Imaging Tissue Microstructure
Sungheon Gene Kim1
1Weill Cornell Medical College, New York, NY, United States

Synopsis

Diffusion MRI is a unique imaging modality suitable for probing tissue microstructure. Unlocking the full potential of diffusion MRI for microstructural imaging requires an adequate diffusion encoding method in terms of diffusion time and gradient strength. dMRI Diffusivity is typically interpreted as a rough measure of the cell density and extracellular water fraction. However, diffusivity and other commonly used dMRI metrics remain non-specific markers, because a diffusion measurement at a fixed diffusion time is affected by multiple factors. The complex microstructural features of tumors and the brain can be probed by adequate sampling of the multi-dimensional space of diffusion encoding.

Highlights

  • Diffusion MRI is a unique imaging modality suitable for probing tissue microstructure.
  • Diffusion time dependency of diffusion MRI parameters can be used to measure microstructural properties of cancer, such as cell size and cell membrane permeability.
  • The complex microstructural features of the brain can be probed by adequate sampling of the multi-dimensional space of diffusion encoding.

Problem Summary

Diffusivity measured from diffusion MRI (dMRI), D, is typically interpreted as a rough measure of the cell density and extracellular water fraction. Consistent with this interpretation, a decrease in D of malignant brain tumors was correlated with increased cell density measured by histology. However, D and other commonly used dMRI metrics remain non-specific markers, because a diffusion measurement at a fixed diffusion time is affected by multiple factors, such as cell size, cell density, composition of the extracellular matrix, and compartmental diffusivities. Since cancer treatment can induce changes in these factors, it remains challenging to understand how these biophysical changes in the tumor tissue affect dMRI metrics. Similar ambiguities are observed when imaging the complex microstructure of the brain. Imaging tissue microstructure using diffusion MRI can be achieved by adequate sampling of the multi-dimensional space of diffusion encoding.

Body

In the case of simple Gaussian diffusion (e.g. in free water), dMRI measurement is characterized by a single parameter, the b-value b = q2t, such that the dMRI signal (the propagator) decays as S=S0exp(-bD). Tissue complexity manifests itself in non-Gaussian diffusion (1, 2), which makes the signal S depend on the diffusion weighting gradient strength q and diffusion time t separately, and is characterized by (i) the presence of the higher-order terms in the cumulant expansion of the signal (3), such as the kurtosis term K (4), and (ii) the time-dependence of all the cumulants including D(t) and K(t) (5). Hence, tissue complexity can be probed in two complementary directions (6): (i) to quantify higher-order cumulants at a given diffusion time, by increasing q (or the b-value at fixed t), and (ii) to probe the time dependence of the cumulants by varying the diffusion time t as cumulants are the signal derivatives at b=0. In both directions, biophysical modeling of diffusion in the tissue microenvironment is required to quantify microstructural changes. In this lecture, we will explore D(t) and K(t) ­– along direction (ii) and discuss about the possibility of measuring cell viability and water exchange time.
Successful treatment results in cell death that can increase the space between cells and also increase the permeability of cell membrane (7). This change can be detected as an increase of D. However, it is often not clear how much of increase in D can be considered sufficient to determine complete response. The characteristic time tr for restrictive effects is ~L2/D where L is a distance typical of the spacing between the barriers and D is the diffusivity of the bulk fluid. As the radius of cancer cells is about 5 mm and D ~ 1 mm2/ms, tr is about 25 ms; diffusion time dependence, D(t), is expected to be observed most clearly when the diffusion time of dMRI scan is around t = 25 ms. In a solid tumor, the extracellular space could exhibit even smaller distances between the restricting membranes, i.e. tr < 25 ms for the extracellular space. This could be the case for tumors before the treatment. In this case, the diffusion times typically used for clinical dMRI scans (50~100ms) are beyond the tr such that the D(t) would not change noticeably.
However, when the cancer is treated with an effective therapy, the treatment-induced cell death can create larger gaps between the membranes that restrict water diffusion. If the average distance is about 10 mm and D ~ 1 mm2/ms, tr is about 100 ms. In this case, D can vary noticeably over the diffusion times used for clinical scans. It is not known which diffusion time is optimal to determine treatment response.
Recent studies have introduced a couple of advanced approaches to use t-dependent dMRI, such as VERDICT (vascular, extracellular, and restricted diffusion for cytometry in tumors) (8, 9), IMPULSED (imaging microstructural parameters using limited spectrally edited diffusion) (10, 11), and POMACE (pulsed and oscillating gradient MRI for assessment of cell size and extracellular space) (12, 13). In addition to ve, VERDICT measures tumor vascular fraction, whereas IMPULSED and POMACE measure cell size. These methods may provide more specific information about the tissue microstructural changes induced by tumor growth or treatment.
In addition to D, the deviation of diffusion signal from the Gaussian behavior is often measured in terms of the diffusional kurtosis K (14). It has been demonstrated that K can be measured in the metastatic nodes of head and neck cancer patients using an extended range of b-values up to 1,500 s/mm2 (15, 16). Non-Gaussian diffusion has also been observed in prostate cancer (17) and hepatocellular carcinomas (18). Furthermore, it was shown that K has greater sensitivity and specificity than D for assessment of hepatocellular carcinoma viability after treatment (19). Wu et al (20) measured D and K before and 7 days after chemotherapy in cervical non-Hodgkin's lymphomas and found that D increased significantly while K decreased significantly. These studies suggest that K is also a promising biomarker for evaluation of treatment response, providing potentially complementary information to D.
However, it is often not fully recognized that K is highly sensitive to diffusion time over the typical range of diffusion times achievable in clinical scanners. K(t) is closely related to the water exchange time tex = veti = (1- ve)te where ti the intracellular water lifetime and te the extracellular water lifetime. The function K(t) changes the most on the scale of t~tex as shown by our earlier studies (14, 21). In our preliminary study, we found that it is possible to measure tex from K(t) in patients with breast cancer and head and neck cancer, as well as in mouse tumor models.
For imaging brain microstructure, dMRI has been widely used for its unique capability of probing the complex brain axonal pathways. There has been a considerable effort to increase the imaging resolution of MRI. A recent study from Wang et al (22) is a good example in such effort in preclinical imaging. In this study, they were able to collect diffusion data with an isotropic resolution of 25 um. Their high resolution images show excellent contrast for small structural details which used to be available only from histology images. However the scan time is close to 4 days, too long to make such high resolution imaging practical at the moment. In addition, extracting the microstructural information content in individual voxels is still not possible even with such high resolution images.
Proving a complex tissue structure requires an adequate sampling of the q-space. Some of popular sampling schemes include diffusion spectrum imaging, high angular resolution diffusion imaging, and multi-shell, sparse hybrid diffusion imaging. In addition, we also need to use an appropriate way to analyze the data. Typically, some type of basis function is used to fit the data to estimate the fiber orientation distribution function FOD. And the fit results are used to extract quantitative data about the tissue microstructure. The restricted spectrum model (23) is one of good examples that aims to reveal the degree of restriction in the brain. As many models are proposed by the research community, there is a growing need to compare the different methods against the ground truth. Schilling et al (24) used confocal microscopy images to estimate 3D FOD as a reference standard data. The confocal FOD was compared with the FODs estimated using different data acquisition and analysis methods shown in this slide. The study showed that each method has its own pros and cons depending on what feature of the tissue complexity is to be measured. But overall, this comparison also supports that what we can measure with diffusion MRI is reasonably close to the actual tissue structure.
Diffusion time dependency of diffusion MRI has been also used for studying brain microstructures. In a recent study by Lee et al (25), they investigated the axonal diffusivity along the axons segmented from the 3D EM images. Their segmentation results showed that there is a remarkable variation of axon caliber along the axon. In a simulation study using the segmented axons, they found that the water diffusivity along the fiber decreases with diffusion time. When the caliber variation is removed, the axonal diffusivity did not decrease with diffusion time. A similar diffusion time dependency was observed in the human brain white matter, suggesting the possibility of using the time dependency as a new way to assess changes in aging and neuro-degenerative diseases.

Summary/recap/follow-up

Diffusion MRI is a powerful tool for imaging tissue microstructure. In order to unlock its potential, it is important to utilize the different information that can be obtained from diffusion MRI with the two dimensional space of q and t, as demonstrated by several examples discussed. Preclinical imaging can provide unique data for validation of various approaches that have been suggested. There are many other promising techniques to utilize the full potential of diffusion MRI for tissue microstructure imaging, including another dimension of utilizing diffusion encoding for tissue microanisotropy.

Acknowledgements

No acknowledgement found.

References

1. Niendorf T, Dijkhuizen RM, Norris DG, Campagne MV, Nicolay K. Biexponential diffusion attenuation in various states of brain tissue: Implications for diffusion-weighted imaging. Magnet Reson Med. 1996;36(6):847-57. doi: DOI 10.1002/mrm.1910360607. PubMed PMID: WOS:A1996VU89000006.

2. Mulkern RV, Gudbjartsson H, Westin CF, Zengingonul HP, Gartner W, Guttmann CRG, Robertson RL, Kyriakos W, Schwartz R, Holtzman D, Jolesz FA, Maier SE. Multi-component apparent diffusion coefficients in human brain. Nmr Biomed. 1999;12(1):51-62. doi: Doi 10.1002/(Sici)1099-1492(199902)12:1<51::Aid-Nbm546>3.0.Co;2-E. PubMed PMID: WOS:000079191200008.

3. Kiselev VG. Ch 10. The Cumulant Expansion: an Overarching Mathematical Framework for Understanding Diffusion NMR. In: Diffusion MRI: theory, methods and applications, by Jones, DK Oxford University Press, New York. 2010.

4. Jensen JH, Helpern JA, Ramani A, Lu HZ, Kaczynski K. Diffusional kurtosis imaging: The quantification of non-Gaussian water diffusion by means of magnetic resonance imaging. Magnet Reson Med. 2005;53(6):1432-40. doi: Doi 10.1002/Mrm.20508. PubMed PMID: WOS:000229468200025.

5. Novikov DS, Kiselev VG. Effective medium theory of a diffusion-weighted signal. Nmr Biomed. 2010;23(7):682-97. doi: Doi 10.1002/Nbm.1584. PubMed PMID: WOS:000283014300003.

6. Burcaw LM, Fieremans E, Novikov DS. Mesoscopic structure of neuronal tracts from time-dependent diffusion. NeuroImage. 2015;114:18-37. doi: 10.1016/j.neuroimage.2015.03.061. PubMed PMID: 25837598; PMCID: 4446209.

7. Thoeny HC, Ross BD. Predicting and monitoring cancer treatment response with diffusion-weighted MRI. Journal of magnetic resonance imaging : JMRI. 2010;32(1):2-16. doi: 10.1002/jmri.22167. PubMed PMID: 20575076; PMCID: 2918419.

8. Panagiotaki E, Chan RW, Dikaios N, Ahmed HU, O'Callaghan J, Freeman A, Atkinson D, Punwani S, Hawkes DJ, Alexander DC. Microstructural characterization of normal and malignant human prostate tissue with vascular, extracellular, and restricted diffusion for cytometry in tumours magnetic resonance imaging. Investigative radiology. 2015;50(4):218-27. doi: 10.1097/RLI.0000000000000115. PubMed PMID: 25426656.

9. Panagiotaki E, Walker-Samuel S, Siow B, Johnson SP, Rajkumar V, Pedley RB, Lythgoe MF, Alexander DC. Noninvasive quantification of solid tumor microstructure using VERDICT MRI. Cancer research. 2014;74(7):1902-12. doi: 10.1158/0008-5472.CAN-13-2511. PubMed PMID: 24491802.

10. Jiang X, Li H, Xie J, McKinley ET, Zhao P, Gore JC, Xu J. In vivo imaging of cancer cell size and cellularity using temporal diffusion spectroscopy. Magnetic resonance in medicine : official journal of the Society of Magnetic Resonance in Medicine / Society of Magnetic Resonance in Medicine. 2017;78(1):156-64. doi: 10.1002/mrm.26356. PubMed PMID: 27495144; PMCID: 5293685.

11. Jiang X, Li H, Xie J, Zhao P, Gore JC, Xu J. Quantification of cell size using temporal diffusion spectroscopy. Magnetic resonance in medicine : official journal of the Society of Magnetic Resonance in Medicine / Society of Magnetic Resonance in Medicine. 2016;75(3):1076-85. doi: 10.1002/mrm.25684. PubMed PMID: 25845851; PMCID: 4592783.

12. Reynaud O, Winters KV, Hoang DM, Wadghiri YZ, Novikov DS, Kim SG. Pulsed and oscillating gradient MRI for assessment of cell size and extracellular space (POMACE) in mouse gliomas. Nmr Biomed. 2016;29(10):1350-63. doi: 10.1002/nbm.3577. PubMed PMID: 27448059; PMCID: 5035213.

13. Reynaud O, Winters KV, Hoang DM, Wadghiri YZ, Novikov DS, Kim SG. Surface-to-volume ratio mapping of tumor microstructure using oscillating gradient diffusion weighted imaging. Magnetic resonance in medicine : official journal of the Society of Magnetic Resonance in Medicine / Society of Magnetic Resonance in Medicine. 2016;76(1):237-47. doi: 10.1002/mrm.25865. PubMed PMID: 26207354; PMCID: 4724565.

14. Jensen JH, Helpern JA, Ramani A, Lu H, Kaczynski K. Diffusional kurtosis imaging: the quantification of non-gaussian water diffusion by means of magnetic resonance imaging. Magn Reson Med. 2005;53(6):1432-40. doi: 10.1002/mrm.20508. PubMed PMID: 15906300.

15. Jansen JF, Stambuk HE, Koutcher JA, Shukla-Dave A. Non-gaussian analysis of diffusion-weighted MR imaging in head and neck squamous cell carcinoma: A feasibility study. AJNR American journal of neuroradiology. 2010;31(4):741-8. doi: 10.3174/ajnr.A1919. PubMed PMID: 20037133; PMCID: 2854270.

16. Yuan J, Yeung DK, Mok GS, Bhatia KS, Wang YX, Ahuja AT, King AD. Non-Gaussian analysis of diffusion weighted imaging in head and neck at 3T: a pilot study in patients with nasopharyngeal carcinoma. PloS one. 2014;9(1):e87024. doi: 10.1371/journal.pone.0087024. PubMed PMID: 24466318; PMCID: 3900693.

17. Rosenkrantz AB, Sigmund EE, Johnson G, Babb JS, Mussi TC, Melamed J, Taneja SS, Lee VS, Jensen JH. Prostate cancer: feasibility and preliminary experience of a diffusional kurtosis model for detection and assessment of aggressiveness of peripheral zone cancer. Radiology. 2012;264(1):126-35. doi: 10.1148/radiol.12112290. PubMed PMID: 22550312.

18. Rosenkrantz AB, Sigmund EE, Winnick A, Niver BE, Spieler B, Morgan GR, Hajdu CH. Assessment of hepatocellular carcinoma using apparent diffusion coefficient and diffusion kurtosis indices: preliminary experience in fresh liver explants. Magnetic resonance imaging. 2012;30(10):1534-40. doi: 10.1016/j.mri.2012.04.020. PubMed PMID: 22819175.

19. Goshima S, Kanematsu M, Noda Y, Kondo H, Watanabe H, Bae KT. Diffusion kurtosis imaging to assess response to treatment in hypervascular hepatocellular carcinoma. AJR American journal of roentgenology. 2015;204(5):W543-9. doi: 10.2214/AJR.14.13235. PubMed PMID: 25905960.

20. Wu R, Suo ST, Wu LM, Yao QY, Gong HX, Xu JR. Assessment of chemotherapy response in non-Hodgkin lymphoma involving the neck utilizing diffusion kurtosis imaging: a preliminary study. Diagnostic and interventional radiology. 2017;23(3):245-9. doi: 10.5152/dir.2017.16184. PubMed PMID: 28381389; PMCID: 5411008.

21. Fieremans E, Novikov DS, Jensen JH, Helpern JA. Monte Carlo study of a two-compartment exchange model of diffusion. Nmr Biomed. 2010;23(7):711-24. doi: 10.1002/nbm.1577. PubMed PMID: 20882537; PMCID: 2997614.

22. Wang N, White LE, Qi Y, Cofer G, Johnson GA. Cytoarchitecture of the mouse brain by high resolution diffusion magnetic resonance imaging. NeuroImage. 2020;216:116876. Epub 2020/04/29. doi: 10.1016/j.neuroimage.2020.116876. PubMed PMID: 32344062; PMCID: PMC7299741.

23. White NS, Leergaard TB, D'Arceuil H, Bjaalie JG, Dale AM. Probing tissue microstructure with restriction spectrum imaging: Histological and theoretical validation. Hum Brain Mapp. 2013;34(2):327-46. Epub 2012/11/22. doi: 10.1002/hbm.21454. PubMed PMID: 23169482; PMCID: PMC3538903.

24. Schilling KG, Janve V, Gao Y, Stepniewska I, Landman BA, Anderson AW. Histological validation of diffusion MRI fiber orientation distributions and dispersion. NeuroImage. 2018;165:200-21. Epub 2017/10/28. doi: 10.1016/j.neuroimage.2017.10.046. PubMed PMID: 29074279; PMCID: PMC5732036.

25. Lee HH, Papaioannou A, Kim SL, Novikov DS, Fieremans E. A time-dependent diffusion MRI signature of axon caliber variations and beading. Commun Biol. 2020;3(1):354. Epub 2020/07/09. doi: 10.1038/s42003-020-1050-x. PubMed PMID: 32636463; PMCID: PMC7341838.

Proc. Intl. Soc. Mag. Reson. Med. 29 (2021)