Purpose/Objective(s): Radiation therapy plays an integral role in the management of gliomas, but has been associated with long-term cognitive decline. Cerebral microbleeds (CMBs) have been identified in patients undergoing cranial radiation, and serve as radiographic markers for microvascular injury thought to contribute to late cognitive decline. While prior studies have established the utility of susceptibility-weighted MRI (SWI) in identifying CMBs, the relationship between CMB formation and radiation dose has not been well established. Materials/Methods: SWI was performed on 14 patients with stable malignant gliomas (WHO III-IV) between 2 and 4 years after chemoradiotherapy, for a total of 19 scans. Scans were characterized as either early (<3 years from treatment) or late (3 years from treatment). CMBs were identified as discrete foci of susceptibility on SWI that did not correspond to vessels. Radiation dosimetry maps were reconstructed and fused with the SWI images. Microbleed density for low (<30 Gy), medium (30-45 Gy) and high (>45 Gy) dose regions was computed as the number of CMBs within each dose region normalized by the region volume, and the microbleed density for each cortical lobe was computed as the number of CMBs within each lobe normalized by the volume of that lobe receiving at least 30 Gy. Results: Thirteen of 14 patients exhibited CMBs, with a total of 311 individual CMBs observed. The number of CMBs observed was significantly higher at late time points (early median: 6.5 CMBs, late median: 27 MBs, P = 0.001). 87% of all CMBs were seen in regions receiving at least 30 Gy, and the CMB density was significantly higher in medium and high dose regions compared to the low dose region at both time points (early: medium vs low P = 0.02, high vs low P = 0.004; late: Medium vs low P<0.001, high vs low P< 0.001). While the density of CMBs in high-dose compared to medium-dose regions was similar at both early and late time points (P = 0.42 and P = 0.9, respectively), late scans demonstrated proportionally more CMBs at lower doses than early scans (P = 0.006 by Kolmogorov-Smirnov test). CMB density varied by location, with a relatively lower density in the frontal lobe compared to other lobes seen at early but not at late time points (P = 0.026 and P = 0.46, respectively). Conclusion: While there appears to be a threshold of approximately 30 Gy for CMB formation initially, we do not find clear evidence of a dose-response relationship at doses higher than 30 Gy. However, the dose distribution of CMBs shifts over time, with a greater frequency of lesions in lower dose regions with increasing time from treatment. The frontal lobe appears to have reduced tendency for CMB formation at early time points compared to other cortical hemispheres. These findings have potentially important implications for understanding the late microvascular sequelae of cranial irradiation, and could inform clinical decisions regarding optimizing avoidance of brain regions sensitive to microvascular damage from high-dose radiation.
Wahl, M., Anwar, M., Hess, C., Chang, S. M., & Lupo, J. M. (2016). Relationship Between Radiation Dose and Microbleed Formation in Patients With Malignant Glioma. International Journal of Radiation Oncology*Biology*Physics, 96(2), E68. https://doi.org/10.1016/j.ijrobp.2016.06.762