Assessing neurotoxicity from the low-dose radiation component of radiosurgery using magnetic resonance spectroscopy

Abstract

The aim of the study was to determine if biochemical changes indicative of injury, assessed using magnetic resonance spectroscopic imaging (SI), are observed after stereotactic radiosurgery (SRS). The study included patients who underwent SI immediately before and 1, 30, and 90 days following SRS. Short TE spectra (TR/TE 1000/35 ms) were acquired at the SRS isocenter with a 2D PRESS-CSI sequence on a single 1.5 T scanner. The SRS isodose lines were overlaid on the magnetic resonance imaging slice utilized for SI data acquisition. N-Acetyl aspartate (NAA)/creatine (Cr) and choline (Cho)/Cr ratios were computed for multiple voxels located between the 25 and 50 cGy isodose lines (low dose) and the 200 and 350 cGy isodose lines (medium dose). An analysis of variance and paired t-tests compared metabolite levels at different time points. Twelve patients were enrolled, although 3 were excluded secondary to poor spectral data quality or deviations from the prescribed SI protocol. The median number of voxels analyzed from the low- and medium-dose region was 7 and 4, respectively. No significant changes in metabolite peak height ratios over time were seen in the low-dose region, for either NAA/Cr (P = .89) or Cho/Cr (P = .85). There was no difference in Cho/Cr peak height ratios in the medium-dose region (P = .62). There was an increase in the NAA/Cr peak height ratio in the medium-dose region between day -1 and day +30 (P = .003), followed by a decline to baseline between days +30 and +90 (P = .03). We did not observe a significant decline in NAA/Cr or change in Cho/Cr peak heights in uninvolved brain parenchyma after SRS.

Fig. 1.

Individual patient data for the low-dose region (25–50 cGy). Median NAA/Cr (solid squares) and Cho/Cr (open squares) values, with error bars representing 1 standard deviation, are shown.

Fig. 2.

Individual patient data for the medium-dose region (200–350 cGy). Median NAA/Cr (solid squares) and Cho/Cr (open squares) values, with error bars representing 1 standard deviation, are shown.

Fig. 3.

Normalized data from all patients for both NAA/Cr (solid squares) and Cho/Cr (open squares), showing changes in metabolite ratios from voxels located between the 25 to 50 cGy isodose lines.

Fig. 4.

Normalized data from all patients for both NAA/Cr (solid squares) and Cho/Cr (open squares), showing changes in metabolite ratios from voxels located between the 200 and 350 cGy isodose lines.