Characterization of Adrenal Masses With Diffusion-Weighted Imaging

Transcription

1 Genitourinary Imaging Original Research Sandrasegaran et al. Diffusion-Weighted Imaging of Adrenal Masses Genitourinary Imaging Original Research Kumaresan Sandrasegaran 1 Aashish A. Patel 1 Raja Ramaswamy 1 Victor P. Samuel 1 Benjamin G. Northcutt 1 Mark S. Frank 1 Isaac R. Francis 2 Sandrasegaran K, Patel AA, Ramaswamy R, et al. Keywords: adenoma, adrenal gland, chemical shift imaging, diffusion-weighted imaging, MRI DOI: /AJR Received March 9, 2010; accepted after revision December 14, K. Sandrasegaran holds a research grant from Siemens Healthcare. 1 Department of Radiology, Indiana University School of Medicine, 550 N University Blvd, UH 0279, Indianapolis, IN Address correspondence to K. Sandrasegaran (ksandras@iupui.edu). 2 Department of Radiology, University of Michigan, Ann Arbor, MI. AJR 2011; 197: X/11/ American Roentgen Ray Society Characterization of Adrenal Masses With Diffusion-Weighted Imaging OBJECTIVE. The purpose of this article is to assess the role of diffusion-weighted MRI in characterizing adrenal masses. MATERIALS AND METHODS. A retrospective review of the MRI database from August 2007 to July 2009 was performed. The MRI examinations of 48 patients, with 49 lesions, were reviewed independently and blindly by two experienced abdominal radiologists who measured the signal intensities on in-phase and opposed-phase T1-weighted imaging and apparent diffusion coefficient (ADC). ADC measurements and quantitative parameters of chemical shift imaging (signal intensity index and adrenal-to-spleen ratio) were assessed separately and in combination. Lesions with indeterminate signal intensity index (< 16.5%) were considered benign if ADC was greater than or equal to mm 2 /s and malignant if ADC was less than mm 2 /s. Stepwise logistic regression analysis and receiver operating characteristic curves analysis were performed. RESULTS. There were 12 malignant and 37 benign lesions. On multivariate analysis, the only significant predictors of lesion status were signal intensity index from reviewer 2 (p = 0.05) and lesion size (p = 0.04); ADC values were not found to be useful. On receiver operating characteristic curve analysis, there was no significant difference in area under the curve for ADC, signal intensity index, adrenal-to-spleen ratio, or the combined signal intensity index and ADC assessment. For lesions that were indeterminate according to signal intensity index, ADC values greater than mm 2 /s were found only in benign lesions, and nine of 11 lesions with ADC less than mm 2 /s were malignant. CONCLUSION. In general, ADC values are not useful in differentiating adrenal lesions. However, when ADC values are applied to lesions that are indeterminate on signal intensity index, they may help in differentiating a subset of benign and malignant lesions. A drenal lesions are seen in up to 6% of patients undergoing imaging studies [1]. The most common adrenal nodule is a nonhyperfunctioning cortical adenoma. Other adrenal lesions to consider include metastases, pheochromocytomas, and adrenal cortical carcinomas. Differentiation of benign and malignant lesions is important. A Hounsfield unit density of less than 10 units on unenhanced CT is highly specific for a benign lesion [2, 3]. However, up to 30% of benign lesions have a higher density [4]. The CT washout on the minute delayed phase, compared with the peak enhancement on the venous phase (the so-called washout scan), is also useful in such patients for further characterizing nodules that are hyperdense on unenhanced CT scans [2, 5, 6]. The three-phase CT entails the use of an iodinated IV contrast agent and additional radiation exposure. There is considerable clinical and public interest, and some concern, in the risk of radiation from imaging and CT in particular [7 9]. Chemical shift MRI has been found to be useful in assessing adrenal tumors [10 20]. However, lipid-poor adenomas cannot be differentiated from malignant lesions using this technique. It would be useful to have an additional MRI technique for characterizing adrenal lesions. Diffusion-weighted imaging (DWI) has been found to be of some use in differentiating benign and malignant lesions in the brain, breast, prostate gland, kidneys, and liver [21 25]. In general, the apparent diffusion coefficient (ADC) values of malignant lesions are lower than those of benign lesions. Possible reasons for this finding include increased cellularity, increased nuclear-to-cytoplasmic ratio, and reduced extracellular space in malignant 132 AJR:197, July 2011

2 Diffusion-Weighted Imaging of Adrenal Masses lesions [26, 27]. To our knowledge, only two articles have assessed the role of DWI in assessing adrenal nodules [28, 29]. We wished to assess the usefulness of ADC values in differentiating benign and malignant lesions and to determine whether there is additional benefit that ADC measurements could provide in the subset of patients with indeterminate adrenal masses on chemical shift imaging. Materials and Methods Patients We performed a retrospective study that was HIPAA compliant. Institutional review board permission was granted for a retrospective review of clinical and imaging data with waiver of informed consent. A search of the departmental MRI database revealed 137 consecutive patients who underwent MRI examinations between August 2007 and July 2009 with a report of an adrenal mass. Patients were excluded if they did not undergo DWI (n = 41) or if the adrenal nodule was smaller than 1 cm (n = 11). In our previous experience, we found that lesions smaller than 1 cm were hard to detect on ADC maps. In addition, a previous study of DWI of adrenal lesions excluded lesions smaller than 1 cm [29]. Patients with adrenal nodules that did not have histology or follow-up imaging over at least 12 months (n = 29) were also excluded. A 12-month follow-up was chosen because lesions that remain stable over this period can reasonably be deemed to be benign. Prior studies have used stability over 6 months [18, 28 30] or 12 months [30 32] to indicate benign lesions. Patients with artifacts on DWI (n = 5) that made accurate ADC measurements impossible were excluded. One patient with myelolipoma and two patients with cystic adrenal masses were also excluded. This left a cohort of 48 patients with 49 lesions (Fig. 1). There were 37 benign lesions, composed of 36 nonfunctioning adenomas, and one pheochromocytoma. There were 12 malignant lesions, including six metastases from renal cell carcinoma (RCC) and three from hepatocellular carcinoma (HCC). Malignant lesions also included two pheochromocytomas and one adrenal cortical carcinoma. The types of malignant lesions reflected the referral pattern for MRI examinations at our institution. One patient with HCC had bilateral metastases. All malignant lesions, except one, had biopsy confirmation of histology. In the one patient without histologic proof, the primary diagnosis was RCC; the adrenal nodule appeared during follow-up, increased in size within 3 months, and was deemed to be a metastasis. Six patients with benign lesions had biopsy confirmation. The remaining lesions were stable over follow-up of more than 12 months (mean, 22.3 months; range, months). 11 lesions with biopsy confirmation 11 patients with 12 malignant lesions 1 new lesion, rapid increase in size* 9 metastases (6 RCC, 3 HCC) 2 malignant pheochromocytomas 1 adrenal cortical carcinoma MRI Examination MRI examinations were performed in a supine position using a 1.5-T MRI scanner (Avanto, Siemens Healthcare) and a body phased-array coil. The standard imaging protocol consisted of in- and opposed-phase gradient-echo imaging and 3D T1- weighted gradient-echo sequences. Ten to twenty milliliters of gadobenate dimeglumine (Multi- Hance, Bracco Diagnostics) was injected IV at a rate of 2 ml/s using a power injector (Spectris Solaris, Medrad). Before the administration of contrast medium, a single-shot echo-planar DWI sequence was acquired according to parameters given in Table 1. The acquisition was performed during respiratory gating, using b values of 0 or 50, 400 or 500, and 800 s/mm 2. From our prior experience, we found that the b value of 800 s/mm 2 gave images with adequate signal-to-noise ratios and without artifacts. Parallel imaging technique (i.e., generalized autocalibrating partially parallel acquisition) was used with an acceleration factor of 2. ADC maps were calculated monoexponentially with all b value sequences using the scanner software (Syngo vb15 or vb13, Siemens Healthcare). 48 patients 31 lesions with > 12- month follow-up 37 patients with 37 benign lesions 1 new lesion, rapid increase in size* 36 nonhyperfunctioning cortical adenomas 1 benign pheochromocytoma Fig. 1 Flow diagram of patients in our study. Asterisk indicates that this adrenal lesion appeared during follow-up of patient with renal cell carcinoma (RCC) and increased in size by 50% in 3 months. It was considered to be metastasis. HCC = hepatocellular carcinoma. Image Review Two abdominal radiologists with 13 and 5 years of experience in reading abdominal MRI independently reviewed the MRI scans for site and size of lesion; they were blinded to the final diagnosis or follow-up. Regions of interest (ROIs) were placed on the adrenal masses. The ROIs were larger than 1 cm 2. In tumors that were larger than 2 cm, multiple ROIs were placed in the parts of adrenal lesions that showed enhancement on contrast-enhanced sequences. Necrotic regions, if any, were avoided. In addition, the signal intensity of the adrenal lesion and spleen were measured on in phase and opposed phase imaging. The adrenal signal intensity index was calculated as follows: [(signal intensity on in phase imaging signal intensity on opposed phase imaging) / (signal intensity on in phase imaging)] 100%. The adrenal-to-spleen ratio (ASR) was calculated as follows: (adrenal signal intensity on opposed phase imaging / spleen signal intensity on opposed phase imaging) / (adrenal signal intensity on in phase imaging / spleen signal intensity on in phase imaging). Lesions with adrenal signal intensity index of less than 16.5% were considered indeterminate [17, 18], and the additional benefit of ADC values in characterizing lesions deemed to be indeterminate on chemical shift imaging (combined assessment) was analyzed. These lesions were categorized as benign if the ADC value was greater than or equal to mm 2 /s and malignant if the ADC value was less than mm 2 /s. Lesions with signal intensity index greater than 16.5% were considered benign, regardless of ADC value, for this combined assessment. Statistical Analysis The measured values were checked for distribution before statistical analysis using the Kolmogorov-Smirnov test. The correlations between the two reviewers were performed using Shrout- Fleiss random effect intraclass correlation coefficient [33] and Lin s concordance coefficient. Univariate analysis for differentiating benign from malignant lesions was performed using the Student t test or Mann-Whitney test for covariates with or without normal distribution, respectively. Logistic regression analysis was performed to find the best predictors of the nature of adrenal lesion (benign vs malignant). The measurements of the two reviewers were treated independently. First, the parameters were assessed univariately as independent AJR:197, July

3 Sandrasegaran et al. TABLE 1: MRI Parameters Sequence TR/TE (ms) Flip Angle ( ) Slice Thickness (mm) Interslice Gap (mm) No. of Excitations Receiver Bandwidth (Hz/Pixel) Matrix T1-weighted 2D gradient-echo (FLASH) 123/2.2 (opposed phase), 4.93 (in phase) x 135 Diffusion-weighted imaging a 1500/ x 115 3D fat-suppressed gradient-echo (VIBE) b 4.98/ NA x 144 Note NA = not applicable, VIBE = volumetric interpolated breath-hold examination. a Three b values used for diffusion-weighted images: 0 or 50, 400 or 500, and 800 s/mm 2. b Sequence was used after IV administration of contrast agent Ben Mal ADC Ben Mal variables in a logistic regression model where the dependent variable was the nature of the adrenal lesion. Parameters with a p value less than or equal to 0.10 at univariate analysis were assessed with multivariate analysis. Stepwise forward analysis was performed in the model, until all remaining variables had p values less than or equal to Receiver operating characteristic (ROC) curve analysis was used to evaluate the usefulness of ADC measurements, both in isolation and in combination with chemical shift imaging, in evaluating the nature of adrenal masses. For ROC curve analysis, lesions were categorized as benign if the ADC value was greater than mm 2 /s, signal intensity index was greater than 16.5%, or if the ASR was less than Lesions with ADC value equal to or lower than mm 2 /s, signal intensity index equal to or lower than 16.5%, or ASR greater than 0.71 were deemed malignant. The categorization of signal intensity index and ASR followed previously reported criteria [18, 20, 34]. A p value less than 0.05 was used to indicate statistical significance. Statistical analysis was performed using MedCalc (version 11.1, MedCalc Software) (for ROC curves) and SPSS (version 17.0, SPSS). SII Ben Mal ASR Fig. 2 Box-and-whisker plot of apparent diffusion coefficient (ADC), adrenal signal intensity index (SII), and adrenalto-spleen ratio (ASR) measurements. Boxes represent interquartile range. Whiskers represent range of all values. Horizontal line within box is median value. Circles and squares indicate outliers and far-outliers, respectively. ADC values are given in 10 3 mm 2 /s. Signal intensity index is given as numeric value (instead of percentage), so that range of values will be similar to other parameters. There is overlap of boxes of benign (BEN) and malignant (MAL) lesions for all parameters measured. Results Patients and Lesions The 48 patients analyzed had 37 benign and 12 malignant lesions. The mean age of all subjects was 61 years (range, years). There were 22 women and 26 men. There was no difference in age or sex distribution between the benign and malignant groups. The mean (± SD) sizes of benign and malignant lesions were 1.9 ± 0.6 cm and 5.0 ± 3.4 cm, respectively (p < 0.01). Thirty-six lesions were on the left, and 13 were on the right. The indications for MRI scans were characterization of a lesion seen on CT (n = 10) and tumor staging (n = 15). In 23 patients, adrenal lesions were found incidentally on scans performed for abdominal pain (n = 8), screening for cancer in patients with cirrhosis (n = 8), follow-up of cystic pancreatic lesions (n = 4), and miscellaneous reasons (n = 3). Reviewer Agreement The intraclass correlation coefficients for ADC value, signal intensity index, and ASR between the two reviewers were 0.47 (95% CI, ), 0.75 (95% CI, ), and 0.61 (95% CI, ), respectively. The Lin s concordance coefficients for ADC, signal intensity index, and ASR between the two reviewers were 0.47 (95% CI, ), 0.74 (95% CI, ), and 0.61 (95% CI, ). The results suggest moderate correlation or concordance between the reviewers for ADC, and good or excellent correlation or concordance for signal intensity index and ASR. Univariate Analysis Figure 2 is a box-and-whisker plot of the range of ADC values for the benign and malignant groups. When all lesions were analyzed, there was a substantial overlap of ADC values between benign and malignant groups. However, all but five of 24 malignant lesions had ADC values less than mm 2 /s (Fig. 3). On the other hand, all 27 lesions with ADC values greater than mm 2 /s were benign (Fig. 4). Analysis of covariance results are given in Table 2. For both reviewers, there was a significant difference between benign and malignant groups for ADC and signal intensity index. For ASR, there was a significant difference between the two groups for reviewer 2, but not for reviewer 1. Multivariate Analysis and ROC Curves On stepwise logistic regression analysis, the only significant parameters that could predict the benignity or malignancy of an adrenal nodule were signal intensity index performed by reviewer 2 (p = 0.05) and the size of the lesion (p = 0.04). The other parameters, including ADC measurements from both reviewers, were eliminated from the model. Figure 5 shows the ROC curves. Table 3 gives the results of ROC curve analysis. The area under the curve (AUC) for all the variables ranged between 0.70 and There was no statistically significant difference between the AUC of ADC and signal intensity index (p = 0.49), of ADC and ASR (p = 0.12), or of signal intensity index and ASR (p = 0.06). 134 AJR:197, July 2011

4 Diffusion-Weighted Imaging of Adrenal Masses A combined evaluation was performed using ADC values to categorize lesions that were indeterminate on the signal intensity index. In lesions with a signal intensity index less than 16.5%, an ADC value of mm 2 /s or more was considered to indicate a benign lesion (Fig. 6), and an ADC value of A D Fig year-old man with hepatocellular carcinoma (HCC) presenting with 2.2-cm left adrenal nodule. A and B, Axial in phase (A) and opposed phase (B) images show left adrenal nodule (arrow) with no convincing signal drop-off on opposed phase image (B). Signal intensity index was estimated as 12% and 8% by reviewers. C, Axial arterial phase contrast-enhanced image shows adrenal nodule (arrow) and ill-defined hypovascular right liver tumor (arrowhead). Note nodular liver and splenomegaly consistent with cirrhosis and portal hypertension. D and E, Axial diffusion-weighted images (DWI) with b values of 50 s/mm 2 (D) and 400 s/mm 2 (E) show minimal reduction in signal as b value increases. This suggested low apparent diffusion coefficient (ADC) value, which was measured as 0.75 and mm 2 /s by reviewers. Mass (arrow) was proven to be HCC metastases on biopsy. Liver mass (HCC) was well seen as hyperintense mass (arrowhead). In our experience, hyperintense solid mass on DWI in cirrhotic liver is almost always HCC, even if contrast-enhanced images are indeterminate. In this patient, signal intensity index was indeterminate but ADC was less than mm 2 /s, suggesting malignant lesion less than mm 2 /s was considered to indicate a malignant lesion (Fig. 3). This evaluation increased the AUC from 0.76 to 0.80, but the difference was not significant (p = 0.40). Table 4 gives the breakdown of lesions that were indeterminate of signal intensity index. B E Discussion The usefulness of CT in characterizing adrenal lesions is well established. Nevertheless, in our practice, a three-phase CT assessing the unenhanced density and washout on the delayed contrast-enhanced phase is rarely performed. The reasons for this are not entirely clear but C A B C Fig year-old man with ampullary cancer found to have 3-cm left adrenal nodule. A and B, Axial in phase (A) and opposed phase (B) images of nodule (arrow) show visible signal drop-off on opposed phase image (B). Signal intensity index was calculated as 73% and 66% by two reviewers. C, Axial apparent diffusion coefficient (ADC) map shows hyperintense nodule (arrow). ADC values were 1.54 and mm 2 /s. Nodule was stable over 2.5 years of follow-up, consistent with adenoma. AJR:197, July

5 Sandrasegaran et al. TABLE 2: Univariate Analysis Measurement Benign Lesions Malignant Lesions p 100 Apparent diffusion coefficient (10 3 mm 2 /s) Reviewer (0.641) (0.285) 0.02 Reviewer (0.732) (0.283) 0.01 Adrenal signal intensity index (%) Reviewer (27.7) 12.8 (14.6) 0.01 Reviewer (25.3) 12.8 (20.2) < 0.01 Adrenal-to-spleen ratio Reviewer (0.55) 0.82 (0.26) 0.15 Reviewer (0.38) 0.88 (0.25) 0.01 Note Except for p values, data are mean (SD). may include concerns about radiation dose or reimbursement for a second CT study performed soon after an initial CT (that showed, but did not characterize, the adrenal lesion). In addition, MRI is increasingly used as the primary abdominal imaging modality, especially for the investigation of abdominal pain thought to be of biliary or pancreatic origin, screening of HCC in patients with cirrhosis, and characterization of renal or hepatic masses seen on sonography. In these clinical situations, there may be no antecedent CT. It is not surprising that 23 of the 48 patients in our cohort had an incidental adrenal nodule that was seen for the first time on MRI. Traditionally, chemical shift imaging has been used for characterizing adrenal nodules. Chemical shift imaging, assessed qualitatively or using quantitative signal intensity ratios, has been reported to have specificity of % and sensitivity of % in characterizing adrenal adenomas [10 17]. More recent studies [18 20] have reported lower sensitivities or specificities of chemical shift imaging in differentiating benign and malignant adrenal lesions. The specificity of this technique depends on the type of malignant lesions in the cohort. HCCs and RCCs are known to contain intracytoplasmic fat [35 38]. Therefore, TABLE 3: Receiver Operating Characteristic Analysis Sensitivity (%) Specificity (%) 100 Fig. 5 Receiver operating characteristic curves plotting sensitivity (y-axis) and 1 specificity (x-axis) of various MRI parameters. Curves were constructed using measurements from both reviewers. Solid black line represents apparent diffusion coefficient (ADC) values. Dotted black line represents signal intensity index. Dashed black line represents adrenal-to-spleen ratio. Dot-and-dash black line represents combined assessment with ADC and signal intensity index (see Materials and Methods for details). Gray line represents area-under-curve (AUC) of There were no significant differences in AUC for four parameters (Table 3). loss of signal on opposed-phase images may be seen in these metastases, mimicking adrenal adenomas [39]. In many centers, DWI has become part of the standard protocol in the abdominal MRI examinations. Therefore, we wished to determine whether DWI, on its own or in combination with quantitative chemical shift imaging, helped in differentiating benign and malignant adrenal lesions. To our knowledge, only two other articles have described the usefulness of DWI in this clinical situation [28, 29]. Those studies assessed the usefulness of DWI in evaluation of adrenal masses, but did not specifically investigate the additional benefit of this technique for indeterminate lesions on chemical shift imaging. The MRI parameters for DWI and chemical shift imaging were similar between the study of Tsushima et al. [28] and the current study. The DWI technique used by Miller et al. [29] was multiple breath-hold technique instead of a free breathing technique. These prior studies did not show a difference in ADC measurements between benign and malignant lesions. In contrast, our study found a significant difference between the mean ADC values of benign and malignant lesions (p = 0.02 or 0.01, for the two reviewers) on univariate analysis. Nevertheless, despite a statistically significant difference in the mean values of benign and malignant lesions, there was substantial overlap between the two groups, making ADC measurements of little value in individual cases. On multivariate analysis, ADC values were not found to reliably predict whether individual lesions were benign or malignant. The AUC values of the ADC curves were not significantly different from those of chemical shift imaging parameters, signal intensity index, and ASR. The study by Miller et al. found a significantly higher AUC for signal intensity index (0.95) compared with ADC (0.55). This difference may be related to the proportion of tumors that were lipid-poor adenomas or fat-containing metastases, such as those from HCC or RCC. If the cohort contained a large proportion of these entities, it would be expect- Parameter Area Under the Curve (95% CI) Best Cutoff a Percentage Sensitivity (95% CI) Percentage Specificity (95% CI) Apparent diffusion coefficient (10 3 mm 2 /s) 0.81 ( ) > ( ) 81.1 ( ) Signal intensity index (%) 0.76 ( ) > ( ) 68.9 ( ) Adrenal-to-spleen ratio 0.70 ( ) < ( ) 59.5 ( ) Combined diffusion-weighted imaging/signal intensity 0.80 ( ) index assessment b Note Dashes ( ) indicate not applicable. a The best cutoff value was generated by the statistical software for value with the best combination of sensitivity and specificity for determining benign lesion. b The combined assessment used apparent diffusion coefficient measurements to categorize lesions that were indeterminate on signal intensity index, as explained in Materials and Methods. 136 AJR:197, July 2011

6 Diffusion-Weighted Imaging of Adrenal Masses A B C Fig year-old woman with pancreatic cystic lesion (proven to be stable on follow-up) and 4-cm left adrenal nodule. A and B, Axial in-phase (A) and opposed-phase (B) images show hypointense left adrenal nodule (arrow) without visible signal drop-off on opposed-phase image (B). Signal intensity index was estimated as 12% and 1% by reviewers. C, Axial apparent diffusion coefficient (ADC) map shows moderately hyperintense nodule (arrow). ADC measurements of reviewers were 1.82 and mm 2 /s. Nodule was stable on 2 years follow-up, consistent with adenoma. Signal intensity index was indeterminate, but high ADC values indicated benign lesion. TABLE 4: Categorization of Lesions Indeterminate on Chemical Shift Imaging Apparent Diffusion Coefficient Value Benign Lesions Malignant Lesions < 1.0 x 10 3 mm 2 /s x 10 3 mm 2 /s 8 3 > 1.5 x 10 3 mm 2 /s 4 0 Note Indeterminate lesions are defined as those with signal intensity index < 16.5%. ed that the value of signal intensity index in differentiating benign and malignant lesions would be reduced. We did not find a statistically significant benefit to using ADC values to assess all lesions that were indeterminate on quantitative chemical shift imaging. The combined assessment of signal intensity index and ADC did not result in a significant increase in AUC, when compared with the AUC of signal intensity index only ( ; p = 0.40). Despite these negative results, we think that ADC measurements may have a role to play in characterizing a subset of lesions that are indeterminate on chemical shift imaging. Lesions that had ADC values greater than mm 2 /s were benign, and nine (82%) of 11 lesions with ADC values less than mm 2 /s were malignant. Although this study was primarily performed to assess the role of ADC measurements, some results of signal intensity index and ASR need to be discussed. A study by Fujiyoshi et al. [17] established 16.5% as the signal intensity index cutoff value to be used for differentiating benign and malignant lesions according to the signal intensity index. Several subsequent studies [18, 20, 29, 34], and our study, used this cutoff value. However, on ROC examination, we found that in our cohort the best cutoff value of signal intensity index was 23%. Differences in the cutoff values for signal intensity index may, in part, reflect the fat content of some of the metastatic lesions in our study group. In addition, quantitative measurements from chemical shift imaging are dependent on parameters used in chemical shift imaging. Increasing T1 weighting, by increasing the flip angle, may result in overestimation of fat content [40]. On stepwise logistic regression, the signal intensity index was found to be the only reliable parameter for predicting whether an adrenal lesion was benign or malignant. ASR was not found to be a useful predictor in this study, which is contrary to earlier reports [12, 13]. Differences in patient population may be the reason for this. Eight of the patients with incidental adrenal adenomas had advanced cirrhosis, and the MRI examination was being performed for HCC screening. Alteration in iron content of spleen as a result of portal hypertension in these patients may have affected ASR measurements. We recognize several limitations of our study. The number of patients with malignancy was small. The referral pattern for MRI resulted in metastatic lesions from primary tumors with potentially higher fat content in our cohort, than the more commonly seen adrenal metastases from lung or breast cancers [41]. We also acknowledge that a standardized method for DWI has not yet been established. Currently, several protocols, including free breathing, respiratory triggered, and breath-hold single-shot echoplanar sequences, are used. The maximum b values used in prior studies have varied from 500 to over 1000 s/mm 2. We assessed ADC monoexponentially and did not obtain images with multiple small b values (< 100 s/ mm 2 ) to separately assess the effects of perfusion and diffusion. The correlation between the ADC measurements of the two reviewers was moderate (ρ = 0.47). This may indicate a difficulty in measuring ADC values in small lesions, given the low resolution of the ADC map. In conclusion, we think that ADC measurements, used in isolation, are not helpful in differentiating benign and malignant lesions. However, lesions that are indeterminate on adrenal signal intensity index and that have an ADC value of more than mm 2 /s may be considered benign. Lesions that are indeterminate on quantitative chemical shift imaging and that have an ADC value of less than mm 2 /s are likely to be malignant. References 1. Boland GW, Blake MA, Hahn PF, Mayo-Smith WW. Incidental adrenal lesions: principles, techniques, and algorithms for imaging characterization. Radiology 2008; 249: Caoili EM, Korobkin M, Francis IR, et al. Adrenal masses: characterization with combined unenhanced and delayed enhanced CT. Radiology 2002; 222: Korobkin M, Brodeur FJ, Yutzy GG, et al. Differentiation of adrenal adenomas from nonadenomas using CT attenuation values. AJR 1996; 166: 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 1998; 171: Blake MA, Kalra MK, Sweeney AT, et al. Distinguishing benign from malignant adrenal masses: multi-detector row CT protocol with 10-minute AJR:197, July

7 Sandrasegaran et al. delay. Radiology 2006; 238: Caoili EM, Korobkin M, Francis IR, Cohan RH, Dunnick NR. Delayed enhanced CT of lipid-poor adrenal adenomas. AJR 2000; 175: Fazel R, Krumholz HM, Wang Y, et al. Exposure to low-dose ionizing radiation from medical imaging procedures. N Engl J Med 2009; 361: Sodickson A, Baeyens PF, Andriole KP, et al. Recurrent CT, cumulative radiation exposure, and associated radiation-induced cancer risks from CT of adults. Radiology 2009; 251: Brenner DJ, Hall EJ. Computed tomography: an increasing source of radiation exposure. N Engl J Med 2007; 357: Tsushima Y, Ishizaka H, Matsumoto M. Adrenal masses: differentiation with chemical shift, fast low-angle shot MR imaging. Radiology 1993; 186: Bilbey JH, McLoughlin RF, Kurkjian PS, et al. MR imaging of adrenal masses: value of chemical-shift imaging for distinguishing adenomas from other tumors. AJR 1995; 164: Korobkin M, Lombardi TJ, Aisen AM, et al. Characterization of adrenal masses with chemical shift and gadolinium-enhanced MR imaging. Radiology 1995; 197: Mayo-Smith WW, Lee MJ, McNicholas MM, Hahn PF, Boland GW, Saini S. Characterization of adrenal masses (< 5 cm) by use of chemical shift MR imaging: observer performance versus quantitative measures. AJR 1995; 165: Outwater EK, Siegelman ES, Radecki PD, Piccoli CW, Mitchell DG. Distinction between benign and malignant adrenal masses: value of T1- weighted chemical-shift MR imaging. AJR 1995; 165: Schwartz LH, Panicek DM, Koutcher JA, et al. Adrenal masses in patients with malignancy: prospective comparison of echo-planar, fast spinecho, and chemical shift MR imaging. Radiology 1995; 197: Heinz-Peer G, Honigschnabl S, Schneider B, Niederle B, Kaserer K, Lechner G. Characterization of adrenal masses using MR imaging with histopathologic correlation. AJR 1999; 173: Fujiyoshi F, Nakajo M, Fukukura Y, Tsuchimochi S. Characterization of adrenal tumors by chemical shift fast low-angle shot MR imaging: comparison of four methods of quantitative evaluation. AJR 2003; 180: Park BK, Kim CK, Kim B, Lee JH. Comparison of delayed enhanced CT and chemical shift MR for evaluating hyperattenuating incidental adrenal masses. Radiology 2007; 243: Jhaveri KS, Wong F, Ghai S, Haider MA. Comparison of CT histogram analysis and chemical shift MRI in the characterization of indeterminate adrenal nodules. AJR 2006; 187: Israel GM, Korobkin M, Wang C, Hecht EN, Krinsky GA. Comparison of unenhanced CT and chemical shift MRI in evaluating lipid-rich adrenal adenomas. AJR 2004; 183: Higano S, Yun X, Kumabe T, et al. Malignant astrocytic tumors: clinical importance of apparent diffusion coefficient in prediction of grade and prognosis. Radiology 2006; 241: Partridge SC, DeMartini WB, Kurland BF, Eby PR, White SW, Lehman CD. Quantitative diffusion-weighted imaging as an adjunct to conventional breast MRI for improved positive predictive value. AJR 2009; 193: Yoshizako T, Wada A, Hayashi T, et al. Usefulness of diffusion-weighted imaging and dynamic contrast-enhanced magnetic resonance imaging in the diagnosis of prostate transition-zone cancer. Acta Radiol 2008; 49: Taouli B, Thakur RK, Mannelli L, et al. Renal lesions: characterization with diffusion-weighted imaging versus contrast-enhanced MR imaging. Radiology 2009; 251: Sandrasegaran K, Akisik FM, Lin C, Tahir B, Rajan J, Aisen AM. The value of diffusion-weighted imaging in characterizing focal liver masses. Acad Radiol 2009; 16: Humphries PD, Sebire NJ, Siegel MJ, Olsen OE. Tumors in pediatric patients at diffusion-weighted MR imaging: apparent diffusion coefficient and tumor cellularity. Radiology 2007; 245: Uhl M, Altehoefer C, Kontny U, Il yasov K, Buchert M, Langer M. MRI-diffusion imaging of neuroblastomas: first results and correlation to histology. Eur Radiol 2002; 12: Tsushima Y, Takahashi-Taketomi A, Endo K. Diagnostic utility of diffusion-weighted MR imaging and apparent diffusion coefficient value for the diagnosis of adrenal tumors. J Magn Reson Imaging 2009; 29: Miller FH, Wang Y, McCarthy RJ, et al. Utility of diffusion-weighted MRI in characterization of adrenal lesions. AJR 2010; 194:459; [web] W179 W Dunnick NR, Korobkin M. Imaging of adrenal incidentalomas: current status. AJR 2002; 179: McNicholas MM, Lee MJ, Mayo-Smith WW, Hahn PF, Boland GW, Mueller PR. An imaging algorithm for the differential diagnosis of adrenal adenomas and metastases. AJR 1995; 165: Szolar DH, Korobkin M, Reittner P, et al. Adrenocortical carcinomas and adrenal pheochromocytomas: mass and enhancement loss evaluation at delayed contrast-enhanced CT. Radiology 2005; 234: Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull 1979; 86: Halefoglu AM, Yasar A, Bas N, Ozel A, Erturk SM, Basak M. Comparison of computed tomography histogram analysis and chemical-shift magnetic resonance imaging for adrenal mass characterization. Acta Radiol 2009; 50: Balci NC, Befeler AS, Bieneman BK, Fattahi R, Saglam S, Havlioglu N. Fat containing HCC: findings on CT and MRI including serial contrast-enhanced imaging. Acad Radiol 2009; 16: Martin J, Sentis M, Zidan A, et al. Fatty metamorphosis of hepatocellular carcinoma: detection with chemical shift gradient-echo MR imaging. Radiology 1995; 195: Helenon O, Chretien Y, Paraf F, Melki P, Denys A, Moreau JF. Renal cell carcinoma containing fat: demonstration with CT. Radiology 1993; 188: Strotzer M, Lehner KB, Becker K. Detection of fat in a renal cell carcinoma mimicking angiomyolipoma. Radiology 1993; 188: Sydow BD, Rosen MA, Siegelman ES. Intracellular lipid within metastatic hepatocellular carcinoma of the adrenal gland: a potential diagnostic pitfall of chemical shift imaging of the adrenal gland. AJR 2006; 187:W550 W Al-Hawary MM, Francis IR, Korobkin M. Noninvasive evaluation of the incidentally detected indeterminate adrenal mass. Best Pract Res Clin Endocrinol Metab 2005; 19: Abrams HL, Spiro RS, Goldstein NT. Metastases in carcinoma; analysis of 1000 autopsied cases. Cancer 1950; 3: AJR:197, July 2011