- Email: [email protected]
Using the modified Dixon technique to evaluate incidental adrenal lesions on 3 T MRI
Radiología. 2018;60(6):485---492
www.elsevier.es/rx
ORIGINAL REPORT
Using the modified Dixon technique to evaluate incidental adrenal lesions on 3 T MRI夽 A. Ecénarro-Montiel a,∗ , S. Baleato-González a , M.I. Santiago-Pérez b , J. Sánchez-González c , P. Montesinos c , R. García-Figueiras a a
Servicio de Radiología, Hospital Clínico Universitario, Santiago de Compostela, Spain Dirección Xeral de Saúde Pública, Consellería de Sanidade, Xunta de Galicia, Santiago de Compostela, Spain c Clinic Scientist, Philips Iberia, Madrid, Spain b
Received 9 January 2018; accepted 1 June 2018
KEYWORDS Adrenal glands; Adrenal adenoma; Adrenal incidentaloma; Magnetic resonance imaging; Adrenal tumors; Differential diagnosis
Abstract Objectives: To use the mDIXON-Quant sequence to quantify the fat fraction of adrenal lesions discovered incidentally on CT studies. To analyze the relation between the signal loss between in-phase and out-of-phase T1-weighted sequences and the fat fraction in mDIXON-Quant. To compare the sensitivity and specificity of the two methods for characterizing adrenal lesions. Material and methods: This prospective descriptive study included 31 patients with incidentally discovered adrenal lesions evaluated with 3 T MRI using in-phase and out-of-phase T1-weighted sequences and mDIXON-Quant; the fat fraction of the adrenal lesions was measured by mDIXONQuant and by calculating the percentage of signal loss between in-phase and out-of-phase T1-weighted sequences. Results: The percentage of signal loss was significantly higher in the group of patients with adenoma (61.3 ± 20.4% vs. 5.1 ± 5.8% in the group without adenoma, p < 0.005). The mean fat fraction measured by mDIXON-Quant was also higher for the adenomas (26.9 ± 10.8% vs. 3.4 ± 3.0%, p < 0.005). The area under the ROC curve was 0.99 (0.96---1.00) for the percentage of signal loss and 0.98 (0.94---1.00) for the fat fraction measured by mDIXON-Quant. The cutoffs obtained were 24.42% for the percentage of signal loss and 9.2% for the fat fraction measured by mDIXON-Quant. The two techniques had the same values for diagnostic accuracy: sensitivity 96% (79.6---99.9), specificity 100% (39.8---100.0), positive predictive value 100% (85.8---100.0), and negative predictive value 80% (28.4---99.5). Conclusion: The fat fraction measured by the modified Dixon technique can differentiate between adenomas and other adrenal lesions with the same sensitivity and specificity as the percentage of signal loss between in-phase and out-of-phase T1-weighted sequences. © 2018 SERAM. Published by Elsevier Espa˜ na, S.L.U. All rights reserved.
夽
Please cite this article as: Ecénarro-Montiel A, Baleato-González S, Santiago-Pérez MI, Sánchez-González J, Montesinos P, García-Figueiras R. Valoración mediante Dixon modificado de las lesiones suprarrenales incidentales en RM 3T. Radiología. 2018;60:485---492. ∗ Corresponding author. E-mail address: [email protected] (A. Ecénarro-Montiel). 2173-5107/© 2018 SERAM. Published by Elsevier Espa˜ na, S.L.U. All rights reserved.
486
PALABRAS CLAVE Glándulas suprarrenales; Adenoma suprarrenal; Incidentaloma suprarrenal; Resonancia magnética; Neoplasias suprarrenales; Diagnóstico diferencial
A. Ecénarro-Montiel et al.
Valoración mediante Dixon modificado de las lesiones suprarrenales incidentales en RM 3T Resumen Objetivo: Cuantificar mediante secuencia mDIXON-Quant la fracción grasa (FG) de las lesiones suprarrenales encontradas incidentalmente en estudios de TC. Analizar la relación de la caída de se˜ nal entre las secuencias potenciadas en T1 en fase y fase opuesta con la FG en mDIXONQuant. Comparar la sensibilidad y especificidad de ambos métodos para caracterizar las lesiones suprarrenales. Material y métodos: Se realizó un estudio prospectivo descriptivo que incluyó a 31 pacientes con lesiones suprarrenales incidentales evaluados mediante RM 3 T con las secuencias T1 en fase y fase opuesta y mDIXON-Quant. Se midió la FG de las lesiones suprarrenales mediante nal en secuencias T1 en fase y fase opuesta, calculando su mDIXON-Quant y la intensidad de se˜ porcentaje de pérdida de se˜ nal (PPS). Resultados: El PPS medio fue significativamente mayor en el grupo adenoma (61,3% ± 20,4%) que en el no adenoma (5,1% ± 5,8%) (p < 0,005). La FG media de los adenomas también fue significativamente mayor (26,9% ± 10,8% vs. 3,4% ± 3,0%) (p < 0,005). El área bajo la curva ROC fue 0,99 (0,96-1,00) para el PPS y 0,98 (0,94-1,00) para la FG. El punto de corte obtenido fue de 24,42% para el PPS y de 9,2% para la FG. Los valores diagnósticos fueron iguales para los dos métodos: sensibilidad del 96% (79,6-99,9), especificidad del 100% (39,8-100,0), valor predictivo positivo del 100% (85,8-100,0) y valor predictivo negativo del 80% (28,4-99,5). Conclusión: La FG obtenida mediante técnica Dixon modificada es capaz de diferenciar adenomas de no adenomas con la misma sensibilidad y especificidad que la PPS. © 2018 SERAM. Publicado por Elsevier Espa˜ na, S.L.U. Todos los derechos reservados.
Introduction Incidental suprarenal lesions are a common finding in radiological studies conducted for different reasons, and most of them are benign.1 The detection of incidental suprarenal lesions has growth during the last few years parallel to the growing use of imaging modalities.2 There is a wide differential diagnosis for incidental suprarenal lesions including both benign and malignant entities, and there can be similar findings in the images too.2 Making the correct diagnosis is essential to be able to determine the best strategy for every patient, especially if they have been cancer patients in the past. In this case the possibility of malignancy is much higher than that of the general population (26---73%).3---5 Within the category of benign suprarenal lesions, the most common entity is the non-functioning adenoma.1,3,6 Adenomas are lesions that predominantly show large contents of intracellular fat.6,7 This finding allows us to distinguish them from all the other suprarenal lesions while conducting computed tomography (CT) scans without contrast and magnetic resonance imaging (MRI) as long as they include chemical shift displacement sequences.6---10 Nevertheless, the amount of fat that adenomas have is variable and there are times that it cannot be detected using CT scans without contrast (lipid-poor adenomas). In these lesions, the sensitivity and specificity of the MRI with chemical shift displacement sequences is higher compared to that of the CT scan.5 The chemical shift displacement is the difference between the effective magnetic fields of water and fat molecules. Thus, when the protons become in-phase, the
water and fat signals combine, but when they are in opposite phases, the water and fat signals repel each other causing signal intensity drops in tissues with intracellular fat content that are proportional to the amount of intracellular fat.5,9,11 That is why the intracellular fat content that adenomas have conditions the signal intensity drop in opposed-phase sequences with respect to in-phase sequences.5---10 Based on chemical shift displacement, the original Dixon method was first described back in 1984 as a technique to separate water signals from fat signals. From the acquisition of two in-phase and opposed-phase spin-echo sequences and through addition and subtraction post-processing, another two sets of images were obtained: water maps and fat maps.12 This model had two limitations: in the first place, the complete separation of water signals from fat signals could not be achieved to be able to quantify precisely the amount of lipids contained in a tissue and, in the second place, since there was a long time of acquisition, the presence of motion artifacts in the maps was a common thing.13 During the last few decades, several updates have been developed based on the original Dixon method. These include improvements based on the acquisition and postprocessing of the sequence while trying to avoid its main limitations.13 The mDIXON-Quant sequence is based on the Dixon method that it modifies by using multiple acquisition echoes, introducing correction factors, and improving postprocessing algorithms. By doing this we can study the correlation between water and fat molecules in several moments of the cycle (multi-echo GRE sequence) and totally separate the signal intensities of every voxel in order
Using the modified Dixon technique to evaluate incidental adrenal lesions on 3 T MRI
487
Contour 5 Area: 104 mm2 Mean: 27.78%
Water
T2∗
Fat fraction
Fat
Figure 1 mDIXON Quant sequence in the color scale with its four maps: water, fat, fat fraction (FF) and T2* in a patient with left suprarenal adenoma with a 27.78% FF.
to know its amount and estimate the FF of the tissues studied.11,14 We are talking about a 3D fast-field echo (FFE) multiple sequence acquired during apnea with which we can obtain two sets of images: ‘‘water only’’ and ‘‘fat only’’ that are processed in order to present another two: the FF and the T2* (transverse magnetization relaxation)11 (Fig. 1). That is why the goal of our study is:
31 patients with 36 incidental suprarenal lesions
7 patients excluded due to lack of diagnostic confirmation
24 patients with 29 lesions
• Use the mDIXON-Quant sequence to quantify the FF of the suprarenal lesions found incidentally on abdominal CT scans. • Analyze the correlation between signal intensity drops and FF in the in-phase and opposed-phase T1-weighted sequences using the mDIXON-Quant sequence. • Compare the sensitivity and specificity of both methods in the characterization of suprarenal lesions.
25 adenomas <10 HU. CT without contrast. No growth. 0 <10% of growth in follow-up study. Histological confirmation.
Figure 2 criteria.
4 nonadenomas with malignant uptake on the PET. Histological confirmation. Characteristicradiological findings of other entities.
Flow charts of the study population and diagnostic
Material and methods Patients This was a descriptive prospective study that included thirty-one (31) patients with suprarenal lesions identified incidentally through abdominal CT scans between 2015 and 2016. The patient’s informed consent and approval from the hospital ethics committee were obtained prior to conducting the study. The lesions were categorized as adenomas whenever they showed any of the following criteria: (a) average density below 10 HU in a CT scan without IV contrast,3,15---17 (b) anatomopathological diagnosis of adenoma and (c) no growth, or growth below 10% in an interval not greater than 6 months comparing the largest diameters.3,7,8 Lesions categorized as nonadenomas (a) lesions with an anatomopathological diagnosis of metastasis or pheochromocytoma, (b) uptake in contrast-enhanced18 18-fluorodeoxyglucose positron emission tomography (PET) and (c) presence of macroscopic fat with or without gross calcifications (findings suggestive of myelolipoma).3,9 Out of the 31 patients who underwent an MRI study, seven (7) were excluded for not being eligible for the diagnostic criteria proposed (Fig. 2).
Table 1 Technical parameters of T1-weighted FFE and mDIXON-Quant sequences.
Slice thickness TR TE Inclination angle Turbo factor NSA
T1-weighted FFE dual
mDIXON-Quant
6 mm 56 ms 1.15---2.3 ms 55◦ 2 1
6 mm 5.8 ms 0.0 3◦ 6 1
Magnetic resonance imaging modality 3 Tesla MRI studies (Philips Achieva 3.0 T) were conducted with SENSE XL Torso multi-transmission surface antenna (16 channels) using coronal and axial T2-weighted Turbo Spin Echo (T2-TSE) images and T1-weighted dual FFE (TR 56 ms, TE 1.15---2.3 ms, 55◦ inclination angle, turbo factor 2, NSA 1 and 6 mm thickness), mDIXON-Quant (TR 5.8 ms, TE 0.0 ms, 3◦ inclination angle, turbo factor 6, NSA 1 and 6 mm thickness) and DWI (Table 1).
488 Table 2
A. Ecénarro-Montiel et al. Characteristics of patients with suprarenal incidentaloma studied. Adenoma (n = 20) n
Nonadenoma (n = 4)
%
n
p-Value
%
Sex Male Female
12 8
60.0 40.0
2 2
50.0 50.0
Age group <65 years old ≥65 years old
9 11
45.0 55.0
2 2
50.0 50.0
Prior history of malignant neoplasm No 11 Yes 9
55.0 45.0
1 3
25.0 75.0
Laterality Unilateral Bilateral
75.0 25.0
4 0
100.0 0.0
1.000
1.000
0.590
0.544 15 5
Quantitative analysis The quantitative analysis was conducted using the Philips ViewForum working station (V 2.6.3.5). The values were obtained after placing one circular region of interest (ROI) of a fixed size of approximately 10 mm2 right in the middle of the suprarenal lesion studied and then spread throughout all the sequences. Using this ROI, the signal intensity (SI) of adrenal lesions was quantified both in-phase and opposed-phase, and the percentage of signal intensity loss (PSIL) was estimated using the following formula: (PSIL) = (SI in phase − SI out-of-phase/SI in phase) × 100. At the same time, the FF of suprarenal lesions was estimated in the map of FF of automatic processing as well as the largest diameter of suprarenal lesions in the axial and coronal planes of the diagnostic CT scan, and the largest of the two was picked. Also, the FF of the subcutaneous cellular tissue and spleen was quantified since these are tissues whose fat component we know.
Statistical analysis The characteristics of the patients were described (sex, age, prior history of malignancy and laterality) for each one of the groups --- adenomas and nonadenomas, and they were compared using Fisher’s exact test or Mann---Whitney U-test as appropriate. In each group, the mean and standard deviation of the PSIL and the FF of the lesion, spleen, and subcutaneous cellular tissue were estimated, and then the means were compared using the Mann---Whitney U-test. For each method, PSIL and FF, the area under the ROC curve was estimated and the optimal cut-off point was determined to be able to identify the adenomas being the only criterion that the Youden index (sensitivity + specificity − 1) was the maximum. With the cut-off points obtained, the indicators of the predictive capability of each method were estimated: sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV), together with a 95% confidence interval. Also, the match between both categorizations and the level of agreement observed were
analyzed, and the degree of linear correlation between PSIL and FF values was quantified using the Pearson’s linear correlation coefficient. p values < 0.05 were considered statistically significant.
Results A total of 24 patients with 29 suprarenal lesions was studied --- 14 men (58.3%) and 10 women (41.7%), among which 5 showed bilateral lesions (20.8%). Out of these patients, 12 (50%) had a prior history of malignant disease being the bronchogenic lung carcinoma the most common entity found in 4 patients (16.7%). Out of the 29 suprarenal lesions included, 25 were categorized as adenomas according to the diagnostic criteria proposed (20 patients with an average age of 64.5 years old) and 4 lesions as nonadenomas (4 patients with an average age of 61.3 years old) (Table 2). The average diameter of adenomas, 28.4 mm (±8.2 mm), was very similar to that of nonadenomas [30.5 mm (±6.9 mm) (p = 0.524)]. The average percentage of signal intensity loss (PSIL) was significantly higher in the group of adenomas (61.3% ± 20.4%) than in the groups of nonadenomas (5.1% ± 5.8%) (p < 0.005). The average FF of adenomas (26.9% ± 10.8%) was also significantly higher than that of nonadenomas (3.4% ± 3.0%) (p = 0.002) (Figs. 3 and 4 and Table 3). The average FF obtained in the arm was 2.8% (±1.4%) in patients from the group of adenomas and 1.3% (±0.7%) in patients from the group of nonadenomas (p = 0.029). The average FF of the subcutaneous cellular tissue was 86.8% (±6.7%) in the group of adenomas compared to 85.5% (±1.0%) in the group of nonadenomas (p = 0.311). The cut-off point obtained was 24.42% for the PSIL and 9.2% for the FF (Fig. 5). Both classifications are consistent with 100% of the lesions, which is why the diagnostic values were the same for the two methods: 96% sensitivity (79.6---99.9), 100% specificity (39.8---100.0), 100% PPV (85.8---100.0) and 80% NPV (28.4---99.5). The correlation coefficient between both methods was 0.746.
Using the modified Dixon technique to evaluate incidental adrenal lesions on 3 T MRI 100
Percentage of signal intensity loss (PSIL)
100
80
Fat fraction (FF)
489
60
40
20
0
80
60
40
20
0
Adenomas
Nonadenomas
Adenomas
Nonadenomas
Figure 3 Box plot of the FF obtained using the mDIXON Quant sequence and percentage of signal intensity loss in the in-phase and opposed-phase T1-weighted sequences of suprarenal lesions: adenomas and nonadenomas.
A
Contour 3 Area:97 mm2 Mean: 1.13 %
Contour 1 Area:108 mm2 Mean: 538.63
Contour 2 Area:108 mm2 Mean: 593.44
B
C
Figure 4 Left adrenal metastasis in a 63-year-old patient with a prior history of bronchogenic lung carcinoma. The lesion shows a significant signal intensity drop in the opposed-phase sequence (B) compared to the in-phase sequence (A), a percentage of signal intensity loss of 9.2% and a 1.13% FF in the mDIXON-Quant sequence (C), indicative of a lesion with poor lipidic content.
Table 3 Characteristics of suprarenal lesions in the magnetic resonance sequences mDIXON-Quant and in the in-phase and opposed-phase T1-weighted sequences. Adenoma (n = 20)
Fat fraction (FF) (%) Percentage of signal intensity loss (PSIL) (%) Size (mm)
Nonadenoma (n = 4)
Median (SD)
Range
Median (SD)
Range
26.9 (10.8) 61.3 (20.4) 28.4 (8.2)
4.5---41.7 10.0---85.2 14---48
3.4 (3.0) 5.1 (5.8) 30.5 (6.9)
0.1---0.6 0.1---13.6 22---37
Discussion During the last few years, measuring the FF of tissues using MRIs has become more and more important for the diagnosis and management of different entities such as liver steatosis and myopathies. The main advantage of this type of fat quantification is that it is a non-invasive method. Both the modifications made in the Dixon method and the automatic correction of confounding factors allow us to accurately know the FF of a tissue between 0% and 100%.19 This has been
p-Value
0.002 0.002 0.524
confirmed in several studies such as Hayashi et al.’s study that exposed the capacity of the modified Dixon method to assess liver steatosis with a sensitivity and specificity of 83.7---100% and 85.3---99.0%, respectively, and a good correlation with the histological study.20,21 In our study, the FF obtained using the modified Dixon technique was able to distinguish adenomas from nonadenomas with 96.0% and 100% sensitivity and specificity, respectively --- findings that are consistent with the medical literature published. Thus, Meng et al. showed 100% sensitivity and 84.0% specificity in their
A. Ecénarro-Montiel et al.
A
B
1.00
1.00
0.75
0.75
Sensitivity
Sensitivity
490
0.50
0.25
0.25
0.00 0.00
0.50
0.00 0.25
0.50
0.75
1.00
0.25
0.00
1-Specificity Fat fraction (FF)
0.50
0.75
1.00
1-Specificity Percentage of signal intensity loss (PSIL)
Figure 5 ROC curves of fat fraction (A) and percentage of signal intensity loss (B) in the magnetic resonance imaging for the discrimination between adenomas and nonadenomas in the suprarenal gland.
77 patients studied.22 However, Marin et al. obtained lower values (86% sensitivity and 82% specificity) in a sample of 69 patients.23 Up until now, and as the studies already published claim, the chemical shift displacement sequences using in-phase and opposed-phase and the PSIL are the best imaging modality for the diagnosis of adenomas, even better than the CT scan, with a high sensitivity (81---100%) and specificity (94---100%), similar to the one obtain in our study.5,8,24 In studies conducted using 1.5 T MRIs, several authors use a minimal PSIL of 16.5% for the diagnosis of adenomas, and with this value they achieve sensitivities and specificities close to 67% and 89---100%, respectively.9 However, the peak value of the PSIL to be able to distinguish correctly adenomas from nonadenomas has not been extensively studied in 3T machines. In a study of 34 patients with 37 adrenal lesions, Marin et al. obtained a cut-off point of 1.8% using double-echo sequences. This value is way below the value described in studies that used 1.5 T machines24 and that we obtained in our study (24.42%). On the other hand, in our study, the sensitivity obtained using both techniques (FF and PSIL) is the same, which would be consistent with other studies published that described a similar diagnostic accuracy with both methods.22,24 We should remember that chemical shift displacement techniques show a limited value in lipid-poor adenomas (defined in the CT scan without IV contrast such as those with an average density above 10 HU). Yet despite the fact that sensitivity remains in values around 89% for those with an average density between 10 and 30 HU, it drops dramatically down to 13% in adenomas with an average density above 30 HU in CT scans without contrast, and these are the lesions that are more difficult to handle with both imaging modalities.5,15 In order to overcome this limitation new studies with large sample sizes need to be conducted in order to be able to determine what values of FF would allow the characterization of lipid-poor adenomas. On the other hand, we should mention here the existence of suprarenal lesions of other nature that may have
intracytoplasmic fat as it is the case with some pheochromocytomas, metastases (especially hepatocellular carcinomas and renal cell carcinoma metastases), adrenocortical carcinomas and the so-called adrenal collision tumors where two lesions are in intimate relation in one suprarenal (adenoma and another lesion), which is why they would show signal intensity drops in the chemical shift displacement sequences.5 Among the study limitations we find the small sample size of nonadenomas and lipid-poor adenomas. Also, the scarce histopathological correlation is another limitation, since in the clinical practice, both adenomas and suprarenal metastases are usually diagnosed using imaging modalities, and the biopsy is only used for cases with an inconclusive study and when there are still chances of changing the therapeutic approach. Among the advantages of the Dixon sequence we have the possibility of characterizing one incidental suprarenal lesion in a short time of acquisition and avoiding the use of ionizing radiation and IV contrast. This is especially useful in young patients and those with renal failure. Also, we can quantify the percentage of fat in other organs and assess the presence of hepatic steatosis (Fig. 6) while we are conducting the suprarenal study. Also, this sequence will allow us to identify the presence of macroscopic fat --a characteristic finding of myelolipomas, which is why we could conduct the differential diagnosis with the suprarenal adenoma in one single sequence. In sum, the mDIXON-Quant sequence using the modified Dixon method allows us to quantify the FF in suprarenal lesion and is capable of differentiating adenomas from nonadenomas with the same sensitivity and specificity than the PSIL.
Authors 1. Manager of the integrity of the study: AEM and SBG. 2. Study Idea: SBG, RGF and AEM. 3. Study Design: SBG, AEM and RGF.
Using the modified Dixon technique to evaluate incidental adrenal lesions on 3 T MRI
491
References
Figure 6 Left suprarenal adenoma in a 57-year-old patient with a prior history of colon adenocarcinoma. The lesion shows a significant signal intensity drop in the in-phase (A) and opposedphase (B) T1-weighted sequences with a percentage of signal intensity loss of 82.54% and a 32.69% FF in the mDIXON-Quant sequence (C). We can also assess the hepatic FF of a patient with known diffuse steatosis and an average FF of 14.7%.
4. Data Mining: SBG, RGF and AEM. 5. Data Analysis and Interpretation: AEM, SBG, MISP, PMSV and JSG. 6. Statistical Analysis: MISP. 7. Reference: AEM, SBG, JSG and PMSV. 8. Writing: AEM, SBG, MISP and RFF. 9. Critical review of the manuscript with intellectually relevant remarks: SBG, AEM, MISP, RFF, OMSV and JSG. 10. Approval of final version: SBG, AEM, MISP, RGF, PMSV and JSG.
Conflicts of interests The authors declare no conflicts of interests associated with this article whatsoever.
1. Kloos RT, Gross MD, Francis IR, Korobkin M, Shapiro B. Incidentally discovered adrenal masses. Endocr Rev. 1995;16:460---84. 2. Boland GWL. Adrenal imaging: why, when, what, and how? Part 1. Why and when to image? AJR Am J Roentgenol. 2010;195:377---81. 3. Mansmann G, Lau J, Balk E, Rothberg M, Miyachi Y, Bornstein SR. The clinically inapparent adrenal mass: update in diagnosis and management. Endocr Rev. 2004;5:309---40. 4. Boland GWL, Blake MA, Hahn PF, Mayo-Smith WW. Incidental adrenal lesions: principles, techniques, and algorithms for imaging characterization. Radiology. 2008;249: 756---75. 5. Adam SZ, Nikolaidis P, Horowitz JM, Gabriel H, Hammond NA, Patel T, et al. Chemical shift MR imaging of the adrenal gland: principles, pitfalls, and applications. Radiographics. 2016;36:414---32. 6. Lattin GE, Sturgill ED, Tujo C, Marko J, Sanchez-Maldonado KW, Craig WD, et al. From the radiologic pathology archives: adrenal tumors and tumor-like conditions in the adult: radiologicpathologic correlation. Radiographics. 2014;34:805---29. 7. Blake MA, Cronin CG, Boland GW. Adrenal imaging. AJR Am J Roentgenol. 2010;194:1450---60. 8. 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 Am J Roentgenol. 2004;108:215---9. 9. 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 Am J Roentgenol. 2003;180:1649---57. 10. Oliveira Caiafa R, Salvador Izquierdo R, Bu˜ nesch Villalba L, Sebastià Cerqueda MC, Nicolau Molina C. Manejo y diagnóstico del incidentaloma suprarrenal. Radiología. 2011;53: 516---30. 11. Del Grande F, Santini F, Herzka DA, Aro MR, Dean CW, Gold GE, et al. Fat-suppression techniques for 3-T MR imaging of the musculoskeletal system. Radiographics. 2014;34: 217---33. 12. Dixon WT. Simple proton spectroscopic imaging. Radiology. 1984;153:189---94. 13. Ma J. Dixon techniques for water and fat imaging. J Magn Reson Imaging. 2008;28:543---58. 14. Kim DH, Yoo HJ, Hong SH, Choi JY, Chae HD, Chung BM. Differentiation of acute osteoporotic and malignant vertebral fractures by quantification of fat fraction with a Dixon MRI sequence. AJR Am J Roentgenol. 2017;209:1---9. 15. Johnson PT, Horton KM, Fishman EK. Adrenal mass imaging with multidetector CT: pathologic conditions, pearls, and pitfalls. Radiographics. 2009;29:1333---51. 16. 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;222:629---33. 17. Bae KT, Fuangtharnthip P, Prasad SR, Joe BN, Heiken JP. Adrenal masses: CT characterization with histogram analysis method. Radiology. 2003;228:735---42. 18. Dong A, Cui Y, Wang Y, Zuo C, Bai Y. 18F-FDG PET/CT of adrenal lesions. AJR Am J Roentgenol. 2014;203:245---52. 19. Reeder SB, Cruite I, Hamilton G, Sirlin CB. Quantitative assessment of liver fat with magnetic resonance. J Magn Reson Imaging. 2011:729---49. 20. Hayashi T, Saitoh S, Takahashi J, Tsuji Y, Ikeda K, Kobayashi M, et al. Hepatic fat quantification using the two-point Dixon method and fat color maps based on non-alcoholic fatty liver disease activity score. Hepatol Res. 2017;47:455---64.
492 21. Wren TA, Bluml S, Tseng-Ong L, Gilsanz V. Three-point technique of fat quantification of muscle tissue as a marker of disease progression in Duchenne muscular dystrophy: preliminary study. AJR Am J Roentgenol. 2008;190:8---12. 22. Meng X, Chen X, Shen Y, Hu X, Tang H, Hu D, et al. Proton-density fat fraction measurement: a viable quantitative biomarker for differentiating adrenal adenomas from non adenomas. Eur J Radiol. 2017;86:112---8.
A. Ecénarro-Montiel et al. 23. Marin D, Dale BM, Bashir MR, Ziemlewicz TJ, Ringe KI, Boll DT. Effectiveness of a three-dimensional dual gradient echo two-point Dixon technique for the characterization of adrenal lesions at 3 Tesla. Eur Radiol. 2012;22:259---68. 24. Marin D, Soher BJ, Dale BM, Boll DT, Youngblood RS, Merkle EM. Characterization of adrenal lesions: comparison of 2D and 3D dual gradient-echo MR imaging at 3 T. Radiology. 2010;254:179---87.
www.elsevier.es/rx
ORIGINAL REPORT
Using the modified Dixon technique to evaluate incidental adrenal lesions on 3 T MRI夽 A. Ecénarro-Montiel a,∗ , S. Baleato-González a , M.I. Santiago-Pérez b , J. Sánchez-González c , P. Montesinos c , R. García-Figueiras a a
Servicio de Radiología, Hospital Clínico Universitario, Santiago de Compostela, Spain Dirección Xeral de Saúde Pública, Consellería de Sanidade, Xunta de Galicia, Santiago de Compostela, Spain c Clinic Scientist, Philips Iberia, Madrid, Spain b
Received 9 January 2018; accepted 1 June 2018
KEYWORDS Adrenal glands; Adrenal adenoma; Adrenal incidentaloma; Magnetic resonance imaging; Adrenal tumors; Differential diagnosis
Abstract Objectives: To use the mDIXON-Quant sequence to quantify the fat fraction of adrenal lesions discovered incidentally on CT studies. To analyze the relation between the signal loss between in-phase and out-of-phase T1-weighted sequences and the fat fraction in mDIXON-Quant. To compare the sensitivity and specificity of the two methods for characterizing adrenal lesions. Material and methods: This prospective descriptive study included 31 patients with incidentally discovered adrenal lesions evaluated with 3 T MRI using in-phase and out-of-phase T1-weighted sequences and mDIXON-Quant; the fat fraction of the adrenal lesions was measured by mDIXONQuant and by calculating the percentage of signal loss between in-phase and out-of-phase T1-weighted sequences. Results: The percentage of signal loss was significantly higher in the group of patients with adenoma (61.3 ± 20.4% vs. 5.1 ± 5.8% in the group without adenoma, p < 0.005). The mean fat fraction measured by mDIXON-Quant was also higher for the adenomas (26.9 ± 10.8% vs. 3.4 ± 3.0%, p < 0.005). The area under the ROC curve was 0.99 (0.96---1.00) for the percentage of signal loss and 0.98 (0.94---1.00) for the fat fraction measured by mDIXON-Quant. The cutoffs obtained were 24.42% for the percentage of signal loss and 9.2% for the fat fraction measured by mDIXON-Quant. The two techniques had the same values for diagnostic accuracy: sensitivity 96% (79.6---99.9), specificity 100% (39.8---100.0), positive predictive value 100% (85.8---100.0), and negative predictive value 80% (28.4---99.5). Conclusion: The fat fraction measured by the modified Dixon technique can differentiate between adenomas and other adrenal lesions with the same sensitivity and specificity as the percentage of signal loss between in-phase and out-of-phase T1-weighted sequences. © 2018 SERAM. Published by Elsevier Espa˜ na, S.L.U. All rights reserved.
夽
Please cite this article as: Ecénarro-Montiel A, Baleato-González S, Santiago-Pérez MI, Sánchez-González J, Montesinos P, García-Figueiras R. Valoración mediante Dixon modificado de las lesiones suprarrenales incidentales en RM 3T. Radiología. 2018;60:485---492. ∗ Corresponding author. E-mail address: [email protected] (A. Ecénarro-Montiel). 2173-5107/© 2018 SERAM. Published by Elsevier Espa˜ na, S.L.U. All rights reserved.
486
PALABRAS CLAVE Glándulas suprarrenales; Adenoma suprarrenal; Incidentaloma suprarrenal; Resonancia magnética; Neoplasias suprarrenales; Diagnóstico diferencial
A. Ecénarro-Montiel et al.
Valoración mediante Dixon modificado de las lesiones suprarrenales incidentales en RM 3T Resumen Objetivo: Cuantificar mediante secuencia mDIXON-Quant la fracción grasa (FG) de las lesiones suprarrenales encontradas incidentalmente en estudios de TC. Analizar la relación de la caída de se˜ nal entre las secuencias potenciadas en T1 en fase y fase opuesta con la FG en mDIXONQuant. Comparar la sensibilidad y especificidad de ambos métodos para caracterizar las lesiones suprarrenales. Material y métodos: Se realizó un estudio prospectivo descriptivo que incluyó a 31 pacientes con lesiones suprarrenales incidentales evaluados mediante RM 3 T con las secuencias T1 en fase y fase opuesta y mDIXON-Quant. Se midió la FG de las lesiones suprarrenales mediante nal en secuencias T1 en fase y fase opuesta, calculando su mDIXON-Quant y la intensidad de se˜ porcentaje de pérdida de se˜ nal (PPS). Resultados: El PPS medio fue significativamente mayor en el grupo adenoma (61,3% ± 20,4%) que en el no adenoma (5,1% ± 5,8%) (p < 0,005). La FG media de los adenomas también fue significativamente mayor (26,9% ± 10,8% vs. 3,4% ± 3,0%) (p < 0,005). El área bajo la curva ROC fue 0,99 (0,96-1,00) para el PPS y 0,98 (0,94-1,00) para la FG. El punto de corte obtenido fue de 24,42% para el PPS y de 9,2% para la FG. Los valores diagnósticos fueron iguales para los dos métodos: sensibilidad del 96% (79,6-99,9), especificidad del 100% (39,8-100,0), valor predictivo positivo del 100% (85,8-100,0) y valor predictivo negativo del 80% (28,4-99,5). Conclusión: La FG obtenida mediante técnica Dixon modificada es capaz de diferenciar adenomas de no adenomas con la misma sensibilidad y especificidad que la PPS. © 2018 SERAM. Publicado por Elsevier Espa˜ na, S.L.U. Todos los derechos reservados.
Introduction Incidental suprarenal lesions are a common finding in radiological studies conducted for different reasons, and most of them are benign.1 The detection of incidental suprarenal lesions has growth during the last few years parallel to the growing use of imaging modalities.2 There is a wide differential diagnosis for incidental suprarenal lesions including both benign and malignant entities, and there can be similar findings in the images too.2 Making the correct diagnosis is essential to be able to determine the best strategy for every patient, especially if they have been cancer patients in the past. In this case the possibility of malignancy is much higher than that of the general population (26---73%).3---5 Within the category of benign suprarenal lesions, the most common entity is the non-functioning adenoma.1,3,6 Adenomas are lesions that predominantly show large contents of intracellular fat.6,7 This finding allows us to distinguish them from all the other suprarenal lesions while conducting computed tomography (CT) scans without contrast and magnetic resonance imaging (MRI) as long as they include chemical shift displacement sequences.6---10 Nevertheless, the amount of fat that adenomas have is variable and there are times that it cannot be detected using CT scans without contrast (lipid-poor adenomas). In these lesions, the sensitivity and specificity of the MRI with chemical shift displacement sequences is higher compared to that of the CT scan.5 The chemical shift displacement is the difference between the effective magnetic fields of water and fat molecules. Thus, when the protons become in-phase, the
water and fat signals combine, but when they are in opposite phases, the water and fat signals repel each other causing signal intensity drops in tissues with intracellular fat content that are proportional to the amount of intracellular fat.5,9,11 That is why the intracellular fat content that adenomas have conditions the signal intensity drop in opposed-phase sequences with respect to in-phase sequences.5---10 Based on chemical shift displacement, the original Dixon method was first described back in 1984 as a technique to separate water signals from fat signals. From the acquisition of two in-phase and opposed-phase spin-echo sequences and through addition and subtraction post-processing, another two sets of images were obtained: water maps and fat maps.12 This model had two limitations: in the first place, the complete separation of water signals from fat signals could not be achieved to be able to quantify precisely the amount of lipids contained in a tissue and, in the second place, since there was a long time of acquisition, the presence of motion artifacts in the maps was a common thing.13 During the last few decades, several updates have been developed based on the original Dixon method. These include improvements based on the acquisition and postprocessing of the sequence while trying to avoid its main limitations.13 The mDIXON-Quant sequence is based on the Dixon method that it modifies by using multiple acquisition echoes, introducing correction factors, and improving postprocessing algorithms. By doing this we can study the correlation between water and fat molecules in several moments of the cycle (multi-echo GRE sequence) and totally separate the signal intensities of every voxel in order
Using the modified Dixon technique to evaluate incidental adrenal lesions on 3 T MRI
487
Contour 5 Area: 104 mm2 Mean: 27.78%
Water
T2∗
Fat fraction
Fat
Figure 1 mDIXON Quant sequence in the color scale with its four maps: water, fat, fat fraction (FF) and T2* in a patient with left suprarenal adenoma with a 27.78% FF.
to know its amount and estimate the FF of the tissues studied.11,14 We are talking about a 3D fast-field echo (FFE) multiple sequence acquired during apnea with which we can obtain two sets of images: ‘‘water only’’ and ‘‘fat only’’ that are processed in order to present another two: the FF and the T2* (transverse magnetization relaxation)11 (Fig. 1). That is why the goal of our study is:
31 patients with 36 incidental suprarenal lesions
7 patients excluded due to lack of diagnostic confirmation
24 patients with 29 lesions
• Use the mDIXON-Quant sequence to quantify the FF of the suprarenal lesions found incidentally on abdominal CT scans. • Analyze the correlation between signal intensity drops and FF in the in-phase and opposed-phase T1-weighted sequences using the mDIXON-Quant sequence. • Compare the sensitivity and specificity of both methods in the characterization of suprarenal lesions.
25 adenomas <10 HU. CT without contrast. No growth. 0 <10% of growth in follow-up study. Histological confirmation.
Figure 2 criteria.
4 nonadenomas with malignant uptake on the PET. Histological confirmation. Characteristicradiological findings of other entities.
Flow charts of the study population and diagnostic
Material and methods Patients This was a descriptive prospective study that included thirty-one (31) patients with suprarenal lesions identified incidentally through abdominal CT scans between 2015 and 2016. The patient’s informed consent and approval from the hospital ethics committee were obtained prior to conducting the study. The lesions were categorized as adenomas whenever they showed any of the following criteria: (a) average density below 10 HU in a CT scan without IV contrast,3,15---17 (b) anatomopathological diagnosis of adenoma and (c) no growth, or growth below 10% in an interval not greater than 6 months comparing the largest diameters.3,7,8 Lesions categorized as nonadenomas (a) lesions with an anatomopathological diagnosis of metastasis or pheochromocytoma, (b) uptake in contrast-enhanced18 18-fluorodeoxyglucose positron emission tomography (PET) and (c) presence of macroscopic fat with or without gross calcifications (findings suggestive of myelolipoma).3,9 Out of the 31 patients who underwent an MRI study, seven (7) were excluded for not being eligible for the diagnostic criteria proposed (Fig. 2).
Table 1 Technical parameters of T1-weighted FFE and mDIXON-Quant sequences.
Slice thickness TR TE Inclination angle Turbo factor NSA
T1-weighted FFE dual
mDIXON-Quant
6 mm 56 ms 1.15---2.3 ms 55◦ 2 1
6 mm 5.8 ms 0.0 3◦ 6 1
Magnetic resonance imaging modality 3 Tesla MRI studies (Philips Achieva 3.0 T) were conducted with SENSE XL Torso multi-transmission surface antenna (16 channels) using coronal and axial T2-weighted Turbo Spin Echo (T2-TSE) images and T1-weighted dual FFE (TR 56 ms, TE 1.15---2.3 ms, 55◦ inclination angle, turbo factor 2, NSA 1 and 6 mm thickness), mDIXON-Quant (TR 5.8 ms, TE 0.0 ms, 3◦ inclination angle, turbo factor 6, NSA 1 and 6 mm thickness) and DWI (Table 1).
488 Table 2
A. Ecénarro-Montiel et al. Characteristics of patients with suprarenal incidentaloma studied. Adenoma (n = 20) n
Nonadenoma (n = 4)
%
n
p-Value
%
Sex Male Female
12 8
60.0 40.0
2 2
50.0 50.0
Age group <65 years old ≥65 years old
9 11
45.0 55.0
2 2
50.0 50.0
Prior history of malignant neoplasm No 11 Yes 9
55.0 45.0
1 3
25.0 75.0
Laterality Unilateral Bilateral
75.0 25.0
4 0
100.0 0.0
1.000
1.000
0.590
0.544 15 5
Quantitative analysis The quantitative analysis was conducted using the Philips ViewForum working station (V 2.6.3.5). The values were obtained after placing one circular region of interest (ROI) of a fixed size of approximately 10 mm2 right in the middle of the suprarenal lesion studied and then spread throughout all the sequences. Using this ROI, the signal intensity (SI) of adrenal lesions was quantified both in-phase and opposed-phase, and the percentage of signal intensity loss (PSIL) was estimated using the following formula: (PSIL) = (SI in phase − SI out-of-phase/SI in phase) × 100. At the same time, the FF of suprarenal lesions was estimated in the map of FF of automatic processing as well as the largest diameter of suprarenal lesions in the axial and coronal planes of the diagnostic CT scan, and the largest of the two was picked. Also, the FF of the subcutaneous cellular tissue and spleen was quantified since these are tissues whose fat component we know.
Statistical analysis The characteristics of the patients were described (sex, age, prior history of malignancy and laterality) for each one of the groups --- adenomas and nonadenomas, and they were compared using Fisher’s exact test or Mann---Whitney U-test as appropriate. In each group, the mean and standard deviation of the PSIL and the FF of the lesion, spleen, and subcutaneous cellular tissue were estimated, and then the means were compared using the Mann---Whitney U-test. For each method, PSIL and FF, the area under the ROC curve was estimated and the optimal cut-off point was determined to be able to identify the adenomas being the only criterion that the Youden index (sensitivity + specificity − 1) was the maximum. With the cut-off points obtained, the indicators of the predictive capability of each method were estimated: sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV), together with a 95% confidence interval. Also, the match between both categorizations and the level of agreement observed were
analyzed, and the degree of linear correlation between PSIL and FF values was quantified using the Pearson’s linear correlation coefficient. p values < 0.05 were considered statistically significant.
Results A total of 24 patients with 29 suprarenal lesions was studied --- 14 men (58.3%) and 10 women (41.7%), among which 5 showed bilateral lesions (20.8%). Out of these patients, 12 (50%) had a prior history of malignant disease being the bronchogenic lung carcinoma the most common entity found in 4 patients (16.7%). Out of the 29 suprarenal lesions included, 25 were categorized as adenomas according to the diagnostic criteria proposed (20 patients with an average age of 64.5 years old) and 4 lesions as nonadenomas (4 patients with an average age of 61.3 years old) (Table 2). The average diameter of adenomas, 28.4 mm (±8.2 mm), was very similar to that of nonadenomas [30.5 mm (±6.9 mm) (p = 0.524)]. The average percentage of signal intensity loss (PSIL) was significantly higher in the group of adenomas (61.3% ± 20.4%) than in the groups of nonadenomas (5.1% ± 5.8%) (p < 0.005). The average FF of adenomas (26.9% ± 10.8%) was also significantly higher than that of nonadenomas (3.4% ± 3.0%) (p = 0.002) (Figs. 3 and 4 and Table 3). The average FF obtained in the arm was 2.8% (±1.4%) in patients from the group of adenomas and 1.3% (±0.7%) in patients from the group of nonadenomas (p = 0.029). The average FF of the subcutaneous cellular tissue was 86.8% (±6.7%) in the group of adenomas compared to 85.5% (±1.0%) in the group of nonadenomas (p = 0.311). The cut-off point obtained was 24.42% for the PSIL and 9.2% for the FF (Fig. 5). Both classifications are consistent with 100% of the lesions, which is why the diagnostic values were the same for the two methods: 96% sensitivity (79.6---99.9), 100% specificity (39.8---100.0), 100% PPV (85.8---100.0) and 80% NPV (28.4---99.5). The correlation coefficient between both methods was 0.746.
Using the modified Dixon technique to evaluate incidental adrenal lesions on 3 T MRI 100
Percentage of signal intensity loss (PSIL)
100
80
Fat fraction (FF)
489
60
40
20
0
80
60
40
20
0
Adenomas
Nonadenomas
Adenomas
Nonadenomas
Figure 3 Box plot of the FF obtained using the mDIXON Quant sequence and percentage of signal intensity loss in the in-phase and opposed-phase T1-weighted sequences of suprarenal lesions: adenomas and nonadenomas.
A
Contour 3 Area:97 mm2 Mean: 1.13 %
Contour 1 Area:108 mm2 Mean: 538.63
Contour 2 Area:108 mm2 Mean: 593.44
B
C
Figure 4 Left adrenal metastasis in a 63-year-old patient with a prior history of bronchogenic lung carcinoma. The lesion shows a significant signal intensity drop in the opposed-phase sequence (B) compared to the in-phase sequence (A), a percentage of signal intensity loss of 9.2% and a 1.13% FF in the mDIXON-Quant sequence (C), indicative of a lesion with poor lipidic content.
Table 3 Characteristics of suprarenal lesions in the magnetic resonance sequences mDIXON-Quant and in the in-phase and opposed-phase T1-weighted sequences. Adenoma (n = 20)
Fat fraction (FF) (%) Percentage of signal intensity loss (PSIL) (%) Size (mm)
Nonadenoma (n = 4)
Median (SD)
Range
Median (SD)
Range
26.9 (10.8) 61.3 (20.4) 28.4 (8.2)
4.5---41.7 10.0---85.2 14---48
3.4 (3.0) 5.1 (5.8) 30.5 (6.9)
0.1---0.6 0.1---13.6 22---37
Discussion During the last few years, measuring the FF of tissues using MRIs has become more and more important for the diagnosis and management of different entities such as liver steatosis and myopathies. The main advantage of this type of fat quantification is that it is a non-invasive method. Both the modifications made in the Dixon method and the automatic correction of confounding factors allow us to accurately know the FF of a tissue between 0% and 100%.19 This has been
p-Value
0.002 0.002 0.524
confirmed in several studies such as Hayashi et al.’s study that exposed the capacity of the modified Dixon method to assess liver steatosis with a sensitivity and specificity of 83.7---100% and 85.3---99.0%, respectively, and a good correlation with the histological study.20,21 In our study, the FF obtained using the modified Dixon technique was able to distinguish adenomas from nonadenomas with 96.0% and 100% sensitivity and specificity, respectively --- findings that are consistent with the medical literature published. Thus, Meng et al. showed 100% sensitivity and 84.0% specificity in their
A. Ecénarro-Montiel et al.
A
B
1.00
1.00
0.75
0.75
Sensitivity
Sensitivity
490
0.50
0.25
0.25
0.00 0.00
0.50
0.00 0.25
0.50
0.75
1.00
0.25
0.00
1-Specificity Fat fraction (FF)
0.50
0.75
1.00
1-Specificity Percentage of signal intensity loss (PSIL)
Figure 5 ROC curves of fat fraction (A) and percentage of signal intensity loss (B) in the magnetic resonance imaging for the discrimination between adenomas and nonadenomas in the suprarenal gland.
77 patients studied.22 However, Marin et al. obtained lower values (86% sensitivity and 82% specificity) in a sample of 69 patients.23 Up until now, and as the studies already published claim, the chemical shift displacement sequences using in-phase and opposed-phase and the PSIL are the best imaging modality for the diagnosis of adenomas, even better than the CT scan, with a high sensitivity (81---100%) and specificity (94---100%), similar to the one obtain in our study.5,8,24 In studies conducted using 1.5 T MRIs, several authors use a minimal PSIL of 16.5% for the diagnosis of adenomas, and with this value they achieve sensitivities and specificities close to 67% and 89---100%, respectively.9 However, the peak value of the PSIL to be able to distinguish correctly adenomas from nonadenomas has not been extensively studied in 3T machines. In a study of 34 patients with 37 adrenal lesions, Marin et al. obtained a cut-off point of 1.8% using double-echo sequences. This value is way below the value described in studies that used 1.5 T machines24 and that we obtained in our study (24.42%). On the other hand, in our study, the sensitivity obtained using both techniques (FF and PSIL) is the same, which would be consistent with other studies published that described a similar diagnostic accuracy with both methods.22,24 We should remember that chemical shift displacement techniques show a limited value in lipid-poor adenomas (defined in the CT scan without IV contrast such as those with an average density above 10 HU). Yet despite the fact that sensitivity remains in values around 89% for those with an average density between 10 and 30 HU, it drops dramatically down to 13% in adenomas with an average density above 30 HU in CT scans without contrast, and these are the lesions that are more difficult to handle with both imaging modalities.5,15 In order to overcome this limitation new studies with large sample sizes need to be conducted in order to be able to determine what values of FF would allow the characterization of lipid-poor adenomas. On the other hand, we should mention here the existence of suprarenal lesions of other nature that may have
intracytoplasmic fat as it is the case with some pheochromocytomas, metastases (especially hepatocellular carcinomas and renal cell carcinoma metastases), adrenocortical carcinomas and the so-called adrenal collision tumors where two lesions are in intimate relation in one suprarenal (adenoma and another lesion), which is why they would show signal intensity drops in the chemical shift displacement sequences.5 Among the study limitations we find the small sample size of nonadenomas and lipid-poor adenomas. Also, the scarce histopathological correlation is another limitation, since in the clinical practice, both adenomas and suprarenal metastases are usually diagnosed using imaging modalities, and the biopsy is only used for cases with an inconclusive study and when there are still chances of changing the therapeutic approach. Among the advantages of the Dixon sequence we have the possibility of characterizing one incidental suprarenal lesion in a short time of acquisition and avoiding the use of ionizing radiation and IV contrast. This is especially useful in young patients and those with renal failure. Also, we can quantify the percentage of fat in other organs and assess the presence of hepatic steatosis (Fig. 6) while we are conducting the suprarenal study. Also, this sequence will allow us to identify the presence of macroscopic fat --a characteristic finding of myelolipomas, which is why we could conduct the differential diagnosis with the suprarenal adenoma in one single sequence. In sum, the mDIXON-Quant sequence using the modified Dixon method allows us to quantify the FF in suprarenal lesion and is capable of differentiating adenomas from nonadenomas with the same sensitivity and specificity than the PSIL.
Authors 1. Manager of the integrity of the study: AEM and SBG. 2. Study Idea: SBG, RGF and AEM. 3. Study Design: SBG, AEM and RGF.
Using the modified Dixon technique to evaluate incidental adrenal lesions on 3 T MRI
491
References
Figure 6 Left suprarenal adenoma in a 57-year-old patient with a prior history of colon adenocarcinoma. The lesion shows a significant signal intensity drop in the in-phase (A) and opposedphase (B) T1-weighted sequences with a percentage of signal intensity loss of 82.54% and a 32.69% FF in the mDIXON-Quant sequence (C). We can also assess the hepatic FF of a patient with known diffuse steatosis and an average FF of 14.7%.
4. Data Mining: SBG, RGF and AEM. 5. Data Analysis and Interpretation: AEM, SBG, MISP, PMSV and JSG. 6. Statistical Analysis: MISP. 7. Reference: AEM, SBG, JSG and PMSV. 8. Writing: AEM, SBG, MISP and RFF. 9. Critical review of the manuscript with intellectually relevant remarks: SBG, AEM, MISP, RFF, OMSV and JSG. 10. Approval of final version: SBG, AEM, MISP, RGF, PMSV and JSG.
Conflicts of interests The authors declare no conflicts of interests associated with this article whatsoever.
1. Kloos RT, Gross MD, Francis IR, Korobkin M, Shapiro B. Incidentally discovered adrenal masses. Endocr Rev. 1995;16:460---84. 2. Boland GWL. Adrenal imaging: why, when, what, and how? Part 1. Why and when to image? AJR Am J Roentgenol. 2010;195:377---81. 3. Mansmann G, Lau J, Balk E, Rothberg M, Miyachi Y, Bornstein SR. The clinically inapparent adrenal mass: update in diagnosis and management. Endocr Rev. 2004;5:309---40. 4. Boland GWL, Blake MA, Hahn PF, Mayo-Smith WW. Incidental adrenal lesions: principles, techniques, and algorithms for imaging characterization. Radiology. 2008;249: 756---75. 5. Adam SZ, Nikolaidis P, Horowitz JM, Gabriel H, Hammond NA, Patel T, et al. Chemical shift MR imaging of the adrenal gland: principles, pitfalls, and applications. Radiographics. 2016;36:414---32. 6. Lattin GE, Sturgill ED, Tujo C, Marko J, Sanchez-Maldonado KW, Craig WD, et al. From the radiologic pathology archives: adrenal tumors and tumor-like conditions in the adult: radiologicpathologic correlation. Radiographics. 2014;34:805---29. 7. Blake MA, Cronin CG, Boland GW. Adrenal imaging. AJR Am J Roentgenol. 2010;194:1450---60. 8. 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 Am J Roentgenol. 2004;108:215---9. 9. 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 Am J Roentgenol. 2003;180:1649---57. 10. Oliveira Caiafa R, Salvador Izquierdo R, Bu˜ nesch Villalba L, Sebastià Cerqueda MC, Nicolau Molina C. Manejo y diagnóstico del incidentaloma suprarrenal. Radiología. 2011;53: 516---30. 11. Del Grande F, Santini F, Herzka DA, Aro MR, Dean CW, Gold GE, et al. Fat-suppression techniques for 3-T MR imaging of the musculoskeletal system. Radiographics. 2014;34: 217---33. 12. Dixon WT. Simple proton spectroscopic imaging. Radiology. 1984;153:189---94. 13. Ma J. Dixon techniques for water and fat imaging. J Magn Reson Imaging. 2008;28:543---58. 14. Kim DH, Yoo HJ, Hong SH, Choi JY, Chae HD, Chung BM. Differentiation of acute osteoporotic and malignant vertebral fractures by quantification of fat fraction with a Dixon MRI sequence. AJR Am J Roentgenol. 2017;209:1---9. 15. Johnson PT, Horton KM, Fishman EK. Adrenal mass imaging with multidetector CT: pathologic conditions, pearls, and pitfalls. Radiographics. 2009;29:1333---51. 16. 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;222:629---33. 17. Bae KT, Fuangtharnthip P, Prasad SR, Joe BN, Heiken JP. Adrenal masses: CT characterization with histogram analysis method. Radiology. 2003;228:735---42. 18. Dong A, Cui Y, Wang Y, Zuo C, Bai Y. 18F-FDG PET/CT of adrenal lesions. AJR Am J Roentgenol. 2014;203:245---52. 19. Reeder SB, Cruite I, Hamilton G, Sirlin CB. Quantitative assessment of liver fat with magnetic resonance. J Magn Reson Imaging. 2011:729---49. 20. Hayashi T, Saitoh S, Takahashi J, Tsuji Y, Ikeda K, Kobayashi M, et al. Hepatic fat quantification using the two-point Dixon method and fat color maps based on non-alcoholic fatty liver disease activity score. Hepatol Res. 2017;47:455---64.
492 21. Wren TA, Bluml S, Tseng-Ong L, Gilsanz V. Three-point technique of fat quantification of muscle tissue as a marker of disease progression in Duchenne muscular dystrophy: preliminary study. AJR Am J Roentgenol. 2008;190:8---12. 22. Meng X, Chen X, Shen Y, Hu X, Tang H, Hu D, et al. Proton-density fat fraction measurement: a viable quantitative biomarker for differentiating adrenal adenomas from non adenomas. Eur J Radiol. 2017;86:112---8.
A. Ecénarro-Montiel et al. 23. Marin D, Dale BM, Bashir MR, Ziemlewicz TJ, Ringe KI, Boll DT. Effectiveness of a three-dimensional dual gradient echo two-point Dixon technique for the characterization of adrenal lesions at 3 Tesla. Eur Radiol. 2012;22:259---68. 24. Marin D, Soher BJ, Dale BM, Boll DT, Youngblood RS, Merkle EM. Characterization of adrenal lesions: comparison of 2D and 3D dual gradient-echo MR imaging at 3 T. Radiology. 2010;254:179---87.