Purpose: To evaluate the presence of residual disease after surgery but before radiotherapy (RT) in patients with high-grade glioma by MRI and magnetic resonance spectroscopy imaging (MRSI) and to estimate the impact of MRSI on the definition of postoperative target volumes for RT treatment planning.
Methods and materials: Thirty patients (27 glioblastoma multiforme, 3 Grade III astrocytoma) underwent MRI and MRSI within 4 weeks after surgery but before the initiation of RT. The MRI data were manually contoured; the regions of interest included T(2)-weighted hyperintensity (T(2)), T(1)-weighted contrast enhancement (T(1)), and the resection cavity (RC). Levels of choline and N-acetyl-aspartate (NAA) in the three-dimensional MRSI data were analyzed on the basis of a choline-to-N-acetyl-aspartate index (CNI). The CNI and other metabolic indexes were superimposed on the MRI data as three-dimensional contours. Composite, conjoint, and disjoint volumes were defined for T(1) and T(2), with/without RC, and within the CNI contour, corresponding to a value of 2. In addition, follow-up MRI studies were examined for new onset contrast enhancement and compared with the initial spectroscopic findings obtained before RT.
Results: Substantial variation was found in the spatial relationship between the MRI and MRSI volumes. Ten patients had no contrast enhancement after surgery, and MRSI revealed abnormal metabolic activity in 8 of 10, averaging 20 cm(3) and extending 11-36 mm beyond the RC. In 20 patients with contrast-enhancing lesions, substantial variation was found between T(1) and CNI2; metabolic activity fell outside the contrast enhancement in 19 patients, averaging 21 cm(3) and extending 8-33 mm beyond the contrast enhancement. For all patients, the T(2) encompassed most of the metabolic volume. However, the CNI2 extended beyond the T(2) in 6 of 10 patients without contrast enhancement (mean, 8 cm(3); maximum, 15-23 mm) and in 13 of 20 patients with contrast enhancement (mean, 7 cm(3); maximum, 8-22 mm), representing an increase in the T(2) volume by as much as 180% (median 13%) and 86% (median 14%) for non-contrast-enhancing and contrast-enhancing patients, respectively. Preliminary evaluation of the MRI follow-up examinations revealed correspondence of areas of new contrast enhancement with initial MRSI abnormalities in 8 of 10 non-contrast-enhancing patients. In addition, CNI volumes correlated inversely with the time to onset of new contrast enhancement.
Conclusion: MRSI is a valuable diagnostic tool for the assessment of residual disease after surgical resection in high-grade glioma. The incorporation of areas of metabolic abnormality into treatment planning for postoperative patients would produce different sizes and shapes of target volumes for both primary and boost volumes. It also may encourage the use of nonuniform margins to define the extent of tumor cell infiltration, rather than the current use of uniform margins.