Pediatric primary spinal cord tumors are exceedingly. Trends in the diagnosis and treatment of pediatric primary spinal cord tumors.

Transcription

1 J Neurosurg Pediatrics 10: , 2012 Trends in the diagnosis and treatment of pediatric primary spinal cord tumors Clinical article Melanie G. Hayden Gephart, M.D., M.A.S., 1 Robert M. Lober, M.D., Ph.D., 1 Robert T. Arrigo, M.S., 1 Corinna C. Zygourakis, M.D., 3 Raphael Guzman, M.D., 1 Maxwell Boakye, M.D., M.P.H., 4 Michael S. B. Edwards, M.D., 1 and Paul G. Fisher, M.D. 2 Departments of 1 Neurosurgery and 2 Neurology, Stanford University School of Medicine, Stanford; 3 Department of Neurosurgery, University of California, San Francisco, California; and 4 Center for Advanced Neurosurgery, University of Louisville, Kentucky Object. Pediatric primary spinal cord tumors (PSCTs) are rare, with limited comprehensive data regarding incidence and patterns of diagnosis and treatment. The authors evaluated trends in the diagnosis and treatment of PSCTs using a nationwide database. Methods. The Surveillance, Epidemiology, and End Results (SEER) registry was queried for the years , evaluating clinical patterns in 330 patients 19 years of age or younger in whom a pediatric PSCT had been diagnosed. Histological diagnoses were grouped into pilocytic astrocytoma, other low-grade astrocytoma, ependymoma, and high-grade glioma. Patient demographics, tumor pathology, use of external beam radiation (EBR), and overall survival were analyzed. Results. The incidence of pediatric PSCT was 0.09 case per 100,000 person-years and did not change over time. Males were more commonly affected than females (58% vs 42%, respectively; p < 0.006). Over the last 3 decades, the specific diagnoses of pilocytic astrocytoma and ependymoma increased, whereas the use of EBR decreased (60.6% from 1975 to 1989 vs 31.3% from 1990 to 2007; p < ). The 5- and 10-year survival rates did not differ between these time periods. Conclusions. While the incidence of pediatric PSCT has not changed over time, the pattern of pathological diagnoses has shifted, and pilocytic astrocytoma and ependymoma have been increasingly diagnosed. The use of EBR over time has declined. Relative survival of patients with low-grade PSCT has remained high regardless of the pathological diagnosis. ( Key Words outcome pediatric radiation spinal cord survival tumor oncology Pediatric primary spinal cord tumors are exceedingly rare. Among children and adults, published incidence rates have ranged from 0.09 to 0.26 case per 100,000 person-years. 11,22 Central nervous system tumors are much less common among children than adults, and in single-institution studies, PSCTs account for 6% 10% of CNS tumors in children. 19,21 Limited population-based data exist to describe incidence, pathological diagnosis, and survival for a large series of pediatric patients with PSCT; therefore, we used the SEER registry to evaluate incidence, diagnosis, and treatment patterns in pediatric patients with PSCT. Abbreviations used in this paper: CBTRUS = Central Brain Tumor Registry of the United States; EBR = external beam radiation; PSCT = primary spinal cord tumor; SEER = Surveillance, Epidemiology, and End Results. Methods Study Parameters Data regarding survival (date of diagnosis, duration of follow-up, and vital status at last contact), patient characteristics (age, race, and sex), tumor characteristics (grade and histology), and receipt of EBR were collected for each patient. Tumors were classified a priori by grade and histology into 1 of 4 groups (Table 1): pilocytic astrocytoma, other low-grade astrocytoma (nonpilocytic astrocytoma and nonependymoma), ependymoma, and high-grade glioma. Queried outcomes were predetermined prior to analysis and were not subsequently modified to ensure validity of the data and prevent systemic analysis bias. To account for the addition of registry areas over time (SEER 17) ( terms.html), changing trends in pathological diagnosis, 555

2 M. G. Hayden Gephart et al. TABLE 1: Primary PSCT classification* Tumor Grade Histology I II III IV Missing pilocytic astrocytoma PA PA HGG HGG PA ependymoma NOS EP EP EP EP EP anaplastic ependymoma EP EP EP EP EP myxopapillary ependymoma EP EP EP EP EP astrocytoma NOS PA LGA HGG HGG LGA anaplastic astrocytoma HGG HGG HGG fibrillary astrocytoma PA LGA HGG HGG LGA GBM NOS HGG HGG HGG mixed glioma PA LGA HGG HGG LGA malignant glioma PA LGA HGG HGG HGG oligodendroglioma NOS PA LGA HGG HGG LGA protoplasmic astrocytoma PA LGA LGA * Note that tumor grade was determined by individual neuropathologists and varies by tissue diagnosis. Abbreviations: EP = ependymoma; GBM = glioblastoma multiforme; HGG = high-grade glioma, nonependymoma; LGA = low-grade astrocytoma, nonependymoma; NOS = not otherwise specified; PA = pilocytic astrocytoma; = no cases. May mean unknown or known but data not entered. and the advent of MRI over the course of the study window, we also evaluated whether overall survival changed depending on the era of diagnosis (that is, vs ). Data Source The SEER registry, maintained by the National Cancer Institute and representing approximately 26% of the US population ( collects patient, disease, and treatment data for reportable tumors. We queried 2 data sets within the registry. First, we assessed incidence trends using SEER 9, which contains tumor registry data for 9 diverse areas across the US consistently recorded between 1975 and 2007 ( seer.cancer.gov/registries/terms.html). Second, we assessed survival outcomes using SEER 17, which includes the SEER 9 sites along with additional locations, which varied depending on the year. The SEER 9 was chosen for the age-adjusted incidence joinpoint analysis because that data set comprises a consistent set of registries over the entire study window, which allowed more accurate estimation of incidence. The SEER 17 was used for the remainder of the study to capture as many possible cases of rare pediatric PSCT, providing statistical power to the survival analyses. Thus, 2 separate cohorts were created: 1) the incidence cohort, which consisted of 214 cases (SEER 9), and 2) the survival cohort, which included 330 cases (SEER 17). Both of these cohorts excluded autopsy and death certificate only cases. For both cohorts, we identified microscopically confirmed tumor cases in all actively followed pediatric patients (age 19 years at diagnosis), using the International Classification of Diseases for Oncology, 3rd Edition (ICD-O-3) primary site codes for spinal meninges (C70.1), spinal cord (C72.1), or cauda equina (C72.1) crossed with ICD-O-3 histology codes for various primary glial tumors (ICD-O ). Twenty cases of primitive neuroectodermal tumor (ICD-O ) and 1 case each of chordoid glioma (ICD-O ) and astroblastoma (ICD-O ) were excluded from the study, as these diagnoses did not fit within the 4 a priori designated primary glial tumor categories (Table 1). Of note, the SEER program asks researchers not to publish tabulated data with cells containing less than 10 cases, and thus limiting the breadth of available subanalyses that could be performed. Statistical Analysis Analyses were performed using SEER*Stat (version 6.6.2, Surveillance Research Program, National Cancer Institute), Joinpoint (version 3.4.3, Statistical Research and Applications Branch, National Cancer Institute), and SAS (version 9.2, SAS Institute, Inc.). Incidence data were age-adjusted to the 2000 US standard population tables (Census P , age group 19 years) provided by SEER. Joinpoint analysis was performed to identify statistically significant changes in annual incidence over any window in the study. Survival analyses were based on relative survival models computed using the Kaplan-Meier method and adjusted to the US by Individual Year expected survival tables provided by SEER. Additional statistical comparisons were conducted using the chi-square test. All tests were 2-sided, and a p value < 0.05 was considered significant. Results Incidence and Characteristics of Pediatric PSCTs The incidence of pediatric PSCT was 0.09 case per 100,000 person-years in the first cohort of 214 cases. This age-adjusted incidence did not change over the 30-year study window (incidence of case from 1975 to 1989 vs case from 1990 to 2007; p value not significant). No significant joinpoints (that is, sharp inflections in incidence over several years) were found, again confirming that there was no overall or sudden change in the incidence of pediatric PSCT over the study period. In the survival cohort, 326 of the 330 tumor cases were located in the spinal cord (C72.1). Characteristics for this cohort are detailed in Table 2. Males accounted for significantly more cases of PSCT than females (58% vs 42%, respectively; p < 0.006), and the majority of tumors occurred in whites (83% vs 10.3% in blacks). Age at diagnosis was 0 4 years in 23.9% of patients, 5 9 in 20.6%, in 31.8%, and in 23.6%. Tumor diagnoses are also summarized in Table 2. Thirty percent of the pediatric PSCTs were classified as pilocytic astrocytoma, 21.2% as ependymoma, 27.3% as low-grade astrocytoma, and 21.2% as high-grade glioma. Twelve percent were classified as Grade I, 21.8% Grade II, 5.8% Grade III, and 13.6% Grade IV. (Note that tumor grade was determined by individual neuropathologists and varies by tissue diagnosis.) Tumor grade was not documented for 46.4% of lesions; however, for some diagnoses (for example, pilocytic astrocytoma) this may 556

3 Pediatric primary spinal cord tumors TABLE 2: Summary of patient characteristics and disease and treatment details along with 5-year relative survival Factor No. of Patients (%) Survival Rate (95% CI) overall 330 (100) 76.0 ( ) sex M 190 (57.6) 74.8 ( ) F 140 (42.4) 77.5 ( ) age at diagnosis in yrs (23.9) 86.5 ( ) (20.6) 85.6 ( ) (31.8) 67.6 ( ) (23.6) 67.9 ( ) race white 274 (83.0) 77.2 ( ) black 34 (10.3) 69.9 ( ) other including unknown 22 (6.7) 72.2 ( ) EBR yes 121 (36.7) 53.7 ( ) no 199 (60.3) 89.0 ( ) unknown 10 (3.0) 89.0 ( ) tumor grade I 41 (12.4) 91.9 ( ) II 72 (21.8) 93.9 ( ) III 19 (5.8) 41.5 ( ) IV 45 (13.6) 28.6 ( ) unknown 153 (46.4) 81.7 ( ) tumor type PA 100 (30.3) 93.1 ( ) EP 70 (21.2) 85.6 ( ) LGA 90 (27.3) 88.5 ( ) HGG 70 (21.2) 26.5 ( ) yr of diagnosis (22.1) 65.6 ( ) (77.9) 79.2 ( ) have been to avoid redundancy (as they are by definition low grade). TABLE 3: Age, radiotherapy, and tumor type by epoch of diagnosis in 330 patients with PSCT No. of Patients (%) Variable p Value* age at diagnosis in yrs (16.4) 67 (26.1) (24.6) 50 (19.4) (35.6) 79 (30.7) (23.3) 61 (23.7) EBR < yes 43 (60.6) 78 (31.3) no 28 (39.4) 171 (68.7) tumor type PA 10 (13.7) 90 (35.0) EP 12 (16.4) 58 (22.6) LGA 31 (42.5) 59 (23.0) HGG 20 (27.4) 50 (19.4) * Chi-square test. Excludes 10 cases with missing data. Changes in EBR Treatment and Tumor Type Over Time During the entire study period, one-third (36.7%) of the children with PSCT received EBR. The use of EBR significantly decreased when comparing the year of diagnosis: 31.3% between 1990 and 2007 and 60.6% between 1975 and 1989 (p < ; Table 3). The diagnosed tumor subtypes changed significantly during the study period. From 1975 to 1989, the most commonly diagnosed PSCT was low-grade astrocytoma (42.5%), excluding pilocytic astrocytoma; only 13.7% of diagnosed tumors were pilocytic astrocytoma. In contrast, in the cohort, the most commonly diagnosed PSCT was pilocytic astrocytoma (35%), and only 23% of diagnosed tumors were other low-grade astrocytomas (Table 3). Age at tumor diagnosis did not change significantly over the entire study period. Survival Rates for Pediatric PSCTs Finally, we examined the survival rates for the identified patient cohort with pediatric PSCT. Table 2 details the 5-year relative survival according to predefined factors of patient characteristics, EBR, and pathological diagnosis. For the entire cohort, overall survival at 5 years was 76%. Note that survival did not differ based on age, sex, or race. Neither did survival change between the and epochs. Note, however, that significantly worse outcomes were associated with the receipt of EBR (53.7% 5-year relative survival with EBR vs 89.0% without EBR), tumor Grade III or IV (41.5% and 28.6% survival for Grades III and IV, respectively, vs 91.9% and 93.9% survival for Grades I and II, respectively), and diagnosis of high-grade glioma (26.5% vs 85.6% 93.1% survival for the other 3 tumor types; Table 2). There was no difference in survival between patients with pilocytic astrocytoma, low-grade astrocytoma, and ependymoma following the use of EBR (Table 4). Discussion Incidence of Pediatric PSCTs We report an overall pediatric PSCT incidence of 0.09 case per 100,000 person-years in our study, which is slightly lower than the previously published 0.26 case per 100,000 person-years from the CBTRUS data 22 and is consistent with the incidence of PSCT in the adult population using the SEER database. 11 The SEER database offers several advantages over the CBTRUS by covering a longer time period and providing follow-up and survival information. 7 Furthermore, the SEER coverage of more 557

4 M. G. Hayden Gephart et al. TABLE 4: Five-year relative survival by tumor type and use of EBR among 320 patients with PSCT* Tumor Type No. of Patients (%) Survival Rate (95% CI) PA EBR 11 (11.3) 74.1 ( ) no EBR 86 (88.7) 94.7 ( ) EP EBR 28 (43.1) 80.6 ( ) no EBR 37 (56.9) 91.0 ( ) LGA EBR 30 (33.7) 83.4 ( ) no EBR 59 (66.3) 91.0 ( ) HGG EBR 52 (75.4) 17.3 ( ) no EBR 17 (24.6) 49.5 ( ) * Ten of 330 patients had missing data. diverse patient populations in the US leads to greater generalizability. There was no difference between the incidence of PSCT from 1975 to 1989 versus 1990 to 2007 (0.078 vs 0.101, p value not significant), suggesting that the advent of MRI and its application in diagnosing spinal cord tumors beginning in the early 1990s did not change the age-adjusted incidence. Tumors were most commonly found in patients between the ages of 10 and 14 years, which was potentially associated with the onset of puberty when scoliosis often becomes evident. The absence of an increased PSCT incidence and no change in the age at diagnosis suggested that diagnosis remains largely clinical and dependent on symptoms. Interestingly, we did find that pediatric PSCTs are more common in male than female children (Table 2), which is consistent with prior studies. 11,22,24 While medulloblastoma 6 and ependymoma 15 have been more common among male children, other glial-based tumors have not been associated with patient sex. Our finding of an increased incidence of PSCT in whites is also consistent with prior studies. 11 This difference in sex and race may have reflected a true biological difference, as the patient sample from the SEER registry covers centers representative of the racial diversity in the US. Differences in the proportions of particular pathological diagnoses were noted over the study periods ( vs ). In the more recent time period, there was an increase in the diagnosis of pilocytic astrocytoma and ependymoma with a concomitant decrease in the diagnosis of other low-grade astrocytoma and high-grade glioma. The combined incidence of the tumors did not change over the 2 study epochs, and this probably reflects variations in operative sampling techniques or changes in histological classification criteria or precision 3 rather than an actual epidemiological shift in the tumor types present in the US population. Survival Rates for Pediatric Patients With PSCTs No association was found between sex, race, or year of diagnosis and overall survival. As expected, our analysis showed a significant effect of tumor type and tumor grade on survival, 18,23 with a very high 5-year survival for low-grade PSCT (85.6% 93%), compared with a very poor outcome for high-grade glioma (26.5%). We also found lower survival rates in patients with tumor Grades III and IV, as compared with those with Grade I and II tumors (41.5% and 28.6% vs 91.9% and 93.9%, respectively). The poor survival associated with pediatric spinal cord high-grade glioma was similar to that reported for children with supratentorial glioblastoma, 25 consistent with prior studies showing that tumor histology and grade are significant predictors of outcome in spinal cord tumors. 10,19 Use of Radiotherapy A striking finding was the decreased use of EBR over time. Given the morbidity associated with radiation of the developing spine and nervous system, 2,4,8,12,16,20 practice patterns have trended away from EBR, particularly when the benefit is unclear. 1,13,14,17 Likewise, new neurosurgical techniques may have allowed an increased extent of resection 5 9 and thus a decreased use of adjuvant radiotherapy. Note, however, that the independent impact of EBR could not be adequately addressed with our methods, as it is unclear from available data whether patients receiving EBR were more likely to have higher-grade tumors. Study Limitations and Future Directions We believe our analysis yields data relevant to the care and treatment of pediatric patients with PSCTs, although we acknowledge several limitations. Incomplete data incorporation into the database may cause apparent incongruities. For example, 30.3% of tumors were classified as pilocytic astrocytoma, almost all of which were Grade I, whereas the percentage of reported Grade I tumors was only 12.4% (Table 2). Given that almost 50% of the tumor types were documented as unknown grade, our analysis is limited by the data completeness, accuracy, and input variability of the SEER registry, and there may be regional reporting and definitional differences as well. Tumors were classified a priori, and queried outcomes were determined prior to analysis without subsequent modification to prevent analysis bias. Unfortunately, the extent of resection, EBR or other adjuvant treatment parameters, and tumor size at the time of diagnosis were not reliably available through SEER. Because of the retrospective nature of our study, relationships between clinical features and survival may be described but cannot demonstrate causality. Our hypothesis-generating analysis lays the foundation for further prospective research and clinical evaluation of pediatric patients with PSCTs, particularly regarding the use of EBR and the impact of the extent of resection. Conclusions In summary, we report the incidence, diagnoses, use of EBR, and survival in pediatric PSCTs using the large, population-based SEER database. While the incidence of 558

5 Pediatric primary spinal cord tumors pediatric PSCT did not change over time, the pathological diagnoses of the lesions have shifted significantly over the past 30 years, with a greater percentage of pilocytic astrocytoma and ependymoma diagnosed. We also noted a decrease in the use of EBR over time. These data may be particularly useful in discussions with families of patients who have PSCT. Disclosure Dr. Fisher serves on the board of directors of the National Brain Tumor Society. Dr. Hayden Gephart is supported by a grant from the California Institute of Regenerative Medicine. This work was supported in part by the Center for Children s Brain Tumors and Lucile Packard Children s Hospital. Author contributions to the study and manuscript preparation include the following. Conception and design: Hayden Gephart, Guzman, Edwards, Boakye, Fisher. Acquisition of data: Hayden Gephart, Arrigo, Zygourakis, Boakye. Analysis and interpretation of data: all authors. Drafting the article: Hayden Gephart, Lober, Arrigo, Zygourakis, Guzman, Edwards, Fisher. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Hayden Gephart. Statistical analysis: Hayden Gephart, Boakye, Fisher. Administrative/technical/material support: Boakye. Study supervision: Hayden Gephart, Fisher. References 1. Al-Halabi H, Montes JL, Atkinson J, Farmer JP, Freeman CR: Adjuvant radiotherapy in the treatment of pediatric myxopapillary ependymomas. Pediatr Blood Cancer 55: , Allen JC, Miller DC, Budzilovich GN, Epstein FJ: Brain and spinal cord hemorrhage in long-term survivors of malignant pediatric brain tumors: a possible late effect of therapy. Neurology 41: , Becker LE: An appraisal of the World Health Organization classification of tumors of the central nervous system. Cancer 56 (7 Suppl): , Broniscer A, Ke W, Fuller CE, Wu J, Gajjar A, Kun LE: Second neoplasms in pediatric patients with primary central nervous system tumors: the St. Jude Children s Research Hospital experience. Cancer 100: , Crawford JR, Zaninovic A, Santi M, Rushing EJ, Olsen CH, Keating RF, et al: Primary spinal cord tumors of childhood: effects of clinical presentation, radiographic features, and pathology on survival. J Neurooncol 95: , Curran EK, Sainani KL, Le GM, Propp JM, Fisher PG: Gender affects survival for medulloblastoma only in older children and adults: a study from the Surveillance Epidemiology and End Results Registry. Pediatr Blood Cancer 52:60 64, Davis FG, McCarthy BJ, Berger MS: Centralized databases available for describing primary brain tumor incidence, survival, and treatment: Central Brain Tumor Registry of the United States; Surveillance, Epidemiology, and End Results; and National Cancer Data Base. Neuro Oncol 1: , de Jonge T, Slullitel H, Dubousset J, Miladi L, Wicart P, Illés T: Late-onset spinal deformities in children treated by laminectomy and radiation therapy for malignant tumours. Eur Spine J 14: , Epstein FJ, Farmer JP: Pediatric spinal cord tumor surgery. Neurosurg Clin N Am 1: , Garcés-Ambrossi GL, McGirt MJ, Mehta VA, Sciubba DM, Witham TF, Bydon A, et al: Factors associated with progression-free survival and long-term neurological outcome after resection of intramedullary spinal cord tumors: analysis of 101 consecutive cases. Clinical article. J Neurosurg Spine 11: , Hsu S, Quattrone M, Ostrom Q, Ryken TC, Sloan AE, Barnholtz-Sloan JS: Incidence patterns for primary malignant spinal cord gliomas: a Surveillance, Epidemiology, and End Results study. Clinical article. J Neurosurg Spine 14: , Isaacson SR: Radiation therapy and the management of intramedullary spinal cord tumors. J Neurooncol 47: , Lober R, Sharma S, Bell B, Free A, Figueroa R, Sheils CW, et al: Pediatric primary intramedullary spinal cord glioblastoma. Rare Tumors 2:e48, Lonjon M, Goh KY, Epstein FJ: Intramedullary spinal cord ependymomas in children: treatment, results and follow-up. Pediatr Neurosurg 29: , McGuire CS, Sainani KL, Fisher PG: Incidence patterns for ependymoma: a Surveillance, Epidemiology, and End Results study. Clinical article. J Neurosurg 110: , Meadows AT, Gallagher JA, Bunin GR: Late effects of early childhood cancer therapy. Br J Cancer Suppl 18:S92 S95, Merchant TE, Haida T, Wang MH, Finlay JL, Leibel SA: Anaplastic ependymoma: treatment of pediatric patients with or without craniospinal radiation therapy. J Neurosurg 86: , Merchant TE, Nguyen D, Thompson SJ, Reardon DA, Kun LE, Sanford RA: High-grade pediatric spinal cord tumors. Pediatr Neurosurg 30:1 5, Milano MT, Johnson MD, Sul J, Mohile NA, Korones DN, Okunieff P, et al: Primary spinal cord glioma: a Surveillance, Epidemiology, and End Results database study. J Neurooncol 98:83 92, Poretti A, Zehnder D, Boltshauser E, Grotzer MA: Long-term complications and quality of life in children with intraspinal tumors. Pediatr Blood Cancer 50: , Rosemberg S, Fujiwara D: Epidemiology of pediatric tumors of the nervous system according to the WHO 2000 classification: a report of 1,195 cases from a single institution. Childs Nerv Syst 21: , Schellinger KA, Propp JM, Villano JL, McCarthy BJ: Descriptive epidemiology of primary spinal cord tumors. J Neurooncol 87: , Tendulkar RD, Pai Panandiker AS, Wu S, Kun LE, Broniscer A, Sanford RA, et al: Irradiation of pediatric high-grade spinal cord tumors. Int J Radiat Oncol Biol Phys 78: , Wilson PE, Oleszek JL, Clayton GH: Pediatric spinal cord tumors and masses. J Spinal Cord Med 30 (Suppl 1):S15 S20, Wolff JE, Driever PH, Erdlenbruch B, Kortmann RD, Rutkowski S, Pietsch T, et al: Intensive chemotherapy improves survival in pediatric high-grade glioma after gross total resection: results of the HIT-GBM-C protocol. Cancer 116: , 2010 Manuscript submitted February 4, Accepted September 4, Please include this information when citing this paper: published online October 12, 2012; DOI: / PEDS1272. Address correspondence to: Melanie G. Hayden Gephart, M.D., M.A.S., Department of Neurosurgery, Stanford University School of Medicine, 300 Pasteur Drive, MC 5327, Stanford, California mghayden@gmail.com. 559