Intracranial ependymomas: analysis of prognostic factors in a population-based series

Prognosis-related histomorphological and immunohistochemical markers in central nervous system tumors of childhood and adolescence

Brain tumors account for approximately 20% of all childhood cancers, and are the leading cause of cancer morbidity and mortality among children. Although numerous demographic, clinical and therapeutic parameters have been identified over the past few years that have significant prognostic bearing for some pediatric brain tumors, predicting the clinical course and outcome among children with central nervous system tumors is still difficult. A survey of publications on prognosis-related histopathological and immunohistochemical features among pediatric brain tumors revealed 172 series, of which 91 presented statistically significant outcome-associated parameters as defined by a P value of less than 0.05. Most investigations revealing significant prognosis-related markers were performed on medulloblastomas (30 publications), ependymomas (25) and astrocytic tumors (18). In total, 16 cohorts consisted of more than 100 cases (5 on ependymomas, 3 each on medulloblastomas and astrocytic tumors). On the other hand, there were also 13 series with fewer than 20 cases (5 on medulloblastomas). Potentially prognostic histopathological markers vary among different entities and consist of assessment of necroses, mitoses, differentiation, vascular proliferation, and growth pattern, whereas immunohistochemical features include proliferation markers (Ki-67, MIB-1), expression of oncogenes/tumor suppressor genes and their proteins (TP53, c-erbB2), growth factor and hormonal receptors (VEGF, EGFR, HER2, HER4, ErbB-2), cell cycle genes (p27, p14ARF) and cell adhesion molecules, as well as factors potentially related to therapeutic resistance (DNA topoisomerase IIalpha, metallothionein, P-glycoprotein, tenascin). This review discusses the prognostic potential of histopathological and immunohistochemical markers that can be investigated by the practicing neuropathologist as part of the routine diagnostic workload, and scrutinizes their benefit for predicting therapy response and patient outcome among children with brain tumors.

Childhood ependymoma: a systematic review of treatment options and strategies

Childhood intracranial ependymoma have a dismal prognosis, especially in young children and when a gross total resection cannot be performed. Even in the absence of a radiologically proven residuum, around two-thirds of these young children will have a recurrence. Adjuvant therapy is therefore necessary for most, if not all, patients. Despite some indication that benign ependymoma (WHO grade II) could show a better outcome, histology cannot be used at present to stratify treatment protocols.Craniospinal irradiation combined with posterior fossa boost has deleterious adverse effects on cognition. Consequently, pediatric oncology teams have, firstly, tried to use chemotherapy to delay or avoid irradiation, and secondly, progressively reduced irradiation fields to the tumor bed without altering the prognosis. Cisplatin, at a dose of 120 mg/m(2) (cumulated response rate of 34% [95% CI 19-54%]) is the only single agent that has reproducibly shown some efficacy in ependymoma. Despite some combinations showing efficacy in the adjuvant setting, childhood intracranial ependymomas can, in general, be considered as chemoresistant. The overexpression of the multidrug resistance-1 gene and the 06-methylguanine-DNA methyltransferase have been implicated as possible mechanisms for this phenomenon. As the use of chemotherapy with current agents is questionable, phase II studies with new agents and combinations are necessary. Since the main problem of this disease is local relapse, it may not be necessary to irradiate the whole posterior fossa. However, local control of the disease by irradiation has to be improved. In this respect, hyperfractionation or radiosensitizers may be valuable therapeutic options. The treatment of children with ependymoma is a challenge for all caregivers. There is no doubt that any possible improvement in the management of this rare tumor will only be the result of well designed cooperative trials.

Ependymoma in childhood: prognostic factors, extent of surgery, and adjuvant therapy

OBJECT

The aim of this study was to investigate the effect of patient-related factors, extent of surgery, and adjuvant therapy on survival in children presenting with intracranial ependymoma.

METHODS

Between 1980 and 1999, 83 children (mean age 36 months) underwent surgery for intracranial ependymomas. Complete resection, verified on postoperative computerized tomography scans, was achieved in 73%. Adjuvant therapy modalities have changed over the years: before 1990 all patients received radiotherapy, whereas after 1990 the children younger than 3 years of age and later those younger than 5 years of age were treated first with chemotherapy and received radiotherapy only after their first tumor recurrence. The follow-up period averaged 70 +/- 49 months.

CONCLUSIONS

Overall survival, intraoperative deaths excluded, was 73 +/- 11% and 51 +/- 14% at 5 and 10 years, respectively. The event-free survival rate at 5 and 10 years was 48 +/- 12% and 46 +/- 12%, respectively. Most of the events were local recurrences. Despite multiple reinterventions, the overall survival rate in this group dropped to 14%. On univariate analysis, the only significant prognostic factors were complete resection and radiotherapy. Both of these factors combined increased the 5- and 10-year survival rates to 93 and 75%, respectively. Age of the patients was not a statistically independent prognostic factor. The patients in the chemotherapy group did not fare as well as those in the radiotherapy group. A subgroup (36%) within the chemotherapy group, however, survived tumor free after a mean follow-up period of 67 months. It is not clear whether this subgroup either responded well to chemotherapy or needed no adjuvant therapy. Further research is warranted to answer this question.

Intracranial ependymoma

OBJECT

An intracranial ependymoma is a relatively rare but very interesting variety of glioma. In this paper, the authors compiled a review of the pathological features, imaging characteristics, and treatment strategies related to this brain tumor.

METHODS

A Medline search was conducted using the term "ependymoma." The bibliographies of papers obtained were also checked for articles and chapters that could provide additional understanding of this tumor. Malignant ependymomas and ependymomas of the spinal cord (including myxopapillary ependymomas) were excluded from this review.

CONCLUSIONS

The posterior fossa is the most frequent site for an intracranial ependymoma. Children are frequently affected. Most authors recommend resecting as much of the tumor as is safely possible. Microscopically, ependymal tumors show both epithelial and glial features. Glial fibrillary acidic protein immunohistochemistry, therefore, helps in identifying ependymomas. Because ependymomas often recur despite surgical intervention, radiotherapy and/or radiosurgery may also play an important role in their treatment. The use of chemotherapy in the treatment of these tumors, especially in the very young, is still being studied.

Immunohistochemical markers for prognosis of ependymal neoplasms

Intracranial ependymomas are the third most common primary brain tumor in children. Although clinical and histological criteria for ependymoma prognosis are recognized, studies have reported contradictory results. Prognostic significance based on immunohistochemistry of ependymomas has been reported in a few studies. One-hundred and twelve patients with intracranial ependymomas were examined retrospectively for immunoexpression of various tumor-associated antigens and apoptosis. The results demonstrated significant preponderance of expression of the tenascin, vascular endothelial growth factor protein (VEGF), epidermal growth factor (EGFR) and p53 protein in high-grade tumors. Also high-grade ependymomas revealed more prominent labeling indices (LI) for proliferative marker Ki-S1 and apoptotic index (AI), and lower LI for cyclin-dependent kinase inhibitors p27/Kipl and pl4ARF. For low-grade ependymomas the progression-free survival time (PFS) was found to be significantly shorter for Ki-S1 LI > 5%, and for tenascin, VEGF and EGFR positivity. For high-grade ependymomas PFS was found to be significantly reduced for p27 LI < 20%, p14ARF LI < 10%, for p53 positivity, and for AI < 1%. The CART modeling process exhibited five final groups of ependymoma patients (1) low-grade and tenascin-negative; (2) low-grade and tenascin-positive; (3) high-grade and p53-negative with p14 LI > 0%; (4) high-grade with combination of either p53 positivity and p14 LI > 10% or p53 negativity and p14 LI < 10%; (5) high-grade and p53-positive with pl4 LI < 10%. In summary, some immunohistochemical variables were found to be the strong predictors of ependymoma recurrence and they seem to be useful for assessing individual tumor prognosis in routinely processed biopsy specimens together with tumor grade. For histologically benign ependymomas immunohistochemical study should be focused on Ki-S1, tenascin, EGFR and VEGF evaluation, whereas p53 expression and number of p27, p14 and ISEL-positive nuclei will be of value in determining PFS from high-grade ependymomas.

Influence of tumor grade on time to progression after irradiation for localized ependymoma in children

PURPOSE

To investigate the influence of histologic grade on progression-free survival (PFS) after irradiation (RT) for pediatric patients with localized ependymoma.

METHODS AND MATERIALS

Fifty patients with localized ependymoma (median age 3.6 years, range 1-18 years at the time of RT) were treated with RT between December 1982 and June 1999. Anaplastic features were identified in 14 of 50 patients. The extent of resection was characterized as gross-total in 36 patients, near-total in 5, and subtotal in 9. The median dose to the primary site was 54 Gy. Of the 50 patients, 23 received pre-RT chemotherapy.

RESULTS

Thirty-nine patients were alive at a median follow-up of 46 months (range 21-214) from diagnosis. Thirty-four patients remained progression free at a median follow-up of 35 months (range 13-183) after the initiation of RT. Progression occurred in 16 patients (12 local and 4 local and distant), with a median time to failure of 21.2 months (range 4.6-65.0). The tumor grade significantly influenced the PFS after RT (p < 0.0005). The estimated 3-year PFS rate was 28% +/- 14% for patients with anaplastic ependymoma compared with 84% +/- 8% for patients with differentiated ependymoma. These results remained significant when corrected for age at diagnosis (<3 years), pre-RT chemotherapy, and extent of resection. Patients who received pre-RT chemotherapy had an inferior 3-year PFS estimate after RT (49 +/- 12%) compared with those who did not (84% +/- 10%; p = 0.056). Anaplastic ependymoma was found more frequently in the supratentorial brain (p = 0.002). Six of 12 patients with supratentorial tumor developed recurrence; recurrence was restricted to patients with anaplastic ependymoma.

CONCLUSION

Tumor grade influences outcome for patients with ependymoma independent of other factors and should be considered in the design and analysis of prospective trials involving pediatric patients treated with RT. Chemotherapy before RT influences the PFS and overall survival after RT. The effect is more pronounced when progression occurs during chemotherapy.

Postoperative radiotherapy for intracranial ependymoma: analysis of prognostic factors and patterns of failure

The long-term results of external beam radiotherapy following surgical resection in patients with intracranial ependymomas were evaluated to identify the prognostic factors and the pattern of recurrence. Between June 1961 and January 1999, 48 patients with intracranial ependymoma were treated with external beam irradiation with >40 Gy following surgery. Total doses of 40.5-63.4Gy were delivered to the tumor site in 22-46 fractions over 33-101 days. Six patients with spinal deposit or positive cerebrospinal fluid cytology received whole spinal axis irradiation, and 4 patients received prophylactic spinal irradiation. The median follow-up of surviving patients was 110 months. The 10-year overall and relapse-free survival rates were 47% and 42%, respectively. In multivariate analysis, female gender, lower tumor grade and total resection were found to be associated with better relapse-free survival. Twenty of 26 recurrences developed at the primary tumor site (inside the irradiation field), two at the unirradiated cerebellum and spinal cord, and four at the spinal cord without intracranial failure. Only one of 34 patients with supratentorial tumors developed isolated spinal metastasis, whereas 3 of 14 patients with infratentorial tumors did so. Regarding the late neurotoxicity of radiotherapy, one of the 15 long-term (>4 years) survivors whose psychosocial status could be evaluated showed marked cognitive impairment. It was suggested that the use of new treatment strategies to improve local control would be warranted, and that prophylactic whole spinal axis irradiation appeared to be of more benefit in patients with infratentorial tumors than in those with supratentorial tumors.

Antineuronal nuclei immunohistochemical staining patterns in childhood ependymomas

NeuN, the mouse-derived monoclonal antibody to the reportedly neuron-specific nuclear protein, has been observed to react with many different types of normal, postmitotic neurons throughout the central and peripheral nervous systems. We retrospectively examined 23 surgical specimens (collected from 20 patients) originally diagnosed at our institution between 1983 and 1999 as ependymoma (9), myxopapillary ependymoma (1), anaplastic/malignant ependymoma (10), and primitive neuroectodermal tumor with ependymal differentiation (3). The ependymomas included lesions from the spine (3), cerebrum (5), and posterior fossa (15). Representative formalin-fixed, paraffin-embedded sections from each tumor were subjected to immunohistochemical staining with antibody against NeuN (Chemicon International, Inc, Temecula, CA). Five astrocytomas, four primitive neuroectodermal tumors, and normal cerebral cortex and ependyma from autopsy brains of premature newborns, term infants, and older children served as controls. Thirteen ependymal tumors had positive nuclear staining ranging from rare tumor cells to numerous groups of cells; of these, 9 were anaplastic ependymomas and had the most staining. These studies suggest that some ependymomas arise from a pluripotential neuroglial cell.

A clinicopathologic study of 81 patients with ependymomas and proposal of diagnostic criteria for anaplastic ependymoma

Optimal histologic criteria for the classification of and grading of ependymomas, including their anaplastic forms, remain elusive. This is especially true because of the poor correlation of these criteria with clinical outcome. The aim of this study was to identify the histopathologic parameters that could distinguish different prognostic groups of patients with ependymomas. Eighty-one patients with ependymal tumors, including those originally diagnosed ependymomas, anaplastic ependymomas and myxopapillary ependymomas, were enrolled in this study. Thirteen histologic parameters, including hypercellularity, nuclear pleomorphism, mitoses, endothelial proliferation, necrosis, clear cell, thrombi, dystrophic calcification, psammoma bodies, bone, cartilage, Rosenthal fibers and MIB-1 labeling index (LI), were evaluated in each patient and correlated with clinical outcome. We assigned one score for each histopathologic parameter evaluated and used a stepwise selection method with entry model based on the significance of the log-rank statistic to formulate a scoring model. Four parameters were chosen in this process, including mitoses > or = 4/10 hpf (1.7/mm2), hypercellularity, endothelial proliferation and necrosis. The sum of these four parameters (scores) was the histopathologic score of the tumor. The progression-free survival (PFS) and overall survival (OS) of patients with histopathologic scores 0 and 1 were significantly better than those with histopathologic scores 2, 3 and 4 (p < 0.001 and p = 0.005, respectively). Because of the latter finding, we proposed that anaplastic ependymoma could be diagnosed by the presence of any two of the aforementioned four parameters. Multivariate analyses including clinical and histopathologic variables showed that histopathologic score > or = 2 and subtotal resection were the factors related to increased risk of recurrence, while histopathologic score > or = 2 was the only factor related to overall survival. Based on the above findings, we concluded that histopathology is an important prognostic indicator for patients with ependymomas.

Prognostic factors in intracranial ependymomas in children

OBJECT

The occurrence of intracranial ependymomas in children is relatively infrequent, and their prognostic factors are still controversial, especially regarding histological composition.

METHODS

A retrospective study was conducted of 37 children treated during the last 20 years for intracranial ependymomas at the Hôpital de la Timone. Both univariate and multivariate statistical analyses were performed to assess the prognostic relevance of patient age and sex, extent of tumor removal, location of the tumor (supratentorial compared with infratentorial, median compared with lateral), tumor histological composition, and adjuvant therapies in affecting the 5-year progression-free survival (PFS) rate and overall survival (OS) rate. The following histopathological features, either alone or in combination, were analyzed: endothelial proliferation, necrosis, loss of differentiating structures (present compared with absent), the number of mitotic figures per 10 hpf, and cellularity (number of nuclei/5 hpf). In addition, immunohistochemical detection of Ki-67 antigen was performed and the Ki-67 labeling index (LI) evaluated in all cases. The 5-year OS and PFS rates were 45% and 25%, respectively (median follow up 34 months). Four patients died of disease without remission (median 163 days) and disease in 21 patients relapsed: 18 in situ and three both in situ and distantly. On univariate analysis total surgical resection and median infratentorial location were associated with a better outcome (p < 0.002) for both OS and PFS. Loss of differentiating structures was associated with poor prognosis (p < 0.008) and the combination of necrosis, endothelial proliferation, and mitotic index greater than 5 was also a negative predictive factor for both OS (p < 0.002) and PFS (p = 0.02). The PFS time was shorter in patients younger than 4 years of age and in patients in whom a Ki-67 LI greater than 1 was found (p = 0.03 and 0.006, respectively). Adjuvant radiotherapy and chemotherapy were not relevant to prognosis. Moreover, among the 15 patients in whom total excision was achieved, OS was better in those who did not receive adjuvant therapies. In contrast, adjuvant therapies significantly enhanced PFS time in patients in whom tumor excision was incomplete.

CONCLUSIONS

This study and analysis of the literature further highlight that total tumor removal is the treatment of choice for ependymomas in children. Postoperative measurement of residual tumor is required, especially because a subgroup of patients might be treated by surgery alone. Median infratentorial ependymomas have to be distinguished from the lateral type. Appropriate and reproducible histological parameters and Ki-67 LI are of interest as predictors of outcome.

Immunohistochemical markers for intracranial ependymoma recurrence. An analysis of 88 cases

Intracranial ependymomas are the third most common primary brain tumor in children. Although clinical and histological criteria for ependymoma prognosis are recognized, studies have reported contradictory results. Prognostic significance based on immunohistochemistry of ependymomas has been reported in a few studies. Eighty-eight patients with intracranial ependymomas were examined retrospectively for immunoexpression of various tumor-associated antigens and apoptosis. The results demonstrated significant preponderance of expression of the tenascin, vascular endothelial growth factor protein (VEGF), epidermal growth factor (EGFR), and p53 protein in high-grade tumors. Also high-grade ependymomas revealed more prominent labeling indices (LI) for proliferative marker Ki-S1 and lower LI for cyclin-dependent kinase inhibitor p27/Kip1. For low-grade ependymomas the progression free survival time (PFS) was found to be significantly shorter for Ki-S1 LI>/=5%, and for tenascin, VEGF, and EGFR positivity. For high-grade ependymomas PFS was found to be significantly reduced for age <16 years, subtotal tumor removal, p27 LI <20%, p53 positivity, and for apoptotic index (AI) <1%. The classification regression tree analysis exhibited four groups of ependymomas; (1) low-grade tenascin negative (32 cases, recurrence rate=0), (2) high-grade with AI >/=1% (21 cases, recurrence rate=57%), (3) low-grade tenascin-positive (10 cases, recurrence rate=89%), and (4) high-grade with AI <1% (25 cases, recurrence rate=100%). So, the immunohistochemical variables were found to be strongest predictors of ependymoma recurrence and they seem to be useful for assessing individual tumor prognosis in routinely processed biopsy specimen.

Prognostic value of immunoexpression of the chemoresistance-related proteins in ependymomas: an analysis of 76 cases

Intracranial ependymomas are the third most common primary brain tumor in children. A variety of chemotherapy protocols have been introduced for the treatment of ependymoma although overall these have not contributed to patients outcome. To our knowledge, data on the prognostic value of immunoexpression of the chemoresistance-related proteins (ChRPs) in ependymomas are absent. Seventy-six patients with intracranial ependymomas who received combined treatment were studied retrospectively. Tumor specimens were immunohistochemically examined with antibodies to metallothioneins (MT), glutathione S-transferase pi (GST pi) and P-glycoprotein (P-GP). The results demonstrated significant preponderance of expression of all the above-mentioned ChRPs for the low-grade tumors. The progression-free survival time was found to be significantly shorter for immunonegative tumors in both tumor grades. Multivariate analysis using a Cox hazard model revealed that recurrence-free survival time is significantly associated with tumor grade, and MT and P-GP expression. Risk of recurrence increased for the high-grade ependymomas (hazard ratio 2.85; P = 0.004), and decreased for the MT-positive tumors (hazard ratio -2.72; P = 0.005) and for the P-GP-positive tumors (hazard ratio -2.02; P = 0.02). The obtained results allow one to conclude that ChRPs expression is closely associated with low-grade ependymomas and immunohistochemical findings may be estimated as a predictor for local tumor progression.

The immunophenotype of ependymomas

The morphologic distinction of ependymomas with epithelial cytology from metastatic carcinoma may pose a significant problem in differential diagnosis. The known presence of keratin in glioma cells further complicates the issue. Using the labeled streptavidin-biotin method with automated staining, we studied epithelial and glial marker expression in 52 ependymomas of varying type and grade, including 20 epithelial-appearing, 14 glial-appearing, eight mixed pattern, and 10 myxopapillary tumors; 38 were low grade and 14 anaplastic. All tumors were immunoreactive for glial fibrillary acidic protein (GFAP), and S-100 protein. Diffuse staining for GFAP was noted in glial-appearing ependymomas featuring perivascular pseudorosettes. Diffuse immunostaining for S-100 protein was seen in cellular lesions exhibiting epithelial-like features. Staining was more diffuse for GFAP than S-100 protein in anaplastic ependymomas. Keratin (AE1/AE3) reactivity was seen in 98% of cases, the pattern being similar to that of GFAP. The frequency of staining for other keratins varied: wide-spectrum keratin (35%), cytokeratin (CK)7 (20%), CAM 5.2 (19%), CK903 (14%), and CK20 (8%); as a rule, it was scant and limited to occasional cells and processes. epithelial membrane antigen (EMA) staining was seen in 36% of all cases and in 67% of epithelial-appearing tumors wherein it often high-lighted microlumina. Aside from AE1/AE3 staining and very infrequent wide-spectrum keratin and EMA reactivity, expression of epithelial markers was not seen in anaplastic ependymomas. No carcinoembryonic antigen (CEA) positivity was noted in any case. Collagen IV reactivity was limited to tumor cell-stroma interfaces. Although variable, S-100 protein and GFAP staining is seen in all ependymomas, particularly in true and perivascular pseudorosettes. Widespread reactivity for keratin AE1/AE3 corresponds closely to the pattern of GFAP staining. Significant staining for other keratins or for CEA is inconsistent with a diagnosis of ependymoma. EMA reactivity is largely limited to luminal staining of rosettes and tubules.

Combined postoperative irradiation and chemotherapy for anaplastic ependymomas in childhood: results of the German prospective trials HIT 88/89 and HIT 91

PURPOSE

To evaluate the outcome in children with anaplastic ependymomas after surgery, irradiation, and chemotherapy; and to identify prognostic factors for survival.

METHODS AND MATERIALS

Fifty-five children (n = 27 girls, 28 boys; median age at diagnosis, 6.2 years) with newly diagnosed anaplastic ependymomas were treated in the multicenter, prospective trials HIT 88/89 and HIT 91. Macroscopic complete resection was achieved in 28 patients; 27 patients underwent incomplete resection. All patients received chemotherapy before (n = 40) or after irradiation (n = 15). The irradiation volume encompassed either the neuraxis followed by a boost to the primary tumor site (n = 40) or the tumor region only (n = 13). No radiotherapy was administered in two patients.

RESULTS

Median follow-up was 38 months. The overall survival rate at 3 years after surgery was 75.6%. Disease progression occurred in 25 children with local progression occurring in 20. The median time to disease progression was 45 months. The only significant prognostic factor was the extent of resection (estimated progression-free survival [EPFS] after 3 years was 83.3% after complete resection and 38.5% after incomplete resection) and the presence of metastases at the time of diagnosis (0% vs. 65.8% 3-year EPFS in localized tumors). Age, sex, tumor site, mode of chemotherapy, and irradiation volume did not influence survival.

CONCLUSIONS

Treatment centers should be meticulous about surgery and diagnostic workup. Because the primary tumor region is the predominant site of failure it is important to intensify local treatment. Dose escalation by hyperfractionation or stereotactic radiotherapy might be a promising approach in macroscopically residual disease. The role of adjuvant chemotherapy requires further study.

A multi-institutional retrospective study of intracranial ependymoma in children: identification of risk factors

PURPOSE

The goal of this multi-institutional retrospective study of children with intracranial ependymoma was to identify risk factors associated with unfavorable overall survival (OS) and event-free survival (EFS).

PATIENTS AND METHODS

Clinical data, including demographics, tumor location, spread, histology, details of surgery, radiation treatment, and chemotherapy were collected. Clinical characteristics and univariate and multivariate analyses of risk factors for OS and EFS are presented.

RESULTS

Eleven U.S. institutions contributed 83 patients treated from 1987 to 1991. The OS at 5 and 7 years was 57% and 46%, and EFS at 5 and 7 years was 42% and 33%. Patients 3 years of age or younger differed from the older group by more common infratentorial location, less common gross total resection (GTR), and postoperative use of chemotherapy rather than radiation. This younger group of patients had worse survival (P < 0.01) than the older age group. Other than young age, less than GTR and World Health Organization (WHO) II grade 3 histology were significant adverse risk factors for EFS in univariate and multivariate analyses. OS shared the same adverse risk factors except for histology in multivariate analysis, which was only of borderline significance (P = 0.05). Progression at the original tumor location, present in 89% of patients, was the major pattern of tumor recurrence. Adjuvant chemotherapy in the group older than 3 years or craniospinal radiation in M0 patients did not significantly change EFS.

CONCLUSIONS

Adverse outcome in childhood intracranial ependymoma is related to age (3 years or younger), histology (grade 3), and degree of surgical resection (less than GTR). New approaches, particularly for local tumor control in younger patients, are needed to improve survival.

Ki-67 immunolabelling index is a prognostic indicator in childhood posterior fossa ependymomas

Conventional histological evaluation and subclassification of childhood ependymomas poorly predict their biological behaviour. The Ki-67 labelling index (Ki-67 LI), a measure of growth fraction, correlates with the biological behaviour of several neoplasms, and this retrospective study tested the hypothesis that Ki-67 LI is a prognostic indicator in childhood posterior fossa ependymomas. Immunocytochemistry using an antibody to Ki-67 was undertaken on 5 microns sections of formalin-fixed, paraffin-embedded tissue from 74 cases of childhood (age < 16 years.) posterior fossa ependymoma. A Ki-67 LI was established by counting the proportion of labelled nuclei in more than 1000 cells from several histological fields. Several clinical and histological variables (including Ki-67 LI) potentially associated with survival were entered into univariate and multivariate analyses using a Cox proportional hazards model. Variables that showed a significant and independent association with survival were Ki-67 LI (P < 0.002), whether total surgical resection had been achieved according to operation records (P < 0.03), and whether no adjuvant therapy had been given (P < 0.01). Age, sex, and the presence of necrosis or microvascular proliferation did not correlate with survival. In our defined population of patients with ependymomas, Ki-67 LI is a strong prognostic indicator. We recommend that Ki-67 LI is used in the histological evaluation of childhood posterior fossa ependymomas during trials of novel adjunctive therapies.

Intracranial ependymomas in children: a critical review of prognostic factors and a plea for cooperation

BACKGROUND

Current controversies in pediatric intracranial ependymoma include histologic categorization and management. Most of our knowledge of this disease comes from single-institution reports.

METHODS

A literature search was done, covering the period 1976-1996. The aim of this review is to analyze the prognostic factors reported in the literature over the last 20 years.

RESULTS

Forty-five series were reviewed, including more than 1,400 children. The largest series reported on 92 patients, and the accrual rate ranged from 0.32-12 patients per year. None of the prognostic factors reported achieved a consensus throughout the different series. Histology remains a major issue, and the range in the incidence of anaplastic ependymo mas (7-89%) highlights the difficulty in agreeing on a histological grading system. The role of surgery on the outcome seems to be determinant. Recent series based on homogeneous imaging-documented extents of resection strongly support the benefit of postoperative radiotherapy. The lack of a proven, effective chemotherapy regimen precludes its use except in prospective pilot studies.

CONCLUSIONS

Limited information is available from single-institution reports in ependymoma. Only large national or international studies can provide enough information to allow a multivariate analysis of the prognostic factors, and thus lead to new therapeutic proposals.