Annals of Oncology 3 (Suppl. 3): S51-S55, 1992. © 1992 Kluwer Academic Publishers. Printed in the Netherlands.
Original article Prophylactic cranial irradiation: Advantages and disadvantages W. T. Sause LDS and Cottonwood Hospitals, University of Utah Medical Center, Salt Lake City, Utah, U.S.A.
Introduction
Prophylactic irradiation to a potential sanctuary site demonstrates several important aspects of the pathophysiology of malignancies like childhood leukemias and small cell lung cancer (SCLC). Features that make treatment of a sanctuary site like the central nervous system (CNS) attractive in a particular disease are the high curability of the disease with systemic therapy, the high incidence of CNS relapse, the effectiveness of cranial irradiation, and the ability of this radiation to sterilize the CNS at doses that are not likely to produce a high incidence of late complications. Theoretically, acute lymphoblastic leukemia in childhood lends itself to such sanctuary-site treatment. Since the 1960s, acute lymphoblastic leukemia in children has been a potentially curable disease. Modern systemic chemotherapy cures more than 50% of children who present with this illness. In the 1960s, CNS relapses were noted to occur in approximately 40% to 60% of children following successful induction therapy. Clinical research conducted at the St Jude Children's Table I. Prevention of CNS relapse: Leukemia. Period
Therapy
CNS relapse (%)
1962-1965 1965-1967 1970-1971
12Gy None 24 Gy
40 60 7
Methotrexate Adapted from Poplack
ment schedules and avoiding the use of doxorubicin with irradiation. Prophylactic cranial irradiation delivered to those patients who obtain a complete remission in response to induction chemotherapy can reduce the incidence of CNS relapse and can usually be delivered with minimal CNS toxicity. Unfortunately, long-term disease control with systemic agents is poor. Until more effective systemic agents are developed, the role of prophylactic cranial irradiation is one of palliation rather than improving survival. Key words: small cell anaplastic carcinoma, prophylactic cranial irradiation, CNS relapse
Research Hospital demonstrated that modest radiation doses delivered to the cranial vault with intrathecal methotrexate reduced the incidence of initial CNS relapse to less than 10% (Table 1) [1]. Over the last 20 years, prognostic factors and radiation doses have been refined to provide maximal CNS sterilization with minimal neurologic toxicity. A similar treatment policy could be applied to anaplastic SCLC. SCLC represents a tumor that responds readily to cytotoxic chemotherapy, exhibits a high incidence of CNS relapse, is relatively radiosensitive, and modest radiation doses to the brain would be unlikely to cause late treatment sequelae in adults. Based on the clinical success observed in the treatment of acute lymphoblastic leukemia, prophylactic cranial irradiation (PCI) has been used in patients with anaplastic SCLC for a number of years. Unfortunately, the utility of PCI remains controversial.
Relapse pattern
Anaplastic SCLC has a high propensity for systemic spread. Early studies using only surgery and radiation were notable for the high incidence of systemic relapse. The development of systemic treatment in the 1970s has been responsible for improved median and 2-year survival in patients with this disease [2]. Ultimately, however, the brain represents a frequent relapse site and the majority of patients relapse. Several randomized and nonrandomized series have been published that attest to the relapse pattern of anaplastic SCLC
Downloaded from http://annonc.oxfordjournals.org/ at University of Chicago on July 13, 2015
Summary. Anaplastic small cell carcinoma of the lung is a disease that responds to multiagent systemic therapy. Unfortunately, a high incidence of central nervous system (CNS) relapse has been observed by many investigators. Cranial irradiation has been successfully used in acute lymphoblastic leukemia to reduce CNS relapse and improve survival. A similar application of prophylactic cranial irradiation in patients undergoing therapy for small cell anaplastic carcinoma of the lung has been noted to decrease CNS relapse from approximately 25% to approximately 5%. However, CNS toxicity has been noted to occur with CNS prophylaxis. We can reduce the incidence of CNS toxicity by using modern treat-
52
Dose-response relationship Anaplastic SCLC is a tumor that responds well to external-beam radiation. In sequential studies conducted by several investigators, a clinical dose-response relationship appears between the radiation dose deliv-
Table 2. PCI randomized trials. References
18] [91 |101 111] [12]
[13] |14]
% CNS relapse
Survival
With PCI
No PCI
0 0 17 9 0 4 5
36 16 24 13 27 18 21
NS NS NS NS NS NS NS
NS = not significant.
ered and intrathoracic tumor control. Arriagada et al. [15] were able to demonstrate improved intrathoracic control when doses escalated from 45 to 55 Gy, with 61% local control at 45 Gy vs. 82% at the 55-Gy dose. The Southwest Oncology Group (SWOG) was able to demonstrate an increase in local control when thoracic doses escalated from 30 to 45 Gy. In SWOG 76-28 [16], an increased number of relapses were noted at the lower dose. Choi [17], who analyzed chest relapses according to radiation dose delivered, also was able to document improved local control with escalating radiation doses (Table 3). Table 3. Intrathoracic control of anaplastic small cell lung cancer. References
[17]
Radiation dose (Gy)
% Control
45 55
61 82
30 50
61
The lowest possible dose necessary to sterilize the CNS remains somewhat controversial. The NCI has analyzed its results with therapeutic radiation for CNS metastases [18]. Thirty-seven of 61 patients achieved an objective response to radiation, with 32% exhibiting CR to treatment for clinically symptomatic brain relapses. Twenty-four of these 37 patients subsequently developed progressive CNS tumors. Patients received radiation in a variety of schedules, but those patients who were treated with more than 40 Gy had a median duration of response of 10 months, as opposed to 7 months for those who received less than 40 Gy. Komaki et al. [19], who analyzed results from the Medical College of Wisconsin, concluded that 10 2.5Gy fractions delivered in 2 weeks were as effective as 10 3-Gy fractions in 2 weeks for prophylactic cranial treatment. The group receiving 25 Gy experienced a 3% relapse rate compared with a 10% relapse rate in those receiving 30 Gy. The results of the NCI analysis suggest, but do not confirm, that 20 Gy in 10 fractions during induction therapy are similar to 24 Gy in 10 fractions delivered after initial chemotherapy. There appears to be a dose-response relationship
Downloaded from http://annonc.oxfordjournals.org/ at University of Chicago on July 13, 2015
and the ability of radiation therapy to reduce the incidence of early relapse [3]. A nonrandomized National Cancer Institute (NCI) series analyzed consecutive cohorts of patients treated with or without PCI [4]; 332 patients were analyzed. The analysis included patients treated from 1974 to 1980. In all, 136 patients had limited disease (LD), and 123 of the 332 had had a complete response (CR) to chemotherapy. A 42% incidence of CNS relapse occurred in the 176 patients who received no PCI. Of those treated with PCI, one group that received 24 Gy in eight fractions exhibited a 34% incidence of CNS relapse, while the group that received 20 Gy with induction chemotherapy experienced a 17% incidence. In the group with CRs to induction therapy, none who had received PCI experienced CNS relapse as the first site of recurrence, whereas 16% of those who received no PCI relapsed initially in the brain. The University of Maryland attempted a randomized trial of patients who initially experienced a CR to therapy [5]. An early analysis of this trial revealed that none of the 17 patients who had received PCI experienced CNS relapse while 31% of those not receiving PCI did. Randomization was discontinued at the time of these early analyses. When all patients who were entered into consecutive trials - conducted from 1975 through 1982 at the University of Maryland - were analyzed, 20% of those who received PCI (compared with 42% of those receiving no PCI) exhibited a CNS relapse. Only 1 patient treated with PCI experienced initial CNS relapse while 23% of the no-PCI group experienced initial CNS relapse [5]. From 1975 to 1979, the Radiation Therapy Oncology Group (RTOG) and Eastern Cooperative Oncology Group (ECOG) conducted a randomized trial of patients with LD anaplastic SCLC [6]. A total of 104 patients experienced CR to radiation therapy and were randomized to PCI vs. no PCI. Of the 104 patients, 11 developed brain metastases; only 2 of these 11 had been randomized to PCI. A recent University of Indiana analysis [7] revealed that of 38 CR patients who received PCI, 1 had an initial CNS relapse. Of 20 patients who did not receive PCI, 4 had brain relapses. Table 2 [8-14] lists various randomized trials and documents the incidence and reduction of CNS relapses with PCI. These trials represent a heterogeneous population of patients, including those with LD and extensive disease (ED), and those who did not as well as those who did experience CR to initial therapy. One consistent aspect of these reported series is a reduction in CNS relapses when PCI is used.
53
between radiation and control of anaplastic SCLC. In the treatment of subclinical CNS disease, the dose necessary for sterilization should be lower than that needed for overt disease. A dose greater than 20 Gy delivered in 10 fractions is probably necessary to prevent CNS relapse, but doses as high as 30 Gy in 10 fractions may be unnecessary. A complicating factor in the delivery of PCI is the timing of systemic chemotherapy, which may have a detrimental effect on normal tissue tolerance. Toxicity
Table 4. Neuret values for a variety of radiation schedules. Total dose (Gy)
Fraction size (Gy)
Tii
50 36 30 20 30 20
2 2.4 3 4 2 2.5
33 47 12 5 19 12
1)
Neuret
983 867 938 894 763 781
Adapted from Turrisi [22|.
Investigators from the University of Indiana recently reported an incidence of CNS toxicity as high as 63% in patients receiving PCI and doxorubicin-based chemotherapy concurrently. Radiation was delivered in fractions of 2.4 to 3 Gy, to a total dose of 30 to 36 Gy [7]. In the same issue of the Journal of Clinical Oncology, investigators from the Princess Margaret Hospital in Toronto analyzed late CNS sequelae following delivery of 20 Gy in five fractions applied between
Schema
Rduction chemotherapy
R E S
T A
G E
CR—-
PR. SD. PD ineligible
R A N D
O M
1 Z \ E
PCI25Gy/10Fx over 2-2 'Awk Follow-up evaluation q3mo years 1-2. q6mo years 3-5 Observation — > CNS — > TCI relapse
Randomization must be
Original article Prophylactic cranial irradiation: Advantages and disadvantages W. T. Sause LDS and Cottonwood Hospitals, University of Utah Medical Center, Salt Lake City, Utah, U.S.A.
Introduction
Prophylactic irradiation to a potential sanctuary site demonstrates several important aspects of the pathophysiology of malignancies like childhood leukemias and small cell lung cancer (SCLC). Features that make treatment of a sanctuary site like the central nervous system (CNS) attractive in a particular disease are the high curability of the disease with systemic therapy, the high incidence of CNS relapse, the effectiveness of cranial irradiation, and the ability of this radiation to sterilize the CNS at doses that are not likely to produce a high incidence of late complications. Theoretically, acute lymphoblastic leukemia in childhood lends itself to such sanctuary-site treatment. Since the 1960s, acute lymphoblastic leukemia in children has been a potentially curable disease. Modern systemic chemotherapy cures more than 50% of children who present with this illness. In the 1960s, CNS relapses were noted to occur in approximately 40% to 60% of children following successful induction therapy. Clinical research conducted at the St Jude Children's Table I. Prevention of CNS relapse: Leukemia. Period
Therapy
CNS relapse (%)
1962-1965 1965-1967 1970-1971
12Gy None 24 Gy
40 60 7
Methotrexate Adapted from Poplack
ment schedules and avoiding the use of doxorubicin with irradiation. Prophylactic cranial irradiation delivered to those patients who obtain a complete remission in response to induction chemotherapy can reduce the incidence of CNS relapse and can usually be delivered with minimal CNS toxicity. Unfortunately, long-term disease control with systemic agents is poor. Until more effective systemic agents are developed, the role of prophylactic cranial irradiation is one of palliation rather than improving survival. Key words: small cell anaplastic carcinoma, prophylactic cranial irradiation, CNS relapse
Research Hospital demonstrated that modest radiation doses delivered to the cranial vault with intrathecal methotrexate reduced the incidence of initial CNS relapse to less than 10% (Table 1) [1]. Over the last 20 years, prognostic factors and radiation doses have been refined to provide maximal CNS sterilization with minimal neurologic toxicity. A similar treatment policy could be applied to anaplastic SCLC. SCLC represents a tumor that responds readily to cytotoxic chemotherapy, exhibits a high incidence of CNS relapse, is relatively radiosensitive, and modest radiation doses to the brain would be unlikely to cause late treatment sequelae in adults. Based on the clinical success observed in the treatment of acute lymphoblastic leukemia, prophylactic cranial irradiation (PCI) has been used in patients with anaplastic SCLC for a number of years. Unfortunately, the utility of PCI remains controversial.
Relapse pattern
Anaplastic SCLC has a high propensity for systemic spread. Early studies using only surgery and radiation were notable for the high incidence of systemic relapse. The development of systemic treatment in the 1970s has been responsible for improved median and 2-year survival in patients with this disease [2]. Ultimately, however, the brain represents a frequent relapse site and the majority of patients relapse. Several randomized and nonrandomized series have been published that attest to the relapse pattern of anaplastic SCLC
Downloaded from http://annonc.oxfordjournals.org/ at University of Chicago on July 13, 2015
Summary. Anaplastic small cell carcinoma of the lung is a disease that responds to multiagent systemic therapy. Unfortunately, a high incidence of central nervous system (CNS) relapse has been observed by many investigators. Cranial irradiation has been successfully used in acute lymphoblastic leukemia to reduce CNS relapse and improve survival. A similar application of prophylactic cranial irradiation in patients undergoing therapy for small cell anaplastic carcinoma of the lung has been noted to decrease CNS relapse from approximately 25% to approximately 5%. However, CNS toxicity has been noted to occur with CNS prophylaxis. We can reduce the incidence of CNS toxicity by using modern treat-
52
Dose-response relationship Anaplastic SCLC is a tumor that responds well to external-beam radiation. In sequential studies conducted by several investigators, a clinical dose-response relationship appears between the radiation dose deliv-
Table 2. PCI randomized trials. References
18] [91 |101 111] [12]
[13] |14]
% CNS relapse
Survival
With PCI
No PCI
0 0 17 9 0 4 5
36 16 24 13 27 18 21
NS NS NS NS NS NS NS
NS = not significant.
ered and intrathoracic tumor control. Arriagada et al. [15] were able to demonstrate improved intrathoracic control when doses escalated from 45 to 55 Gy, with 61% local control at 45 Gy vs. 82% at the 55-Gy dose. The Southwest Oncology Group (SWOG) was able to demonstrate an increase in local control when thoracic doses escalated from 30 to 45 Gy. In SWOG 76-28 [16], an increased number of relapses were noted at the lower dose. Choi [17], who analyzed chest relapses according to radiation dose delivered, also was able to document improved local control with escalating radiation doses (Table 3). Table 3. Intrathoracic control of anaplastic small cell lung cancer. References
[17]
Radiation dose (Gy)
% Control
45 55
61 82
30 50
61
The lowest possible dose necessary to sterilize the CNS remains somewhat controversial. The NCI has analyzed its results with therapeutic radiation for CNS metastases [18]. Thirty-seven of 61 patients achieved an objective response to radiation, with 32% exhibiting CR to treatment for clinically symptomatic brain relapses. Twenty-four of these 37 patients subsequently developed progressive CNS tumors. Patients received radiation in a variety of schedules, but those patients who were treated with more than 40 Gy had a median duration of response of 10 months, as opposed to 7 months for those who received less than 40 Gy. Komaki et al. [19], who analyzed results from the Medical College of Wisconsin, concluded that 10 2.5Gy fractions delivered in 2 weeks were as effective as 10 3-Gy fractions in 2 weeks for prophylactic cranial treatment. The group receiving 25 Gy experienced a 3% relapse rate compared with a 10% relapse rate in those receiving 30 Gy. The results of the NCI analysis suggest, but do not confirm, that 20 Gy in 10 fractions during induction therapy are similar to 24 Gy in 10 fractions delivered after initial chemotherapy. There appears to be a dose-response relationship
Downloaded from http://annonc.oxfordjournals.org/ at University of Chicago on July 13, 2015
and the ability of radiation therapy to reduce the incidence of early relapse [3]. A nonrandomized National Cancer Institute (NCI) series analyzed consecutive cohorts of patients treated with or without PCI [4]; 332 patients were analyzed. The analysis included patients treated from 1974 to 1980. In all, 136 patients had limited disease (LD), and 123 of the 332 had had a complete response (CR) to chemotherapy. A 42% incidence of CNS relapse occurred in the 176 patients who received no PCI. Of those treated with PCI, one group that received 24 Gy in eight fractions exhibited a 34% incidence of CNS relapse, while the group that received 20 Gy with induction chemotherapy experienced a 17% incidence. In the group with CRs to induction therapy, none who had received PCI experienced CNS relapse as the first site of recurrence, whereas 16% of those who received no PCI relapsed initially in the brain. The University of Maryland attempted a randomized trial of patients who initially experienced a CR to therapy [5]. An early analysis of this trial revealed that none of the 17 patients who had received PCI experienced CNS relapse while 31% of those not receiving PCI did. Randomization was discontinued at the time of these early analyses. When all patients who were entered into consecutive trials - conducted from 1975 through 1982 at the University of Maryland - were analyzed, 20% of those who received PCI (compared with 42% of those receiving no PCI) exhibited a CNS relapse. Only 1 patient treated with PCI experienced initial CNS relapse while 23% of the no-PCI group experienced initial CNS relapse [5]. From 1975 to 1979, the Radiation Therapy Oncology Group (RTOG) and Eastern Cooperative Oncology Group (ECOG) conducted a randomized trial of patients with LD anaplastic SCLC [6]. A total of 104 patients experienced CR to radiation therapy and were randomized to PCI vs. no PCI. Of the 104 patients, 11 developed brain metastases; only 2 of these 11 had been randomized to PCI. A recent University of Indiana analysis [7] revealed that of 38 CR patients who received PCI, 1 had an initial CNS relapse. Of 20 patients who did not receive PCI, 4 had brain relapses. Table 2 [8-14] lists various randomized trials and documents the incidence and reduction of CNS relapses with PCI. These trials represent a heterogeneous population of patients, including those with LD and extensive disease (ED), and those who did not as well as those who did experience CR to initial therapy. One consistent aspect of these reported series is a reduction in CNS relapses when PCI is used.
53
between radiation and control of anaplastic SCLC. In the treatment of subclinical CNS disease, the dose necessary for sterilization should be lower than that needed for overt disease. A dose greater than 20 Gy delivered in 10 fractions is probably necessary to prevent CNS relapse, but doses as high as 30 Gy in 10 fractions may be unnecessary. A complicating factor in the delivery of PCI is the timing of systemic chemotherapy, which may have a detrimental effect on normal tissue tolerance. Toxicity
Table 4. Neuret values for a variety of radiation schedules. Total dose (Gy)
Fraction size (Gy)
Tii
50 36 30 20 30 20
2 2.4 3 4 2 2.5
33 47 12 5 19 12
1)
Neuret
983 867 938 894 763 781
Adapted from Turrisi [22|.
Investigators from the University of Indiana recently reported an incidence of CNS toxicity as high as 63% in patients receiving PCI and doxorubicin-based chemotherapy concurrently. Radiation was delivered in fractions of 2.4 to 3 Gy, to a total dose of 30 to 36 Gy [7]. In the same issue of the Journal of Clinical Oncology, investigators from the Princess Margaret Hospital in Toronto analyzed late CNS sequelae following delivery of 20 Gy in five fractions applied between
Schema
Rduction chemotherapy
R E S
T A
G E
CR—-
PR. SD. PD ineligible
R A N D
O M
1 Z \ E
PCI25Gy/10Fx over 2-2 'Awk Follow-up evaluation q3mo years 1-2. q6mo years 3-5 Observation — > CNS — > TCI relapse
Randomization must be
