T2*-weighted imaging and DCE-MRI as complementary tools to characterize the continuous process of radionecrosis and neovascularization
Jérémie P. Fouquet1, Julie Constanzo1, Laurence Masson-Côté1,2, Luc Tremblay1, Philippe Sarret3, Sameh Geha4, Kevin Whittingstall5, Benoit Paquette1, and Martin Lepage1

1Department of Nuclear Medicine and Radiobiology, Université de Sherbrooke, Sherbrooke, QC, Canada, 2Service of Radiation Oncology, Centre Hospitalier Universitaire de Sherbrooke, Sherbrooke, QC, Canada, 3Department of Pharmacology-Physiology, Université de Sherbrooke, Sherbrooke, QC, Canada, 4Department of Pathology, Université de Sherbrooke, Sherbrooke, QC, Canada, 5Department of Diagnostic Radiology, Université de Sherbrooke, Sherbrooke, QC, Canada

Synopsis

Radiation dose delivered to healthy tissues during brain tumors radiosurgery can cause important side effects. We imaged an animal model of brain irradiation with DCE-MRI and T2*-weighted imaging at different time points after treatment. DCE-MRI allowed the discrimination of areas with high vessel permeability and necrotic regions. T2*-weighted imaging enabled the visualization of a necrotic core and micro-lesions at its periphery. Micro-lesions were initially co-localized with permeable vessels and later evolved into necrosis. Together, DCE-MRI and T2*-weighted images provided a coherent picture on the phenomena involved in radionecrosis progression, which could help in the management of associated problems.

Introduction

During radiosurgery of brain tumors, surrounding healthy tissue may receive significant radiation dose. This can induce important side effects such as cognitive decline and stroke-like symptoms. A better characterization of the underlying processes could lead to more appropriate patient management 1. We characterized radiation necrosis in an animal model and showed how T2*‑weighted (T2*w) imaging combined with dynamic contrast‑enhanced MRI (DCE‑MRI) can be used to monitor and predict physiological changes after radiosurgery.

Methods and materials

Eighteen Fischer rats were irradiated using Leksell Gamma Knife Perfexion® and a custom-designed stereotactic bed. A single isocenter treatment plan was used to deliver ≥ 100 Gy to the brain primary somatosensory forelimb area (S1FL, right hemisphere) 2. Rats were scanned with a small-animal 7T MRI scanner (Varian Inc., Palo Alto, CA) and a dedicated rat head-coil (RAPID MR International, OH) before treatment and 16, 21, 54, 82 and 110 days following irradiation. At every imaging session, rats underwent T2-weighted (T2w), Gd-DTPA (Magnevist, Berlex Canada Inc., Montreal, Canada) DCE-MRI and T2*w sequences, in this order (see Table 1 for sequence parameters). DCE-MRI concentrations were estimated using a variable flip angle approach and were normalized to the surrounding muscle concentration. For display purposes, the minimum intensity projection (mIP) of the T2*w volume was computed for each corresponding DCE-MRI slice. On day 110, histopathology was performed on brain sections stained with chicken anti-GFAP antibody (#AB5541, Millipore, CA, USA), rabbit anti-Iba1 (#019-19741, Wako Chemicals, VA, USA), luxol fast blue counterstained with cresyl violet and hematoxylin and eosin, to characterize respectively astrocytosis reaction, inflammatory responses, myelin damage and necrosis.

Results

From day 1 to day 21, no alterations were observed on all T2w, T2*w and DCE-MR images (data not shown). On day 54, hyperintense areas on T2w images indicated the presence of edema (Fig. 1). DCE-MRI revealed contrast agent accumulation in a volume of about 100 mm3 in the irradiated area using, indicating the formation of highly permeable vessels. The DCE-MRI time curve enabled the discrimination of two different sub-regions: a “slow-enhancing” region and a “fast-enhancing” region, respectively located in the center and at the periphery of the contrast agent exposed volume (Fig. 2). On T2*w images, several hypointense regions were observed in the irradiated area, with a large region in the center of the affected area and smaller ones at the periphery (Fig. 1). On day 110, similar phenomena were observed on all MR images but at a larger scale. The difference between “slow-enhancing” and “fast-enhancing” regions on DCE-MRI time curves was more obvious (Fig. 2). Multiple and larger hypointense regions were detected on T2*w images. Hypointense areas were also visible on T2w images (Fig. 1). Necrosis, neovessels and an inflammatory response were corroborated by histopathology on day 110.

Discussion and conclusion

From day 54 to day 110, the volume of hypointense T2*w signal increased dramatically, revealing important field inhomogeneities. These were attributed to micro-lesions in brain tissue where vascularization was developing. We cannot confirm the exact nature of these micro‑lesions, but it would be plausible to relate them to hemorrhage 3. The DCE-MRI time curves confirmed that necrosis occurred at the center of the affected region and that permeable neovessels at the periphery were responsible for contrast agent accumulation. Histopathology confirmed the presence of neovessels and inflammation. These observations suggest that radionecrosis progressed from a small region and then expanded, presumably through an unresolved inflammatory process. Micro‑lesions detected by T2*w imaging initially co-existed with neovascularization eventually leading to necrosis. The inflammatory reaction slowly moved outwards and radionecrosis proceeded until an endpoint was reached. Together, DCE-MRI and T2*w images provided a coherent picture on the phenomena involved in the progression of radionecrosis. Combined with other observations 4,5, this increases our understanding of radionecrosis and could lead to better management of related problems.

Acknowledgements

This work was supported by the Fonds de Recherche Québécois Nature et Technologies (grant no 172009). Martin Lepage, Laurence Masson-Côté, Benoit Paquette, Philippe Sarret and Kevin Whittingstall are members of the FRQS-funded Centre de recherche du CHUS. The authors thank the Electron Microscopy & Histology Research Core at the Université de Sherbrooke for their histology services.

References

1. Parvez K, Parvez A, Zadeh G. The Diagnosis and Treatment of Pseudoprogression, Radiation Necrosis and Brain Tumor Recurrence. Int J Mol Sci. 2014;15(7):11832–11846.

2. Constanzo J, Paquette B, Charest G, et al. Gamma Knife irradiation method based on dosimetric controls to target small areas in rat brains Gamma Knife irradiation method based on dosimetric controls to target small areas in rat brains. Med Phys. 2015;42(5):2311–2316.

3. Zeng Q-S, Kang X-S, Li C-F, et al. Detection of hemorrhagic hypointense foci in radiation injury region using susceptibility-weighted imaging. Acta radiol. 2011;52:115–119.

4. Chan KC, Khong P-L, Cheung MM, et al. MRI of late microstructural and metabolic alterations in radiation-induced brain injuries. J Magn Reson Imaging. 2009;29(5):1013–20.

5. Perez-Torres CJ, Yuan L, Schmidt RE, et al. Perilesional edema in radiation necrosis reflects axonal degeneration. Radiat Oncol. 2015;10:10–13.

Figures

Table 1 - MRI sequence parameters

Figure 1 - Typical images obtained 54 and 110 days after irradiation. The mIP of the T2*w substack corresponding to the DCE-MRI slice is displayed. The DCE-MRI first time frame with significant contrast enhancement is displayed.

Figure 2 - DCE-MRI time curves for two regions in the affected area 54 and 110 days after irradiation, with corresponding images for three time frames. Dashed lines indicate at which time points the images correspond. A shift in the maximum of the concentration curve indicates necrosis.



Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
2795