Fax: (206) 598-3786
Evaluation of the use of prophylactic cranial irradiation in small cell lung cancer†
Version of Record online: 31 DEC 2008
Copyright © 2009 American Cancer Society
Volume 115, Issue 4, pages 842–850, 15 February 2009
How to Cite
Patel, S., Macdonald, O. K. and Suntharalingam, M. (2009), Evaluation of the use of prophylactic cranial irradiation in small cell lung cancer. Cancer, 115: 842–850. doi: 10.1002/cncr.24105
Presented in part at the European Cancer Conference meeting, Barcelona, Spain, September 2007.
- Issue online: 2 FEB 2009
- Version of Record online: 31 DEC 2008
- Manuscript Accepted: 15 SEP 2008
- Manuscript Revised: 19 AUG 2008
- Manuscript Received: 19 MAR 2008
- lung cancer;
- cranial irradiation;
- overall survival;
- cause-specific survival
Prophylactic cranial irradiation has been used in patients with small cell lung cancer to reduce the incidence of brain metastasis after primary therapy. The purpose of this study was to evaluate the effects of prophylactic cranial irradiation (PCI) on overall survival and cause-specific survival.
A total of 7995 patients with limited stage small cell lung cancer diagnosed between 1988 and 1997 were retrospectively identified from centers participating in the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) Program. Of them, 670 were identified as having received PCI as a component of their first course of therapy. Overall survival and cause-specific survival were estimated by the Kaplan-Meier method, comparing patients treated with or without prophylactic whole-brain radiotherapy. The Cox proportional hazards model was used in the multivariate analysis to evaluate potential prognostic factors.
The median follow-up time was 13 months (range, 1 month to 180 months). Overall survival at 2 years, 5 years, and 10 years was 23%, 11%, and 6%, respectively, in patients who did not receive PCI. In patients who received PCI, the 2-year, 5-year, and 10-year overall survival rates were 42%, 19%, and 9%, respectively (P = <.001). The cause-specific survival rate at 2 years, 5 years, and 10 years was 28%, 15%, 11%, respectively, in patients who did not receive PCI and 45%, 24%, 17%, respectively, in patients who did receive PCI (P = <.001). On multivariate analysis of cause-specific and overall survival, age at diagnosis, sex, grade, extent of primary disease, size of disease, extent of lymph node involvement, and PCI were found to be significant (P = <.001). The hazards ratios for disease-specific and all cause mortality were 1.13 and 1.11, respectively, for those not receiving PCI.
Significantly improved overall and cause-specific survival was observed in patients treated with prophylactic cranial irradiation on unadjusted and adjusted analyses. This study concurs with the previously published European experience. Prophylactic cranial irradiation should be considered for patients with limited stage small cell lung cancer. Cancer 2009. © 2008 American Cancer Society.
An estimated 213,380 new cases of primary cancer of the lung and bronchus will be diagnosed in 2007, with an estimated 160,390 patient deaths resulting from lung and bronchial primary malignancies.1 Small cell lung cancer has recently been observed to account for approximately 13% of all newly diagnosed lung and bronchial cancers.2 The majority of patients with small cell cancer of the lung present with extensive disease that, without treatment, generally leads rapidly to mortality. Recent analysis regarding the changing epidemiology of small cell lung cancer within the US over the past 30 years using data from the Survival, Epidemiology, and End Results (SEER) database indicates patients diagnosed with limited stage small cell disease had an approximately 10% 2-year survival rate in the year 1998. Randomized analysis of an aggressive combined chemotherapy and radiation strategy for patients with limited stage disease observed a median survival of 23 months with a 5-year survival rate of 26%.3
Brain metastases are common in patients diagnosed with small cell lung cancer. Reports indicate that nearly 20% of patients will have brain metastasis at the time of diagnosis but that the incidence increases considerably thereafter.4, 5 The efficacy of systemic chemotherapy for asymptomatic and symptomatic brain metastasis is modest at best, and to our knowledge maintenance chemotherapy has not demonstrated an ability to reduce the incidence of subsequent metastasis after primary therapy.6-8 Although the majority of patients do not present with brain metastases, the significant risk of subsequently developing them can lend to a poor prognosis and increase morbidity, adversely affecting quality of life.
Many prospective studies were initiated in the late 1980s and early 1990s investigating the use of prophylactic cranial irradiation (PCI) in an effort to decrease the risk of symptomatic brain metastasis in patients with primary small cell carcinoma of the lung. Many of those studies and subsequent meta-analyses reported a reduced risk of symptomatic brain metastasis with PCI and improved survival.9-12 The meta-analysis reported by Auperin et al, which included 7 trials, 1 with unpublished data, demonstrated an absolute improvement in survival from 15.3% to 20.7% at 3 years.9 The 5.4% benefit represents a >30% improvement in survival for the treated versus nontreated group. Furthermore, an absolute decrease of 25.3% in subsequent development of brain metastasis was observed. Meert et al, in a similar meta-analysis composed of 12 randomized studies, reported a hazards ratio for the incidence of brain metastasis of 0.48 for those who underwent PCI.11 Recent data in the extensive stage setting suggest that PCI is also indicated in patients with advanced stages of disease, because a survival advantage was observed in patients who received PCI after any response to induction chemotherapy.13 Although this recently reported study continues to shed more light on the use of PCI, it is worth noting that this study was criticized for the fraction schemes used as well as the lack of brain imaging performed before the start of PCI.
Despite data substantiating a survival benefit with the application of PCI in patients with limited stage disease, there still is a reticence among providers to recommend it for fear of subsequent neurologic toxicity as well as a question of the long–term benefit of this therapy. In an effort to further characterize the potential benefits of prophylactic cranial irradiation in the population of patients with limited stage disease, we performed an analysis using a large registry database supported by the National Cancer Institute evaluating the endpoints of overall and cancer-specific survival among those who did or did not receive PCI.
MATERIALS AND METHODS
The data for this analysis were obtained from the National Cancer Institute SEER program using the SEER 17 registries, November 2005 edition (1973-2003 data sets). SEER is composed of a set of geographically defined, population-based central cancer registries in the US and is operated under contract with the National Cancer Institute by local nonprofit organizations. Serial registry data were submitted electronically without personal identifiers.
The analyzed population consisted of patients with primary, nonmetastatic small cell carcinoma of the lung for the years 1988 through 1997. SEER registry requirements for the period 1988 through 1997 incorporated the recording of whole-brain radiotherapy when delivered as a component of the first course of therapy. The practice of recording whole-brain irradiation for the time period preceding 1988 and the time period after 1997 was not observed. Specifics regarding the delivery of radiotherapy, including fraction size, total dose, and other technical factors, were not recorded in the SEER database.
The information gathered from the database regarding eligible patients included year of diagnosis, ethnicity, age at diagnosis, sex, marital status, primary disease site, extent of lymph nodes involved, extent of localized disease, size of primary disease, lateral site of disease, histologic grade, the presence or absence of surgery, the presence or absence of primary radiotherapy, the presence or absence of whole-brain radiotherapy, survival time in months from the date of diagnosis to last contact or death, and vital status. Clinical staging using the current American Joint Committee on Cancer (AJCC) criteria (6th edition)14 was retrospectively applied using the information within the registry, including disease site, histology, extent of invasion, lymph node status, and distant metastasis status. Specifics regarding the staging studies performed for each patient were not available; rather, information regarding the clinical characteristics of disease was recorded based on information in the patient's medical record at the time of data submission. Chemotherapy data were not available within the SEER and therefore were not a component of the analysis. Analyses for the endpoints of overall survival (OS) and cause-specific (CSS) survival were performed using commonly accepted statistical models. Cause-specific death was defined as a recorded death from lung cancer or brain tumor. The actuarial rates of OS and CSS were estimated using the Kaplan-Meyer method.15 Adjusted, multivariate analysis was performed using the Cox proportional hazards model for the following factors: age at diagnosis (as a continuous variable), sex, ethnicity, T category, N category, size (cm, continuous), laterality (ie, right or left), presence of surgical resection, presence of radiation (to the primary disease), and the presence of whole-brain radiotherapy as first course of therapy (in M0) disease.16 Statistical significance was declared for P < .05.
For the years 1988 through 1997, the SEER contains information on 22,674 patients with a diagnosis of small cell carcinoma, of which 7995 patients had complete demographic and clinical information and were included in the analysis. The median age was 67 years (range, 16-99 years). Of all patients, 51% (n = 4093) were male; the remainder were female (n = 3902). The distribution of patients for the years 1988 through 1997, inclusive, was generally equal with a low of 557 patients included from the year 1990 and a high of 1012 patients from the year 1995. On average, 800 patients per year were available for analysis. Prophylactic whole-brain radiotherapy as a first course of therapy for nonmetastatic disease was delivered in 670 (8%) patients. Patient demographic, clinical, and pathologic data are available in Table 1.
|Median age (range), y||67 (16-99)|
|Tumor size, cm|
|Median (range)||4.0 (0.1-20)|
|Lymph node disease|
|AJCC stage of disease (6th ed)|
|Prophylactic cranial irradiation|
|Thoracic external beam radiotherapy|
|Median follow-up (range), mo||13 (1-180)|
|Dead due to disease||6382||(80)|
The median follow-up was 13 months (range, 1-180 months). Rates of overall survival at 2 years, 5 years, and 10 years were 23%, 11%, and 6%, respectively, in patients who did not receive PCI. Those patients who received PCI experienced 2-year, 5-year, and 10-year survival rates of 42%, 19%, and 9%, respectively (P < .001 for all time points) (Fig. 1A). Similarly, those patients who underwent PCI experienced improved cause-specific survival compared with those who did not, with 2-year, 5-year, and 10-year survival rates of 45%, 24%, and 17%, respectively, compared with 28%, 15%, and 11%, respectively (P < .001 for all time points) (Fig. 1B). All but 2 (<1%) of the total deaths (n = 7049) occurring within the first 5 years among all patients were recorded as cancer-related. Analysis of the survival endpoints was performed adjusting for disease stage. Stages I and II were combined owing to the relative paucity of patients receiving PCI in those stage groupings. OS and CSS was observed to be significantly improved with PCI for stages I-II, stage IIIA, and stage IIIB disease (Figs. 2-4).
The potential benefit of prophylactic cranial irradiation was also evaluated for sex and ethnicity. The results indicate a similarly significant magnitude of benefit with PCI for both the CSS and OS endpoints regardless of sex (Table 2). When analyzed by race, African–American patients were noted to have significantly worse OS, with 2-year and 5-year rates of 21% and 8% compared with 27% and 12% among whites (P < .001 for both time points). Similar, significant findings were observed for the CSS end point when analyzed by race.
Adjusted analysis revealed age, sex, extent of primary disease, extent of lymph node involvement, size of primary disease, and the use of PCI to independently predict for the overall survival endpoint. Age, sex, grade, extent of primary disease, extent of lymph node involvement, primary tumor size, and PCI were also found to predict for the CSS endpoint (Table 3).
|Cause-specific Survival||Overall Survival|
|HR (95% CI)||P||HR (95% CI)||P|
|Age at diagnosis*||1.02 (1.01-1.03)||<.0001||1.03 (1.02-1.03)||<.0001|
|African American||1.10 (1.01-1.20)||0.04†||1.11 (1.03-1.20)||0.008†|
|Other‡||0.97 (0.90-1.06)||0.97 (0.90-1.05)|
|Female||0.93 (0.90-0.97)||<.0001||0.92 (0.89-0.95)||<.0001|
|T2||0.94 (0.85-1.03)||<.0001†||0.98 (0.90-1.07)||<.0001†|
|T3||1.31 (1.15-1.49)||1.24 (1.09-1.40)|
|T4||1.06 (0.99-1.13)||1.06 (0.99-1.13)|
|N1||1.09 (0.97-1.21)||<.0001†||1.06 (0.96-1.17)||<.0001†|
|N2||1.10 (1.03-1.17)||1.08 (1.02-1.15)|
|N3||1.26 (1.12-1.42)||1.22 (1.09-1.37)|
|LN, NOS||0.96 (0.89-1.04)||0.96 (0.90-1.03)|
|Primary tumor size*||1.04 (1.02-1.05)||<.0001||1.03 (1.02-1.04)||<.0001|
|Left||1.01 (0.97-1.05)||.61||1.00 (0.97-1.04)||.77|
|Wedge||1.06 (0.90-1.24)||<.0001†||1.08 (0.94-1.24)||<.0001†|
|Lobectomy||0.65 (0.56-0.75)||0.71 (0.63-0.80)|
|Pneumonectomy||0.81 (0.60-1.08)||0.75 (0.57-0.98)|
|Yes||0.82 (0.79-0.86)||<.0001||0.83 (0.80-0.86)||<.0001|
|None||1.09 (1.02-1.16)||.01||1.07 (1.01-1.13)||.03|
This analysis suggests that prophylactic cranial irradiation substantially improves not only overall survival but also CSS in patients with limited stage small cell lung cancer. Although it is well established that PCI reduces the likelihood of cranial recurrence,10, 12, 13, 17 to our knowledge the current study is 1 of the few studies that reveals an improvement in OS and is the first study to report that this improvement persists at 10 years. These favorable survival results are consistent with other studies that evaluated PCI in patients with limited stage small cell lung cancer, including the studies by Auperin et al and Arriagada et al, which revealed a 2-year OS rate of 22% versus 29% in the treatment group and the meta-analysis for which the 3-year OS rate increased from 15.3% to 20.7%.9, 17 Worth noting are the results from the recently reported randomized study of PCI in patients with extensive stage disease with any response to chemotherapy, which indicate a 1-year survival rate of 27.1% with PCI compared with 13.3% in the control group (P = .003). Whereas the benefit in survival was noted in this most recent European Organization for Research and Treatment of Cancer (EORTC) PCI study, it is worth noting that this study has limitations, including the finding that brain imaging was not required before patient study participation and that the radiation dose regimens used would, by US standards, be considered definitive and not prophylactic. The most common fractionation in this study was 20 grays (Gy)/5 fractions, which is not commonly used for PCI in North America.
The observed benefit in OS and CSS remained significant when individual stages were separated and evaluated. Furthermore, the significant benefit with PCI was observed in males and females. In addition, we observed a significant difference in OS and CSS among whites compared with African Americans in this cohort. Causes for the relative decrement in survival among African Americans can only be hypothesized owing to the limitations of the data from the SEER. However, this finding is not unique insofar as other reports have observed similarly poorer outcomes affecting African Americans with breast cancer, soft tissue sarcoma, and gynecologic malignancies.18-21 Further investigations are necessary to determine whether these differences are a result of tumor biology, differences in screening and access to care, the type of care received, or other factors or combinations of factors.
One of the limitations of the current study is that only those patients with primary, nonmetastatic small cell carcinoma of the lung for the years 1988 through 1997 were included as the analyzable population, because for the periods before and after, the use of whole-brain radiotherapy was not recorded in the SEER database. This limits this current study because changes in diagnostic testing, systemic therapy, and radiation treatment planning/delivery since that time could impact the outcomes in patients treated today, or could allow for the overstatement of observed findings in the population of patients studied in this work. In addition, although the authors were unable to ascertain such from the SEER database, it is worth mentioning that patients selected for PCI may be a different population from those not selected for PCI. In addition, patients were identified as having the nonmetastatic disease solely by the individual inputting the data. The type of brain imaging used was not reported. Another weakness included in this study is the lack of information regarding the use of chemotherapy. This truly limits the analysis because one cannot be certain whether the patients received the standard of care for small cell lung cancer.
As survival continues to increase with the advent of newer chemotherapies and radiation treatment techniques for small cell lung cancer, it is reasonable to conclude that the chance of brain metastases will also increase. With the increased propensity of metastases, one should focus on prevention. Prophylactic cranial irradiation is instrumental not only in prevention but also potentially improves quality of life.22 Slotman et al report that cranial irradiation did not significantly deteriorate patients' self-assessment of their global health status and was reasonably tolerated in the acute setting. Robust analysis of their quality of life data was not possible however, because of the commonly rapid demise of many patients with extensive stage disease.13
The detriment on neurocognitive effects continues to be debated among oncologists. Crossen et al conducted a wide-ranging review of neurologic and neurocognitive adverse effects noted in patients after cranial irradiation in a wide range of clinical settings, including dementia and brain necrosis after high-dose radiotherapy for primary gliomas, an interaction between chemotherapy and radiation, and abnormalities in psychometric testing being detected more readily than positive findings on clinical examination.23 Johnson et al studied long-term survivors of small cell lung cancer and found a decline in neurocognitive function with some correlation with abnormalities on computed tomography.24 In both of these studies, problems were more frequent when chemotherapy was given at the time of cranial irradiation, or large radiation fraction size was used. Fleck et al had recommended against PCI, finding it had limited efficacy at reducing brain metastases and the occurrence of significant complications in neurologic and cognitive functions.25 Examination of their treatment details reveals treatment plans that would not be recommended today; 1 plan called for split-course irradiation in proximity to chemotherapy, and the other delivered 10 fractions of 3 Gy each concurrent with a doxorubicin-containing chemotherapy regimen.
To our knowledge, there have been few published studies in which psychometric testing was performed in conjunction with PCI.26-28 These studies reported a neurocognitive deficit before PCI that did not deteriorate after completion of cranial irradiation. In addition, what to our knowledge are 2 of the largest randomized studies published to date incorporated neurocognitive function by psychometric testing before and after PCI.17, 29 To our knowledge, the PCI85 trial was the first major randomized study that addressed this issue.17 In each treatment center, initial neurocognitive tests and follow-up assessments were performed by the same neurologist. The results of testing over 5 years of follow-up in the PCI85 trial demonstrated no differences in neurocognitive function between patients receiving PCI and those who did not. The UK02 trial was another randomized trial to include a formal assessment of neurocognitive functioning.29 Testing was done at randomization and every 6 months afterward. Impairment was defined using standard deviations from age-matched controls. Up to 40% of patients demonstrated significant abnormalities on individual tests before PCI. Psychometric evaluations performed in follow-up demonstrated no adverse effects of PCI compared with the control arm of the study. Data from the SEER are not adequate to investigate the impact of PCI or other primary treatments on the cognitive function of patients with limited stage small cell lung cancer.
Theoretically, one would believe that prevention of brain metastasis, with its potential side effects, would be less morbid and life-threatening than delaying therapy until brain metastasis are identified. Unfortunately, the information regarding this in the literature is inadequate and inconsistent, mostly owing to the rapid progression of disease in most patients with small cell lung cancer. Furthermore, this analysis is unable to provide a corollary insofar as cognitive function, radiation dose, quality of life, and other parameters have not been included in the SEER. Data exist that support a dose-response relation with radiobiologically equivalent doses of 30 to 35 Gy with conventional fractionation (2 Gy) and Meert et al found a nonsignificant decrease in the rate of brain metastases with higher radiation doses.11, 30 Given this lack of robust data concerning the optimum dose and schedule, the authors encourage others to enroll patients on available protocols and await the final results of the Radiation Therapy Oncology Group (RTOG) 0212 trial.
In conclusion, we would recommend that PCI be considered in all patients with limited small cell lung cancer who experience a complete response to primary treatment. Further investigation is recommended to derive the best dosage schedules.
Conflict of Interest Disclosures
The authors made no disclosures.
- 14GreeneFL, PageDL, FlemingID, et al, eds. American Joint Committee on Cancer staging manual. 6th ed. New York: Springer-Verlag; 2002.
- 16Regression models and life tables. J R Stat Soc B. 1972; 34: 187-220..
- 18Breast cancer characteristics at diagnosis and survival among Arab-American women compared to European- and African-American women. Breast Cancer Res Treat. 2008; 15: 15., , , , , .
- 22Dose-escalating conformal thoracic radiation therapy with induction and concurrent carboplatin/paclitaxel in unresectable stage IIIA/B nonsmall cell lung carcinoma: a modified phase I/II trial. Cancer. 2001; 92: 1213-1223., , , et al.