Predictors of Survival for Patients With Brain Metastases: Results of a Randomized Phase III Trial

Reviewer: Ryan Smith, MD
Last Modified: October 8, 2002

Presenter: Walter Curran
Presenter's Affiliation: Thomas Jefferson University
Type of Session: Scientific


    Brain metastases occur in 20-40% of all patients with cancer. There is a general life expectancy of these patients, though it may vary widely depending on the patient's prognostic factors. This study attempts to evaluate clinical and neurocognitive data to determine their correlation with survival. This data was obtained from a phase III study reporting on the effect of motexafin gadolinium on patients with brain metastases.

Materials and Methods

  • This phase III study randomized patients to receive either whole brain radiation therapy alone (30 Gy) vs. radiation therapy + motexafin gadolinium.
  • Patient characteristics and data were obtained from this database
  • Patients with any primary tumor and KPS greater than or equal to 70 were included
  • Patients with liver metastases, >2 sites of extracranial metastases, small cell lung cancer, lymphoma, and germ cell tumors were excluded
  • Various clinical and neurocognitive factors were scored prior to the patient beginning treatment
  • All patients had a minimum of 6 months of follow up


  • Median survival was 7.17 months for breast cancer patients, 4.4 months for lung cancer patients, and 5.03 months for patients with other primaries
  • On multivariate analysis, best predictors for survival were female gender, number of brain metastases (>2 vs. 2 or less), higher KPS, lower LDH, breast primary, and better motor dexterity as measured by the Hopkins Verbal Learning Test
  • RPA class was not a prognostic factor (median survival for RPA class I-5.43 months, class II-4.93 months)

Author's Conclusions

  • Best predictors for survival in patients with brain metastases were gender, number of metastases, KPS, LDH, and site of primary
  • Given that RPA class was not predictive, should it continue to be used?
  • Should future trials stratify by number of brain metastases present and via lung vs. breast primary?

Clinical/Scientific Implications

    As stated, 20-40% of all patients with cancer develop brain metastases. Management, except in very selected cases, is with radiation therapy. However, patients with CNS metastatic disease are likely to have other systemic disease, of which the main treatment is chemotherapy. Hence, there needs to be an obvious balance between the use of radiation to control CNS disease and the use of chemotherapy to control systemic disease. Part of this decision process should include an idea of an individual patient's projected survival. This study elucidates some prognostic factors that can assist in this prediction. However, there is little new information. Breast histology (vs. lung), KPS, neurological symptoms, and other findings indicating better health of the patient (albumin, LDH, etc) have long been associated with increased survival. Surprising is the fact that RPA class was not predictive for survival, as it has been in other studies. Possible explanations for this are the fact that these patients had no prior resection, a lower incidence of a single metastasis, and all were treated to a standard dose of 30 Gy, which is lower than some other studies. Perhaps this skewed the results, though other explanations could be made as well. One explanation is the fact that lung patients did worse than breast patients, and these are the very patients that are more likely to have extracranial metastases. Since the presence of extracranial metastases is a large part of the definition of RPA class, this may have overshadowed the RPA classification in predicting survival. Regardless, as RPA has been verified before, it should not be discarded based on this single study. Rather, other stratification parameters should be incorporated, such as primary site and number of brain metastases, as shown in this study.

Oncolink's ASTRO Coverage made possible by an unrestricted Educational Grant from Ortho Biotech.