Paul Thompson's Research Publications

Growth Rates, MR Spectroscopy: Prognostic Markers for Glioblastoma Multiforme

Society for Neuro-Oncology, Chicago, Nov. 2000

Haney S, Thompson PM, Cloughesy TF, Alger JR, Frew AJ, Toga AW
Laboratory of Neuro Imaging, Department of Neurology, Division of Brain Mapping;
Neuro-Oncology Program, The Henry E. Singleton Brain Cancer Research Program;
Dept. of Radiological Sciences; UCLA School of Medicine, Los Angeles, CA 90095


Mapping Rates of Tumor Growth. Rates of tumor growth in GBM patients can be tracked using a combination of image analysis algorithms. Maps of contrast-enhancing tumor can be generated using tissue classifiers. Growth rates can be calculated from these maps, and monitored during therapy. Here growth rates are expressed in terms of equivalent doubling times, which indicate how long it would take for the tumor to double in volume, based on the changes observed in an interval between MRI scans.


As part of an ongoing study of patients with malignant gliomas, we sought to determine whether imaging markers such as contrast enhancing tissue, edema and necrosis, MR spectroscopic measures, and other factors were prognostically significant.   Prognostic markers offer valuable information for the physician providing treatment, for the patient himself, especially in terms of making decisions regarding quality of life issues, and for the researcher designing clinical studies targeting specific subgroups of patients. A variety of potential prognostic information, such as tumor location, presenting symptoms, and time to recurrence were also tracked as  part of this study.   Imaging measures may function as dynamic markers, providing prognostic data which may be revised during the course of therapy.


16 patients with histopathologically confirmed diagnosis of glioblastoma multiforme (GBM), and with tumors that were neither midlinear or infratentorial, were serially scanned. Measures of edema, necrosis and contrast-enhancing tissue were derived from T2-weighted and gadolinium enhanced T1-weighted SPGR (spoiled Grass; 256x256, 3mm spacing, no gap) MRIs.  MR spectroscopy measures were also acquired. Volumetric analysis was performed using a 3D tissue segmentation algorithm to guide manual segmentation.  Volumetric data was then compared across time.  Growth rates for contrast enhancing tissue and CHO/CRE (choline/creatine) measures were examined.


Preliminarily, multivariate analysis indicated a significant negative relationship between growth of contrast enhancing tumor tissue and survival (p<0.03).  Values of CHO/CRE > 1.8 were significantly negatively related to survival.  Age, a well-known prognostic marker, was also significantly negatively related to survival (p<0.01). Analysis of other factors is ongoing.


Multivariate analysis indicates that growth rates of contrast-enhancing tissue, CHO/CRE measures and age are prognostically significant.  These prognostic factors, as well as others, are valuable for the patient and physician in making decisions regarding therapy choice, and for reserchers designing clinical trials.

Related Publications

  • (back to main list)

    Contact Information

  • Mail:

    Paul Thompson, Ph.D.
    Assistant Professor of Neurology
    Lab of Neuro Imaging and Brain Mapping Division
    4238 Reed Neurology
    710 Westwood Plaza
    UCLA Medical Center
    Los Angeles CA 90095-1761, USA.

  • E-mail:
  • Tel: (310)206-2101
  • Fax: (310)206-5518