Magnetic resonance in neuro-oncology: Achievements and challenges - Signal Processing in MRS with Biomedical Applications - page 260

8.1.4.3 Characterization of pediatric brain tumors

Proton MRS has also provided some help in characterizing childhood brain
tumors. Thirty five children with cerebellar tumors were examined in a single
voxel study using a 1.5T scanner and short TE [268]. Eighteen nonglial
tumors (medulloblastomas) were identified on the basis of high taurine, phos
phocholine and glutamate together with low glutamine. High myoinositol
and glycerophosphocholine characterized five ependymomas. In vitro MRS
was used for corroboration [268, 269]. In an earlier report of these children
from that same group of investigators [270], the combination of a creatine to
choline ratio < 0.75 and a myoinositol to NAA ratio < 2.1 distinguished the
medulloblastomas from the ependymomas.

8.1.5 MRSI for target planning for brain tumors
It has been suggested that incorporation of MRS and MRSI into RT planning
for highgrade gliomas could improve control while reducing complications
[271]. Most often, target definition in planning radiation therapy has been
based upon ratios of choline to creatine or choline to NAA. The clinical target
volume for RT of gliomas was previously generated by adding uniform margins
of 2 to 3 cm to the area of T 2 hyperintensity [272]. This is because the peri
tumoral “uncertain” zone, while appearing normal on MRI is often infiltrated
by tumor when examined histopathologically [235]. Using MRSI to determine
areas of high choline to NAA ratio, the shape and size of the clinical target
volume can be better identified, with more confident sparing of uninvolved
brain tissue [273]. Choline to creatine ratios have been used to tailor radiation
dose levels to the glioma grade within a map of clinical target volumes [274].
Not only have maps of choline to creatine ratios as well as choline to NAA
ratios been helpful in refining RT dose contouring, they have also improved the
effectiveness of surgical planning to treat brain tumors [207]. A phase I dose
intensification trial for gliomas using intensity modulated radiation therapy
(IMRT) is being planned using the additional information provided by choline
to creatine ratios from MRSI as well as functional MRI [275].

8.1.6 Assessing response of brain tumors to therapy and

prognosis via MRSI

8.1.6.1 Distinguishing recurrent tumor from response to therapy

Brain tumors frequently return after treatment, and often do so at a higher
grade. In addition, the treatment itself, especially RT, provokes changes in
brain tissue that are di cult to distinguish from tumor recurrence using MRI.
Radiation necrosis, local response to immunotherapy, as well as the tumor
itself all can show contrast enhancement. MRSI improves specificity by help
ing to differentiate radiation necrosis or enhancement phenomena after lo
cal immunotherapy, from tumor recurrence. Distinction between recurrent