Effectiveness of whole brain radiotherapy in the treatment of brain metastases: a systematic review.

BACKGROUND: Brain metastases are the most common intracranial tumour in adults, estimated to occur in up to 40% of patients with cancer. Despite being used in clinical practice for 50 years, the effectiveness of whole brain radiotherapy for the treatment of brain metastases remains uncertain.

OBJECTIVES: To assess the effectiveness of whole brain radiotherapy (WBRT) on survival and quality of life. To identify whether patient performance status, number of brain metastases, extent of extracranial disease and primary site of cancer are important effect modifiers.

DESIGN: Systematic literature review.

METHODS: Electronic searches of four databases, augmented by hand searches of the most frequently encountered journal and assessment of the reference lists of consensus statements and all retrieved papers. Included papers underwent structured data extraction, assessment and qualitative synthesis.

RESULTS: Thirty-two primary studies were included, with a range of study designs, methodological quality, pre-treatment variables, interventions and outcome measures. From the limited evidence available, survival appeared to increase when patients were selected by performance status (survival increasing from approximately three to seven months in high performance status groups, as defined by Karnofsky performance status or Recursive Partitioning Analysis classification). The evidence suggests no survival benefit when patients with poor performance status were treated with whole brain radiotherapy. No studies undertook direct measurement of patients' quality of life. Surrogate measures of patients' quality of life, such as improvement in neurological function or improvement/maintenance of KPS > or =70, produced response rates ranging from 7 to 90%.

CONCLUSION: The heterogeneity of study designs, quality and outcomes necessitates caution in interpreting the review findings. WBRT appears to be of benefit in higher performance status patients but not in low performance status patients. This suggests a basis for current practice, however further robust trial evidence is required.