Diffusion-weighted magnetic resonance imaging in the differentiation of angiomyolipoma with minimal fat from clear cell renal cell carcinoma


Yasuhisa Fujii M.D., Ph.D., Department of Urology, Tokyo Medical and Dental University Graduate School, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8519, Japan. Email: y-fujii.uro@tmd.ac.jp


The aim of the present study was to evaluate diffusion-weighted (DW) magnetic resonance imaging (MRI) in differentiating between minimal fat angiomyolipoma (MFAML) and clear cell renal cell carcinoma (CCRCC). Forty-one solid renal tumors without visible macroscopic fat on unenhanced computed tomography images were evaluated by MRI, including DW-MRI, and were diagnosed pathologically as CCRCC (n = 36) or MFAML (n = 5). To evaluate the heterogeneity of diffusion in each tumor, the signals of the tumors on DW-MRI were analyzed subjectively and the apparent diffusion coefficient (ADC) values and histograms assessed objectively. Thirty-three of 36 CCRCC (92%) exhibited a heterogeneous signal on DW-MRI and several peaks in the ADC value histogram, whereas four of five MFAML exhibited a homogeneous signal on DW-MRI and a single prominent peak in the histogram. The standard deviations of the ADC values were significantly smaller for MFAML than for CCRCC (P = 0.0015). In conclusion, DW-MRI can be considered a useful and noninvasive addition to the preoperative differentiation of CCRCC and MFAML.


Renal angiomyolipoma (AML) is one of the most common benign renal neoplasms.1 It is important to differentiate AML from renal cell carcinomas (RCC) because AML, particularly if they are small and asymptomatic, can be observed without any treatment,2 whereas RCC generally require surgical resection. Most AML can be easily diagnosed on imaging by their fat component, but approximately 5% contain no fat or an insufficient amount of fat to be detected on imaging (minimal fat AML; MFAML), making them difficult to differentiate from RCC.

Previous investigators have described the imaging findings of MFAML as follows: homogeneously high attenuation on unenhanced computed tomography (CT), homogeneous enhancement and prolonged enhancement pattern on dynamic contrast-enhanced CT (DCE-CT), hypointense on T2-weighted magnetic resonance imaging (MRI), and isoechoic on ultrasonography.3,4 However, these findings lack objectivity, especially with regard to the homogeneity, which has been reported to be one of the most significant predictors of MFAML.4

Diffusion-weighted (DW) MRI is a functional imaging that derives image contrast from the motion of water molecules in tissue. On DW-MRI, lesions with dense cellularity and poor interstitium that restrict the mobility of water molecules exhibit high signal intensity. This imaging technique has been applied to the diagnosis of cancer.5–7 The extent of the diffusion of water molecules can be assessed quantitatively by the apparent diffusion coefficient (ADC) value. Although the ADC value has been conventionally calculated from a small region of interest (ROI) arbitrarily positioned in a small part of the targeted lesion, it was advocated that detailed profiles of tumor diffusion environments should be analyzed from the ROI contoured around the targeted lesion.8

In the present study, based on the hypothesis that MFAML exhibit homogeneity of diffusion on DW-MRI, we attempted to evaluate the findings of MFAML by DW-MRI, both visually and objectively, with an ADC value histogram. As the first attempt to differentiate MFAML by DW-MRI, we limited the control to clear cell RCC (CCRCC) because RCC exhibit different features according to tumor subtype and CCRCC account for a large percentage of small renal masses.


Between April 2006 and March 2009, 41 solid renal tumors without visible macroscopic fat on unenhanced CT images were evaluated by MRI, including DW-MRI, and were diagnosed pathologically as CCRCC (n = 36) or MFAML (n = 5) at partial or radical nephrectomy. All 41 tumors were presumed to be RCC before surgery. The characteristics of the patients and tumors are presented in Table 1. The Mann–Whitney U-test or Chi-squared test revealed no significant differences between the two groups in terms of age, sex, and tumor size.

Table 1.  Characteristics of patients and tumors
 CCRCC (n = 36)MFAML (n = 5)P-value
  • One male patient had bilateral clear cell renal cell carcinoma (CCRCC).

  • Relative to normal renal cortex.

  • ADC, apparent diffusion coefficient; DW-MRI, diffusion-weighted magnetic resonance imaging; MFAML, minimal fat angiomyolipoma; T1W, T1-weighted imaging; T2W, T2-weighted imaging.

Median (range) age (years)57 (38–78)45 (39–64)NS
Median (range) tumor size (cm)3.2 (1.3–9.6)2.8 (1.1–3.4)NS
Findings of MRI and ADC value histogram   
 ADC value histogram   
  Number of peaks   
   Two or more331

In the present study, MRI was performed using a 1.5-T MR imager (Intera Achieva; Philips, Best, The Netherlands) with a four-channel sensitivity-encoding body coil and without breath-holding. The maximal gradient strength was 33 mT/m and the slew rate was 160 T/m per s. Following routine T1- and T2-weighted imaging (T1W and T2W, respectively), DW-MRI was performed. The imaging parameters of DW-MRI with a single-shot echo planar imaging sequence were set as follows: repetition time: 1500 msec; echo time: 65 msec; matrix: 160 × 160; field of view: 36 cm × 24 cm; slice thickness: 7 mm; interslice gap: 0.7 mm; slice number: 24; bandwidth: 1704.2 Hz per pixel; b value: 0 and 800 s/mm2); fat suppression: spectral presaturation inversion recovery. The total acquisition time was approximately 100 s.

The DW-MRI findings were analyzed subjectively according to the heterogeneity of the signal. Furthermore, detailed profiles of ADC values were analyzed to assess the degree of diffusion objectively. The ADC maps of the tumors were reconstructed at a workstation (Philip View Forum R4.1; Philips) and, using OsiriX software (http://www.osirix-viewer.com, accessed 27 July 2011), circular ROI that were not extending over the tumor were positioned on the transverse ADC map at the slice that showed maximal tumor diameter. Circular ROI with a diameter of 4 mm were also set in the ipsilateral normal renal cortexes. The ADC values were calculated using the formula: ADC = –ln (S/S0)/(b – b0), where S0 and S are the signal intensities that were obtained with two different diffusion gradient values (b0 and b; 0 and 800 s/mm2, respectively). The minimum, mean, maximum, and standard deviation (SD) of the ADC values of each tumor were calculated. Furthermore, the ADC value histograms of the tumor were plotted.

One urologist (HT) who was blinded to all the clinical information reviewed the images retrospectively. These MRI findings were compared with the pathological diagnosis. Differences in ADC values among CCRCC, MFAML, and normal renal cortex were evaluated by the Mann–Whitney U-test using JMP software version 6.0 (SAS Institute Inc., Cary, NC, USA). P < 0.05 was considered significant.


All 41 tumors were detected by T1W, T2W, and DW-MRI. The MRI findings according to pathological diagnosis are shown in Figures 1 and 2 and listed in Table 1. Based on the T1W findings, MFAML could not be discriminated from CCRCC. On fat-saturated T1W, no tumors showed macroscopic fat. On T2W, all tumors exhibiting high signal intensity were diagnosed as CCRCC, whereas two and five of the seven tumors exhibiting low signal intensity were diagnosed as CCRCC and MFAML, respectively.

Figure 1.

T2-weighted image (T2W), diffusion weighted magnetic resonance imaging (DW-MRI), apparent diffusion coefficient (ADC) map and ADC value histogram of a representative case of clear cell renal cell carcinoma (CCRCC) 5 cm in diameter. The CCRCC exhibits heterogeneous signal intensity on DW-MRI and several peaks on the histogram.

Figure 2.

T2-weighted image (T2W), diffusion weighted magnetic resonance imaging (DW-MRI), apparent diffusion coefficient (ADC) map and ADC value histogram of a representative case of minimal fat angiomyolipoma (MFAML) 2.8 cm in diameter. The MFAML exhibits homogeneous signal intensity on DW-MRI and a single prominent peak on the histogram.

As indicated in Table 1, three of 36 CCRCC (8%) exhibited a visually homogeneous signal on DW-MRI, compared with four of five MFAML. The ADC values of renal tumors and normal renal cortex are given in Table 2. The minimum ADC value of the renal tumors was significantly lower than that of normal renal cortex (P < 0.0001). Although there was no significant difference in the minimum ADC values between MFAML and CCRCC, the mean and maximum ADC values of MFAML were significantly lower than those of CCRCC (P = 0.0030 and 0.0009, respectively). Analyses of the ADC value histograms revealed that there were two patterns for the tumor histograms: (i) histograms with several peaks; and (ii) histograms with a single prominent peak (Figs 1,2). The subjectively evaluated heterogeneity corresponded to the presence of several peaks in the histogram. To objectively assess heterogeneity, the SD of the ADC values of the tumor was evaluated. The SD of the ADC values of MFAML was significantly smaller than that of CCRCC (P = 0.0015).

Table 2.  Apparent diffusion coefficient values of renal tumors and normal renal cortex
 CCRCC (n = 36)MFAML (n = 5)Normal renal cortex (n = 41)P-value*
  • Data show median values with the range given in parentheses.

  • *

    Between clear cell renal cell carcinoma (CCRCC) and minimal fat angiomyolipoma (MFAML).

  • ADC, apparent diffusion coefficient; SD, standard deviation.

ADC values (×10−3 mm2/s)    
 Minimum0.87 (0.00–2.09)0.67 (0.46–0.97)1.53 (0.90–2.17)NS
 Mean1.54 (0.75–2.44)0.80 (0.58–1.29)1.64 (0.99–2.35)0.0030
 Maximum2.15 (1.22–3.53)0.93 (0.76–1.81)1.81 (1.13–2.60)0.0009
 SD0.27 (0.07–0.71)0.07 (0.05–0.20)0.0015

Because larger renal tumors tend to have necrotic lesions, tumor size may influence the heterogeneity of diffusion. To rule out this possibility, we performed a subgroup analysis using a size-matched cohort with tumors < 4 cm in diameter. Of the five MFAML and 23 CCRCC <4 cm in diameter, the mean, maximum, and SD of the ADC values still differed significantly (P = 0.0069, 0.0022, and 0.0048, respectively).


This is the first study to demonstrate that DW-MRI may be a useful and noninvasive addition in differentiating MFAML from CCRCC. Most MFAML exhibit homogeneous diffusion on DW-MRI, whereas many CCRCC exhibit heterogeneity. Evaluating the SD and histogram pattern of ADC values objectively confirmed these findings. Previous studies have reported on the applicability of DW-MRI in the diagnosis of renal tumors.9,10 The results of these studies revealed that solid renal tumors exhibited high signal intensities compared with normal renal cortex on DW-MRI, but the signal variability and heterogeneity were not adequately addressed. This limitation may be due to the conventional method of arbitrarily positioning the ROI in the targeted lesions on the ADC map.8 In the present study, we evaluated not only the intensity, but also the heterogeneity of signals on DW-MRI by positioning the largest possible ROI in the tumors. Furthermore, ADC value histograms were used to evaluate the heterogeneity of signals more closely. Comparison of the size-matched cohort revealed the same differences in DW-MRI findings for MFAML and CCRCC. Thus, the heterogeneity of CCRCC is likely due to the histological characteristics, not tumor size.

The present study was a retrospective study and we only included MFAML and CCRCC based on final pathological diagnosis. There were only five MFAML in the present study. Prospective larger-scale studies are needed to validate the preliminary results reported herein.


The authors thank the members of the Department of Radiology in Ochanomizu Surugadai Clinic for their radiological advice.

Conflict of interest

None declared.