Value of intravoxel incoherent motion for differential diagnosis of renal tumors

Abstract

Background

Few studies have reported on the use of intravoxel incoherent motion (IVIM) for renal tumors.

Purpose

To investigate the value of IVIM for distinguishing renal tumors.

Material and Methods

Thirty-one patients with clear cell renal cell carcinomas (CCRCCs), 13 patients with renal angiomyolipomas with minimal fat (RAMFs), eight patients with chromophobe renal cell carcinomas (ChRCCs), and ten patients with papillary renal cell carcinomas (PRCCs) were examined. The tissue diffusivity (D), pseudodiffusivity (D*), and perfusion fraction (f) were calculated.

Results

The D and f values were highest for CCRCCs, lowest for PRCCs, and intermediate for ChRCCs and RAMFs (P < 0.05). The D values of CCRCCs differed significantly from those of ChRCCs and PRCCs (P < 0.05). The D* values were highest for RAMFs, lowest for ChRCCs, and intermediate for CCRCCs and PRCCs (P < 0.05). Statistically significant differences were observed between the D* values of CCRCCs and RAMFs (P < 0.05). The D* values of the CCRCCs differed significantly from the D* values of the ChRCCs (P < 0.05). Using the D and f values of 1.10 and 0.41, respectively, as the threshold values for differentiating CCRCCs from RAMFs, ChRCCs, and PRCCs, the best results had sensitivities of 81.0% and 66.8% and specificities of 85.7% and 81.0%, respectively. Using the D* value of 0.038 as the threshold value for differentiating RAMFs from CCRCCs, ChRCCs, and PRCCs, the best result obtained had a sensitivity of 90.5% and specificity of 76.2%.

Conclusion

IVIM may provide information for differentiating renal tumor types.

Keywords

Diffusion kidney tumor intravoxel incoherent motion

Introduction

Clear cell renal cell carcinoma (CCRCC) is the most common primary malignant epithelial tumor of the kidney and accounts for 70% of all renal tumors (1). CCRCCs are predominantly sporadic (95%) but can be familial with other renal tumors (2), such as renal angiomyolipomas with minimal fat (RAMFs), chromophobe renal cell carcinomas (ChRCCs), and papillary renal cell carcinomas (PRCCs). Interestingly, CCRCCs and other renal tumors share similar ontogenic and histologic features as well as some imaging features (3). However, different renal tumors exhibit differences in prognoses and treatments. RAMF is a benign tumor with a favorable prognosis without surgery. ChRCC and PRCC are renal cell carcinoma (RCC) subtypes with overall favorable prognoses compared with CCRCC that include slower growth and significantly lower rates of metastasis, progression, and death (4). ChRCC and PRCC are generally managed via partial nephrectomy when technically possible; routine radical nephrectomy for this entity would be gross mismanagement that would needlessly expose the patient to a greater risk of reduced renal function and an associated higher rate of cardiovascular events from the more extensive surgery.

Intravoxel incoherent motion (IVIM) can provide a unique view of tissue perfusion without the use of exogenous contrast agents. Chandarana et al. (5) demonstrated higher perfusion fractions (fs) and lower tissue diffusivities (Ds) in enhancing renal tumors compared with benign non-enhancing lesions. These authors also found that the parameters f and D can be used to discriminate between renal tumor subtypes. However, their studies only used lower b-values (0–800 s/mm²). Notohamiprodjo et al. (6) reported that, with the higher b-values, more accurate measurements of diffusion were possible. In our study, we demonstrated that b-values in the range of approximately 0–1500 s/mm² are sufficient for abdominal IVIM to enable the observation of the departure of the diffusion signal from mono-exponential behavior. Our purpose was to characterize the IVIM characteristics of renal tumor types.

Material and Methods

Patients

A total of 62 patients who met the criteria were identified. Our study was approved by our institutional review board and written informed consent was obtained from all patients before the study. A total of 62 patients with confirmed pathology and immunohistochemistry from January 2012 to July 2017 were recruited, and the patients underwent 3.0-T kidney magnetic resonance imaging (MRI). The patients included 31 patients with CCRCCs, 13 patients with RAMFs, eight patients with ChRCCs, and ten patients with PRCCs.

Magnetic resonance imaging

The MRI examinations were performed using a 3.0-T MR scanner (GE Signa EXCITE HD, Milwaukee, WI, USA) with an eight-channel array body coil and a 24-channel phased array spine coil integrated into the scanner table. For IVIM, a single-shot echo-planar imaging sequence was applied in the axial plane using respiratory triggering via a respiratory belt, nine b-values (0, 30, 50, 80, 150, 300, 500, 800, and 1500 s/mm²), and eight signal averages. The other imaging parameters were as follows: 24 axial slices covering both kidneys; TE = 33.2 × 86.6 ms; TR = 1000 ms; matrix = 96 × 128; and field of view (FOV) = 36 × 36 cm.

Image analysis

The acquired images were transferred to an off-line workstation for processing. Before IVIM quantification, image co-registration and smoothing were performed using the automated image registration (AIR) software v4.6.4. With our IVIM protocol, we obtained parameters related to D, D*, and f. The assessment of the renal tumors and region of interest (ROI) positioning were conducted by a radiologist (with five years of clinical experience in interpreting MR images). ROIs of the mass (range=100–225 mm²) were manually drawn using ImageJ (National Institutes of Health, Bethesda, MD, USA) on the axial slices. Attention was given to avoiding areas with aliasing artifacts when they appeared in the image. Moreover, intratumoral calcification and cystic components were excluded from the ROI if central necrosis or calcifications were considered. Each IVIM parameter was measured three times and the mean value was used.

Statistical analysis

The statistical analyses were undertaken using the SPSS version 17.0 statistical software (SPSS, Chicago, IL, USA). Numeric data are expressed as the means and the standard deviations (±SDs), and categorical data are expressed as percentages. The evaluated IVIM parameters were compared with the renal tumor types by applying analysis of variance (ANOVA) and post-hoc tests (Tukey). P values < 0.05 were considered statistically significant. The specificities of the IVIM values for the diagnoses of the renal tumor types and the best cut-offs were calculated from the areas under the ROC curves. P < 0.05 was considered statistically significant.

Results

The D values (Fig. 1) were highest for the CCRCCs, lowest for the PRCCs, and intermediate for the ChRCCs and RAMFs (F = 37.229, P < 0.05; Fig. 1e, Table 1). The D values of the CCRCCs were significantly different from those of the ChRCCs and PRCCs (F = 35.783, P < 0.05); however, the D values of the RAMFs did not differ significantly from the D values of the ChRCCs and PRCCs (t = 2.637, 4.329, P > 0.05). Using the D value (Fig. 1f) of 1.10 as the threshold value for differentiating CCRCCs from RAMFs, ChRCCs, and PRCCs, the best result obtained had a sensitivity of 81.0% and a specificity of 85.7%.

Table 1.

The IVIM parameters of CCRCC, RAMF, ChRCC, and PRCC.

Type	D (×10^–3 mm²/s)	D* (mm²/s)	f (%)
CCRCC	1.36 ± 0.23	0.033 ± 0.007	0.45 ± 0.13
RAMF	0.71 ± 0.05	0.051 ± 0.01	0.30 ± 0.06
ChRCC	0.73 ± 0.07	0.021 ± 0.006	0.36 ± 0.07
PRCC	0.62 ± 0.04	0.028 ± 0.006	0.33 ± 0.08
F values	37.229	30.746	11.080
P values	0.000	0.000	0.002

The D values of the CCRCCs significantly different from those of the ChRCCs and PRCCs (F = 35.783, P < 0.05); however, the D values of the RAMFs did not differ significantly from the D values of the ChRCCs and PRCCs (t = 2.637, 4.329, P > 0.05). The D* values of the CCRCCs significantly differed from the D* values of the ChRCCs (t = 6.732, P < 0.05); however, there was no significant difference between the ChRCCs and PRCCs (t = 3.128, P > 0.05). The f values of the RAMFs did not differ significantly from the f values of the ChRCCs and PRCCs (F = 4.692, P > 0.05).

The D* values (Fig. 2) were highest for the RAMFs, lowest for the ChRCCs, and intermediate for the CCRCCs and PRCCs (F = 30.746, P < 0.05; Fig. 2e, Table 1). The D* values of the CCRCCs differed significantly from the D* values of the ChRCCs (t = 6.732, P < 0.05); however, there was no significant difference between the ChRCCs and PRCCs (t = 3.128, P > 0.05). Using the D* (Fig. 2f) value of 0.038 as the threshold value for differentiating RAMFs from CCRCCs, ChRCCs, and PRCCs, the best result obtained had a sensitivity of 90.5% and a specificity of 76.2%.

The f values (Fig. 3) were highest for the CCRCCs, lowest for the PRCCs, and intermediate for the ChRCCs and RAMFs (F = 11.080, P < 0.05; Fig. 3e, Table 1). However, there was no significant difference between the RAMFs, ChRCCs, and PRCCs, and the value was only slightly and non-significantly increased for the ChRCCs (F = 4.692, P > 0.05). Using the f (Fig. 3f) value of 0.41 as the threshold value for differentiating the CCRCCs from the RAMFs, ChRCCs, and PRCCs, the best result obtained had a sensitivity of 66.8% and a specificity 81.0%.

Discussion

The cure rate for patients with localized RCCs who are treated with radical nephrectomy is only 50–65% (7). The main reason for cancer death is the presence of micrometastases that are undetectable at the time of diagnosis. A major problem in the management of patients with renal tumors is predicting the tumors’ malignancy potential. A variety of prognostic features, such as grade, size, cell type, tumor pattern, deoxyribonucleic acid (DNA) ploidy, and nuclear morphometry, have been proposed as prognostically useful parameters but have achieved only limited clinical application. Currently, only few studies have reported the use of IVIM for renal tumors.

Diffusion-weighted imaging (DWI) can provide physiologic information (8) about RCCs. Standard DWI and diffusion tensor imaging (DTI) techniques assume a Gaussian diffusion distribution (9), and they can indicate a small diffusion distance in the tissue, but the reason (high viscosity or more cell membranes) for the distance remains unclear. IVIM can provide a unique view of the tissue perfusion without the use of exogenous contrast agents. Because renal tumor types differ in the degree of vascularity and the cellularity of the tumor architecture, the diffusion IVIM parameters D, D*, and f reflect distinct processes that contribute to ADC and may provide more sensitive and specific characterization of renal tumors.

IVIM is analyzed by bi-exponential fitting of the signal decay to account for the D component (10). The D values are influenced by tissue perfusion and tissue cellularity. The contribution of perfusion to the diffusion signal was investigated by Le Bihan et al. (11) in their study of IVIM. In our study, the D values were highest for the CCRCCs, lowest for the PRCCs, and intermediate for the ChRCCs and RAMFs (P < 0.05). The D values of the CCRCCs significantly differed from those of the ChRCCs and PRCCs (P < 0.05); however, there were no significant differences between the RAMFs, ChRCCs, and PRCCs, and the values were only slightly and non-significantly increased for the ChRCCs (P > 0.05). Many studies have attributed lower D values to higher cellularity (12). The reason for the reduced D values can either be an increased viscosity of the tumor tissue (13) or a mechanical hindrance of water motion due to barriers, such as cell membranes. The cells of CCRCCs are rich in lipid content, and cholesterol, neutral lipids, and phospholipids are abundant on pathology (14). In our study, a D value of 1.10 was used as the threshold value for differentiating CCRCCs from RAMFs, ChRCCs, and PRCCs, and the best result obtained had a sensitivity of 81.0% and a specificity of 85.7%.

The blood flow in randomly oriented microvasculature, referred to as D* (15), contributes to diffusion signal decay predominantly at low b values (<200 s/mm²). Capillary density can influence D* values. Increasing D* values are likely caused by increasing capillary density in the tissue (16). RAMFs and CCRCCs are rich in capillaries on pathology. In our study, the D* values were highest for the RAMFs, lowest for the ChRCCs, and intermediate for the CCRCCs and PRCCs (P < 0.05). Statistically significant differences were observed between the D* values of the CCRCCs and RAMFs. Moreover, the D* values of the CCRCCs significantly differed from the D* values of the ChRCCs. On pathology, capillaries are frequently observed in RAMFs and CCRCCs (16). Huang et al. (17) applied diffusional kurtosis imaging to a normal human kidney and reported preliminary diffusional kurtosis imaging measurements. The results of these authors indicated that the rich vasculature complicates the interpretation of the diffusion pattern in the kidney. There was a good consistency between microscopic appearances of the RAMFs and CCRCCs and the D* characteristics.

Rheinheimer et al (18) applied the IVIM technique to renal tumor types and found that the f value provided significant added value, particularly for the differentiation of CRCCs from non-CCRCCs. Notohamiprodjo (19) found a significant corticomedullary difference with a higher average f value in the cortex than in the medulla. This finding appears reasonable because 90% of the renal blood volume is distributed in the cortex (20,21). Sigmund et al. (22) conducted both IVIM and DTI of a human kidney under hydration and furosemide flow challenges and revealed that the corticomedullary contrast of the kidney on diffusion imaging is sensitive to vascular flow. In our study, we demonstrated that the f values were highest for the CCRCCs, lowest for the PRCCs, and intermediate for the ChRCCs and RAMFs (P < 0.05). However, there were no significant difference between the RAMFs, ChRCCs, and PRCCs, and there was only a slight and non-significant increase for the ChRCCs (P > 0.05). On pathology, CCRCCs originate from the renal cortex and are rich in blood flow (16). There was a good consistency between the microscopic appearances of the CCRCCs and the f characteristics.

The main limitation of our study is the small number of patients with each renal tumor type. We recommend further studies with larger populations to validate the results of our study. Furthermore, respiratory movements of the kidney mainly occur in the craniocaudal direction and do not always coincide with abdominal wall movements.

In conclusion, the results of this study demonstrated significant differences in IVIM between renal tumor types. This technique can potentially be used as another non-invasive biomarker for the differential diagnoses of renal tumor types. However, percutaneous biopsy cannot be skipped with these imaging features because of much overlap between renal tumors.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors received: Jiangsu Provincial Medical Youth Talent (QNRC2016349). Jiangsu Provincial Commission of Health and Family Planning (H201551); Six Talent Peaks Project in Jiangsu Province (WSW-079). Social development project of Yangzhou City(YZ2017066).

References

Denton

Magee

Ovuworie

et al.

Prevalence of renal cell carcinoma in patients with ESRD pre-transplantation: a pathologic analysis. Kidney Int 2002; 61: 2201–2209.

Kucharz

Michałowska-Kaczmarczyk

et al.

The changes in complete blood count in patients treated with sunitinib malate for metastatic clear cell renal cell carcinoma. Przegl Lek 2013; 70: 712–714.

Yap

Ong

et al.

Clinical prognostic factors and survival outcome in renal cell carcinoma patients – a malaysian single centre perspective. Asian Pac J Cancer Prev 2013; 14: 7497–7500.

Coquia

Johnson

Ahmed

et al.

MDCT imaging following nephrectomy for renal cell carcinoma: Protocol optimization and patterns of tumor recurrence. World J Radiol 2013; 5: 436–445.

Chandarana

Lee

Hecht

et al.

Comparison of biexponential and monoexponential model of diffusion weighted imaging in evaluation of renal lesions: preliminary experience. Invest Radiol 2011; 46: 285–291.

Notohamiprodjo

Dietrich

Horger

et al.

Diffusion tensor imaging (DTI) of the kidney at 3 Tesla – feasibility, protocol evaluation and comparison to 1.5 Tesla. Invest Radiol 2010; 45: 245–254.

Kinouchi

Mano

Matsuoka

et al.

Immature tumor angiogenesis in high-grade and high-stage renal cell carcinoma. Urology 2003; 62: 765–770.

Doğanay

Kocakoç

Ciçekçi

et al.

Ability and utility of diffusion weighted MRI with different b values in the evaluation of benign and malignant renal lesions. Clin Radiol 2011; 66: 420–425.

Cova

Squillaci

Stacul

et al.

Diffusion-weighted MRI in the evaluation of renal lesions: preliminary results. Br J Radiol 2004; 77: 851–857.

10.

Le Bihan

. Intravoxel incoherent motion perfusion MR imaging: a wake-up call. Radiology 2008; 249: 748–752.

11.

Le Bihan

Breton

Lallemand

et al.

Separation of diffusion and perfusion in intravoxel incoherent motion MR imaging. Radiology 1988; 168: 497–505.

12.

Eckerbom

Hansell

Bjerner

et al.

Intravoxel incoherent motion MR imaging of the kidney: pilot study. Adv Exp Med Biol 2013; 765: 55–58.

13.

MacLennan

Bostwick

. Microvessel density in renal cell carcinoma: lack of prognostic significance. Urology 1995; 46: 27–30.

14.

Koh

Lim

Choi

et al.

Renal histologic parameters influencing postoperative renal function in renal cell carcinoma patients. Korean J Pathol 2013; 47: 557–562.

15.

Ichikawa

Motosugi

Ichikawa

et al.

Intravoxel incoherent motion imaging of the kidney: alterations in diffusion and perfusion in patients with renal dysfunction. Magn Reson Imaging 2013; 31: 414–417.

16.

Wang

Cheng

Zhang

. Renal cell carcinoma: diffusion-weighted MR imaging for subtype differentiation at 3.0 T. Radiology 2010; 257: 135–143.

17.

Huang

Chen

Zhang

et al.

MRI quantification of non-Gaussian water diffusion in normal human kidney: a diffusional kurtosis imaging study. NMR Biomed 2015; 28: 154–161.

18.

Rheinheimer

Stieltjes

Schneider

et al.

Investigation of renal lesions by diffusion-weighted magnetic resonance imaging applying intravoxel incoherent motion-derived parameters–initial experience. Eur J Radiol 2012; 81: e310–316.

19.

Notohamiprodjo

Chandarana

Mikheev

et al.

Combined intravoxel incoherent motion and diffusion tensor imaging of renal diffusion and flow anisotropy. Magn Reson Med 2015; 73: 1526–1532.

20.

Lee

Rusinek

Bokacheva

et al.

Renal function measurements from MR renography and a simplified multicompartmental model. Am J Physiol Renal Physiol 2007; 292: F1548–1559.

21.

Notohamiprodjo

Sourbron

Staehler

et al.

Measuring perfusion and permeability in renal cell carcinoma with dynamic contrastenhanced MRI: a pilot study. J Magn Reson Imaging 2010; 31: 490–501.

22.

Sigmund

Vivier

Sui

et al.

Intravoxel incoherent motion and diffusion-tensor imaging in renal tissue under hydration and furosemide flow challenges. Radiology 2012; 263: 758–769.