Adrenal incidentalomas: imaging challenges—role of MDCT scan versus MRI in evaluating adrenal incidentalomas

The term adrenal incidentaloma (AI), by definition, is an adrenal mass that is unexpectedly detected through an imaging procedure performed for reasons unrelated to adrenal dysfunction or suspected dysfunction. Despite their frequent appearance, the challenge remains in recognizing and treating the small percentage of AI that poses a significant risk, either because of their hormonal activity or because of their malignant histology. The aim of this study is to study the role of MRI, specifically chemical shift imaging (CSI), against various MDCT scans (non-enhanced, enhanced, and delayed) in the characterization of incidentally discovered adrenal masses to offer a way for the patients to avoid unnecessary time and money-wasting imaging modalities used to reach a diagnosis of their incidentally discovered adrenal lesions. We examined a total number of 20 patients with total of 22 adrenal lesions. The mean age was 51.1 ± 15.27. In our study, we found that among CT parameters, APW and RPW showed the highest sensitivity and specificity for detection of lipid-rich adenomas. CSI has also proven to be the best MR technique. However, there is no statistically significant difference in the diagnostic capability of CSI versus the CT washout technique. Both modalities could be conducted, according to specific patient preferences and/or limitations, with comparable highly accurate outcomes. This study demonstrates that a similar diagnostic outcome is obtained from contrast-enhanced CT (CECT) and MRI with CSI of adrenal lesions.


Background
Adrenal masses discovered accidentally in the course of imaging for other abdominal conditions are known as adrenal incidentalomas (AIs). Those patients who are having these incidentalomas-by definition-have no history or symptoms of adrenal diseases prior to imaging procedures that lead to their discovery [1]. Adrenal metastasis, which is a common radiological finding in 6-20% of patients with extra-adrenal malignancy, should not be considered AI [2].
The prevalence of AI increases with age. It reaches up to 7% of people above the age of 70 compared to only 0.5% of people in their 20s. On the other hand, AIs are present in 4-5% 0f abdominal CT examinations [3].
Adrenal incidentalomas comprise a wide range of pathologies but the vast majority of cases are adrenal adenomas. Although in most cases, this is not proved pathologically, yet the clinical course can depict its benign nature. However, still there is a percentage of these AIs which could be either malignant or hormone secreting lesions which are not clinically evident yet. Table 1 demonstrates imaging findings in common adrenal pathologies. Based on that, both the physicians and the radiologists should do every effort to exclude or prove the non-benign nature of these lesions such as adrenocortical carcinoma (ACC) as well as other AI with secretory potentials [2].
The main target in dealing with AI is to prove or exclude the presence of malignancy especially ACC or the presence of secretory potential. This entails careful clinical history and examination, biochemical testing, and detailed radiographic examination.
Careful clinical history and examination might reveal symptoms and signs of adrenal dysfunction in hormone secreting AI, and this needs high index of suspicion among physicians [4].
Currently, CT scan and MRI examination are the imaging modalities of choice with high diagnostic accuracy regarding the depiction of the nature of the lesion and the presence of malignant potential. The size of the lesion is an important determinant of the malignant potential of AI as the ACC is larger than adenomas. The risk of malignancy increases with the size, in one meta-analysis it was found that ACCs represent 2% of all tumors < 4 cm and up to 25% in tumors > 6 cm [5]. Definitely, the size alone has low specificity in diagnosis of malignant potential. Other criteria, which can suggest the malignant potential, include irregular shape and borders, tumor heterogeneity with central necrosis or hemorrhage and invasion into surrounding structures. On the other hand, the main features of benign adrenal adenomas are their smaller size, regular outline, and slower growth [2].
There are some imaging properties of AIs, depending on the modality used, which can be helpful for differentiation between benign and malignant adrenal lesions are shown in (Table 1).

Methods
This cross-sectional analytic study was conducted on 20 patients at the Radiology Department in collaboration with the Internal Medicine Department using MDCT and MRI. This was done after approval from the institutional review board and obtaining informed consent from all patients. The mean age of the participants was 51.1 ± 15.27 with 2:3 female to male ratio. Twenty patients were included in our study with a total number of 22 lesions.

Inclusion criteria
Adult patients with accidentally discovered adrenal lesions by a previous imaging modality.

Exclusion criteria
People performed adrenal surgical procedures, people known to have known endocrine problems, pediatric age group, and pregnant patients are all excluded. Claustrophobic patients were excluded as well.
The following measures have been done: Informed written consent with detailed explanation of the study and its purpose was obtained from each patient. This is followed by an inquiry of all patients about how they discovered their adrenal lesion(s) and whether they performed subsequent laboratory tests or not.
Unenhanced and if possible, contrast-enhanced CT and MRI protocols (T1-T2-DUAL phase and DWI) for characterizing their adrenal lesions. Six months and/or 12 months follow-up CT and MRI studies were performed by contacting the patients' phones and arranging meeting or through their treating physician. Also, a histopathological confirmation was obtained if available. A combination of pathological results and/or follow-up was considered as a standard of reference. Protocol for MDCT MDCT imaging of the abdomen, of 20 patients with 4 of them failed to endure contrast due to raised creatinine level, was performed with a multi-detector row helical CT scanner (GE; Medical systems). Imaging parameters included section collimation of 3-5 mm, high-quality mode, table speed of 7.5 mm/s, 50% overlap reconstruct. The first set of images (precontrast phase) was taken before intravenous contrast administration. The second set (portal venous phase) was taken 75 s after intravenous injection of 120 ml of contrast medium. Delayed images, the last set of images, were obtained after 15 min from the time of injection using 120 KV and 280-380 mA.

Post-processing calculations
We had three sets of images: pre attenuation (PA), early contrast enhanced, early attenuation (EA), and delayed contrast enhanced, delayed attenuation (DA) images. All these images were recorded. A circular region of interest, ROI was placed in the center of the adrenal mass on the section where it appeared the largest. The ROI covered approximately half to two-thirds of the mass excluding any region of calcification or degeneration, which luckily was not common in our cases. Then, absolute percentage washout (APW) and the relative percentage washout (RPW) were calculated using the following formulae:   Protocol for MRI MR imaging was performed in 20 patients using a 1.5 T unit (Philips GyroscanIntera) with a phased-array body coil. T1-weighted spin echo axial images were acquired with a TR of 425 ms and a TE of 15 ms. T2-weighted axial and coronal images were acquired with the fast spin echo technique using a TR of 1800 ms and a TE of 100 ms. The section thickness was 4 mm with an intersection gap of 0.4 mm for both T1-and T2-weighted imaging. The matrix size was 204 × 256. Respiratory gating was used for spin echo imaging with flow compensation being used for most of the spin echo images.

Chemical shift imaging (CSI)
To obtain the chemical shift imaging (CSI), we used a breath-holding T1-weighted gradient recalled echo sequence, a TR of 68-160 ms, a TE of 4.9 ms for in-phase images, and a TE of 6.3 ms for opposed-phase images. The flip angle was 30°, the matrix size 128 × 256, and the field of view varied from 320 to 375 cm.  As seen in Fig. 3, the ROC curve, and The obtained chemical shift images were qualitatively assessed to find out if there was significant drop of signal in the lesion on the opposed-phase images compared to the in-phase images or not [7].

Statistical data analysis
Data analysis packages are SPSS version 21 and Medcalc. Qualitative data was presented by number and percentage, while quantitative data was presented by mean, standard deviation.
Statistical tests were done for qualitative data: chisquare test with Fisher's exact. Level of significance was considered if p equal to or below 0.05.
ROC curve was used to detect predictive value of MRI chemical shift and MDCT in adrenal incidentalomas. Youden index test was used to detect cut-off point, specificity, sensitivity, positive, and negative likelihood ratio.

Results
This cross-sectional study included a total of 20 patients with adrenal lesions. Their ages ranged from 19 to 73 years with a mean value of 51.1 ± 15.27. All these patients were found to have either unilateral (18 patients) or bilateral (2 patients) adrenal lesions. Patients known to have a prior malignancy constitute 30% (6 patients) and patients with no history of known malignancy constitute 70% (14 patients) performing the exam due to other reasons like preoperative assessment of abdominal hernia repair, etc.
All the 22 lesions were subjected to size measurement along maximum axial dimensions. Positive group (< 4 cm ) was formed of 13 lesions (59.09%) and negative group (> 4 cm) of 9 lesions (40.91%). Benign adrenal lesions ranged in size from 1 to 10 cm with a mean of 3.05 while malignant lesions ranged from 2 to 6.5 with a mean of 4.64. Fig. 1 and Table 2 show the CT and MRI diagnosis of adrenal lesions.
Thirteen lipid-rich adenomas (59.1%), 7 malignant looking lesions confirmed by a further tissue diagnosis (31.8%) and 2 other lesions (9%) including adrenal cyst and a myelolipoma. Six months follow-up MDCT and MRI studies were performed for all cases.
As shown in the ROC curve (Fig. 2) and area under the curve (Table 3), the CT_RPW are the best indicators  used for CT detection of adrenal lipid-rich adenomas with sensitivity and specificity reaching 100%.
As seen in Fig. 3, the ROC curve, and Table 4, area under the curve, we conclude that the MRI, CSI, and T2 signal are the best indicators used for MRI detecting intracytoplasmic fat content and confirming high lipid content in lipid-rich adenomas with a sensitivity of 100% and specificity of 88.89%. Table 5 shows that a similar diagnostic outcome is obtained from CECT and MRI with CSI of adrenal lesions with a net result of 13 typical lipid-rich adenomas (Fig.  4), 7 malignant lesions (Figs. 5 and 6), 1 simple adrenal cyst (Fig. 7), and 1 suprarenal myelolipoma (Fig. 8).

Discussion
Unexpectedly discovered adrenal masses are found in about 5% of all abdominal CT examinations done in patients not known to have malignancy or endocrine abnormality. The incidence of AIs increases to 9-13% in patients imaged for a known malignancy [8].
The most common AI is benign adrenal adenoma [9]. The aim of this work is to study the role of MRI, specifically chemical shift imaging (CSI), against various MDCT scans (non-enhanced, enhanced, and delayed) in the characterization of incidentally discovered adrenal masses to offer a way for the patients to avoid unnecessary time and money-wasting imaging modalities used to reach a diagnosis of their incidentally discovered adrenal lesions.
In previous studies, a considerable overlap in size between the suprarenal adenomas and the nonadenomatous lesions, especially metastases was demonstrated. This was also noted in our study. We found that the size of < 4 cm has 100.00% sensitivity and 66.6% specificity for being a benign suprarenal adenoma. This is keeping with Angelelli et al. 2013 stating that the size   images (e, f). Diagnosis was right suprarenal malignant neoplastic lesion and was confirmed pathologically to be an adreno-cortical adenocarcinoma criterion yielded 98% sensitivity, but a specificity of 53%, which was too low to be accepted as a discriminator [6]. We concluded that the differentiation between benign and malignant suprarenal lesions cannot rely on the lesions size alone. The size could be a helpful factor to reach the precise diagnosis but still we need more reliable diagnostic criteria.
In our study, we demonstrated that unenhanced CT (UCT) attenuation value of adrenal lesions is a useful imaging modality to characterize lipid-rich adenomas that measure 10 HU or less. A threshold of 10 HU on UCT can be used. The sensitivity and specificity for adenoma characterization were 72.7% and 77.78%, respectively.
This partially agrees with Park et al. 2016 and Kamiyama et al. 2009 who came out with that an attenuation value of a suprarenal mass of less than 10 HU has a sensitivity and specificity for adenoma of 71% and 98%, respectively [10,11].
In this study, the unenhanced value is considered significant to distinguish between adenomas and nonadenomas. This is keeping with Angelelli et al. 2013 stating that no case of non-adenoma lesion with the exception of myelolipomas and cysts has a basal density of less than 10 HU [6].
Previous investigators have demonstrated that the washout attenuation is much larger in adenomas than in non-adenomas [12,13]. In our study, washout values  were also significantly larger in lipid-rich adenomas. We concluded that relative percentage washout (RPW) is a good indicator of the lipid-rich content and benignity of adenomas with area under the curve (AUC) 1.000, sensitivity and specificity of 100% (Table 2) in comparison to absolute percentage washout (APW) with (AUC) 0.883, sensitivity and specificity of 90% and 85%, respectively. This agrees with Park et al. 2016 who stated that RPW is more accurate than APW for diagnosing an adrenal adenoma. If an adenoma measures less than 0 HU in unenhanced CT images, RPW becomes higher than APW. Therefore, it is not uncommon for APW to be below 60% in lipid-rich adenoma [10].
The best MR imaging technique for to characterize suprarenal adenoma is found to be chemical shift imaging (CSI). Its sensitivity and specificity in diagnosis of indeterminate adrenal lesions are reported as 67% and 89-100%, respectively [7]. Subgroup analysis reveals that for adrenal lesions measuring 10-30 HU in unenhanced CT studies, the sensitivity of CSI is as high as 89%, which is comparable to the data retrieved from washout CT. However, its sensitivity declines markedly to be as low as 13% for suprarenal lesions measuring more than 30 HU in unenhanced CT images. This concludes that MR imaging is not a reliable tool for these lesions, although the specificity can be as high as 100% [14]. On that account, adrenal lesions measuring 10-30 HU can be further evaluated with either washout CT or chemical shift MR imaging according to personal and institutional circumstances; lesions with CT attenuation measuring higher than 30 HU are best imaged with washout CT [7].
In our study, CSI, also referred to by dual phase imaging, has proven to be the best indicator for evaluating adenomas from non-adenomas  [15]. There was no signal loss in any of the malignant masses. Some adrenocortical carcinomas have been reported to contain intracellular lipid [16]; however, no cases of lipid-rich adrenocortical carcinomas were included in our study population.

CT versus CS MRI
No statistically significant difference was seen in our study regarding the preference of using CT over CSI (Table 4) unless in certain situations when the use of one diagnostic modality may be superior. This is keeping with literature stating that the sensitivity and specificity of CSI for the differentiation of incidental adrenal lesion are similar to those of unenhanced CT densitometry, at 81-100% and 94-100%, respectively [14,17,18].
Studies have shown that for lipid-rich adenomas, there is effectively no difference between CECT and MRI, as in our study, but CSI might be superior when evaluating lipid-poor adenomas [14]. However, Park et al. 2016 showed that CSI might be useful only when the unenhanced CT attenuation is less than 30 HU and that delayed enhanced CT can characterize additional hyperattenuating adrenal masses that cannot be identified with chemical shift MRI [10].

Limitations
Some limitations that were encountered in our study included our study design which depended on accidentally discovered adrenal masses in patients undergoing abdominal studies for irrelevant causes which limited our sample size. Also, other limitations included the wide varieties of pathologies that cause incidentally discovered adrenal masses. This variation affected the sample size in each pathological category. Further study on a larger number of patients may be of value in retrieving better statistical results.

Conclusion
There is no statistically significant difference in the diagnostic capability of the CSI versus the CT washout technique. Both modalities could be conducted according to specific patient preferences and/or limitations with comparable highly accurate outcomes.