminate (lipid-poor) adrenal masses: use of washout characteristics at contrast-enhanced CT. Radiology 2000;217(3):798–802. [9] Graser A, Johnson TRRC, Chandarana H, Macari M. Dual energy CT: preliminary observations and potential clinical applications in the abdomen. Eur Radiol 2009;19(1):13–23. [10] Coursey CA, Nelson RC, Boll DT, Paulson EK, Ho LM, Neville AM, et al. Dualenergy multidetector CT: how does it work, what can it tell us, and when can we use it in abdominopelvic imaging? Radiographics 2010;30(4):1037–55. [11] Gnannt R, Fischer M, Goetti R, Karlo C, Leschka S, Alkadhi H. Dual-energy CT for characterization of the incidental adrenal mass: preliminary observations. AJR Am J Roentgenol 2012;198(1):138–44. [12] Kim YK, Park BK, Kim CK, Park SY. Adenoma characterization: adrenal protocol with dual-energy CT. Radiology 2013;267(1):155–63. [13] Gupta RT, Ho LM, Marin D, Boll DT, Barnhart HX, Nelson RC. Dual-energy CT for characterization of adrenal nodules: initial experience. AJR Am J Roentgenol 2010;194(6):1479–83. [14] Shi JWJ, Dai HZ, Shen LH, Xu DFD. Dual-energy CT: clinical application in differentiating an adrenal adenoma from a metastasis. Acta Radiol 2013;55(4):505–12. [15] Johnson TRC, Krauss B, Sedlmair M, Grasruck M, Bruder H, Morhard D, et al. Material differentiation by dual energy CT: initial experience. Eur Radiol 2007;17(6):1510–7. [16] Remy Jardin M, Faivre JBJB, Pontana FJB, Remy JJB. Dual energy CT for thoracic applications. Rev Mal Respir 2012;29(10):1268–71. [17] Morgan DE. Dual-energy CT of the abdomen. Abdom Imaging 2014;39(1):108–34. [18] Vlahos I, Chung R, Nair A, Morgan R. Dual-energy CT: vascular applications. AJR Am J Roentgenol 2012;199(5 Suppl.):S87–97. [19] Kaza RK, Platt JF, Cohan RH, Caoili EM, Al Hawary M, Wasnik A. Dual-energy CT with single- and dual-source scanners: current applications in evaluating the genitourinary tract. Radiographics 2012;32(2):353–69. [20] Botsikas D, Hansen C, Stefanelli S, Becker CD, Montet X. Urinary stone detection and characterisation with dual-energy CT urography after furosemide intravenous injection: preliminary results. Eur Radiol 2014;24(3):709–14. [21] Primak AN, Fletcher JG, Vrtiska TJ, Dzyubak OP, Lieske JC, Jackson ME, et al. Noninvasive differentiation of uric acid versus non-uric acid kidney stones using dual-energy CT. Acad Radiol 2007;14(12):1441–7. [22] Stolzmann P, Kozomara M, Chuck N, Müntener M, Leschka S, Scheffel H, et al. In vivo identification of uric acid stones with dual-energy CT: diagnostic performance evaluation in patients. Abdom Imaging 2010;35(5):629–35. [23] Leschka S, Stolzmann P, Baumüller S, Scheffel H, Desbiolles L, Schmid B, et al. Performance of dual-energy CT with tin filter technology for the discrimination of renal cysts and enhancing masses. Acad Radiol 2010;17(4):526–34. [24] Neville AM, Gupta RT, Miller CM, Merkle EM, Paulson EK, Boll DT. Detection of renal lesion enhancement with dual-energy multidetector CT. Radiology 2011;259(1):173–83. [25] Arndt N, Staehler M, Siegert S, Reiser MF, Graser A. Dual energy CT in patients with polycystic kidney disease. Eur Radiol 2012;22(10):2125–9. [26] Ho LM, Marin D, Neville AM, Barnhart HX, Gupta RT, Paulson EK, et al. Characterization of adrenal nodules with dual-energy CT: can virtual unenhanced attenuation values replace true unenhanced attenuation values? AJR Am J Roentgenol 2012;198(4):840–5. [27] Morgan DE, Weber AC, Lockhart ME, Weber TM, Fineberg NS, Berland LL. Differentiation of high lipid content from low lipid content adrenal lesions using single-source rapid kilovolt (peak)-switching dual-energy multidetector CT. J Comput Assist Tomogr 2013;37(6):937–43. [28] Forbes GB. Splenic lipidosis after administration of intravenous fat emulsions. J Clin Pathol 1978;31(8):765–71. [29] Hahn PF, Blake MA, Boland GW. Adrenal lesions: attenuation measurement differences between CT scanners. Radiology 2006;240(2):458–63. [30] Stadler A, Schima W, Prager G, Homolka P, Heinz G, Saini S, et al. CT density measurements for characterization of adrenal tumors ex vivo: variability among three CT scanners. AJR Am J Roentgenol 2004;182(3):671–5.