Dose escalation beyond 30 grays in 10 fractions for patients with multiple brain metastases

Authors


Abstract

BACKGROUND.

Whole-brain radiotherapy (WBRT) to 30 grays (Gy) in 10 fractions is the standard treatment in patients with multiple brain metastases in the majority of treatment centers worldwide. The current study investigated the potential benefit of dose escalation beyond 30 Gy.

METHODS.

Data regarding 416 patients who were treated with WBRT for multiple brain metastases were evaluated retrospectively. Survival and freedom from recurrent brain metastasis (local control) of 257 patients who were treated with 10 fractions of 3 Gy each for 2 weeks were compared with those of 159 patients treated with 45 Gy in 15 fractions for 3 weeks or 40 Gy in 20 fractions for 4 weeks. Eight additional potential prognostic factors were investigated including age, gender, Karnofsky performance score (KPS), tumor type, interval between tumor diagnosis and RT, number of metastases, extracranial metastases, and Recursive Partitioning Analysis (RPA) class.

RESULTS.

On multivariate analysis, improved survival was found to be associated with lower RPA class (P < .001), age <60 years (P = .026), KPS ≥70 (P < .001), and absence of extracranial metastases (P = .003). A trend was observed for number of metastases (2–3 vs ≥4; P = .07). Improved local control was associated with a KPS ≥70 (P < .001) and breast cancer (P < .001). A trend was observed for number of metastases (P = .059). The RT schedule did not appear to have any significant impact on survival (P = .86) or local control (P = .61). The subgroup analyses, performed for each of the 3 RPA classes, did not demonstrate a significantly better outcome with dose escalation.

CONCLUSIONS.

Dose escalation beyond 30 Gy in 10 fractions does not appear to improve survival or local control in patients with multiple brain metastases but does increase the treatment time and cost of therapy. Cancer 2007. © 2007 American Cancer Society.

Brain metastases represent the most common intracranial malignancy in adults, reportedly occurring in 10% to 30% of all adult cancer patients during the course of their disease.1 More than 150,000 new cases of brain metastases occur each year in the U.S., and the incidence is rising.2, 3 The majority of patients with single brain metastasis are treated with surgery followed by whole-brain radiotherapy (WBRT) or with radiosurgery with or without WBRT, whereas the majority of patients with multiple brain metastases receive WBRT alone. The optimal RT schedule of WBRT alone is controversial. A total dose of 30 grays (Gy) in 10 fractions (overall treatment time of 2 weeks) is considered the standard dose-fractionation schedule used in the majority of centers worldwide. Although this schedule proved to be an effective treatment of brain metastases, it is possible that these patients would benefit from an escalation of the RT dose beyond 30 Gy. Such a dose escalation may be achieved with RT schedules such as 45 Gy in 15 fractions (15 fractions of 3 Gy each given over 3 weeks) or 40 Gy in 20 fractions (20 fractions of 2 Gy each given over 4 weeks). However, higher dose programs can only be recommended if they provide a better outcome than the dose of 30 Gy in 10 fractions. Longer RT programs increase the treatment time, which is less convenient for the patients. Furthermore, longer programs increase the cost of therapy.4

The median survival of untreated patients with multiple brain metastases is approximately 1 month.5 Even with treatment, most patients with multiple brain metastasis have a median life expectancy of only 4 months.6 The survival of patients with brain metastases may be predicted with the help of the recursive partitioning analysis (RPA) based on 3 Radiation Therapy Oncology Group (RTOG) brain metastasis trials.7 In this RTOG analysis, 3 prognostic classes (RPA classes) were defined. RPA class 1 patients have a Karnofsky performance score (KPS) ≥70 and 3 other favorable prognostic factors: age <65 years, no extracranial metastases, and controlled primary tumor. RPA class 2 patients have a KPS ≥70 and at least 1 unfavorable prognostic factor (age ≥65 years, extracranial metastases, or uncontrolled primary tumor). RPA class 3 includes all patients with a KPS <70. In the RTOG analysis, the median survival for the 3 RPA classes was 7.1 months, 4.2 months, and 2.3 months, respectively.7

Improved treatment results for patients with multiple brain metastases is needed. This possibly could be achieved by escalating the RT dose beyond the “standard” of 30 Gy in 10 fractions. Previous studies that compared 30 Gy in 10 fractions with higher doses did not demonstrate a significant difference in survival.8–10 However, those studies from the literature did not take account the different RPA classes or were comparably inhomogeneous, including patients with single brain metastasis. With regard to the various survival prognoses in the RPA classes, the appropriate treatment for patients may vary with RPA class. To our knowledge, the current study is the first to compare “standard” RT with 30 Gy in 10 fractions with treatment schedules using higher doses such as 45 Gy in 15 fractions and 40 Gy in 20 fractions in patients with multiple brain metastases with regard to treatment outcome for the entire cohort and for each RPA class.

MATERIALS AND METHODS

A total of 416 patients who were treated with WBRT alone for ≥2 brain metastases at the university hospitals of Luebeck and Hamburg between 1989 and 2005 were included in this retrospective analysis. The criteria for inclusion were as follows: ≥2 brain metastases treated with WBRT alone (6–10-megaelectron volt photons), no prior RT to the brain, confirmation of metastases by computed tomography (CT) or magnetic resonance imaging (MRI), and the administration of dexamethasone (at a dose of 12–32 mg/day) during RT. The data were obtained from the patients, their general practitioners, treating oncologists, and patient files. The patient characteristics are summarized in Table 1. The 257 patients who received 30 Gy in 10 fractions (10 fractions of 3 Gy each, with an overall treatment time of 2 weeks) were compared with 159 patients who received higher doses such as 45 Gy in 15 fractions (15 fractions of 2.5 Gy each over 3 weeks [57 patients]) and 40 Gy in 20 fractions (20 fractions of 2 Gy each over 4 weeks [102 patients]) with respect to survival and freedom from recurrent brain metastasis (local control). The patients treated at the University Hospital Schleswig-Holstein in Luebeck received 10 fractions of 3 Gy each, whereas at the University Hospital Hamburg-Eppendorf in Hamburg, the RT schedule varied with period of time and with the treating physician.

Table 1. Patient Characteristics of the 2 Treatment Groups
 Entire series (N = 416)30 Gy/10 fractions (N = 257)Higher doses (N = 159) 
 No. (%)No. (%)No. (%)P
  1. Gy indicates grays; RT, radiotherapy; RPA, recursive partitioning analysis.

Age, y
 <60224 (54)137 (53)87 (55) 
 ≥60192 (46)120 (47)72 (45).89
Gender
 Female227 (55)141 (55)86 (54) 
 Male189 (45)116 (45)73 (46).93
Karnofsky performance score
 <70198 (48)119 (46)79 (50) 
 ≥70218 (52)138 (54)80 (50).70
Primary tumor
 Breast cancer108 (26)68 (26)40 (25) 
 Lung cancer186 (45)115 (45)71 (45) 
 Other tumors122 (29)74 (29)48 (30).98
No. of metastases
 2–372 (17)45 (18)27 (17) 
 4344 (83)212 (82)132 (83).95
Extracranial metastases
 No122 (29)72 (28)50 (31) 
 Yes294 (71)185 (72)109 (69).73
Interval from tumor diagnosis to RT, mo
 ≤12218 (52)135 (53)83 (52) 
 >12198 (48)122 (47)76 (48).98
RPA class
 157 (14)37 (14)20 (13) 
 2161 (39)101 (39)60 (38) 
 3198 (48)119 (46)79 (50).90

Local control was defined as the absence of progressive disease or recurrent brain metastasis. The diagnosis of disease progression or recurrence was confirmed by CT or MRI. Recurrence was defined either as recurrent metastasis if RT led to a complete response or as a progression of brain metastasis if RT resulted in partial response or no change. Time to disease recurrence and time to death were measured from the completion of RT.

In addition to the RT schedule, the following potential prognostic factors were evaluated: age (<60 years vs ≥60 years; median age: 59 years), gender, KPS (<70 vs ≥70), primary tumor type (breast cancer vs lung cancer vs other tumors), number of metastases (2–3 vs ≥4), presence of extracranial metastases at the time of RT, interval between tumor diagnosis and RT (≤12 months vs >12 months; median interval: 12 months), and RPA class (RPA class 1 vs RPA class 2 vs RPA class 3). Both treatment groups (30 Gy in 10 fractions vs higher dose RT) were well balanced with regard to these factors (Table 1). Subgroup analyses were performed for each of the 3 RPA classes to evaluate whether patients in individual RPA classes benefit from a dose escalation beyond 30 Gy in 10 fractions in terms of better local control and survival.

Local control and survival rates were calculated using the Kaplan-Meier-method.11 The differences between the Kaplan-Meier curves were determined using the log-rank test (univariate analysis). The prognostic factors found to be significant (P < .05) were included in a multivariate analysis, which was performed with the Cox proportional hazards model.

RESULTS

Patients were followed until death or from 6 to 49 months (median: 13 months) in survivors.

The median survival of the entire cohort was 7.5 months after RT. The impact of the potential prognostic factors on survival is shown in Table 2 (using univariate analysis). On univariate analysis, improved survival was found to be significantly associated with age <60 years, a KPS ≥70, only 2 to 3 brain metastases, a lack of extracranial metastases, and RPA class 1. The RT schedule did not appear to have any significant impact on survival. The multivariate analysis of survival included all prognostic factors found to be significant on the univariate analysis but only the RPA class maintained significance (P < .001). However, the RPA class includes the variables of age, KPS, and extracranial metastases. To avoid confounding variables, a second multivariate analysis was performed without the RPA class. On this second multivariate analysis, age (P = .026), KPS (P < .001), and lack of extracranial metastases (P = .003) were found to maintain significant associations with survival. The number of brain metastases was found to be associated with a strong trend (P = .07).

Table 2. Results of the Univariate Analysis of Survival
 At 6 months, %At 12 months, %P
  1. Gy indicates grays; RT, radiotherapy, RPA, recursive partitioning analysis.

Radiation schedule
 30 Gy/10 fractions (n = 257)3320 
 Higher doses (n = 159)2918.86
Age, y
 <60 (n = 224)4023 
 ≥60 (n = 192)2214<.001
Gender
 Female (n = 227)3723 
 Male (n = 189)2615.052
Karnofsky performance score
 <70 (n = 198)62 
 ≥70 (n = 218)5634<.001
Primary tumor
 Breast cancer (n = 108)4429 
 Lung cancer (n = 186)2816 
 Other tumors (n = 122)2615.09
No. of metastases
 2–3 (n = 72)5738 
 4 (n = 344)2615<.001
Extracranial metastases
 No (n = 122)4833 
 Yes (n = 294)2513<.001
Interval from tumor diagnosis to RT, mo
 ≤12 (n = 218)2917 
 >12 (n = 198)3521.06
RPA class
 1 (n = 57)8663 
 2 (n = 161)4523 
 3 (n = 198)62<.001
All patients (n = 416)3219 

Recurrence of brain metastases was observed in 256 patients (62%) after a median interval of 3 months (range, 0–31 months). The potential prognostic factors in relation to freedom from disease recurrence are summarized in Table 3 (univariate analysis). The univariate analysis revealed that improved local control was significantly associated with female gender, a KPS ≥70, a favorable primary tumor (breast cancer), only 2 to 3 brain metastases, an interval between tumor diagnosis and RT of ≥12 months, and RPA class 1.

Table 3. Results of the Univariate Analysis for Local Control of Brain Metastases
 At 6 months, %At 12 months, %P
  1. Gy indicates grays; RT, radiotherapy, RPA, recursive partitioning analysis.

Radiation schedule
 30 Gy/10 fractions (n = 257)3924 
 Higher doses (n = 159)4130.61
Age, y
 <60 (n = 224)4429 
 ≥60 (n = 192)3523.06
Gender
 Female (n = 227)4833 
 Male (n = 189)3018.003
Karnofsky performance score
 <70 (n = 198)147 
 ≥70 (n = 218)5737<.001
Primary tumor
 Breast cancer (n = 108)6144 
 Lung cancer (n = 186)3621 
 Other tumors (n = 122)2619<.001
No. of metastases
 2–3 (n = 72)5938 
 4 (n = 344)3524.001
Extracranial metastases
 No (n = 122)4224 
 Yes (n = 294)3928.87
Interval from tumor diagnosis to RT, mo
 ≤12 (n = 218)3322 
 >12 (n = 198)4731.004
RPA class
 1 (n = 57)7242 
 2 (n = 161)5138 
 3 (n = 198)147<.001
All patients (n = 416)4026 

The multivariate analysis of local control demonstrated that only KPS (P = .005) and type of primary tumor (P = .012) maintained statistical significance. A trend was observed for the number of metastases (P = .099). The RT schedule was found to have no significant impact on local control. On the multivariate analysis without the RPA class, KPS (P < .001) and type of primary tumor (P = .012) again were found to maintain statistical significance, and the number of metastases was found to be of borderline significance. Gender (P = .44 and P = .48, respectively) and the interval between tumor diagnosis and RT (P = .16 and P = .17, respectively) lost significance in both multivariate analyses. The grade 3 acute toxicity rates (according to the National Cancer Institute's Common Toxicity Criteria [version 2.0]) were 5.8% after 30 Gy and 5.0% after higher doses (P = .92). Neurocognitive dysfunction/dementia was noted in 6 patients (2.3%) who were treated with 10 fractions of 3 Gy each and in 8 patients (5.0%) treated with higher doses (P = .24).

The subgroup analyses of each of the 3 RPA classes revealed no significant impact of the RT schedule on survival or local control (Table 4).

Table 4. Subgroup Analyses of the 3 RPA Classes for Survival and Local Control: Potential Impact of the Radiation Schedule
 At 6 months, %At 12 months, %P
  1. RPA indicates recursive partitioning analysis; Gy, grays; NA, not available.

Survival
RPA class 1
 30 Gy/10 fractions (n = 37)8462 
 Higher doses (n = 20)9066.32
RPA class 2
 30 Gy/10 fractions (n = 101)4622 
 Higher doses (n = 60)4326.25
RPA class 3
 30 Gy/10 fractions (n = 119)62 
 Higher doses (n = 79)20.21
Local Control
RPA class 1
 30 Gy/10 fractions (n = 37)6837 
 Higher doses (n = 20)7951.24
RPA class 2
 30 Gy/10 fractions (n = 101)4835 
 Higher doses (n = 60)5545.23
RPA class 3
 30 Gy/10 fractions (n = 119)179 
 Higher doses (n = 79)8NA.40

DISCUSSION

The majority of patients with multiple brain metastases are treated with WBRT alone. RT with 30 Gy in 10 fractions is considered the “standard of care.” However, the results of WBRT alone with 30 Gy are associated with a relatively poor median survival of 3 to 4 months and should be considered suboptimal.

Several approaches have been attempted in an effort to improve these results. Four randomized studies compared 30 Gy in 10 fractions alone versus 30 Gy in 10 fractions plus radiosensitizing agents including misonidazole, metronidazole, lonidamine, and motexafin gadolinium.12–15 However, the administration of radiosensitizers failed to improve survival. Chemotherapy in addition to WBRT also did not result in better survival compared with WBRT alone.16–20 An alternative approach to improve the results of WBRT is the escalation of the RT dose beyond “standard” treatment with 30 Gy in 10 fractions. The current study compared this “standard” regimen with regimens with higher total doses such as 45 Gy in 15 fractions or 40 Gy in 20 fractions. One would expect a better treatment effect in terms of survival and local control of brain metastases with higher RT doses. However, the biologic effectiveness of irradiation depends on the total dose and the dose per fraction. Different RT schedules can be compared with the Equivalent Dose in 2 Gray Fractions (EQD2), which takes into account both the total dose and the dose per fraction.21 The EQD2 is calculated with the equation EQD2 = D × [(d + α/β)/(2 Gy + α/β)], as derived from the linear-quadratic model; D = total dose, d = dose per fraction, α = the linear (first-order dose-dependent) component of cell killing, β = the quadratic (second-order dose-dependent) component of cell killing, and the α/β ratio = the dose at which both components of cell killing are equal. Assuming an αβ ratio of 10 Gy for tumor cell kill, the EQD2 of the radiation schedules are 32.5 Gy (10 fractions of 3 Gy each), 56.25 Gy (15 fractions of 3 Gy each), and 40 Gy (20 fractions of 2 Gy each), respectively.

In contrast to these expectations, the data did not demonstrate an improved outcome with doses >30 Gy in 10 fractions (EQD2 >32.5 Gy). Because the survival of patients with brain metastases is dependent on RPA class, we wanted to determine whether these results were also true for each RPA class of patients. Therefore, we performed subgroup analyses for each of the 3 RPA classes. The subgroup analyses confirmed the results for the entire cohort. The RT schedules were similarly effective with regard to survival and local control. Thus, longer RT programs with doses >30 Gy in 10 fractions cannot be recommended because of the increase in cost and treatment time.4 These patients are often debilitated and benefit by spending as little time as possible receiving therapy. Both treatment groups were well balanced with respect to the other potential prognostic factors, thereby reducing the risk of selection bias. In spite of this apparent balance of prognostic factors, the retrospective nature of the current study must be taken into account when interpreting the results.

The results of the current study are in accordance with the few studies available to date that compared 30 Gy in 10 fractions with higher doses for the treatment of patients with multiple brain metastases. Kurtz et al. compared 30 Gy in 10 fractions with 50 Gy in 20 fractions in a series of 259 patients.8 The median survivals were 4.2 months and 3.9 months, respectively, and the 6-month survival rates were 38% and 36%, respectively (P value not stated). Chatani et al. compared 30 Gy in 10 fractions with 50 Gy in 20 fractions in a series of 92 lung cancer patients.9 The median survivals were 5.4 months and 4.8 months, respectively, and the 1-year survival rates were 27% and 17%, respectively (P = .84). However, neither study evaluated the different RPA classes. Murray et al. compared 30 Gy in 10 fractions with accelerated hyperfractionation with 54.4 Gy in 34 fractions (twice daily over 17 days) in a series of 429 patients with a KPS ≥70.10 The median survivals in that series were 4.5 months and 4.5 months, respectively, and the 1-year survival rates were 19% and 16%, respectively (P = .52). The authors evaluated the 119 RPA class 1 patients, who had 1-year survival rates of 34% and 25%, respectively (P = 0.95), separately. However, approximately 28% of their patients had a single metastasis. To our knowledge, the current study is the first to compare 30 Gy in 10 fractions with higher doses in patients with multiple metastases and that evaluated each RPA class separately in addition to the entire cohort.

Acute toxicity rates were similar in both groups. The incidence of WBRT-induced neurocognitive dysfunction was not found to be significantly different in both treatment groups. The rates of RT-induced dementia of 2.3% and 5.0% noted in the current series are similar to those observed in the retrospective analysis by DeAngelis et al.22 Those authors reported dementia rates of 1.9% to 5.1% after total doses of 25 to 39 Gy given in 3-Gy to 6-Gy fractions. However, in the series of DeAngelis et al., as well as in the current series, the rates of WBRT-induced neurocognitive dysfunction are likely to be underestimated due to the fact that less severe symptoms may not have been reported. According to the results of the current study, a reduction in overall treatment time did not appear to be associated with an increase in RT-induced dementia.

In contrast to the RT schedule, which appeared to have no impact on survival or local control, the treatment outcome was found to be significantly associated with RPA class, age, KPS, and lack of extracranial metastases. The type of primary tumor was found to be significantly associated with local control, but not with survival. A trend was observed for the number of metastases. These findings are in accordance with data regarding the survival of patients with brain metastases presented by Gaspar et al.7 In that study, age, KPS, and lack of extracranial metastases were found to be significant on both univariate analysis and RPA. The type of primary tumor and the number of metastases were found to be significant on the univariate analysis.

In conclusion, according the results of the current study, 30 Gy in 10 fractions (with a treatment time of 2 weeks) provides similar survival and local control and appears not to be associated with increased toxicity compared with longer RT programs with higher doses such as 45 Gy in 15 fractions and 40 Gy in 20 fractions. Thus, escalation of the RT dose beyond 30 Gy in 10 fractions does not appear to benefit these patients. In addition, longer RT programs increase the duration and cost of therapy. Therefore, doses >30 Gy in 10 fractions should not be recommended for patients with multiple brain metastases.

Ancillary