To evaluate renal dysfunction using diffusion weighted magnetic resonance imaging based on intra-voxel incoherent motion (IVIM) and a mono-exponential model – a comparison study
Jiule Ding1, Jie Chen1, Zhenxing Jiang1, Hua Zhou2, Jia Di2, Wei Xing1, and Yongming Dai3

1Department of Radiology, Third Affiliated Hospital of Suzhou University, Changzhou, China, People's Republic of, 2Department of Nephrology, Third Affiliated Hospital of Suzhou University, Changzhou, China, People's Republic of, 3Philips Healthcare, Shanghai, China, People's Republic of

Synopsis

The IVIM model was compared with the mono-exponential model to be used to differentiate sRI from non-sRI in this study. Results indicated that IVIM contributed little to improving the differentiation, therefore the mono-exponential model based ADC, a combination of fast and slow diffusion, might be more suitable as a biomarker image for assessing renal dysfunction.

Background

Renal dysfunction is a worldwide problem; and a non-invasive assessment is required for planning treatment and follow-up. Human kidney is one of the well-perfused organs,1 and its major function is related to water transportation (glomerular filtration, active and passive tubular reabsorption) under normal conditions.2 Intra-voxel incoherent motion (IVIM) can isolate fast from slow water motion in well-perfused organs and may be helpful in assessing renal dysfunction compared with a mono-exponential model.

Purpose

To evaluate IVIM in the assessment of renal dysfunction compared with a mono-exponential model.

Methods

Fifty-four subjects were enrolled into this study. The estimated glomerular filtration rate (eGFR) was assessed to classify the subjects as having severe renal injury group (sRI, eGFR < 30 ml/min/1.73 m2) or not (non-sRI). A free-breathing DWI with 7 b-factors was performed. A radiologist conducted Image analysis. An apparent diffusion coefficient map (ADCmon) was generated in mono-exponential model, and diffusion coefficient (Dslow and Dfast), and fraction of fast diffusion (Ffast) maps were generated in IVIM. The small circular regions of interest were placed at the interface between the cortex and medulla for parameter measurements.

Results

All four parameters (ADCmon, Dslow, Dfast, and Ffast) were less in sRI than non-sRI (P<0.05). However, only both ADCmon and Dslow were linearly related with eGFR (P<0.05). For differentiating sRI from non-sRI using the four parameters, receiver operating characteristic curve analysis shown no significant difference was present between them (P>0.05). Furthermore, the correlation was 0.93 between ADCmon and Dslow, larger than 0.57 between Dfast and Ffast, 0.48 between ADCmon and Dfast, and 0.34 between ADCmon and Ffast (P<0.05).

Conclusion

IVIM model can differentiate sRI from non-sRI, but ADCmon, a combination of fast and slow diffusion, might be more suitable as a biomarker image for assessing renal dysfunction.

Acknowledgements

No acknowledgement found.

References

1. Brezis M, Rosen S. Hypoxia of the renal medulla–its implications for disease. N Engl J Med. 1995; 332(10):647–655.

2. Yang D, Ye Q, Williams DS, et al. Normal and transplanted rat kidneys: diffusion MR imaging at 7 T. Radiology. 2004;231(3):702–709.



Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
1973