Summarizing the current evidence regarding the usefulness of the previous markers for predicting patient outcome, the most promising proliferation marker for predicting patient outcome in patients with brain tumors appears to be KI-67/MIB-1. Its potential usefulness appears to be greatest (1) for grade II and III astrocytic and oligodendroglial tumors in adults, where it may potentially predict length of survival; (2) for nonpilocytic gliomas in children, where it may also potentially predict survival; and (3) for benign, completely resected meningiomas, where it may potentially predict tumor recurrence. Although MIB-1 shows potential, we do not believe there is evidence yet definitely to consider MIB-1 labeling index a predictor of prognosis in these tumors due to the lack of published prospective studies validating the preliminary findings of multiple investigators who have performed only single-center retrospective studies. We also believe that the exact use of MIB-1 labeling index as an independent prognostic indicator for any type and grade of tumor may be complex, given the fact that all of these brain tumors have multiple independent prognostic factors contributing to patient outcome, many of which are readily available clinical factors. In the present managed care environment, MIB-1 labeling should probably await demonstration that it significantly contributes to the physician's ability accurately to predict patient outcome. The presence of telomerase RNA or telomerase activity appears to correlate with degree of malignancy in multiple types of brain tumors, including gliomas; however, it currently has no use as an independent prognostic indicator of patient outcome. It may instead be a marker for malignant tumor initiation or progression. p53 and EGFR are molecular markers that show promise as prognostic indicators of recurrence-free and overall survival in patients with GBMs, but further prospective studies are needed to confirm the retrospective findings. Postsurgical evaluation of these markers is potentially helpful in planning follow-up and treatment for these patients, those patients having tumors expressing relatively high levels of these markers requiring closer follow-up and, when possible, more aggressive therapies. Despite intensive investigation into the expression of molecules regulating apoptosis in brain tumors, no evidence presently exists to support their usefulness as markers of patient outcome. This also applies to measurements of the apoptotic rate itself in human brain tumors. Our overall impression, therefore, is that despite our great desire to find the Holy Grail of patient prognosis in the measurement of a single molecular marker with a precise cut-off value, one has not been identified ... yet! If we allow ourselves to think in terms of organismal biology, it is not surprising that the attempt directly to correlate one or even five (MIB-1, telomerase, BCL-2, p53, EGFR) gene products with a phenomenon as complex as a patient's long-term survival is unrealistic. More likely, multiple pieces of clinical information (which we already know significantly impact patient outcome in patients with brain tumors) will be considered in conjunction with new, scientifically proved molecular information as it becomes available, allowing us to predict, ever more accurately, a given patient's clinical course and outcome.