RATIONALE AND OBJECTIVES. Phosphorus magnetic resonance spectroscopy h
as been used noninvasively to determine characteristic spectral parame
ters for untreated human brain tumors as a prelude to its use in clini
cal diagnosis. METHODS. The spectra, which reflect the relative amount
s of phosphorus-containing compounds, and the pH within and surroundin
g the tumors, were obtained in vivo using the localization technique o
f one-dimensional chemical shift imaging applied with a surface coil,
Phosphorus-31 chemical shift imaging was performed successfully in viv
o on 9 volunteers and 27 patients with untreated brain tumors, includi
ng 7 with astrocytoma, 4 with glioblastoma, 3 with meningioma, and 11
with metastases, This study provides spectra from within and surroundi
ng the brain tumors, and allows accountability for the heterogeneity o
f brain tumors by the selection of the maximum data point for each par
ameter. RESULTS. The ratios of resonance areas, phosphodiesters over n
ucleoside triphosphate (NTP), and phosphomonoesters over NTP, were fou
nd to be higher in glioblastomas (2.55 +/- 0.22, 1.06 +/- 0.09) and as
trocytomas (3.04 +/- 0.36, 1.28 +/- 0.36) than in normal brain (2.00 /- 0.32, 0.79 +/- 0.22), The ratios of areas due to inorganic phosphat
e and NTP, and phosphocreatine and NTP, also were higher in astrocytom
as (1.16 +/- 0.40, 1.17 +/- 0.41) compared with glioblastomas (0.68 +/
- 0.01, 0.88 +/- 0.19) and normal brain (0.61 +/- 0.03, 0.77 +/- 0.03)
, The pH of brain tumors ranged from alkaline to neutral, with meningi
omas consistently having alkaline pH. CONCLUSIONS. These data show tha
t there are statistically significant differences in the magnetic reso
nance parameters of the affected brain hemispheres of patients with as
trocytomas, glioblastomas, meningiomas, and normal brain tissue, and u
nderline the need for a multisite clinical trial to establish clinical
criteria for diagnosis.