MR-spectroscopy imaging for resected high-grade gliomas prior to radiation therapy: tumor extent according to metabolic activity in relation to MRI

Abstract

Purpose/Objective: We reported earlier on the value of 3D MR-Spectroscopy imaging (MRSI) in defining the extent of glioma infiltration as assessed in patients prior to definitive surgery. We have extended this study to evaluate the presence of residual disease following surgery but prior to radiation therapy (RT) in patients with high-grade gliomas. The goal was to investigate how postoperative target volume definition for treatment planning might be impacted by the incorporation of functional imaging such as MRSI.Materials/Methods: Thirty patients (27 GBM, 3 grade III astrocytomas) were evaluated, each of whom had undergone MRI and MRSI studies within 4 weeks post surgery but prior to the initiation of RT. For each patient, the MRI data set was manually contoured; regions of interest included the T2 region of hyperintensity (T2), the T1 region of contrast enhancement (T1), and the resection cavity. The 3D-MRSI peak parameters for choline (Cho) and N-acetyl-aspartate (NAA), acquired on a voxel-by-voxel basis, were categorized based on a Cho-to-NAA Index (CNI), a tool for quantitative assessment of tissue metabolite levels. CNI data were aligned to the MRI and displayed as 3D contours. T1 is classically considered to be the region of gross or active disease, T2 the region of microscopic and/or suspected disease, and CNI can be viewed as a continuous measure of disease activity with a CNI of 2 defined as the lowest value corresponding to biopsy confirmed tumor. Composite, as well as conjoint and disjoint volumes, were defined for T1 and T2, both including the resection cavity, vs. CNI 2, these were then compared to T1 and T2, respectively.Results: There was substantial variation in the relationship between the MRI and MRSI designated volumes (see table below). Ten patients did not show any contrast enhancement (CE) post surgery. In 8 of these patients, MRSI revealed metabolic activity (with a CNI 2 > 1.5 cc) averaging 20 cc that extended as much as 11-36 mm beyond the resection cavity. In the 20 patients with CE lesions, substantial variation between the T1 and CNI 2 volume was found, with the majority of metabolic activity falling outside the CE by as much as 8-33 mm in 19 patients, averaging 21 cc. For all patients, T2 encompassed the majority of the metabolic volume. However, CNI 2 extended beyond T2 in 6 of the 10 patients without CE (mean 8 cc, max. 15-23 mm) and 13 of the 20 patients with CE (mean 7 cc, max. 8-22 mm), representing an increase of as much as 180% (median 13%) and 86% (median 14%) for non-CE and CE patients, respectively. In addition, preliminary evaluation of MRSI follow-up exams performed post RT revealed isolated cases where areas of MRSI abnormality were predictive for regions of focal recurrence.Conclusions: MRSI appears to be a very valuable diagnostic tool for the assessment of residual disease following surgical resection in high-grade gliomas. The 3D data acquisition allows areas of tumor cell infiltration to be mapped beyond even the anatomic imaging capability of MRI. The incorporation of areas of metabolic abnormality into the treatment planning process for post surgical patients would produce different sizes and shapes of target volumes for both primary and boost volumes. It also may encourage the use of non-uniform margins to define the extent of tumor cell infiltration, rather than the current use of uniform margins. However, continued evaluation of areas of focal recurrence after RT are required in order to determine the spectroscopic signatures that would be most indicative of tissue volumes that might benefit from alternative dose schedules. Tabled 1T1T1T2T2no CE (n=10)CE (n=20)no CE (n=10)CE (n=20)Mean Volume T [cc]18.630.449.665.2Mean Volume CNI 2 [cc]20.927.420.927.4Mean (CNI 2 + T) [cc]43.351.958.375.0CNI 2 outside Y [cc]20.0 (n=8)20.7 (n=19)8.0 (n=6)6.7 (n=13)Conjoint CNI 2 and T [cc]0.99.011.221.5T outside CNI 2 [cc]17.922.239.145.8Median increase (T + CNI 2) vs. T [%]124621314 Open table in a new tab Purpose/Objective: We reported earlier on the value of 3D MR-Spectroscopy imaging (MRSI) in defining the extent of glioma infiltration as assessed in patients prior to definitive surgery. We have extended this study to evaluate the presence of residual disease following surgery but prior to radiation therapy (RT) in patients with high-grade gliomas. The goal was to investigate how postoperative target volume definition for treatment planning might be impacted by the incorporation of functional imaging such as MRSI. Materials/Methods: Thirty patients (27 GBM, 3 grade III astrocytomas) were evaluated, each of whom had undergone MRI and MRSI studies within 4 weeks post surgery but prior to the initiation of RT. For each patient, the MRI data set was manually contoured; regions of interest included the T2 region of hyperintensity (T2), the T1 region of contrast enhancement (T1), and the resection cavity. The 3D-MRSI peak parameters for choline (Cho) and N-acetyl-aspartate (NAA), acquired on a voxel-by-voxel basis, were categorized based on a Cho-to-NAA Index (CNI), a tool for quantitative assessment of tissue metabolite levels. CNI data were aligned to the MRI and displayed as 3D contours. T1 is classically considered to be the region of gross or active disease, T2 the region of microscopic and/or suspected disease, and CNI can be viewed as a continuous measure of disease activity with a CNI of 2 defined as the lowest value corresponding to biopsy confirmed tumor. Composite, as well as conjoint and disjoint volumes, were defined for T1 and T2, both including the resection cavity, vs. CNI 2, these were then compared to T1 and T2, respectively. Results: There was substantial variation in the relationship between the MRI and MRSI designated volumes (see table below). Ten patients did not show any contrast enhancement (CE) post surgery. In 8 of these patients, MRSI revealed metabolic activity (with a CNI 2 > 1.5 cc) averaging 20 cc that extended as much as 11-36 mm beyond the resection cavity. In the 20 patients with CE lesions, substantial variation between the T1 and CNI 2 volume was found, with the majority of metabolic activity falling outside the CE by as much as 8-33 mm in 19 patients, averaging 21 cc. For all patients, T2 encompassed the majority of the metabolic volume. However, CNI 2 extended beyond T2 in 6 of the 10 patients without CE (mean 8 cc, max. 15-23 mm) and 13 of the 20 patients with CE (mean 7 cc, max. 8-22 mm), representing an increase of as much as 180% (median 13%) and 86% (median 14%) for non-CE and CE patients, respectively. In addition, preliminary evaluation of MRSI follow-up exams performed post RT revealed isolated cases where areas of MRSI abnormality were predictive for regions of focal recurrence. Conclusions: MRSI appears to be a very valuable diagnostic tool for the assessment of residual disease following surgical resection in high-grade gliomas. The 3D data acquisition allows areas of tumor cell infiltration to be mapped beyond even the anatomic imaging capability of MRI. The incorporation of areas of metabolic abnormality into the treatment planning process for post surgical patients would produce different sizes and shapes of target volumes for both primary and boost volumes. It also may encourage the use of non-uniform margins to define the extent of tumor cell infiltration, rather than the current use of uniform margins. However, continued evaluation of areas of focal recurrence after RT are required in order to determine the spectroscopic signatures that would be most indicative of tissue volumes that might benefit from alternative dose schedules. Tabled 1T1T1T2T2no CE (n=10)CE (n=20)no CE (n=10)CE (n=20)Mean Volume T [cc]18.630.449.665.2Mean Volume CNI 2 [cc]20.927.420.927.4Mean (CNI 2 + T) [cc]43.351.958.375.0CNI 2 outside Y [cc]20.0 (n=8)20.7 (n=19)8.0 (n=6)6.7 (n=13)Conjoint CNI 2 and T [cc]0.99.011.221.5T outside CNI 2 [cc]17.922.239.145.8Median increase (T + CNI 2) vs. T [%]124621314 Open table in a new tab