The value of 15-minute delayed contrast-enhanced CT to differentiate hyperattenuating adrenal masses compared with chemical shift MR imaging
Hyun Jung Koo . Hyuck Jae Choi . Hwa Jung Kim . Sun-Ok Kim . Kyoung-Sik Cho
Received: 28 August 2013 /Revised: 12 November 2013 / Accepted: 28 November 2013 C European Society of Radiology 2013
Abstract
Objectives To investigate the diagnostic performance of 15- min delayed contrast-enhanced computed tomography (15- DECT) compared with that of chemical shift magnetic reso- nance (CSMR) imaging in differentiating hyperattenuating adrenal masses and to perform subgroup analysis in underly- ing malignancy and non-malignancy.
Methods This study included 478 adrenal masses in 453 patients examined with 15-DECT and 235 masses in 217 patients examined with CSMR. Relative percentage washout (RPW) and absolute percentage washout (APW) on 15- DECT, and signal intensity index (SII) and adrenal-to-spleen ratio (ASR) on CSMR were measured. Sensitivity, specificity and accuracy of 15-DECT and CSMR were analysed for characterisation of adrenal adenoma. Subgroup analyses were performed in patients with and without underlying malignan- cy. Attenuation and size of the masses on unenhanced CT correlated with the risk of non-adenoma.
Results RPW calculated from 15-DECT showed the highest diagnostic performance for characterising hyperattenuating adrenal masses regardless of underlying malignancy, and the sensitivity, specificity and accuracy were 91.7 %, 74.8 % and 88.1 %, respectively in all patients. The risk of non-adenoma increased approximately threefold as mass size increased 1 cm or as its attenuation value increased by 10 Hounsfield units.
H. J. Koo . H. J. Choi () . K .- S. Cho Department of Radiology and Research Institute of Radiology, Asan Medical Center, University of Ulsan College of Medicine, 388-1 Poongnap-dong, Songpa-gu, Seoul 138-736, Korea e-mail: hjchoi@ncc.re.kr
H. J. Kim · S .- O. Kim
Department of Clinical Epidemiology and Biostatistics, Asan Medical Center, University of Ulsan College of Medicine, Cancer Center, Seoul, South Korea
Conclusions 15-DECT was more accurate than CSMR in characterising hyperattenuating adrenal masses regardless of underlying malignancy.
Key Points
· Delayed contrast-enhanced CTand chemical shift magnetic resonance (CSMR) characterise adrenal lesions.
· 15-min DECT is more accurate than CSMR in characterising hyperattenuating adrenal masses.
· Sensitivity of CSMR decreases as the CT attenuation of adenomas increases.
· Risk of non-adenoma is increased 2.9-fold as size increased by 1 cm.
· Risk of non-adenoma is increased 2.9-fold as attenuation increased by 10 HU.
Keywords Adrenal gland . Hyperattenuating adrenal mass . Delayed contrast-enhanced CT · Chemical shift magnetic resonance imaging · Accuracy
Introduction
Adrenal masses are often detected incidentally on routine abdominal computed tomography (CT) examinations, with a reported prevalence of approximately 5 % [1]. In patients with hyperattenuating adrenal masses, clinical decision-making whether to perform biopsy, surgery or follow-up is challeng- ing, especially in patients with a history of extra-adrenal malignancy. Non-invasive diagnostic strategies for character- isation of hyperattenuating adrenal masses are required to reduce unnecessary imaging studies or procedures. Although attenuation on unenhanced CT is frequently utilised, differen- tiating among adrenal lesions with attenuation values greater than 10 Hounsfield units (HU) remains challenging. Fifteen-
minute delayed contrast-enhanced CT (15-DECT) and chem- ical shift magnetic resonance (CSMR) are useful for the characterisation of these lesions, although the diagnostic value of CSMR is reportedly lower than that of 15-DECT for differentiating among hyperattenuating adrenal masses [2]. To date, however, there are no studies that compare the diag- nostic accuracy of 15-DECT and CSMR in differentiating among hyperattenuating adrenal masses in a large number of patients.
The likelihood of an adrenal lesion being malignant de- pends on the presence of underlying malignancy. The possi- bility of metastasis in an adrenal lesion is approximately 25- 72 % in patients with a history of malignancy and approxi- mately 27 % in oncology patients with microscopic metasta- ses [3-6]. Conversely, the likelihood of adrenal malignancy is extremely low in patients without a history of malignancy [1, 7]. For example, of 1,049 adrenal masses in 973 patients without a history of underlying malignancy, none was malig- nant [1]. We therefore hypothesised that the prevalence of underlying malignancy could affect the diagnostic perfor- mance of imaging techniques and that subgroup analysis would be necessary to reduce selection bias. However, no study to date on the diagnostic accuracy of adrenal imaging for identifying hyperattenuating adrenal lesions has included subgroup analysis regarding the presence of underlying ma- lignancy. In view of these needs, we retrospectively compared the diagnostic performances of 15-DECT and CSMR in characterising hyperattenuating adrenal masses in a large group of patients, including subgroup analyses regarding the presence of underlying malignancy.
Materials and methods
Patients
This study was conducted according to the STARD guidelines for studies of diagnostic accuracy. The protocol was approved by the institutional review board of our hospital, which waived the requirement for patient informed consent owing to the retrospective design of the study. We searched our hospital’s picture archiving and communication system (PACS) for the period from January 2000 to March 2012 to find patients who had undergone 15-DECT or CSMR, using the keywords “adrenal metastasis”, “adrenal metastases”, “ad- renal mass”, “adrenal masses”, “adrenal lesion”, “adrenal lesions” or “pheochromocytoma”. A total of 1,548 consecu- tive patients who underwent 15-DECT (n=907) or CSMR (n=641) were initially identified. Thirty-six patients who underwent both 15-DECT and CSMR were assigned on the basis of their initial study technique. Exclusion criteria were as follows: previous adrenal surgery; hypoattenuating adrenal mass (≤10 HU); absence of unenhanced CT images; adrenal
hyperplasia; adrenal cyst; haemorrhagic or calcified sequelae; and adrenal lesions due to direct invasion from adjacent cancer, such as renal cell carcinoma or pancreatic cancer (Fig. 1). Patients with no repeat imaging studies to identify adrenal masses after the 6-month follow-up were also exclud- ed [8-10]. Finally, the study included 478 hyperattenuating adrenal masses in 453 patients (male/female=262:191, mean age 51.8±12.8 years) who underwent 15-DECT and 235 masses in 217 patients (male/female=174:43, mean age 57.0 ±12.0 years) who underwent CSMR.
Imaging protocols
CT technique
Multi-detector CT images were obtained using five multi- detector units: LightSpeed QX/I; LightSpeed plus 16; LightSpeed Ultra 16; LightSpeed VCT (GE Medical Systems, Milwaukee, WI, USA); and Somatom Sensation 16 (Siemens Medical Solutions, Erlangen, Germany). The imaging param- eters for unenhanced and contrast-enhanced CT examinations included 3.0-5.0-mm reconstruction slice thickness, 0.75- 0.875:1 pitch, 0.5-s rotation time, 120 kVp and 200- 400 mA. Enhanced images were obtained after intravenous administration of 150 mL iopromide (Ultravist 370; Schering, Germany) at a rate of 3.0 mL/s using a power injector. All patients imaged using 15-DECT underwent unenhanced CT and three-phase, contrast-enhanced CT, at 60 s (arterial phase), 120 s (portal phase) and 15 min (delayed phase) after the beginning of the bolus injection of contrast agent.
MR imaging technique
All MR imaging was performed on 1.5-T imaging systems including Magnetom Avanto (Siemens Medical Solutions, Erlangen, Germany), Achieva (Philips Medical Solutions, Eindhoven, the Netherlands) and Signa (GE Medical Systems, Milwaukee, WI, USA), with the images available on the institutional PACS. MR imaging parameters varied slightly, depending on the clinical protocols used at the time. All MR imaging included transverse T1-weighted, dual-echo, in- phase and out-of-phase chemical shift images. The ranges of MR parameters for all sequence types that were used for this study are listed in Table 1.
Image analysis
The CT and MR results were reviewed in consensus by two radiologists (H.J.C. with 11 years of clinical experience in genitourinary imaging and H.J.K. with 3 years of clinical experience in general radiology) who were blinded to the patients’ clinical or histopathological data, the presence of an extra-adrenal underlying malignancy, and the original CT
Springer
1548 patients included in this study
891 patients with DECT
621 patients with CSMR
36 patients with both DECT and CSMR
16 had initial DECT before underwent CSMR
☐ 20 had initial CSMR before underwent DECT
454 excluded in DECT (n = 907) 299 had hypoattenuating adrenal mass (≤ 10 HU) 87 adrenal cyst, hemorrhagic or tuberculosis sequelae, hyperplasia 50 were postoperative state 18 had less than 6 months follow-up
424 excluded in CSMR (n = 641) 139 had hypoattenuating adrenal mass (≤ 10 HU)
115 did not have unenhanced CT
74 had extra-adrenal malignancy which invades adrenal gland directly 69 had less than 6 months follow-up
17 inadequate image quality (3-T MR, outside MR, artifact, no spleen) 10 adrenal cyst, hemorrhagic or tuberculosis sequelae, hyperplasia
453 patients with DECT
217 patients with CSMR
478 adrenal lesions in 453 patients
235 adrenal lesions in 217 patients
109 patients with underlying malignancy ☐ 344 patients with non-malignancy
185 patients with underlying malignancy
☐ 32 patients with non-malignancy
375 adenoma
☐ 103 nonadenoma
164 adenoma
☐ 71 nonadenoma
or MR imaging findings. For each tumour, a region-of-interest (ROI) was placed within the soft-tissue region without calci- fication, haemorrhage or necrosis, and the mean attenuation or signal intensity of the tumour was obtained by averaging three measurements using the round-shape ROI automatic tools on PACS. The ROIs were drawn as large as possible on the adrenal masses to cover them, although the peripheral areas
| Image orientation | Axial |
|---|---|
| Repetition time (ms) | 131-174 |
| Echo time (ms) | 2.2 and 4.4 |
| Flip angle (°) | 70 |
| Number of signals acquired | 1 |
| Matrix size | 188×161 to 384×512 |
| Field of view (mm) | 350×299 to 285×380 |
| Number of slices | 20-55 |
| Slice thickness (mm) | 3-6 |
of the lesions were avoided to prevent a partial-volume effect. The low signal lines seen on the opposed-phase MR images were not included in the ROIs. Identical ROIs were placed on the same region on the unenhanced and dynamic CT phases and the in-phase or opposed-phase MR images.
Absolute percentage washout (APW) and relative percent- age washout (RPW) were determined for patients who underwent 15-DECT and adrenal-to-spleen ratio (ASR) and signal intensity index (SII) for patients who underwent CSMR. The equations for APW, RPW, ASR and SII for diagnosing adrenal adenoma are shown in Fig. 2 [11-13]. Adenoma was diagnosed on 15-DECT if the APW was greater than 60 % or the RPW was greater than 40 %; adenoma was diagnosed on CSMR if the ASR was less than 0.71 or the SII was greater than 16.5 % [14-17].
The sizes (the largest diameter on axial images) of all adrenal masses were measured on unenhanced CT images, and the mean tumour sizes between adenoma and non-adenoma were com- pared. The risk of non-adenoma based on the size of adrenal mass was assessed. In the same way, the attenuation (HU) of the mass on unenhanced CT was measured in all adrenal masses,
| Delayed contrast- enhanced CT | APW (absolute percentage washout) = (CTee - CTde) x 100 / (CTee - CTue) |
| CTee, CTde, and CTue: the atteuation values on early enhanced, delayed enhanced and unenhanced images | |
| RPW (relative percentage washout) = (CTee - CTde) x 100 / CTee | |
| CTee and CTde: the atteuation values on early and delayed enhanced images | |
| Chemical shift MR imaging | ASR (adrenal to spleen ratio) = (SIao / SIso) / (SIai / SIsi) |
| SIao and SIso: the signal intensities of an adrenal mass and the spleen on opposed-phase images | |
| SIai and SIsi: the signal intensities of an adrenal mass and the spleen on in-phase images | |
| SII (signal intensity index) = (SIai - SIao) × 100 / SIai | |
| SIai and Slao: the signal intensities of an adrenal mass on in- and opposed-phase images |
and the mean attenuation values of the adenoma and non- adenoma were compared. The risk of non-adenoma according to the attenuation of the adrenal mass was also analysed. In addition, the changes in sensitivity of RPW and SII were dem- onstrated by increasing the attenuation of adrenal masses.
Standards of reference
Standards of reference were based on a final diagnosis of adrenal adenoma or non-adenoma in each patient’s medical records, based on the histological diagnosis on surgery or biopsy (n=250), or change in size and shape over at least 6 months after initial adrenal imaging [8-10]. A mass was diagnosed as an adenoma if its size had not changed for at least 6 months, whereas a mass was diagnosed as a non- adenoma if it increased in size within 6 months or decreased in size after chemotherapy.
Statistical analysis
The sensitivity, specificity and accuracy of the four parameters of the CT and MR were obtained. To compare the values of RPW and SII, we used the Chi-squared test or Fisher’s exact test. Statistical analysis was performed using SAS for Win- dows (version 9.2; SAS Institute, Cary, NC, USA). Multiple masses in the same patient were regarded as different lesions, resulting in lesion-based comparisons for calculated parame- ters. A P value less than 0.05 was considered statistically significant. Odds ratios (ORs) and 95 % confidence intervals (CIs) were calculated using logistic regression. As the per- centages of patients with a history of extra-adrenal malignan- cy differed in the CT and MR groups, subgroup analyses were performed in patients with underlying malignancy and those without malignancy. Patients with underlying hepatocellular carcinoma (HCC) or renal cell carcinoma (RCC), and patients with pheochromocytomas were excluded from additional sub- group analysis, because adrenal metastases from HCC or RCC, and pheochromocytoma are known to mimic adrenal adenoma because of their intracellular fat component, early enhancement, and washout pattern [18-21]. The mean sizes and mean attenuations of adenoma and non-adenoma were
compared using Student’s t test. To obtain the risk of non- adenoma according to the size or attenuation of adrenal mass, multivariate analysis was used.
Results
Patients and baseline characteristics
In this study, 478 hyperattenuating adrenal masses in 453 patients (male/female=262:191, mean age 51.8±12.8 years) who underwent 15-DECT and 235 masses in 217 patients (male/female=174:43, mean age 57.0±12.0 years) who underwent CSMR were included (Fig. 1). Among these pa- tients, 212 (46.6 %) in the DECT group and 38 (17.5 %) in the CSMR group underwent percutaneous biopsy or definitive surgery to confirm the histopathological diagnoses of the adrenal masses. The mean follow-up period in the 420 patients who did not undergo biopsy or surgery was 29.2±29.6 months (range 6.0-409.3 months). In the DECT group, 19 patients had two adrenal lesions each and 3 patients had three lesions each; in the CSMR group, 18 patients had two lesions each. The demographic and clinical characteristics of the DECT and CSMR groups are shown in Table 2.
Among the 713 adrenal masses, 539 were adenomas, includ- ing 375 in the 15-DECT group (Fig. 3) and 164 in the CSMR group, and 174 were non-adenomas, including 103 in the 15- DECT group (Fig. 4) and 71 in the CSMR group (Fig. 5). The non-adenomas were pheochromocytoma, adrenal cortical carci- noma, lymphoma and metastases. Extra-adrenal malignancies were observed in 109 patients (24.1 %) assessed by 15-DECT and in 185 (85.3 %) assessed by CSMR; the types of underlying malignancy in each group are shown in Table 2.
Quantitative parameters of DECT and CSMR
In all patients, the sensitivity, specificity and accuracy of APW on 15-DECT were 84.3 % (316/375), 78.6 % (81/103) and 83.1 % (397/478), respectively, while those for RPW were 91.7 % (344/375), 74.8 % (77/103) and 88.1 % (421/478), respectively. On CSMR, the sensitivity, specificity and
| 478 lesions detected with 15-DECT (n=453) | 235 lesions detected with CSMR (n=217) | |
|---|---|---|
| Mean age (years)±SD | 51.8±12.8 | 57.0±12.0 |
| Male/female | 262:191 | 174:43 |
| Underlying malignancy | 109 (24.1 %) | 185 (85.3 %) |
| HCC | 3 | 66 |
| Gastric cancer | 18 | 15 |
| Breast cancer | 10 | 1 |
| Lung cancer | 19 | 11 |
| Bile-duct cancer | 1 | 27 |
| RCC | 9 | 17 |
| Colon cancer | 9 | 18 |
| Thyroid cancer | 11 | 2 |
| Others | 29 | 28 |
| Surgery or biopsy (n) | 212 (46.6 %) | 38 (17.5 %) |
| Adenoma | 375 (78.5 %) | 164 (69.8 %) |
| Non-adenoma | 103 (21.5 %) | 71 (30.2 %) |
| Pheochromocytoma | 70 (68.0 %) | 15 (21.1 %) |
| Metastasis | 20 (19.4 %) | 51 (71.8 %) |
| Adrenal cortical carcinoma | 8 (7.8 %) | 2 (2.8 %) |
| Lymphoma | 3 (2.9 %) | 3 (4.2 %) |
| Othersª | 2 (1.9 %) | – |
Unless otherwise indicated, data represent the numbers of patients, except for the number of lesions proven to be adenomas and non-adenomas SD standard deviation, HCC hepatocellular carcinoma, RCC renal cell carcinoma
a One malignant rhabdoid tumour and one malignant fibrohistiocytoma
accuracy of SII were 67.1 % (110/164), 88.7 % (63/71) and 73.6 % (173/235), respectively, while those of ASR were 44.5 % (73/164), 97.2 % (69/71) and 60.4 % (142/235), respectively (Table 3). RPW on 15-DECT was significantly more accurate and sensitive than SII on CSMR (P< 0.001 each). However, in all patients, the specificity of SII on CSMR
was higher than that of RPW on 15-DECT (P=0.022). Both for lesions with underlying malignancy and for those with no malignancy, the sensitivity of RPW on 15-DECT was higher than that of SII on CSMR, whereas their specificities did not differ significantly in either subgroup (Table 3).
Additional subgroup analyses, after excluding the patients with HCC (Fig. 4), RCC (Fig. 5) and pheochromocytoma (Fig. 6), showed that the accuracy and sensitivity of RPW on 15-DECT were higher than those of SII on CSMR, although the specificities of the two techniques did not differ signifi- cantly (Table 4).
Multivariate logistic regression analyses were performed to compare the diagnostic accuracy of RPW on 15-DECT and SII on CSMR regarding the presence of underlying malignancy (Table 5). RPW on 15-DECT was more accurate (OR 2.72, P<0.001) and sensitive (OR 4.71, P<0.001), but not more specific (P=0.348), than SII on CSMR. Additional subgroup analysis, after excluding the patients with adrenal adenoma mimicking tumours, again showed that RPW on 15-DECT was significantly more accurate and sensitive, but not more specific, than SII on CSMR (Table 6).
Tumour size and attenuation on unenhanced CT
Mean tumour size was significantly smaller for adenomas than for non-adenomas (1.8 cm vs. 4.2 cm, P<0.001). Mul- tivariate analysis showed that a 1-cm increase in adrenal mass size increased the risk of non-adenoma 2.9-fold (Table 7). The mean attenuation of masses on unenhanced CT was 25.4 HU for adenomas and 36.2 HU for non-adenomas. Multivariate analysis showed that an increase in attenuation by 10 HU also increased the risk of non-adenoma 2.9-fold. In addition, al- though the sensitivity of RPW on 15-DECT is only slightly reduced, the sensitivity of SII on CSMR is markedly de- creased when the attenuation adrenal masses is high on unenhanced CT (Fig. 7).
A
A
B
C
D
Discussion
Although approximately 70 % of adrenal adenomas contain a substantial amount of intracellular fat, about 30 % do not, with
attenuation values greater than 10 HU on unenhanced CTs, making it challenging to identify these lesions using only unenhanced CT [22]. As both unenhanced CT and CSMR characterise adrenal masses by determining the presence of
A
B
C
D
| n | Accuracy | Sensitivity | Specificity | ||||||
|---|---|---|---|---|---|---|---|---|---|
| n (%) | 95 % CI | n (%) | 95 % CI | n (%) | 95 % CI | ||||
| Total patient population | CT | APW | 478 | 397/478 (83.1) | (79.7-86.4) | 316/375 (84.3) | (80.6-88.0) | 81/103 (78.6) | (70.7-86.6) |
| RPW | 421/478 (88.1) | (85.2-91.0) | 344/375 (91.7) | (89.0-94.5) | 77/103 (74.8) | (66.4-83.2) | |||
| MR | SII | 235 | 173/235 (73.6) | (68.0-79.3) | 110/164 (67.1) | (59.9-74.3) | 63/71 (88.7) | (81.4-96.1) | |
| ASR | 142/235 (60.4) | (54.2-66.7) | 73/164 (44.5) | (36.9-52.1) | 69/71 (97.2) | (90.2-99.7) | |||
| P value* | <0.001 | <0.001 | 0.022 | ||||||
| Underlying malignancy | CT | APW | 109 | 96/109 (88.1) | (82.0-94.2) | 68/77 (88.3) | (81.1-95.5) | 28/32 (87.5) | (71.0-96.5) |
| RPW | 99/109 (90.8) | (85.4-96.2) | 70/77 (90.9) | (84.5-97.3) | 29/32 (90.6) | (75.0-98.0) | |||
| MR | SII | 185 | 134/185 (72.4) | (66.0-78.9) | 80/123 (65.0) | (56.6-73.5) | 54/62 (87.1) | (78.8-95.4) | |
| ASR | 111/185 (60.0) | (52.9-67.1) | 51/123 (41.5) | (32.8-50.2) | 60/62 (96.8) | (88.8-99.6) | |||
| P value* | <0.001 | <0.001 | 0.744 | ||||||
| Non-malignancy | CT | APW | 369 | 301/369 (81.6) | (77.6-85.5) | 248/298 (83.2) | (79.0-87.5) | 53/71 (74.7) | (64.5-84.8) |
| RPW | 322/369 (87.3) | (83.9-90.7) | 274/298 (92.0) | (88.9-95.0) | 48/71 (67.6) | (56.7-78.5) | |||
| MR | SII | 50 | 39/50 (78.0) | (66.5-89.5) | 30/41 (73.2) | (59.6-86.7) | 9/9 (100) | (66.4-100.0) | |
| ASR | 31/50 (62.0) | (48.6-75.5) | 22/41 (53.7) | (38.4-68.9) | 9/9 (100) | (66.4-100.0) | |||
| P value* | 0.075 | 0.001 | 0.053 | ||||||
*P values were obtained between the RPW and SII values in each group using the x2 test or Fisher’s exact test
intracellular fat, the usefulness of CSMR in characterising hyperattenuating adrenal masses that could not be diagnosed by unenhanced CT remains unclear [23-25].
Delayed CT has been shown to be useful for characterising hyperattenuating adrenal masses [8-10, 20, 22]. DECT uti- lises the faster washout pattern of adrenal adenomas compared with the prolonged enhancement of malignant adrenal lesions;
therefore, DECT is relatively independent of the amount of intracellular fat used to characterise masses. In particular, 15- DECT is more frequently used to evaluate adrenal masses, because a 10-min or shorter delay may decrease its diagnostic performance [17, 26]. Indeed, we had hypothesised that 15- DECT might characterise hyperattenuating adrenal masses more successfully than CSMR.
A
B
C
D
Superior
Right
Left
Inferior
respectively, with the SII suggesting an adrenal adenoma. d However, the mass showed strong uptake on I-131 metaiodobenzylguanidine (MIBG) single-photon emission computed tomography (SPECT), indi- cating a pheochromocytoma. After right adrenalectomy, the mass was confirmed to be a pheochromocytoma
| n | Accuracy | Sensitivity | Specificity | ||||||
|---|---|---|---|---|---|---|---|---|---|
| n (%) | 95 % CI | n (%) | 95 % CI | n (%) | 95 % CI | ||||
| Total patient population | CT | APW | 398 | 339/398 (85.2) | (81.7-88.7) | 311/369 (84.3) | (80.6-88.0) | 28/29 (96.6) | (82.2-99.9) |
| RPW | 367/398 (92.2) | (89.6-94.8) | 339/369 (91.9) | (89.1-94.7) | 28/29 (96.6) | (82.2-99.9) | |||
| MR | SII | 140 | 108/140 (77.1) | (70.2-84.1) | 80/109 (73.4) | (65.1-81.7) | 28/31 (90.3) | (74.3-98.0) | |
| ASR | 85/140 (60.7) | (52.6-68.8) | 55/109 (50.5) | (41.1-59.9) | 30/31 (96.8) | (83.3-99.9) | |||
| P value* | <0.001 | <0.001 | 0.613 | ||||||
| Underlying malignancy | CT | APW | 93 | 85/93 (91.4) | (85.7-97.1) | 63/71 (88.7) | (81.4-96.1) | 22/22 (100) | (84.6-100.0) |
| RPW | 87/93 (93.6) | (88.6-98.5) | 65/71 (91.6) | (85.1-98.0) | 22/22 (100) | (84.6-100.0) | |||
| MR | SII | 98 | 77/98 (78.6) | (70.5-86.7) | 50/68 (73.5) | (63.0-84.0) | 27/30 (90.0) | (73.5-97.9) | |
| ASR | 62/98 (63.3) | (53.7-72.8) | 33/68 (48.5) | (36.7-60.4) | 29/30 (96.7) | (82.8-99.9) | |||
| P value* | <0.001 | 0.005 | 0.25 | ||||||
| Non-malignancy | CT | APW | 305 | 254/305 (83.3) | (79.1-87.5) | 248/298 (83.2) | (79.0-87.5) | 6/7 (85.7) | (42.1-99.6) |
| RPW | 280/305 (91.8) | (88.7-94.9) | 274/298 (92.0) | (88.9-95.0) | 6/7 (85.7) | (42.1-99.6) | |||
| MR | SII | 42 | 31/42 (73.8) | (60.5-87.1) | 30/41 (73.2) | (59.6-86.7) | 1/1 (100) | (2.5-100.0) | |
| ASR | 23/42 (54.8) | (39.7-69.8) | 22/41 (53.7) | (38.4-68.9) | 1/1 (100) | (2.5-100.0) | |||
| P value* | 0.001 | 0.001 | 1.000 | ||||||
15-DECT 15-minute delayed contrast-enhanced computed tomography, CSMR chemical shift magnetic resonance
*Between RPW and SII values, calculated using the x2 test or Fisher’s exact test.
A case series of 34 patients who underwent both 15-DECT and CSMR found that 15-DECT identified five adrenal ade- nomas and three adrenal metastases not detected by CSMR [24]. However, because of the small sample size, the differ- ence in diagnostic accuracy between the two techniques was not statistically significant. The larger number of patients in our study resulted in statistically significant differences be- tween the two techniques.
Although CSMR is useful in patients with pregnancy, poor renal function or iodine hypersensitivity, the sensitivity of chemical shift MR in diagnosing adenoma decreases with increasing attenuation of hyperattenuating adrenal mass on unenhanced CT. Haider et al evaluated the performance of CSMR in identifying adrenal masses of between 10 and 30 HU as adenomas, and reported a sensitivity of 89 % and a specificity of 100 % [24]. Conversely, CSMR was found to
| Outcome | Total patients (n=713) | Underlying malignancy (n=294) | Non-malignancy (n=419) | |||||
|---|---|---|---|---|---|---|---|---|
| OR (95 % CI) | P value | OR (95 % CI) | P value | OR (95 % CI) | P value | |||
| Accuracy | Techniques | CT | 2.72 (1.68-4.42) | <0.001 | 3.77 (1.82-7.79) | <0.001 | 1.93 (0.93-4.03) | 0.079 |
| MR | 1 | 1 | 1 | |||||
| (594/713) | Underlying malignancy | (+) | 1.05 (0.65-1.70) | 0.843 | – | – | ||
| (-) | 1 | – | – | |||||
| Sensitivity | Techniques | CT | 4.71 (2.65-8.39) | <0.001 | 5.38 (2.27-12.71) | <0.001 | 4.19 (1.87-9.38) | 0.001 |
| MR | 1 | 1 | 1 | |||||
| (454/539) | Underlying malignancy | (+) | 0.76 (0.43-1.35) | 0.354 | – | – | ||
| (-) | 1 | – | – | |||||
| Specificity | Techniques | CT | 0.61 (0.22-1.70) | 0.348 | 1.43 (0.35-5.82) | 0.615 | Infiniteª | 0.964 |
| MR | 1 | 1 | 1 | |||||
| (140/174) | Underlying malignancy | (+) | 2.37 (0.93-6.04) | 0.071 | – | – | ||
| (-) | 1 | – | – | |||||
OR odds ratio
ª The specificity could not be estimated
| Outcome | Total patients (n=713) | Underlying malignancy (n=294) | Non-malignancy (n=419) | |||||
|---|---|---|---|---|---|---|---|---|
| OR (95 % CI) | P value | OR (95 % CI) | P value | OR (95 % CI) | P value | |||
| Accuracy | Techniques | CT | 3.97 (2.15-7.33) | <0.001 | 3.96 (1.52-10.30) | 0.005 | 3.97 (1.79-8.85) | 0.001 |
| MR | 1 | 1 | 1 | |||||
| (475/538) | Underlying malignancy | (+) | 1.30 (0.70-2.42) | 0.411 | ||||
| (-) | 1 | |||||||
| Sensitivity | Techniques | CT | 4.07 (2.17-7.64) | <0.001 | 3.90 (1.44-10.55) | 0.007 | 4.19 (1.87-9.38) | 0.001 |
| MR | 1 | 1 | 1 | |||||
| (419/478) | Underlying malignancy | (+) | 0.99 (0.52-1.87) | 0.964 | ||||
| (-) | 1 | |||||||
| Specificity | Techniques | CT | 5.73 (0.34-97.83) | 0.228 | Infiniteª | 0.999 | Infiniteª | |
| MR | 1 | 1 | 1 | |||||
| (56/60) | Underlying malignancy | (+) | 5.94 (0.30-117.09) | 0.241 | ||||
| (-) | 1 | |||||||
OR odds ratio
ª The specificity could not be estimated
have decreased sensitivity in identifying adrenal masses of greater than 30 HU [25]. Similarly, we also found that the sensitivity of CSMR decreased significantly as the attenuation of adrenal lesions increased, suggesting that CSMR is of limited value in evaluating hyperattenuating adrenal masses. On the other hand, the sensitivity of 15-DECT was less affected by the increased attenuation of adrenal lesions. In- verse relationships have been reported between the fat con- centration of an adrenal mass and its attenuation on unenhanced CT and between the percentage of lipid-rich cells and relative changes in signal intensity on CSMR [27]. This could explain why the sensitivity of CSMR decreased as the attenuation of an adrenal mass on unenhanced CT increased.
However, we showed that the sensitivity of MR (80.6 % for 10-19 HU and 88.4 % for 20-29 HU) was relatively high for characterising adrenal adenoma measuring 10-29 HU on unenhanced CT. Therefore, considering radiation exposure of DECT, CSMR could be performed for evaluating hyperattenuating adrenal mass especially in patients who can- not undergo contrast-enhanced CT because of pregnancy, poor renal function or iodine hypersensitivity.
By comparing the CT parameters, RPW was more diag- nostically accurate than APW for assessing hyperattenuating adrenal lesions, and that finding was similar to those of
previous reports [25, 28, 29]. We believe that the greater diagnostic accuracy of RPW is due to the inclusion of unenhanced CT attenuation (CTue) in the formula for APW ((CTee - CTde)×100/(CTee - CTue)), but not for RPW ((CTee-CTde)×100/CTee); therefore, RPW was not affected by the increasing attenuation of adrenal masses on unenhanced CT. This may be one of the explanations for the higher diagnostic accuracy of RPW compared with that of APW for differentiating hyperattenuating adrenal lesions.
With regard to adrenal adenoma-mimicking lesions, adre- nal metastases from RCC or HCC, and pheochromocytoma may contain intracellular fat, causing signal loss during the opposed phase of CSMR, as well as mimicking the early washout pattern of adrenal adenoma seen on DECT [11, 18-21]. To reduce bias resulting from the differing prevalence of these tumours in the 15-DECT and CSMR groups, we performed an additional analysis after excluding patients di- agnosed with HCC, RCC or pheochromocytoma. Before ex- clusion, 15-DECT showed significantly lower specificity than CSMR; however, after excluding these lesions, the specific- ities of these two techniques were similar.
Although adrenal masses larger than 3-4 cm have a greater potential to be malignant than smaller lesions [22, 30], the substantial size overlap between adenomas and malignant
| Odds ratio (95 % CI) | P value | |||
|---|---|---|---|---|
| Univariate | Attenuation (HU) | Increase per 10 HU | 3.71 (2.92-4.70) | <0.001 |
| Size (cm) | Increase per 1 cm | 3.19 (2.61-3.90) | <0.001 | |
| Multivariable | Attenuation (HU) | Increase per 10 HU | 2.93 (2.23-3.84) | <0.001 |
| Size (cm) | Increase per 1 cm | 2.88 (2.33-3.55) | <0.001 | |
HU Hounsfield unit
100
94.5
88.4
93.3
88.6
87.5
80.6
80
60
40
39.5
20
9.1
0
11-19
20-29
30-39
≥40
lesions limits the clinical usefulness of size alone. We there- fore quantified the relationship between increased size and attenuation value of an adrenal mass on unenhanced CT and the degree of increased risk of non-adenoma. We found that a 1-cm or 10-HU increase increased the risk of non-adenoma approximately threefold. Additional studies are needed to investigate the relationships between the combination of size and attenuation on unenhanced CT and DECT and CSMR parameters.
Our study had several limitations. First, because of its retrospective design, the prevalence and types of underlying malignancy differed in the 15-DECT and CSMR groups. To reduce the selection bias, we performed subgroup analyses in lesions with and without underlying malignancy, as well as after excluding adrenal lesions that mimic adrenal adenoma on both 15-DECT and CSMR. Second, imaging studies were done in two separate patient populations and therefore could not provide an opportunity to compare the two techniques in the same group. In addition, the retrospective nature and the long study period resulted in heterogeneity in the techniques and parameters used for MR imaging. However, as we con- fined the study population only to patients who underwent 1.5-T MR and assessed percentage signal intensity change rather that actual signal intensity, the measurement bias may be minimal. The possibility of poor MR quality with outdated MR systems could be reduced by excluding the patients who had inadequate image quality. Finally, the mean duration of follow-up (29.2±29.6 months; range 6.0-409.3 months) was relatively short. Although the follow-up periods greater than 6 months were thought to be valid for correctly diagnosing adrenal adenoma [8-10], a lack of tumour growth over at least a 6-month follow-up may not be an accurate method for distinguishing between adenoma and non-adenoma in some patients [31].
In conclusion, we showed that 15-DECT was more accurate diagnostically than CSMR in characterising hyperattenuating
adrenal masses, regardless of the presence of underlying ma- lignancy. The risk of non-adenoma was increased approximate- ly threefold as size increased by 1 cm or unenhanced CT attenuation increased by 10 HU.
References
1. Song JH, Chaudhry FS, Mayo-Smith WW (2008) The incidental adrenal mass on CT: prevalence of adrenal disease in 1,049 consec- utive adrenal masses in patients with no known malignancy. AJR Am J Roentgenol 190:1163-1168
2. Park BK, Kim CK, Kim B, Lee JH (2007) Comparison of delayed enhanced CT and chemical shift MR for evaluating hyperattenuating incidental adrenal masses. Radiology 243:760-765
3. Blake MA, Kalra MK, Sweeney AT et al (2006) Distinguishing benign from malignant adrenal masses: multi-detector row CT pro- tocol with 10-minute delay. Radiology 238:578-585
4. Francis IR, Smid A, Gross MD, Shapiro B, Naylor B, Glazer GM (1988) Adrenal masses in oncologic patients: functional and mor- phologic evaluation. Radiology 166:353-356
5. Mayo-Smith WW, Boland GW, Noto RB, Lee MJ (2001) State-of- the-art adrenal imaging. Radiographics 21:995-1012
6. Young WF Jr (2007) Clinical practice. The incidentally discovered adrenal mass. N Engl J Med 356:601-610
7. Zeiger MA, Siegelman SS, Hamrahian AH (2011) Medical and surgical evaluation and treatment of adrenal incidentalomas. J Clin Endocrinol Metab 96:2004-2015
8. Caoili EM, Korobkin M, Francis IR, Cohan RH, Dunnick NR (2000) Delayed enhanced CT of lipid-poor adrenal adenomas. AJR Am J Roentgenol 175:1411-1415
9. Caoili EM, Korobkin M, Francis IR et al (2002) Adrenal masses: characterization with combined unenhanced and delayed enhanced CT. Radiology 222:629-633
10. Muth A, Hammarstedt L, Hellstrom M, Sigurjonsdottir HA, Almqvist E, Wangberg B (2011) Cohort study of patients with adrenal lesions discovered incidentally. Br J Surg 98:1383-1391
11. Merkle EM, Schindera ST (2007) MR imaging of the adrenal glands: 1.5T versus 3T. Magn Reson Imaging Clin N Am 15:365-372
12. Nakamura S, Namimoto T, Morita K et al (2012) Characterization of adrenal lesions using chemical shift MRI: comparison between 1.5 Tesla and two echo time pair selection at 3.0 Tesla MRI. J Magn Reson Imaging 35:95-102
13. Siegelman ES (2012) Adrenal MRI: techniques and clinical applica- tions. J Magn Reson Imaging 36:272-285
14. Blake MA, Cronin CG, Boland GW (2010) Adrenal imaging. AJR Am J Roentgenol 194:1450-1460
15. Low G, Dhliwayo H, Lomas DJ (2012) Adrenal neoplasms. Clin Radiol 67:988-1000
16. Low G, Sahi K (2012) Clinical and imaging overview of functional adrenal neoplasms. Int J Urol 19:697-708
17. Taffel M, Haji-Momenian S, Nikolaidis P, Miller FH (2012) Adrenal imaging: a comprehensive review. Radiol Clin North Am 50:219- 243
18. Choi YA, Kim CK, Park BK, Kim B (2013) Evaluation of adrenal metastases from renal cell carcinoma and hepatocellular carcinoma: use of delayed contrast-enhanced CT. Radiology 266:514-520
19. Park BK, Kim B, Ko K, Jeong SY, Kwon GY (2006) Adrenal masses falsely diagnosed as adenomas on unenhanced and delayed contrast- enhanced computed tomography: pathological correlation. Eur Radiol 16:642-647
20. Park BK, Kim CK, Kwon GY, Kim JH (2007) Re-evaluation of pheochromocytomas on delayed contrast-enhanced CT: washout en- hancement and other imaging features. Eur Radiol 17:2804-2809
21. Ramsay JA, Asa SL, van Nostrand AW, Hassaram ST, de Harven EP (1987) Lipid degeneration in pheochromocytomas mimicking adre- nal cortical tumors. Am J Surg Pathol 11:480-486
22. van Erkel AR, van Gils AP, Lequin M, Kruitwagen C, Bloem JL, Falke TH (1994) CT and MR distinction of adenomas and nonadenomas of the adrenal gland. J Comput Assist Tomogr 18: 432-438
23. Outwater EK, Siegelman ES, Huang AB, Birnbaum BA (1996) Adrenal masses: correlation between CT attenuation value and chem- ical shift ratio at MR imaging with in-phase and opposed-phase sequences. Radiology 200:749-752
24. Haider MA, Ghai S, Jhaveri K, Lockwood G (2004) Chemical shift MR imaging of hyperattenuating (>10 HU) adrenal masses: does it still have a role? Radiology 231:711-716
25. Park BK, Kim CK, Kim B, Choi JY (2007) Comparison of delayed enhanced CT and 18F-FDG PET/CT in the evaluation of adrenal masses in oncology patients. J Comput Assist Tomogr 31:550-556
26. Sangwaiya MJ, Boland GW, Cronin CG, Blake MA, Halpern EF, Hahn PF (2010) Incidental adrenal lesions: accuracy of
characterization with contrast-enhanced washout multidetector CT-10-minute delayed imaging protocol revisited in a large patient cohort. Radiology 256:504-510
27. McGahan JP (1988) Adrenal gland: MR imaging. Radiology 166: 284-285
28. Fujiyoshi F, Nakajo M, Fukukura Y, Tsuchimochi S (2003) Characterization of adrenal tumors by chemical shift fast low-angle shot MR imaging: comparison of four methods of quantitative eval- uation. AJR Am J Roentgenol 180:1649-1657
29. Pena CS, Boland GW, Hahn PF, Lee MJ, Mueller PR (2000) Characterization of indeterminate (lipid-poor) adrenal masses: use of washout characteristics at contrast-enhanced CT. Radiology 217: 798-802
30. Candel AG, Gattuso P, Reyes CV, Prinz RA, Castelli MJ (1993) Fine- needle aspiration biopsy of adrenal masses in patients with extraadrenal malignancy. Surgery 114:1132-1136
31. Pantalone KM, Gopan T, Remer EM et al (2010) Change in adrenal mass size as a predictor of a malignant tumor. Endocr Pract 16:577- 587