The objective of this study was to find out whether the worse prognosis of older patients with primary or metastatic brain tumors can be explained by different patterns of care compared with younger patients.
The objective of this study was to find out whether the worse prognosis of older patients with primary or metastatic brain tumors can be explained by different patterns of care compared with younger patients.
A data base that included 430 patients with glioblastomas and 916 patients with brain metastases who underwent radiotherapy at the author's hospital between 1980 and 2000 was analyzed. Patterns of care were compared for different age groups using the chi-square test.
In both patient groups, age turned out to be an independent risk factor. Older age was associated with worse overall survival. Independent of the cut-off age (< 50 years vs. ≥ 50 years, < 60 years vs. ≥ 60 years, < 65 years vs. ≥ 65 years, and < 70 years vs. ≥ 70 years), there were no statistically significant differences between the age groups concerning the use of different imaging modalities (computed tomography scans vs. magnetic resonance imaging), type of surgery (none vs. biopsy vs. resection), waiting time for radiotherapy (< median vs. ≥ median), radiotherapy treatment planning (simulator-based vs. computer-based), use of radiation sources (cobalt unit vs. linear accelerator), and fractionation protocols (conventional vs. modified). When the recruitment period of 21 years was divided into 3 intervals, impressive changes with regard to the patterns of care became apparent. However, the changes were seen similarly throughout the different age groups.
Older age did not limit access to state-of-the-art patterns of care in neurooncology. Patients participated in medical progress irrespective of their age. The worse prognosis of older patients with glioblastomas or brain metastases was not determined by age-related differences in access to health care. Cancer 2005. © 2005 American Cancer Society.
Age is a prognostic factor of paramount importance in patients with primary and metastatic brain tumors. Elderly patients have a worse outcome. This has been shown in numerous retrospective, prospective, and epidemiologic studies.1 Accordingly, age usually represents one of the first nodes in decision-tree models, e.g., the recursive partitioning analysis (RPA)-based classifications for malignant gliomas or for brain metastases.2, 3
At the same time, older age may be a barrier for cancer patients who are seeking medical care. Some facets of this issue have been highlighted previously, ranging from less informational support after first diagnosis, to a lower referral rate, to hospice care in terminally ill patients.4, 5 In the academic setting, a substantial under representation of patients age ≥ 65 years was found in studies of treatment for cancer.6
An aspect that has not yet attracted much attention in this context is the degree to which elderly patients take part in medical progress. Populations in the United States, Europe, and other developed countries are ageing dramatically.7 Due to improvements in public health, more individuals now are living longer, and the proportion of those living beyond 60 years has increased and will increase further over the next 20 years.8 The question whether elderly patients benefit from the introduction of new technologies and treatment strategies as much as their younger counterparts has not been investigated sufficiently to date.
The issue of age seems to be particularly problematic in patients with incurable malignancies.9 Therefore, we decided to examine these aspects of medical progress in a large cohort of patients with primary or metastatic brain tumors. We wanted to answer the following questions: 1) Is older age associated with a worse prognosis? 2) Do patterns of care differ between younger and older patients? 3) Is there an association between changes in patterns of care and age?
For the current study, 1346 consecutive patients who received radiotherapy for glioblastomas or brain metastases at our institution between 1980 and 2000 were analyzed. Patient-related, tumor-related, and treatment-related variables were extracted from the patients' charts and were entered into a computerized data base.
In 430 patients, a histologically proven glioblastoma was diagnosed. Details of treatment have been reported previously.10, 11 Briefly, after biopsy (n = 344 patients) or resection (n = 86 patients), the patients received either conventionally fractionated radiotherapy (1 daily fraction of 2 grays [Gy] up to a total dose of 60 Gy; n = 97 patients), hypofractionated radiotherapy (1 daily fraction of 3.5 Gy up to a total dose of 42 Gy; n = 104 patients), or hyperfractionated accelerated radiotherapy (3 daily fractions of 1.5 Gy up to a total dose of 54 Gy; n = 229 patients). None of the patients received chemotherapy.
Nine hundred sixteen patients with brain metastases received postoperative (n = 257) or definite (n = 659) whole-brain radiotherapy.12 Two fractionation schemes were used: conventional fractionation (1 daily fraction of 2 Gy up to a total dose of 50 Gy; n = 688 patients) or hypofractionation (1 daily fraction of 3 Gy up to a total dose of 30 Gy; n = 228 patients).
Overall survival was calculated from the first day of radiotherapy to death. Patients with incomplete follow-up were censored with the last date they were known alive. Survival rates were estimated using the method of Kaplan and Meier. In univariate analyses of survival, the different categories were compared using the log-rank test. Variables with a P value < 0.1 in univariate analysis were analyzed multivariately using a Cox proportional hazards model. For a detailed analysis of the patterns of care, we divided the treatment process as a whole into several separate steps (see Fig. 1).
The following items were studied in patients with glioblastomas: type of imaging (computed tomography [CT] scans vs. magnetic resonance imaging [MRI]), type of surgery (biopsy vs. resection), time from surgery to the beginning of radiotherapy (< median vs. ≥ median), type of radiotherapy treatment planning (simulator-based vs. computer-based), type of radiation source (cobalt unit vs. linear accelerator), and fractionation protocol used (conventional fractionation vs. modified fractionation). Patterns of care studied in patients with brain metastases were as follows: type of imaging (CT vs. MRI), type of surgery (none vs. biopsy vs. resection), time from surgery to the beginning of radiotherapy (limited to patients who had undergone biopsy or resection; < median vs. ≥ median), and fractionation protocol used (conventional fractionation vs. modified fractionation). The variables “type of treatment planning” and “type of radiation source” were not analyzed for patients with brain metastases, because whole-brain radiotherapy using lateral opposed fields required no simulation, and it was administered uniformly with a cobalt unit.
To simplify interpretation and comparisons with the literature, we analyzed age as a dichotomous variable. Different cut-off values were used, as shown in Tables 3 and 4. The correlation between the above-mentioned diagnostic and therapeutic procedures and age was analyzed by using contingency table chi-square statistics.
A change of the patterns of care during the recruitment period of 21 years was investigated by dividing it in 3 intervals of the same length. The correlation between patterns of care and age also was analyzed separately for each interval by using contingency table chi-square statistics.
Patient characteristics are shown in Table 1. No patients were lost to follow-up. All patients had died by the end of follow-up (September, 2004). The median overall survival for the whole group was 8.1 months. The median overall survival was 8.8 months for patients age < 65 years compared with 6.0 months for patients age ≥ 65 years (P < 0.0001) (Fig. 2). In a multivariate model, seven variables with complete data were investigated. Age, Karnofsky performance status (KPS), and nonlobar tumor location turned out to be independent prognostic variables. The relative risk (RR) of death was 1.3 for patients age ≥ 65 years compared with younger patients (95% confidence interval [95%CI], 1.16–1.45; P < 0.0001).
|Variable||No. of patients||%|
|< 50 yrs||86||20|
|Mean (range)||59 (19–81)|
|Karnofsky performance status|
|Central tumor location|
|Greatest tumor dimension|
|≤ 4 cm||136||32|
|> 4 cm||116||27|
Patient characteristics are shown in Table 2. Fifteen patients were lost to follow-up, and 878 patients had died by the end of follow-up (September, 2004). The median overall survival for the whole group was 3.5 months. The median overall survival was 3.8 months for patients age < 65 years compared with 2.6 months for patients age ≥ 65 years (P = 0.0002) (Fig. 3). In a multivariate model, six variables with complete data were investigated. Age, KPS, the number of brain metastases, and status of the primary tumor (controlled vs. uncontrolled) were identified as independent variables. The RR of death was 1.12 for patients age ≥ 65 years compared with younger patients (95%CI, 1.04–1.20; P = 0.003).
|Variable||No. of patients||%|
|< 50 yrs||206||23|
|≥ 65 yrs||275||30|
|Karnofsky performance status|
|Primary lesion site|
|No. of brain metastases|
Table 3 shows the results of the patterns of care analysis. With regard to imaging, the majority of patients had CT scans. Resection was the neurosurgical method of choice. Many patients received simulator-based radiotherapy planning and were treated with linear accelerators. Accelerated hyperfractionation was the preferred fractionation protocol.
Irrespective of the different cut-off values, there was no correlation between patient age and the examined procedures. This also held true for the variable “waiting time for radiotherapy.” We identified 1 exception: CT-based treatment planning was used significantly more often in patients age < 50 years compared with patients age ≥ 50 years (38% vs. 27%, respectively; P = 0.03).
|Variable||No. of patients (%) (n = 430)||Age group (%)|
|< 50 yrs (n = 86)||≥ 50 yrs (n = 344)||P value||< 60 yrs (n = 211)||≥ 60 yrs (n = 219)||P value||< 65 yrs (n = 303)||≥ 65 yrs (n = 127)||P value||< 70 yrs (n = 367)||≥ 70 yrs (n = 63)||P value|
|< Median (14 days)||231 (54)||51||54||55||53||52||58||54||52|
|≥ Median (14 days)||199 (46)||49||46||0.59||45||47||0.61||48||42||0.22||46||48||0.81|
|Cobalt unit||46 (11)||7||12||12||9||12||8||11||10|
|Linear accelerator||384 (89)||93||88||0.55||88||91||0.47||88||92||0.59||89||90||0.11|
|Conventional fractionation||97 (23)||20||23||21||24||22||24||22||29|
|Accelerated hyperfractionation||229 (53)||59||52||0.45||55||51||0.62||54||52||0.99||53||52||0.37|
Approximately two-thirds of the patients with brain metastases had CT scans (Table 4). The majority of patients received radiotherapy without prior biopsy or resection. Whole-brain radiotherapy usually was given with conventional fractionation. Irrespective of the cut-off values, no differences with regard to the use of any of the examined procedures and patient age could be identified. Patients age < 65 years underwent MRI more often than the older patients (34% vs. 28%, respectively), and they also were treated more often with conventional fractionation (77% vs. 71%, respectively). However, the differences were not statistically significant (P = 0.12 and P = 0.10, respectively).
|Variable||No. of patients (%) (n = 916)||Age group (%)|
|< 50 yrs (n = 206)||≥ 50 yrs (n = 710)||P value||< 60 yrs (n = 484)||≥ 60 yrs (n = 432)||P value||< 65 yrs (n = 641)||≥ 65 yrs (n = 275)||P value||< 70 yrs (n = 778)||≥ 70 yrs (n = 138)||P value|
|< Median (13 days)||155 (52)||63||49||58||45||57||38||56||32|
|≥ Median (13 days)||146 (48)||37||51||0.05||42||55||0.04||43||62||0.005||44||68||0.04|
|Conventional fractionation||688 (75)||77||75||76||74||77||71||76||71|
Table 5 shows that the ratio of patients age < 65 years to patients age ≥ 65 years was 1.0:0.7 in the first period, 1.0:1.0 in the second period, and 1.0:0.4 in the third period, respectively. Patterns of care changed dramatically during the recruitment period. Whereas CT scans virtually were the only imaging modality in the early 1980s, this proportion dropped below 40% in the late 1990s. Biopsy was used in 10–14% of patients during 1980–1986 and 1987–1993 but in only 38–45% of patients during 1994–2000. Treatment planning was entirely simulator-based in the first period but predominantly was computer-based in the third period. Whereas approximately one-third of patients were treated with a cobalt source in the first period, radiotherapy was given with a linear accelerator to virtually all patients at the end of the recruitment period. The preferred fractionation scheme was hypofractionation during 1980–1986, accelerated hyperfractionation during 1987–1993, and conventional fractionation during 1994–2000. It is noteworthy that none of those changes were age-related (Fig. 3). We repeated the analysis with other cut-off values (age < 50 years vs. age ≥ 50 years; age < 70 years vs. age ≥ 70 years); however, again, no statistically significant difference was found.
|Variable||Percentage of patients|
|Age < 65 yrs (n = 80)||Age ≥ 65 yrs (n = 59)||P value||Age < 65 yrs (n = 85)||Age ≥ 65 yrs (n = 84)||P value||Age < 65 yrs (n = 88)||Age ≥ 65 yrs (n = 34)||P value|
|Surgery — radiotherapy|
|< Median (14 days)||69||63||62||65||23||27|
|≥ Median (14 days)||31||37||0.46||38||35||0.67||77||73||0.60|
Table 6 shows that the ratio of patients age < 65 years to patients age ≥ 65 years was approximately constant over the whole recruitment period (1.0:0.4 in the first and second periods, 1.0:0.5 in the third period). The use of CT imaging decreased from about 90% during 1980–1986 to < 50% during 1994–2000. The percentage of patients who underwent resection for their metastases dropped from approximately 35% to 20%. Whereas almost all patients received whole-brain radiotherapy with conventional fractionation in the first period, hypofractionation became the preferred protocol during 1994–2000. Patients age < 65 years tended to undergo MRI more often than the older patients in the first period (13% vs. 6%; P = 0.16) and in the second period (39% vs. 27%; P = 0.13), but not in the third period. In the third period, older patients with metastases underwent resection more often than younger patients (26% vs. 15%, respectively; P = 0.05) (Fig. 4). During 1987–1993, hypofractionation was used more often in older patients than in younger patients (7% vs. 2%; P = 0.06); however, this age difference was not seen during 1980–1986 or during 1994–2000 (Fig. 4).
|Variable||Percentage of patients|
|Age < 65 yrs (n = 216)||Age ≥ 65 yrs (n = 80)||P value||Age < 65 yrs (n = 119)||Age ≥ 65 yrs (n = 74)||P value||Age < 65 yrs (n = 226)||Age ≥ 65 yrs (n = 121)||P value|
|< Median (13 days)||71||54||47||32||43||29|
|≥ Median (13 days)||29||46||0.11||53||68||0.23||57||71||0.24|
The issue of age has become a frequently discussed topic in modern oncology. Data from the literature suggest that, for a variety of reasons, elderly patients with primary or metastatic brain tumors have a poorer outcome than younger patients.
Krzyzanowska et al. showed convincingly that receipt of treatment is correlated strongly with nondisease-related factors, especially sociodemographic characteristics, indicating possible disparities in access to care.13 Older age, lower socioeconomic status, and the presence of comorbid illnesses are associated with a lower likelihood of receiving treatment. Brandes and Monfardini postulated that elderly patients with brain tumors frequently are treated suboptimally.9 Thus, discrimination in the delivery of health care, based on the age of the patient, increasingly is attracting attention.
In the current study, older age was associated with decreased survival. When patients age < 65 years and age ≥ 65 years were compared, median survival was 8.8 months versus 6.0 months, respectively, in patients with glioblastomas and 3.8 months versus 2.6 months, respectively, in patients with brain metastases.
An analysis of > 18,000 patients with glioblastomas from the National Cancer Data Base revealed that 5-year overall survival in patients ages 15–24 years, 35–44 years, and 65–74 years was 21.0 months, 10.3 months, and 0.2 months, respectively.14 In the recursive partitioning analysis of the Radiation Therapy Oncology Group (RTOG) that included 1578 patients with malignant gliomas, the most important split was by age, with age 50 years the most prominent breakpoint.2 The median overall survival for patients age < 50 years was 18 months, compared with 8.8 months in patients age ≥ 50 years. The report of the Medical Research Council Brain Tumor Working Group divided patients into 3 age groups: < 45 years, 45–59 years, and > 59 years.15 The median overall survival in those age groups was 12 months, 9 months, and < 5 months, respectively. In a population-based study that included 715 Swiss patients with glioblastomas, the median survival in patients ages < 50 years, 50–59 years, 60–69 years, 70–79 years, and ≥ 80 years was 8.8 months, 7.3 months, 4.4 months, 2.9 months, and 1.6 months, respectively.16 Ohgaki et al. underlined that age was the most significant prognostic factor in univariate and multivariate analyses. Lacroix et al. published a retrospective study on 416 patients with glioblastomas who were treated at The University of Texas M. D. Anderson Cancer Center.17 In that study, compared with patients age < 45 years, the RR was 1.8 (95%CI, 1.3–2.3) for patients age 45–64 years and 3.0 (95%CI, 2.1–4.2) for patients age ≥ 65 years. In the current study, the RR was 1.3 (95%CI, 1.16–1.45) for patients age ≥ 65 years compared with younger patients.
Next to KPS and extracranial tumor activity, age is the most important prognostic factor in patients with brain metastases. This was shown in the recursive partitioning analysis of the RTOG that included 1200 patients from 3 consecutive RTOG trials.3 Given a KPS ≥ 70, the median survival of patients age < 65 years without extracerebral metastases was 7.1 months compared with 4.2 months in patients age ≥ 65 years. In the study by Lagerwaard et al. that included 1292 patients who were treated at the Daniel den Hoed Cancer Center, performance status, response to steroid treatment, systemic tumor activity, and serum lactate dehydrogenase were independent prognostic factors that had the strongest impact on survival.18 Age also was identified as a prognostic factor, although it had lesser importance. This finding is in accordance with the results of our analysis. The RR was 1.56 for patients who had brain metastases and a KPS < 70 compared with patients who had a KPS ≥ 70, and it was 1.12 for patients age ≥ 65 years compared with younger patients (95%CI, 1.04–1.20).
With few exceptions, the various diagnostic and therapeutic procedures examined were not administered in an age-dependent manner at our institution. Radiotherapy treatment planning less often was CT-based in elderly patients with glioblastomas than in younger patients, but only when the cut-off age of 50 years was chosen (27% vs. 38%). For all other cut-off ages, the proportion of patients who underwent CT-based treatment planning was virtually the same (29% for patients age < 65 years and patients age ≥ 65 years). From these data, we conclude that there was no systematic difference in the use of treatment planning procedures with regard to age.
The median waiting time for radiotherapy was significantly longer in elderly patients with brain metastases compared with younger patients. This item may be examined only in a subset of the study population, i.e., those patients who underwent biopsy or resection (n = 301 patients). It became more prominent with higher cut-off values for age: The percentage of patients who started radiotherapy later than the median waiting time of 13 days was 42% in patients age < 60 years (compared with 55% in patients age ≥ 60 years), 43% in patients age < 65 years (compared with 62% in patients age ≥ 65 years), and 44% in patients age < 70 years (compared with 68% in patients age ≥ 70 years). We do not interpret this finding as a statistical artifact. Table 2 shows that many patients with brain metastases presented with a low performance status, multiple cerebral lesions, and extracerebral metastases. The median survival in this poor prognostic group was < 3 months. We believe that it is not age, per se, but cancer in its terminal stage and comorbidity that account for postoperative complications, finally resulting in the longer waiting time for radiotherapy. It is noteworthy that, within the same institution, no significant differences with regard to waiting time were found between younger and older patients with glioblastomas.
A literature search revealed that the issue of age-related differences in patterns of care in patients with brain tumor has not been investigated systematically to date. Few studies explicitly have addressed the question of possible age discrimination, and studies provide conflicting results.
Steinfeld et al. investigated whether there is an age-related delay in the diagnosis of patients with glioblastomas.19 In their study, a total of 204 charts from patients who were treated from 1972 to 1992 were evaluated for type and duration of symptoms. The median duration of symptoms for the entire group was 28 days. No age-related difference in duration of symptoms was noted. This view was confirmed in a Minnesota cohort study that included 354 patients with glioblastomas.20 In that study, the interval between onset of symptoms and diagnosis was longest for patients age 18–24 years (median, 28 days). It exceeded the duration of symptoms when analyzed both for patients age ≥ 65 years (median, 20 days) and for patients age ≥ 75 years (median, 21 days). Both of those study groups concluded that elderly patients with malignant brain tumors are diagnosed as promptly as younger patients.
Cowan et al. analyzed the impact of provider volume on mortality after resection in patients with various malignant intracranial tumors (n = 7547 patients).21 It is noteworthy that the mean patient age at admission in their study was significantly lower in hospitals that had high case loads compared with hospitals that had low case loads (52 years vs. 60 years; P < 0.001). Those results were confirmed in a study by Long et al. that included 4723 patients22 in which the results may be interpreted to show that elderly patients with brain tumors were less likely to be referred to specialized centers with high case loads compared with their younger counterparts.
A population-based cohort study of all patients with glioblastomas residing in Ontario between 1982 and 1994 (n = 3279 patients) revealed that the proportion of patients undergoing resection compared with biopsy alone varied with age (resection: patients age 50–59 years, 74%; patients age 70–79 years, 63%; biopsy: patients age 50–59 years, 15%; patients age 70–79 years, 22%; P < 0.0001).23 Patients age ≥ 70 years were less likely to undergo resection and were more likely to undergo biopsy or to have no record of surgery. Only a minority of patients age ≥ 70 years received radiotherapy.
Those findings were confirmed in a Swiss population-based study.16 In that study, the mean age of patients who underwent resection was significantly lower compared with the mean age of patients who did not undergo surgery (56 years vs. 68 years; P < 0.0001). The survival rates of patients who underwent resection were significantly lower compared with patients who did not undergo surgery (7.9 months vs. 2.5 months; P < 0.001). The mean age of patients who received radiotherapy was 55 years, significantly younger than the mean age of patients who did not receive radiotherapy (68 years; P < 0.0001). The survival rate of patients who received radiotherapy was significantly longer compared with patients who did not receive radiotherapy (10 months vs. 2 months; P < 0.001). According to the authors, the inclusion of a high percentage of older patients with low KPS in population-based studies is the main reason for the low survival rates compared with the results of clinical studies. In their study, Ohgaki et al. reported that, even in a country with unrestricted access to a sophisticated health care system, the prognosis for older patients with glioblastomas remains depressingly poor.16
During the last 20 years, much has changed in neurooncology. Some of these trends can be identified clearly in our study population. The management of patients with glioblastomas was characterized by an increased use of MRI (1980–1986, < 2%; 1994–2000, > 60%), a higher frequency of biopsies in the later period (1980–1986, < 14%; 1994–2000, > 38%), the introduction of computer-based treatment planning (1980–1986, 1%; 1994–2000, > 55%), the definite break-through of linear accelerator technology (1980–1986, < 78%; 1994–2000, > 96%), and the use of various subsequent fractionation protocols.
In patients with brain metastases, the increasing use of MRI (1980–1986, < 13%; 1994–2000, > 43%) and the advent of hypofractionation (1980–1986, 4%; 1994–2000, 60%) constituted the most obvious changes in patterns of care. Younger patients and older patients participated equally in the introduction of new technologies and treatment concepts. We found no evidence for age discrimination as a function of time. In this context, it has to be stressed that our study covered a full 21-year recruitment period, allowing a long-term analysis of changes in patterns of care.
The trends discussed above also were described by other investigators. A national survey of patterns of care for brain tumor patients revealed an increased use of MRI as well as an increased use of stereotactic biopsy between 1980 and 1985.24 However, some authors found that changes in patterns of care were age-dependent. In the Minnesota cohort study, elderly patients who were diagnosed between 1990 and 1995 were less likely to undergo a multimodal treatment consisting of surgery, radiotherapy, and chemotherapy (7%) compared with elderly patients who were diagnosed between 1980 and 1989 (24%; P < 0.01).20 This finding suggests a less aggressive treatment of elderly patients with glioblastomas in the last decade compared with the previous period. Moreover, the analysis of treatment modalities used for patients age ≥ 65 years revealed that patients who were diagnosed between 1990 and 1995 were significantly more likely to receive no treatment (18%) compared with patients who were diagnosed before 1990 (4%; P = 0.004).20 In the current study, the ratio of patients with glioblastomas age < 65 years to patients with glioblastomas age ≥ 65 years changed from 1.0:1.0 in 1987–1993 to 1.0:0.4 in 1994–2000. The lower proportion of elderly patients may be interpreted as a consequence of an increasingly strict indication for radiotherapy in this subset of patients.
In view of reports from the literature, some characteristics of our study deserve special attention. First, we analyzed consecutive patients who were treated outside clinical trials. Age discrimination can become apparent only in a setting that allows (relatively) free decisions. It is less likely to occur in prospective trials with strict guidelines for patient enrollment. Second, we examined a large cohort of patients with the most frequent malignant brain tumors. This allows a robust interpretation of the data and cross-checks between the two indications (e.g., differences in waiting time; see above). Third, and most important, this is the only study to date investigating the relation between age and all steps in the treatment process that are considered relevant, i.e., imaging studies, surgery, and the various aspects of radiotherapy. Compared with other investigators, we did not focus on a single item but analyzed the whole supply chain.
One criticism that may be leveled at the current study is that it included only patients who received at least one fraction of radiotherapy. The numbers of patients who had imaging studies that were suspicious for a malignant brain tumor but who were not referred for neurosurgery are unknown; likewise, the numbers of patients who had a malignant brain tumor diagnosed after undergoing biopsy or resection but who were not referred for radiotherapy also are unknown. Thus, we cannot exclude the possibility that age-related decisions were made at an earlier stage in the treatment process. However, the patients who finally were included in our study had equal access to the available resources irrespective of their age. This also held true for the steps in the treatment process before the initiation of radiotherapy (imaging, surgery) as well as for the various components of radiotherapy itself. It is noteworthy that, within this relatively homogeneous group, older patients clearly had a worse prognosis compared with younger patients.
In recent years, a number of molecular and genetic alterations have been identified in human glioblastomas, underlining their biologic heterogeneity.16 In clinicopathologic studies, the correlation between these changes and age has been investigated along with their potential prognostic impact. McKeever et al. analyzed the correlation between age and proliferation of glioblastoma cells.25 The Molecular Immunology Borstel number 1 proliferation index was significantly lower in younger patients compared with older patients. Barker et al. correlated radiation response in patients with glioblastomas with epidermal growth factor receptor (EGFR) immunoreactivity.26 Thirty-three percent of patients who had tumors with no EGFR immunoreactivity had a good response to radiation (> 50% reduction in tumor size on CT or MRI studies) compared with 18% of patients who had tumors with intermediate EGFR staining and 9% of patients who had tumors with strong staining. Significant correlations were noted between EGFR score and older age. Shono et al. found that high cyclooxygenase-2 (Cox-2) expression in glioblastoma cells was associated with poor survival.27 Other authors demonstrated that Cox-2 inhibitors can sensitize human gliomas to ionizing radiation.28 In view of the literature and our own data, it seems that biology, and not technology, determines the worse prognosis of elderly patients with glioblastomas.
Brandes and Monfardini stressed that treatment in elderly patients with brain tumors should be based on sound scientific data and not on ad hoc decisions.9 Consequently, the number of controlled clinical trials designed for elderly patients with brain tumors has increased continuously over the last years, including neurosurgical contributions.29 Future therapies tailored to the needs of elderly patients will have to consider both clinical characteristics and biologic features.