Using proton magnetic resonance spectroscopic imaging to predict in vivo the response of recurrent malignant gliomas to tamoxifen chemotherapy

Citation
Mc. Preul et al., Using proton magnetic resonance spectroscopic imaging to predict in vivo the response of recurrent malignant gliomas to tamoxifen chemotherapy, NEUROSURGER, 46(2), 2000, pp. 306-318
Citations number
58
Categorie Soggetti
Neurology,"Neurosciences & Behavoir
Journal title
NEUROSURGERY
ISSN journal
0148396X → ACNP
Volume
46
Issue
2
Year of publication
2000
Pages
306 - 318
Database
ISI
SICI code
0148-396X(200002)46:2<306:UPMRSI>2.0.ZU;2-8
Abstract
OBJECTIVE: Most patients with a malignant glioma spend considerable time on a treatment protocol before their response (or nonresponse) to the therapy can be determined. Because survival time in the absence of effective thera py is short, the ability to predict the potential chemosensitivity of indiv idual brain tumors noninvasively would represent a significant advance in c hemotherapy planning. METHODS: Using proton magnetic resonance spectroscopic imaging (H-1 MRSI), we studied 16 patients with a recurrent malignant glioma before and during treatment with high-dose orally administered tamoxifen. We evaluated whethe r 1H MRSI data could predict eventual therapeutic response to tamoxifen at the pretreatment and early treatment stages. RESULTS: Seven patients responded to tamoxifen therapy (three with glioblas tomas multiforme; four with anaplastic astrocytomas), and nine did not (six with glioblastomas multiforme; three with anaplastic astrocytomas). Respon ders and nonresponders exhibited no differences in their age, sex, tumor ty pe, mean tumor volume, mean Karnofsky scale score, mean number of weeks pos tradiotherapy, or mean amount of prior radiation exposure. Resonance profil es across the five metabolites measured on H-1 MRSI spectra (choline-contai ning compounds, creatine and phosphocreatine, N-acetyl groups, lactate, and lipids) differed significantly between these two groups before and during treatment. Furthermore, linear discriminant analyses based on patients' in vivo biochemical information accurately predicted individual response to ta moxifen both before and at very early treatment stages (2 and 4 wk). Simila r analyses based on patient sex, age, Karnofsky scale score, tumor type, an d tumor volume could not reliably predict the response to tamoxifen treatme nt at the same time periods. CONCLUSION: It is possible to accurately predict the response of a tumor to tamoxifen on the basis of noninvasively acquired in vivo biochemical infor mation. H-1 MRSI has potential as a prognostic tool in the pharmacological treatment of recurrent malignant gliomas.