Pediatric Chordomas: Results of a Multicentric Study of 40 Children and Proposal for a Histopathological Prognostic Grading System and New Therapeutic Strategies

Pediatric Chordomas: Results of a Multicentric Study of 40 Children and Proposal for a... Abstract Pediatric chordomas are rare malignant neoplasms, and few data are available for optimizing therapeutic strategies and outcome. This study aimed at evaluating how best to manage them and to identify prognostic factors. This multicentric retrospective study included 40 children diagnosed with chordomas between 1966 and 2012. Clinical, radiological, and histopathological data, treatment modalities, and outcomes were reviewed. The median age was 12 years old. Most chordomas were histologically classical forms (45.5%) and were mostly located at the skull base (72.5%). The overall survival (OS) was 66.6% and 58.6%, and progression-free survival (PFS) was 55.7% and 52% at 5 and 10 years, respectively. Total resection was correlated with a better outcome (p = 0.04 for OS and PFS, log-rank). A histopathological/immunohistochemical grading system recently crafted for adults was applied. In a multivariate analysis, it significantly correlated with outcome (PFS and OS, p = 0.004), and the loss of BAF47 immunoexpression appeared to be a significant independent prognostic factor (PFS, p = 0.033). We also identified clinical and histopathological parameters that correlated with prognosis. A new grading system combined with the quality of surgical resection could help classify patients to postpone radiotherapy in case of low risk. Targeted therapy and reirradiation at recurrence may be considered as potential therapeutic strategies. Histopathological grading, Pediatric chordomas, Prognostic algorithm, Radiotherapy INTRODUCTION Chordomas are rare slow-growing malignant neoplasms believed to derive from notochordal remnants. They are most frequently located in the sacrococcygeal and in the skull base regions and particularly affect adults with a peak incidence at diagnosis between 50 and 60 years of age. Pediatric cases represent less than 5% of cases. Prognosis of these tumors essentially depends on their local evolution and their high capacity for recurrence. In terms of treatment, maximal surgical resection is the first step, followed by local irradiation, ideally with proton beam therapy. To date, chemotherapy is not considered as a standard treatment. Publications on pediatric chordomas are rare (250 cases in the literature) (1). While age and extent of resection seem to influence the prognosis in children, no histopathological prognostic factors have yet been clearly identified (2). The aim of this multicentric retrospective study was to evaluate the clinical, neurosurgical, and immunohistochemical prognostic factors of 40 pediatric chordomas. The expression of potentially useful markers for targeted molecular therapies such as VEGF, EGFR, PTEN, pSTAT3, and the loss of BAF47 expression were also investigated. MATERIALS AND METHODS Patients This multicentric and retrospective study included 40 patients younger than 18 years of age from 6 centers, with chordomas diagnosed between 1966 and 2012. Informed consent for the surgery and translational research program was obtained from the parents or guardian according to the IRB approved protocol (number DC-2009-955 for tumor banking). Clinical and radiological findings, the treatment type, and outcome (progression-free survival [PFS] and overall survival [OS]) were analyzed by 2 pediatric neurosurgeons (K.B. and S.P.). The location of chordomas was reported as skull base (tumors of the clivus and of the first cervical vertebra) and non-skull base (tumors from the second cervical vertebra to the sacrum). The quality of surgical resection was assessed from CT and MR scans when available or from surgery, hospitalization, or radiology reports when imaging was not available. The diagnosis of chordoma was based on histopathology, confirmed by Brachyury immunopositivity. Histopathological Review Central pathology review was performed by 2 neuropathologists (P.V. and A.T-E.). Tumors were subtyped into well-differentiated chordomas (classical, chondroid, or mixed subtypes) and poorly differentiated chordomas, as previously defined (2). Histopathological parameters (nuclear pleomorphism, prominent nucleoli, apoptosis, necrosis, hemorrhage, inflammatory cell infiltrates, and mitotic activity) were evaluated as in adult series (2). Immunohistochemistry Immunohistochemical studies were performed using antibodies directed against Ki67 (clone K2; Zytomed, Berlin, Germany, 1:400), p53 (clone Do-7; Dako, Santa Clara, CA, 1:200), E-cadherin (clone EP700Y; Cell Marque, Rocklin, CA, 1:100), EGFR (clone 3C6; Ventana Medical Systems, Tucson, AZ, prediluted), PTEN (clone 6H2-1;Dako, 1: 300), pSTAT3 (clone F2; Santa Cruz Biotechnology, Dallas, TX, 1:100), VEGF (polyclonal; Millipore, Billerica, MA, 1:1000), and BAF47 (clone 25/BAF47; BD-Biosciences, San Jose, CA, 1:200). Positive controls were used and immunostainings were scored independently by 2 observers (P.V. and A.T-E.) as previously described in the adult series (2). Cytoplasmic staining for PTEN was considered homogeneously positive if cytoplasmic staining was present in more than 90% of tumor cells. Molecular Analysis of the SMARCB1 Gene DNA extraction from formaldehyde-fixed paraffin-embedded (FFPE) tumors was performed using the Nucleospin Tissue kit (Macherey-Nagel, Düren, Germany). Six-micrometer-thick dewaxed sections were used for DNA extraction. Samples that underwent decalcification were excluded. Copy number alterations were searched for using a multiplex ligation-dependent probe amplification (MLPA) assay (Salsa MLPA KIT P258-B1 SMARCB1; MRC Holland, Amsterdam, The Netherlands). Overall, 50 ng of tumor DNAs were used for hybridization, ligation, and amplification according to the manufacturer’s instructions in conjunction with a normal control DNA with 2 copies of each exon of the SMARCB1 gene. Two microliters of amplification products were analyzed on an ABI 3130 DNA Analyzer (Applied Biosystems, Foster City, CA). Data analysis was performed using GeneMarker software (SoftGenetics, State College, PA). Small mutation screenings of each coding exon of the SMARCB1 gene were carried out using Sanger sequencing. Sequences were analyzed on an ABI 3130 DNA sequencer (Applied Biosystems) and further aligned with Seqscape analysis software (Applied Biosystems). FISH SMARCB1 was carried out on cases with BAF47 immunoexpression loss but without any deletion in the SMARCB1 gene using MLPA analysis. FISH procedures were performed as previously described. Copy number abnormalities were assessed on 100 nuclei by 2 independent observers. Statistical Analysis We provide the mean and median values and frequency (percentages) for the description of continuous and categorical variables, respectively. Histopathological and immunohistochemical data were combined to compute a grading score as previously described (2). OS was defined as the time in months from the date of diagnosis to death from any cause or to the date of the last follow-up. PFS was defined as the time in months between the diagnosis and the earliest date of disease progression (local and distant), death from any cause if no progression was observed or data cutoff. OS and PFS were estimated using the Kaplan-Meier method and expressed by median with a 95% confidence interval (CI). Survival curves were compared using log-rank test. All variables with p ≤ 0.05 observed in univariate analysis were included in the multivariate Cox regression model with stepwise backward elimination to estimate hazard ratio with a 95% CI (except variables that were used to build the grading system when it was introduced in the model). All statistical tests were 2-sided, and p < 0.05 were regarded as significant. Data were analyzed by using PASW Statistics for Windows, Version 20.0 (SPSS Inc., Chicago, IL). RESULTS Clinical and Surgical Data The median age was 12 years (mean = 11.1, range from 2.3 to 17.8 years) with a male predominance (24 M/16 F). This series included 29 skull base and 11 non-skull base chordomas; there were no sacrococcygeal chordomas, and no patient initially presented with metastases at the time of primary diagnosis. Among the 40 included patients, a total resection of the tumor was initially performed for 4 patients and subtotal resection for 28 patients. In 8 cases, the extent of resection was not available. After total resection, no recurrence was observed at 9 months, 5, 9, and 10 years after diagnosis. Among patients with subtotal resection, 21 patients presented a progression of residual tumor (10 of them died of their disease); 7 patients presented with a stable disease. Adjuvant Treatment Radiotherapy was the principal adjuvant treatment performed after a maximal tumor resection. All except 6 patients were irradiated by photon beam therapy (n = 7), proton beam therapy (n = 6), or a combination of photon and proton beam therapy (n = 20). For 1 patient, the type of radiotherapy was unknown. Mean doses of irradiation were 54 Gy (minimum = 50 Gy and maximum = 55.8 Gy) for photon beam therapy, 70.8 Gy (minimum = 66.6 Gy and maximum = 74 Gy) for proton beam therapy, and 69.4 Gy (minimum = 60 Gy and maximum = 76 Gy) for the combination therapy. Four patients received different protocols of chemotherapy, either in cases of progressive metastatic disease (n = 2) or in cases of initial misdiagnosis (n = 2). Disease control was never obtained, and all patients died some months after the beginning of the treatment. Histopathological Findings Of the 40 patients in the clinical study, we obtained sufficient histopathological material for 33 of them. Furthermore, 45 tumors (33 primary and 12 recurrences) were included for histopathological assessment. Of the 33 primary chordomas, 15 (45.5%) were classical, 1 (3.0%) was chondroid, 12 (36.4%) were mixed, and 5 (15.1%) were poorly differentiated (Fig. 1). In all recurrent chordomas, the histopathological subtype remained unchanged between the first surgery and recurrence. Among the 33 primary chordomas, prominent nucleoli, hemorrhage, apoptosis, necrosis, nuclear pleomorphism, and inflammatory cell infiltrates were present in 6 (18.2%), 7 (21.2%), 11 (33.3%), 14 (42.4%), 20 (60.6%), and 24 (72.7%) of the initial tumors, respectively. Mitotic activity was very low. Figure 1. View largeDownload slide Histopathological findings. (A) Classical (myxoid) chordoma (H&E, 120×). (B) Anti-Brachyury labels the nuclei of the tumor cells but not the stroma (120×). (C) Well-differentiated chordoma with classical (myxoid) and chordoid features (H&E, 120×). (D) Anti-Brachyury labels the nuclei of the tumor cells of the 2 components, 120×). (E) Poorly differentiated pseudosarcomatous chordoma composed of spindle cells (H&E, 90×). (F) Anti-Brachyury labels the nuclei of the majority of tumor cells (90×). (G) Preserved nuclear immunoexpression of BAF47 (120×). (H) Loss of BAF47 nuclear expression in tumor cells of this poorly differentiated pseudosarcomatous chordoma. Note preserved immunoreactivity in endothelial cells (120×). Figure 1. View largeDownload slide Histopathological findings. (A) Classical (myxoid) chordoma (H&E, 120×). (B) Anti-Brachyury labels the nuclei of the tumor cells but not the stroma (120×). (C) Well-differentiated chordoma with classical (myxoid) and chordoid features (H&E, 120×). (D) Anti-Brachyury labels the nuclei of the tumor cells of the 2 components, 120×). (E) Poorly differentiated pseudosarcomatous chordoma composed of spindle cells (H&E, 90×). (F) Anti-Brachyury labels the nuclei of the majority of tumor cells (90×). (G) Preserved nuclear immunoexpression of BAF47 (120×). (H) Loss of BAF47 nuclear expression in tumor cells of this poorly differentiated pseudosarcomatous chordoma. Note preserved immunoreactivity in endothelial cells (120×). Immunohistochemical Findings Insufficient tumor tissue was available to perform all the immunostains in the 45 tumor samples. Nuclear expression of BAF47 was lost in 4 poorly differentiated chordomas and in 3 classical/mixed chordomas (18.9%). The mean Ki67 labeling index (LI) was 14% (1%–60%) and equal or greater than 6% in 14/24 (58.3%) primary chordomas. p53 nuclear accumulation was present in 18/25 (72%) primary chordomas. The mean p53 LI was 29% (1%–91%) and equal to or greater than 25% in 12 primary chordomas. VEGF was intensely expressed in 94.9% of all tumors (37/39). pSTAT3 was expressed in 31/36 (86.1%) of all chordomas with intense staining in 76.7% of the tumors. EGFR was expressed in 35/38 (92.1%) of all chordomas. Immunohistochemical score was greater than 150 in 47.4% of tumors (18/38). Loss of PTEN was observed in 17/23 of tested tumors (73.9%) (Fig. 2). Figure 2. View largeDownload slide Prognostic and therapeutic biomarkers in chordomas. (A) High Ki67 labeling index in a poorly differentiated chordoma (200×). (B) High p53 labeling index up to 25% (150×). (C) Strong immunoreactivity for pSTAT3 in a classical chordoma (200×). (D) Diffuse membranous immunopositivity of E-cadherin in a classical myxoid chordoma (150×). (E) Loss of the immunoexpression of E-cadherin in a poorly differentiated chordoma (150×). (F) Diffuse and strong expression of VEGF in a classical chordoma (150×). (G) Diffuse and strong immunopositivity for EGFR (score 300) by a poorly differentiated chordoma (200×). (H) Diffuse and moderate immunoreactivity for EGFR (score 200) in a classical chordoma (200×). (I) Loss of the expression of PTEN in tumor cells. Note preserved expression in endothelial cells (200×). Figure 2. View largeDownload slide Prognostic and therapeutic biomarkers in chordomas. (A) High Ki67 labeling index in a poorly differentiated chordoma (200×). (B) High p53 labeling index up to 25% (150×). (C) Strong immunoreactivity for pSTAT3 in a classical chordoma (200×). (D) Diffuse membranous immunopositivity of E-cadherin in a classical myxoid chordoma (150×). (E) Loss of the immunoexpression of E-cadherin in a poorly differentiated chordoma (150×). (F) Diffuse and strong expression of VEGF in a classical chordoma (150×). (G) Diffuse and strong immunopositivity for EGFR (score 300) by a poorly differentiated chordoma (200×). (H) Diffuse and moderate immunoreactivity for EGFR (score 200) in a classical chordoma (200×). (I) Loss of the expression of PTEN in tumor cells. Note preserved expression in endothelial cells (200×). Molecular Analysis of the SMARCB1 Gene MLPA and sequencing of the 9 exons of the SMARCB1 gene were performed on FFPE samples from 5 BAF47-deficient tumors. Only 1 case displayed a biallelic loss of function of the SMARCB1 gene with complete loss of all exons of the SMARCB1 gene on 1 allele and an intragenic deletion of exons 7–9 on the other allele resulting in a heterozygous deletion of exons 1–6 and a homozygous deletion of exons 7–9. In 2 other cases, MLPA analysis revealed a decrease of the probes hybridized on the SMARCB1 gene, including 1 case with a very small decrease of the probes suggesting a loss of at least 1 copy of the gene, but the sample analyzed contained a low fraction of tumor cells. No alteration was detected in the SMARCB1 gene in 1 case, but it could be hypothesized that the fraction of tumor cells was too small and below the sensitivity rate of the MLPA. In that last case, no coherent results were obtained due to the poor quality of extracted DNA from the FFPE sample. FISH analysis of these 2 latter cases was also not contributory due to insufficient number of tumor cells. Follow-up and Survival Analysis The OS was 66.6% and 58.6% at 5 and 10 years, respectively, and the PFS was 55.7% and 52% at 5 and 10 years, respectively. After a median follow-up of 5.3 years (min 0.1 years to max 22.3 years), 9 patients (22.5%) were in complete remission and 16 had a stable disease (40%). Fifteen patients (37.5%) died from their disease after a median evolution of 2.3 years (from 0.14 to 5 years). Three patients developed metastasis during follow-up. No recurrence was observed after complete resection. For the others, disease progression occurred earlier after a treatment by surgery alone (mean 453 days) than after an association of surgery and radiotherapy (mean 769 days). The prognosis was better in case of total resection compared to subtotal resection (p = 0.04 for both OS and PFS, log-rank; Table). We observed a better 5-year PFS for skull base chordomas compared to the others (p = 0.01 log-rank; Table). Prognostic factors correlated with OS and PFS in univariate and multivariate analyses are listed in the Table. Grading System As previously described, histopathological and immunohistochemical parameters were combined to establish a grading system (2): Ki67 LI ≥6%, p53 LI ≥25%, presence of a poorly differentiated component, significantly abundant necrosis (≥3% of the tumor surface), prominent nucleoli or apoptosis significantly observed (≥3% of the tumor cells) or mitoses in more than 3/10 HPF were scored 1, absence of these criteria was scored 0. The sum of these criteria defined a score. The low- (scores 0–3) and high-grade (scores 4–7) groups included 17 and 14 tumors, respectively. Two chordomas were excluded from the grading system due to missing data. The classification into 2 groups has proved to be significantly correlated with PFS and OS (Table; Fig. 3). Figure 3. View largeDownload slide Prognostic value of the proposed histopathological and immunohistochemical grading system. Grades significantly correlated to PFS and OS: mean 5 years PFS = 52.9 months for low-grade and 14.3 months for high-grade, p = 0.03; mean 5 years OS = 64.7 months for low-grade and 35.7 months for high-grade, p = 0.002). OS, overall survival; PFS, progression-free survival. Figure 3. View largeDownload slide Prognostic value of the proposed histopathological and immunohistochemical grading system. Grades significantly correlated to PFS and OS: mean 5 years PFS = 52.9 months for low-grade and 14.3 months for high-grade, p = 0.03; mean 5 years OS = 64.7 months for low-grade and 35.7 months for high-grade, p = 0.002). OS, overall survival; PFS, progression-free survival. DISCUSSION Our results agree with those previously published, as classical chordoma is the most frequent subtype in children, followed by pure chondroid chordoma. Poorly differentiated chordomas (46/222, 20.7%), including dedifferentiated chordomas (12/222, 5.4%), and “atypical” forms (34/222, 15.3%), including epithelioid or solid forms (12 cases), have also been described (3–35) but this histopathological subtype is not included in the 2014 WHO classification of tumors of soft tissue and bone (3, 5, 12, 36, 37). In our series, a poorly differentiated component represented a significant prognostic factor in terms of PFS and OS as previously assumed (2, 36) (Table). Among other histopathological parameters, we observed that nuclear pleomorphism, necrosis, and apoptosis were significantly correlated to OS and PFS (Table). The poorer prognostic value of nuclear pleomorphism and necrosis had previously been described in adults (2). However, apoptosis correlated with a shorter survival in our series, contrary to previous studies (2). Ki67 LI remains a controversial prognostic factor in the literature (3). Our results along with those obtained in a previous study indicate that Ki67 LI is higher in pediatric chordomas than in adult cases (2). We also observed a higher p53 LI with no prognostic significance, as opposed to other adult and pediatric studies (2). Finally, loss of E-cadherin expression significantly correlated with a better OS (p = 0.03, log-rank; Table), according to a previous smaller series (2). Loss of BAF47 (SMARCB1) immunoexpression represents a powerful prognostic factor in pediatric chordomas. Originally described in atypical teratoid and rhabdoid tumors, alterations of SMARCB1 gene were also described in chordomas and correlated with more aggressive behavior (38–40). The mean age of previously reported patients with BAF47-deficient chordomas was 6.9 years (1–17 years) (4, 38, 39, 41, 42). Chordomas were predominantly located within the skull base, and 19/26 cases (73.1%) were poorly differentiated with sarcoma-like features (4, 38–42). This loss of expression is due to mutations (38), or large deletions of the SMARCB1 gene (40, 42), as observed in 22.6% of cases in our series. Treatments were extremely variable and progression was observed in 13/15 cases (38, 40–42). The loss of BAF47 immunoreactivity correlated with a poorer prognosis and death occurred within the first 3 years after the initial presentation in 14/20 children (4, 38, 40–42). This poorer prognosis was confirmed in our series (PFS, p < 0.001, log-rank; Table), and has led to considering BAF47 immunoexpression as a major prognostic factor for pediatric chordomas. The prognostic value of the histopathological grading system suggested by our team in the adult series has been confirmed by our pediatric series: patients with a low-grade tumor have a significantly better outcome than patients with a high-grade tumor, both in PFS and OS (Table). Loss of BAF47 immunoexpression was an independent prognostic factor in term of PFS, and it led us to create a prognostic algorithm based on its immunoexpression and data obtained from our grading system (Fig. 4). Considering the significant impact of radical surgery and histopathological grading on outcome, we infer that patients with low-grade tumors for whom a complete resection has been performed have a low risk of recurrence. For these patients, we wonder if radiotherapy is necessary as the first line of treatment and if it could be postponed and subsequently offered only for recurrent tumors. Figure 4. View largeDownload slide Proposed prognostic algorithm of pediatric chordomas. The first step evaluates the status of the expression of BAF47, the worst prognosis being associated with a loss of this immunoexpression. The second step includes the results of the histopathological and immunohistochemical grading system. Figure 4. View largeDownload slide Proposed prognostic algorithm of pediatric chordomas. The first step evaluates the status of the expression of BAF47, the worst prognosis being associated with a loss of this immunoexpression. The second step includes the results of the histopathological and immunohistochemical grading system. We further investigated the expression of molecules that may influence the growth and survival of cancer cells and could lead to targeted therapies. Hypoxia-driven VEGF, a major inducer of tumor angiogenesis, was highly expressed in this series (94.9% of chordomas), as well as in the adult series (2). While EGFR immunolabeling is highly variable in the literature (8%–81%) (2), it was intensely expressed in 47.4% of tumors in our study. These results are in accordance with a previous but small pediatric series (12). Previous studies have reported partial response to both anti-VEGF (43) and anti-EGFR drugs (2) in adults, and anti-EGFR therapy has been shown to inhibit growth in a patient-derived chordoma xenograft (2); therefore, these drugs could potentially be useful in the treatment of chordomas. We also found pSTAT3 expression in 87.5% of primary pediatric chordomas, that is, higher than in adult studies (2). Inhibitors of the STAT3 pathway increased the rate of apoptosis in chemoresistant cancer cell lines and decreased the growth rate of chordoma cell lines (2), so this biomarker could be considered an interesting therapeutic target. Loss of PTEN expression has previously been observed in chordoma samples and is correlated with tumor invasion (44). In our study, a loss of expression in the majority of chordomas (73.9% of all cases), was demonstrated, concordant with other studies in which PTEN deletion and loss of the PTEN gene were reported (45, 46). Treatment of chordomas is based on multidisciplinary management, and all authors agree on the fact that surgery is the first step of this treatment (47–50). Surgery has 2 aims: to reduce the tumor volume and, if possible, to obtain complete resection, and to keep any residue away from vascular or nervous anatomical structures for adjuvant radiotherapy. Limited data exist on pediatric series due to the small size of cohorts but Ridenour et al observed a tendency for a better OS in case of total resection compared to subtotal resection (n = 35, p = 0.374) (3). In our series, this difference was significant in terms of OS and PFS (p = 0.04, log-rank; Table). Complete resection of these tumors remains challenging and varies from 0% to 36.4% in the main pediatric studies (22.5% in our series) (3, 47, 49, 51). High-dose radiotherapy is considered to be the standard treatment after surgery (30, 52–54). Limited data are available for children but in a review on intracranial chordomas, the benefit of radiotherapy was clearly established (OS, p = 0.004) (36). Proton therapy has become the technique of choice (55, 56) because it allows for higher doses of radiation on tumor tissue while sparing adjacent healthy tissue (5, 57, 58). In our cohort, patients who were irradiated had a better outcome in terms of PFS and OS (p < 0.001, log-rank, Table), but patients who were not irradiated died rapidly or presented a rapid growth of the tumor. Thus, it is difficult to conclude on the direct consequences of the absence of irradiation of our patients. Considering the high potential for local recurrence of chordomas, the benefits of reirradiation may also have to be evaluated in the case of progressive disease, as is recommended for other pediatric tumors (59). In conclusion, our results provide evidence that the aforementioned clinical and histopathological parameters are correlated with prognosis. Our grading system combined with the quality of surgical resection could help in categorizing patients and postponing radiotherapy in low-risk cases. The immunoexpression of biomarkers implicated in targeted therapy (EGFR, VEGF, and STAT3 pathways) could help in the selection of therapeutic strategies among those highly expressed. Table. Univariate and Multivariate Analysis for Prognostic Factors     Univariate Analysis       n  5-Year PFS (%)  p Value  5-Year OS (%)  p Value  Location  SB  29  56.2  0.01  0.9  0.42    NSB  11  27.3    60.6    Extent of surgery                STR  33  40.7  0.04  61.3  0.04    TR  7  85.7    100    Treatment                No RxT  6  0  <0.001  22.2  <0.001    Photon Rxt  7  71.4    71.4      PT ± Photon Rxt  26  55.1    79.3    Nuclear pleomorphism                Absent  13  65.8  0.04  91.7  0.04    Present  18  31.1    47.6    Histopathological subtype                Well differentiated  24  50.7  0.01  77.2  <0.01    Poorly differentiated  9  22.2    25.4    Apoptosis                Absent  20  61.7  <0.01  83.9  <0.01    Present  11  18.2    34.1    Necrosis                Absent  19  65.7  <0.01  88.5  <0.01    Present  12  12.5    31.3    Grade                Low  17  52.9  0.03  64.7  0.002    High  14  14.3    35.7    pSTAT3 expression                Absent or weak  7  14.3  0.02  21.4  0.01    High  23  53.4    77.3    E-cadherin expression                < 50%  28  46.6  0.21  73.4  0.03    ≥ 50%  5  20    20    BAF47 expression                Lost  7  14.3  <0.01  51.4  0.27    Preserved  24  50.9    68.7        Multivariate analysis  PFS    Grade  <0.01            Extent of surgery  0.01            Nuclear pleomorphism  0.02            BAF47 expression  0.03        OS    Grade  <0.01            E-cadherin expression  0.02            Univariate Analysis       n  5-Year PFS (%)  p Value  5-Year OS (%)  p Value  Location  SB  29  56.2  0.01  0.9  0.42    NSB  11  27.3    60.6    Extent of surgery                STR  33  40.7  0.04  61.3  0.04    TR  7  85.7    100    Treatment                No RxT  6  0  <0.001  22.2  <0.001    Photon Rxt  7  71.4    71.4      PT ± Photon Rxt  26  55.1    79.3    Nuclear pleomorphism                Absent  13  65.8  0.04  91.7  0.04    Present  18  31.1    47.6    Histopathological subtype                Well differentiated  24  50.7  0.01  77.2  <0.01    Poorly differentiated  9  22.2    25.4    Apoptosis                Absent  20  61.7  <0.01  83.9  <0.01    Present  11  18.2    34.1    Necrosis                Absent  19  65.7  <0.01  88.5  <0.01    Present  12  12.5    31.3    Grade                Low  17  52.9  0.03  64.7  0.002    High  14  14.3    35.7    pSTAT3 expression                Absent or weak  7  14.3  0.02  21.4  0.01    High  23  53.4    77.3    E-cadherin expression                < 50%  28  46.6  0.21  73.4  0.03    ≥ 50%  5  20    20    BAF47 expression                Lost  7  14.3  <0.01  51.4  0.27    Preserved  24  50.9    68.7        Multivariate analysis  PFS    Grade  <0.01            Extent of surgery  0.01            Nuclear pleomorphism  0.02            BAF47 expression  0.03        OS    Grade  <0.01            E-cadherin expression  0.02        NSB, non-skull base; OS, overall survival; PFS, progression-free survival; Photon RxT, Photon beam radiotherapy; PT, Proton beam radiotherapy; RxT, radiotherapy; SB, skull base; STR, subtotal resection; TR, total resection. ACKNOWLEDGMENTS The authors are grateful to all laboratory technicians for their technical assistance. They thank Albane Gareton for assistance in translation. Part of this study has been presented at the 24th ESPN Congress in Rome, Italy, and at the ISPN 2016 44th Annual Meeting in Kobe, Japan. REFERENCES 1 Beccaria K, Sainte-Rose C, Zerah M, et al.   Paediatric chordomas. Orphanet J Rare Dis  2015; 10: 116 http://dx.doi.org/10.1186/s13023-015-0340-8 Google Scholar CrossRef Search ADS PubMed  2 Tauziède-Espariat A, Bresson D, Polivka M, et al.   Prognostic and therapeutic markers in chordomas: A study of 287 tumors. J Neuropathol Exp Neurol  2016; 75: 111– 20 Google Scholar CrossRef Search ADS PubMed  3 Ridenour RVIII, Ahrens WA, Folpe AL, et al.   Clinical and histopathologic features of chordomas in children and young adults. Pediatr Dev Pathol  2010; 13: 9– 17 http://dx.doi.org/10.2350/09-01-0584.1 Google Scholar CrossRef Search ADS PubMed  4 Yadav R, Sharma MC, Malgulwar PB, et al.   Prognostic value of MIB-1, p53, epidermal growth factor receptor, and INI1 in childhood chordomas. Neuro-oncology  2014; 16: 372– 81 http://dx.doi.org/10.1093/neuonc/not228 Google Scholar CrossRef Search ADS PubMed  5 Hoch BL, Nielsen GP, Liebsch NJ, et al.   Base of skull chordomas in children and adolescents: A clinicopathologic study of 73 cases. Am J Surg Pathol  2006; 30: 811– 8 http://dx.doi.org/10.1097/01.pas.0000209828.39477.ab Google Scholar CrossRef Search ADS PubMed  6 Moore KA, Bohnstedt BN, Shah SU, et al.   Intracranial chordoma presenting as acute hemorrhage in a child: Case report and literature review. Surg Neurol Int  2015; 6: 63 http://dx.doi.org/10.4103/2152-7806.155445 Google Scholar CrossRef Search ADS PubMed  7 Hashim H, Rosman AK, Abdul Aziz A, et al.   Atypical clival chordoma in an adolescent without imaging evidence of bone involvement. Malays J Med Sci  2014; 21: 78– 82 Google Scholar PubMed  8 Dhall G, Traverso M, Finlay JL, et al.   The role of chemotherapy in pediatric clival chordomas. J Neurooncol  2011; 103: 657– 62 http://dx.doi.org/10.1007/s11060-010-0441-0 Google Scholar CrossRef Search ADS PubMed  9 Schechter MM, Liebeskind AL, Azar-Kia B. Intracranial chordomas. Neuroradiol  1974; 8: 67– 82 http://dx.doi.org/10.1007/BF00345038 Google Scholar CrossRef Search ADS   10 Fink FM, Ausserer B, Schröcksnadel W, et al.   Clivus chordoma in a 9-year-old child: Case report and review of the literature. Pediatr Hematol Oncol  1987; 4: 91– 100 Google Scholar CrossRef Search ADS PubMed  11 Inagaki H, Anno Y, Hori T, et al.   [Clival chordoma in an infant; case report and review of the literature]. No Shinkei Geka  1992; 20: 809– 13 Google Scholar PubMed  12 Yadav YR, Kak VK, Khosla VK, et al.   Cranial chordoma in the first decade. Clin Neurol Neurosurg  1992; 94: 241– 6 http://dx.doi.org/10.1016/0303-8467(92)90096-L Google Scholar CrossRef Search ADS PubMed  13 Schamschula RG, Soo MY. Clival chordomas. Australas Radiol  1993; 37: 259– 64 http://dx.doi.org/10.1111/j.1440-1673.1993.tb00069.x Google Scholar CrossRef Search ADS PubMed  14 Auger M, Raney B, Callender D, et al.   Metastatic intracranial chordoma in a child with massive pulmonary tumor emboli. Pediatr Pathol  1994; 14: 763– 70 http://dx.doi.org/10.3109/15513819409037673 Google Scholar CrossRef Search ADS PubMed  15 Scimeca PG, James-Herry AG, Black KS, et al.   Chemotherapeutic treatment of malignant chordoma in children. J Pediatr Hematol Oncol  1996; 18: 237– 40 http://dx.doi.org/10.1097/00043426-199605000-00032 Google Scholar CrossRef Search ADS PubMed  16 Soo MY. Chordoma: Review of clinicoradiological features and factors affecting survival. Australas Radiol  2001; 45: 427– 34 http://dx.doi.org/10.1046/j.1440-1673.2001.00950.x Google Scholar CrossRef Search ADS PubMed  17 Lountzis NI, Hogarty MD, Kim HJ, et al.   Cutaneous metastatic chordoma with concomitant tuberous sclerosis. J Am Acad Dermatol  2006; 55: S6– S10 Google Scholar CrossRef Search ADS PubMed  18 Yoneoka Y, Tsumanuma I, Fukuda M, et al.   Cranial base chordoma—long term outcome and review of the literature. Acta Neurochir (Wien)  2008; 150: 773– 8 Google Scholar CrossRef Search ADS PubMed  19 Chang SW, Gore PA, Nakaji P, et al.   Juvenile intradural chordoma: Case report. Neurosurgery  2008; 62:E525–6; discussion: E527 20 Al-Rahawan MM, Siebert JD, Mitchell CS, et al.   Durable complete response to chemotherapy in an infant with a clival chordoma. Pediatr Blood Cancer  2011; 59: 323– 5 Google Scholar CrossRef Search ADS PubMed  21 Richards AT, Stricke L, Spitz L. Sacrococcygeal chordomas in children. J Pediatr Surg  1973; 8: 911– 4 http://dx.doi.org/10.1016/0022-3468(73)90010-9 Google Scholar CrossRef Search ADS PubMed  22 Garofalo E, Minerva A, Baltieri G. [Sacrococcygeal chordoma in a 19-month-old boy]. Minerva Pediatr  1976; 28: 1909– 14 Google Scholar PubMed  23 Nix WL, Steuber CP, Hawkins EP, et al.   Sacrococcygeal chordoma in a neonate with multiple anomalies. J Pediat  1978; 93: 995– 8 http://dx.doi.org/10.1016/S0022-3476(78)81235-9 Google Scholar CrossRef Search ADS PubMed  24 Kozlowski K, Barylak A, Campbell J, et al.   Primary sacral bone tumours in children (report of 16 cases with a short literature review). Australas Radiol  1990; 34: 142– 9 http://dx.doi.org/10.1111/j.1440-1673.1990.tb02830.x Google Scholar CrossRef Search ADS PubMed  25 Cable DG, Moir C. Pediatric sacrococcygeal chordomas: A rare tumor to be differentiated from sacrococcygeal teratoma. J Pediatr Surg  1997; 32: 759– 61 http://dx.doi.org/10.1016/S0022-3468(97)90028-2 Google Scholar CrossRef Search ADS PubMed  26 Shinmura Y, Miura K, Yajima S, et al.   Sacrococcygeal chordoma in infancy showing an aggressive clinical course: An autopsy case report. Pathol Int  2003; 53: 473– 7 http://dx.doi.org/10.1046/j.1440-1827.2003.01496.x Google Scholar CrossRef Search ADS PubMed  27 Al-Adra D, Bennett A, Gill R, et al.   Pediatric metastatic sacrococcygeal chordoma treated with surgery. Eur J Pediatr Surg  2011; 21: 196– 8 http://dx.doi.org/10.1055/s-0031-1271635 Google Scholar CrossRef Search ADS PubMed  28 Windeyer BW. Chordoma. Proc R Soc Med  1959; 52: 1088– 100 Google Scholar PubMed  29 Proyard G, DONY H. [Chordoma of the cervical spine]. Acta Chir Belg  1965; 64: 132– 9 Google Scholar PubMed  30 Higinbotham NL, Phillips RF, Farr HW, et al.   Chordoma. Thirty-five-year study at Memorial Hospital. Cancer  1967; 20: 1841– 50 http://dx.doi.org/10.1002/1097-0142(196711)20:11<1841::AID-CNCR2820201107>3.0.CO;2-2 Google Scholar CrossRef Search ADS PubMed  31 Occhipinti E, Mastrostefano R, Pompili A, et al.   Spinal chordomas in infancy. Report of a case and analysis of the literature. Childs Brain  1981; 8: 198– 206 Google Scholar PubMed  32 Huang S-M, Chen C-C, Chiu P-C, et al.   Unusual presentation of posterior mediastinal chordoma in a 2-year-old boy. J Pediatr Hematol Oncol  2003; 25: 743– 6 http://dx.doi.org/10.1097/00043426-200309000-00014 Google Scholar CrossRef Search ADS PubMed  33 Killampalli VV, Power D, Stirling AJ. Preadolescent presentation of a lumbar chordoma: Results of vertebrectomy and fibula strut graft reconstruction at 8 years. Eur Spine J  2006; 15: 621– 5 http://dx.doi.org/10.1007/s00586-006-0138-4 Google Scholar CrossRef Search ADS PubMed  34 Wang Y, Xiao J, Wu Z, et al.   Primary chordomas of the cervical spine: A consecutive series of 14 surgically managed cases. J Neurosurg Spine  2012; 17: 292– 9 http://dx.doi.org/10.3171/2012.7.SPINE12175 Google Scholar CrossRef Search ADS PubMed  35 Hart ES. Cervical spine chordoma. Orthop Nurs  2012; 31: 355– 6 http://dx.doi.org/10.1097/NOR.0b013e318274272f Google Scholar CrossRef Search ADS PubMed  36 Borba LA, Al-Mefty O, Mrak RE, et al.   Cranial chordomas in children and adolescents. J Neurosurg  1996; 84: 584– 91 http://dx.doi.org/10.3171/jns.1996.84.4.0584 Google Scholar CrossRef Search ADS PubMed  37 Jo VY, Fletcher CDM. WHO classification of soft tissue tumours: An update based on the 2013 (4th) edition. Pathology  2014; 46: 95– 104 http://dx.doi.org/10.1097/PAT.0000000000000050 Google Scholar CrossRef Search ADS PubMed  38 Antonelli M, Raso A, Mascelli S, et al.   SMARCB1/INI1 Involvement in pediatric chordoma: A mutational and immunohistochemical analysis. Am J Surg Pathol  2017; 41: 56– 61 http://dx.doi.org/10.1097/PAS.0000000000000741 Google Scholar CrossRef Search ADS PubMed  39 Chavez JA, Nasir Ud D, Memon A, et al.   Anaplastic chordoma with loss of INI1 and Brachyury expression in a 2-year-old girl. Clin Neuropathol  2014; 33: 418– 20 http://dx.doi.org/10.5414/NP300724 Google Scholar CrossRef Search ADS PubMed  40 Hasselblatt M, Thomas C, Hovestadt V, et al.   Poorly differentiated chordoma with SMARCB1/INI1 loss: A distinct molecular entity with dismal prognosis. Acta Neuropathol  2016; 132: 149– 51 http://dx.doi.org/10.1007/s00401-016-1574-9 Google Scholar CrossRef Search ADS PubMed  41 Renard C, Pissaloux D, Decouvelaere AV, et al.   Non-rhabdoid pediatric SMARCB1-deficient tumors: Overlap between chordomas and malignant rhabdoid tumors? Cancer Genet  2014; 207: 384– 9 Google Scholar CrossRef Search ADS PubMed  42 Mobley BC, McKenney JK, Bangs CD, et al.   Loss of SMARCB1/INI1 expression in poorly differentiated chordomas. Acta Neuropathol  2010; 120: 745– 53 http://dx.doi.org/10.1007/s00401-010-0767-x Google Scholar CrossRef Search ADS PubMed  43 Asklund T, Sandström M, Shahidi S, et al.   Durable stabilization of three chordoma cases by bevacizumab and erlotinib. Acta Oncol  2014; 53: 980– 4 Google Scholar CrossRef Search ADS PubMed  44 Chen K, Mo J, Zhou M, et al.   Expression of PTEN and mTOR in sacral chordoma and association with poor prognosis. Med Oncol  2014; 31: 886– 5 http://dx.doi.org/10.1007/s12032-014-0886-7 Google Scholar CrossRef Search ADS PubMed  45 Choy E, MacConaill LE, Cote GM, et al.   Genotyping cancer-associated genes in chordoma identifies mutations in oncogenes and areas of chromosomal loss involving CDKN2A, PTEN, and SMARCB1. PLoS ONE  2014; 9: e101283 Google Scholar CrossRef Search ADS PubMed  46 Horbinski C, Oakley GJ, Cieply K, et al.   The prognostic value of Ki-67, p53, epidermal growth factor receptor, 1p36, 9p21, 10q23, and 17p13 in skull base chordomas. Arch Pathol Lab Med  2010; 134: 1170– 6 Google Scholar PubMed  47 Wold LE, Laws ER. Cranial chordomas in children and young adults. J Neurosurg  1983; 59: 1043– 7 http://dx.doi.org/10.3171/jns.1983.59.6.1043 Google Scholar CrossRef Search ADS PubMed  48 Heffelfinger MJ, Dahlin DC, MacCarty CS, et al.   Chordomas and cartilaginous tumors at the skull base. Cancer  1973; 32: 410– 20 http://dx.doi.org/10.1002/1097-0142(197308)32:2<410::AID-CNCR2820320219>3.0.CO;2-S Google Scholar CrossRef Search ADS PubMed  49 Benk V, Liebsch NJ, Munzenrider JE, et al.   Base of skull and cervical spine chordomas in children treated by high-dose irradiation. Int J Radiat Oncol Biol Phys  1995; 31: 577– 81 http://dx.doi.org/10.1016/0360-3016(94)00395-2 Google Scholar CrossRef Search ADS PubMed  50 Carpentier A, Polivka M, Blanquet A, et al.   Suboccipital and cervical chordomas: The value of aggressive treatment at first presentation of the disease. J Neurosurg  2002; 97: 1070– 7 http://dx.doi.org/10.3171/jns.2002.97.5.1070 Google Scholar CrossRef Search ADS PubMed  51 Coffin CM, Swanson PE, Wick MR, et al.   Chordoma in childhood and adolescence. A clinicopathologic analysis of 12 cases. Arch Pathol Lab Med  1993; 117: 927– 33 Google Scholar PubMed  52 Dahlin DC, MacCarty CS. Chordoma. Cancer  1952; 5: 1170– 8 Google Scholar CrossRef Search ADS PubMed  53 Tewfik HH, McGinnis WL, Nordstrom DG, et al.   Chordoma: Evaluation of clinical behavior and treatment modalities. Int J Radiat Oncol Biol Phys  1977; 2: 959– 62 http://dx.doi.org/10.1016/0360-3016(77)90194-8 Google Scholar CrossRef Search ADS PubMed  54 Reddy EK, Mansfield CM, Hartman GV. Chordoma. Int J Radiat Oncol Biol Phys  1981; 7: 1709– 11 http://dx.doi.org/10.1016/0360-3016(81)90197-8 Google Scholar CrossRef Search ADS PubMed  55 Hug EB, Sweeney RA, Nurre PM, et al.   Proton radiotherapy in management of pediatric base of skull tumors. Int J Radiat Oncol Biol Phys  2002; 52: 1017– 24 http://dx.doi.org/10.1016/S0360-3016(01)02725-0 Google Scholar CrossRef Search ADS PubMed  56 Igaki H, Tokuuye K, Okumura T, et al.   Clinical results of proton beam therapy for skull base chordoma. Int J Rad Oncol Biol Phys  2004; 60: 1120– 6 http://dx.doi.org/10.1016/j.ijrobp.2004.05.064 Google Scholar CrossRef Search ADS   57 Habrand JL, Bolle S, Datchary J, et al.   La protonthérapie en radiothérapie pédiatrique. Cancer Radiothér  2009; 13: 550– 5 Google Scholar CrossRef Search ADS PubMed  58 Habrand JL, Bondiau PY, Dupuis O, et al.   [Late effects of radiotherapy in children]. Cancer Radiother  1997; 1: 810– 6 Google Scholar CrossRef Search ADS PubMed  59 Bouffet E, Hawkins CE, Ballourah W, et al.   Survival benefit for pediatric patients with recurrent ependymoma treated with reirradiation. Int J Radiat Oncol Biol Phys  2012; 83: 1541– 8 http://dx.doi.org/10.1016/j.ijrobp.2011.10.039 Google Scholar CrossRef Search ADS PubMed  © 2018 American Association of Neuropathologists, Inc. All rights reserved. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Neuropathology & Experimental Neurology Oxford University Press

Loading next page...
 
/lp/ou_press/pediatric-chordomas-results-of-a-multicentric-study-of-40-children-and-eiljimZ0e0
Publisher
American Association of Neuropathologists, Inc.
Copyright
© 2018 American Association of Neuropathologists, Inc. All rights reserved.
ISSN
0022-3069
eISSN
1554-6578
D.O.I.
10.1093/jnen/nlx118
Publisher site
See Article on Publisher Site

Abstract

Abstract Pediatric chordomas are rare malignant neoplasms, and few data are available for optimizing therapeutic strategies and outcome. This study aimed at evaluating how best to manage them and to identify prognostic factors. This multicentric retrospective study included 40 children diagnosed with chordomas between 1966 and 2012. Clinical, radiological, and histopathological data, treatment modalities, and outcomes were reviewed. The median age was 12 years old. Most chordomas were histologically classical forms (45.5%) and were mostly located at the skull base (72.5%). The overall survival (OS) was 66.6% and 58.6%, and progression-free survival (PFS) was 55.7% and 52% at 5 and 10 years, respectively. Total resection was correlated with a better outcome (p = 0.04 for OS and PFS, log-rank). A histopathological/immunohistochemical grading system recently crafted for adults was applied. In a multivariate analysis, it significantly correlated with outcome (PFS and OS, p = 0.004), and the loss of BAF47 immunoexpression appeared to be a significant independent prognostic factor (PFS, p = 0.033). We also identified clinical and histopathological parameters that correlated with prognosis. A new grading system combined with the quality of surgical resection could help classify patients to postpone radiotherapy in case of low risk. Targeted therapy and reirradiation at recurrence may be considered as potential therapeutic strategies. Histopathological grading, Pediatric chordomas, Prognostic algorithm, Radiotherapy INTRODUCTION Chordomas are rare slow-growing malignant neoplasms believed to derive from notochordal remnants. They are most frequently located in the sacrococcygeal and in the skull base regions and particularly affect adults with a peak incidence at diagnosis between 50 and 60 years of age. Pediatric cases represent less than 5% of cases. Prognosis of these tumors essentially depends on their local evolution and their high capacity for recurrence. In terms of treatment, maximal surgical resection is the first step, followed by local irradiation, ideally with proton beam therapy. To date, chemotherapy is not considered as a standard treatment. Publications on pediatric chordomas are rare (250 cases in the literature) (1). While age and extent of resection seem to influence the prognosis in children, no histopathological prognostic factors have yet been clearly identified (2). The aim of this multicentric retrospective study was to evaluate the clinical, neurosurgical, and immunohistochemical prognostic factors of 40 pediatric chordomas. The expression of potentially useful markers for targeted molecular therapies such as VEGF, EGFR, PTEN, pSTAT3, and the loss of BAF47 expression were also investigated. MATERIALS AND METHODS Patients This multicentric and retrospective study included 40 patients younger than 18 years of age from 6 centers, with chordomas diagnosed between 1966 and 2012. Informed consent for the surgery and translational research program was obtained from the parents or guardian according to the IRB approved protocol (number DC-2009-955 for tumor banking). Clinical and radiological findings, the treatment type, and outcome (progression-free survival [PFS] and overall survival [OS]) were analyzed by 2 pediatric neurosurgeons (K.B. and S.P.). The location of chordomas was reported as skull base (tumors of the clivus and of the first cervical vertebra) and non-skull base (tumors from the second cervical vertebra to the sacrum). The quality of surgical resection was assessed from CT and MR scans when available or from surgery, hospitalization, or radiology reports when imaging was not available. The diagnosis of chordoma was based on histopathology, confirmed by Brachyury immunopositivity. Histopathological Review Central pathology review was performed by 2 neuropathologists (P.V. and A.T-E.). Tumors were subtyped into well-differentiated chordomas (classical, chondroid, or mixed subtypes) and poorly differentiated chordomas, as previously defined (2). Histopathological parameters (nuclear pleomorphism, prominent nucleoli, apoptosis, necrosis, hemorrhage, inflammatory cell infiltrates, and mitotic activity) were evaluated as in adult series (2). Immunohistochemistry Immunohistochemical studies were performed using antibodies directed against Ki67 (clone K2; Zytomed, Berlin, Germany, 1:400), p53 (clone Do-7; Dako, Santa Clara, CA, 1:200), E-cadherin (clone EP700Y; Cell Marque, Rocklin, CA, 1:100), EGFR (clone 3C6; Ventana Medical Systems, Tucson, AZ, prediluted), PTEN (clone 6H2-1;Dako, 1: 300), pSTAT3 (clone F2; Santa Cruz Biotechnology, Dallas, TX, 1:100), VEGF (polyclonal; Millipore, Billerica, MA, 1:1000), and BAF47 (clone 25/BAF47; BD-Biosciences, San Jose, CA, 1:200). Positive controls were used and immunostainings were scored independently by 2 observers (P.V. and A.T-E.) as previously described in the adult series (2). Cytoplasmic staining for PTEN was considered homogeneously positive if cytoplasmic staining was present in more than 90% of tumor cells. Molecular Analysis of the SMARCB1 Gene DNA extraction from formaldehyde-fixed paraffin-embedded (FFPE) tumors was performed using the Nucleospin Tissue kit (Macherey-Nagel, Düren, Germany). Six-micrometer-thick dewaxed sections were used for DNA extraction. Samples that underwent decalcification were excluded. Copy number alterations were searched for using a multiplex ligation-dependent probe amplification (MLPA) assay (Salsa MLPA KIT P258-B1 SMARCB1; MRC Holland, Amsterdam, The Netherlands). Overall, 50 ng of tumor DNAs were used for hybridization, ligation, and amplification according to the manufacturer’s instructions in conjunction with a normal control DNA with 2 copies of each exon of the SMARCB1 gene. Two microliters of amplification products were analyzed on an ABI 3130 DNA Analyzer (Applied Biosystems, Foster City, CA). Data analysis was performed using GeneMarker software (SoftGenetics, State College, PA). Small mutation screenings of each coding exon of the SMARCB1 gene were carried out using Sanger sequencing. Sequences were analyzed on an ABI 3130 DNA sequencer (Applied Biosystems) and further aligned with Seqscape analysis software (Applied Biosystems). FISH SMARCB1 was carried out on cases with BAF47 immunoexpression loss but without any deletion in the SMARCB1 gene using MLPA analysis. FISH procedures were performed as previously described. Copy number abnormalities were assessed on 100 nuclei by 2 independent observers. Statistical Analysis We provide the mean and median values and frequency (percentages) for the description of continuous and categorical variables, respectively. Histopathological and immunohistochemical data were combined to compute a grading score as previously described (2). OS was defined as the time in months from the date of diagnosis to death from any cause or to the date of the last follow-up. PFS was defined as the time in months between the diagnosis and the earliest date of disease progression (local and distant), death from any cause if no progression was observed or data cutoff. OS and PFS were estimated using the Kaplan-Meier method and expressed by median with a 95% confidence interval (CI). Survival curves were compared using log-rank test. All variables with p ≤ 0.05 observed in univariate analysis were included in the multivariate Cox regression model with stepwise backward elimination to estimate hazard ratio with a 95% CI (except variables that were used to build the grading system when it was introduced in the model). All statistical tests were 2-sided, and p < 0.05 were regarded as significant. Data were analyzed by using PASW Statistics for Windows, Version 20.0 (SPSS Inc., Chicago, IL). RESULTS Clinical and Surgical Data The median age was 12 years (mean = 11.1, range from 2.3 to 17.8 years) with a male predominance (24 M/16 F). This series included 29 skull base and 11 non-skull base chordomas; there were no sacrococcygeal chordomas, and no patient initially presented with metastases at the time of primary diagnosis. Among the 40 included patients, a total resection of the tumor was initially performed for 4 patients and subtotal resection for 28 patients. In 8 cases, the extent of resection was not available. After total resection, no recurrence was observed at 9 months, 5, 9, and 10 years after diagnosis. Among patients with subtotal resection, 21 patients presented a progression of residual tumor (10 of them died of their disease); 7 patients presented with a stable disease. Adjuvant Treatment Radiotherapy was the principal adjuvant treatment performed after a maximal tumor resection. All except 6 patients were irradiated by photon beam therapy (n = 7), proton beam therapy (n = 6), or a combination of photon and proton beam therapy (n = 20). For 1 patient, the type of radiotherapy was unknown. Mean doses of irradiation were 54 Gy (minimum = 50 Gy and maximum = 55.8 Gy) for photon beam therapy, 70.8 Gy (minimum = 66.6 Gy and maximum = 74 Gy) for proton beam therapy, and 69.4 Gy (minimum = 60 Gy and maximum = 76 Gy) for the combination therapy. Four patients received different protocols of chemotherapy, either in cases of progressive metastatic disease (n = 2) or in cases of initial misdiagnosis (n = 2). Disease control was never obtained, and all patients died some months after the beginning of the treatment. Histopathological Findings Of the 40 patients in the clinical study, we obtained sufficient histopathological material for 33 of them. Furthermore, 45 tumors (33 primary and 12 recurrences) were included for histopathological assessment. Of the 33 primary chordomas, 15 (45.5%) were classical, 1 (3.0%) was chondroid, 12 (36.4%) were mixed, and 5 (15.1%) were poorly differentiated (Fig. 1). In all recurrent chordomas, the histopathological subtype remained unchanged between the first surgery and recurrence. Among the 33 primary chordomas, prominent nucleoli, hemorrhage, apoptosis, necrosis, nuclear pleomorphism, and inflammatory cell infiltrates were present in 6 (18.2%), 7 (21.2%), 11 (33.3%), 14 (42.4%), 20 (60.6%), and 24 (72.7%) of the initial tumors, respectively. Mitotic activity was very low. Figure 1. View largeDownload slide Histopathological findings. (A) Classical (myxoid) chordoma (H&E, 120×). (B) Anti-Brachyury labels the nuclei of the tumor cells but not the stroma (120×). (C) Well-differentiated chordoma with classical (myxoid) and chordoid features (H&E, 120×). (D) Anti-Brachyury labels the nuclei of the tumor cells of the 2 components, 120×). (E) Poorly differentiated pseudosarcomatous chordoma composed of spindle cells (H&E, 90×). (F) Anti-Brachyury labels the nuclei of the majority of tumor cells (90×). (G) Preserved nuclear immunoexpression of BAF47 (120×). (H) Loss of BAF47 nuclear expression in tumor cells of this poorly differentiated pseudosarcomatous chordoma. Note preserved immunoreactivity in endothelial cells (120×). Figure 1. View largeDownload slide Histopathological findings. (A) Classical (myxoid) chordoma (H&E, 120×). (B) Anti-Brachyury labels the nuclei of the tumor cells but not the stroma (120×). (C) Well-differentiated chordoma with classical (myxoid) and chordoid features (H&E, 120×). (D) Anti-Brachyury labels the nuclei of the tumor cells of the 2 components, 120×). (E) Poorly differentiated pseudosarcomatous chordoma composed of spindle cells (H&E, 90×). (F) Anti-Brachyury labels the nuclei of the majority of tumor cells (90×). (G) Preserved nuclear immunoexpression of BAF47 (120×). (H) Loss of BAF47 nuclear expression in tumor cells of this poorly differentiated pseudosarcomatous chordoma. Note preserved immunoreactivity in endothelial cells (120×). Immunohistochemical Findings Insufficient tumor tissue was available to perform all the immunostains in the 45 tumor samples. Nuclear expression of BAF47 was lost in 4 poorly differentiated chordomas and in 3 classical/mixed chordomas (18.9%). The mean Ki67 labeling index (LI) was 14% (1%–60%) and equal or greater than 6% in 14/24 (58.3%) primary chordomas. p53 nuclear accumulation was present in 18/25 (72%) primary chordomas. The mean p53 LI was 29% (1%–91%) and equal to or greater than 25% in 12 primary chordomas. VEGF was intensely expressed in 94.9% of all tumors (37/39). pSTAT3 was expressed in 31/36 (86.1%) of all chordomas with intense staining in 76.7% of the tumors. EGFR was expressed in 35/38 (92.1%) of all chordomas. Immunohistochemical score was greater than 150 in 47.4% of tumors (18/38). Loss of PTEN was observed in 17/23 of tested tumors (73.9%) (Fig. 2). Figure 2. View largeDownload slide Prognostic and therapeutic biomarkers in chordomas. (A) High Ki67 labeling index in a poorly differentiated chordoma (200×). (B) High p53 labeling index up to 25% (150×). (C) Strong immunoreactivity for pSTAT3 in a classical chordoma (200×). (D) Diffuse membranous immunopositivity of E-cadherin in a classical myxoid chordoma (150×). (E) Loss of the immunoexpression of E-cadherin in a poorly differentiated chordoma (150×). (F) Diffuse and strong expression of VEGF in a classical chordoma (150×). (G) Diffuse and strong immunopositivity for EGFR (score 300) by a poorly differentiated chordoma (200×). (H) Diffuse and moderate immunoreactivity for EGFR (score 200) in a classical chordoma (200×). (I) Loss of the expression of PTEN in tumor cells. Note preserved expression in endothelial cells (200×). Figure 2. View largeDownload slide Prognostic and therapeutic biomarkers in chordomas. (A) High Ki67 labeling index in a poorly differentiated chordoma (200×). (B) High p53 labeling index up to 25% (150×). (C) Strong immunoreactivity for pSTAT3 in a classical chordoma (200×). (D) Diffuse membranous immunopositivity of E-cadherin in a classical myxoid chordoma (150×). (E) Loss of the immunoexpression of E-cadherin in a poorly differentiated chordoma (150×). (F) Diffuse and strong expression of VEGF in a classical chordoma (150×). (G) Diffuse and strong immunopositivity for EGFR (score 300) by a poorly differentiated chordoma (200×). (H) Diffuse and moderate immunoreactivity for EGFR (score 200) in a classical chordoma (200×). (I) Loss of the expression of PTEN in tumor cells. Note preserved expression in endothelial cells (200×). Molecular Analysis of the SMARCB1 Gene MLPA and sequencing of the 9 exons of the SMARCB1 gene were performed on FFPE samples from 5 BAF47-deficient tumors. Only 1 case displayed a biallelic loss of function of the SMARCB1 gene with complete loss of all exons of the SMARCB1 gene on 1 allele and an intragenic deletion of exons 7–9 on the other allele resulting in a heterozygous deletion of exons 1–6 and a homozygous deletion of exons 7–9. In 2 other cases, MLPA analysis revealed a decrease of the probes hybridized on the SMARCB1 gene, including 1 case with a very small decrease of the probes suggesting a loss of at least 1 copy of the gene, but the sample analyzed contained a low fraction of tumor cells. No alteration was detected in the SMARCB1 gene in 1 case, but it could be hypothesized that the fraction of tumor cells was too small and below the sensitivity rate of the MLPA. In that last case, no coherent results were obtained due to the poor quality of extracted DNA from the FFPE sample. FISH analysis of these 2 latter cases was also not contributory due to insufficient number of tumor cells. Follow-up and Survival Analysis The OS was 66.6% and 58.6% at 5 and 10 years, respectively, and the PFS was 55.7% and 52% at 5 and 10 years, respectively. After a median follow-up of 5.3 years (min 0.1 years to max 22.3 years), 9 patients (22.5%) were in complete remission and 16 had a stable disease (40%). Fifteen patients (37.5%) died from their disease after a median evolution of 2.3 years (from 0.14 to 5 years). Three patients developed metastasis during follow-up. No recurrence was observed after complete resection. For the others, disease progression occurred earlier after a treatment by surgery alone (mean 453 days) than after an association of surgery and radiotherapy (mean 769 days). The prognosis was better in case of total resection compared to subtotal resection (p = 0.04 for both OS and PFS, log-rank; Table). We observed a better 5-year PFS for skull base chordomas compared to the others (p = 0.01 log-rank; Table). Prognostic factors correlated with OS and PFS in univariate and multivariate analyses are listed in the Table. Grading System As previously described, histopathological and immunohistochemical parameters were combined to establish a grading system (2): Ki67 LI ≥6%, p53 LI ≥25%, presence of a poorly differentiated component, significantly abundant necrosis (≥3% of the tumor surface), prominent nucleoli or apoptosis significantly observed (≥3% of the tumor cells) or mitoses in more than 3/10 HPF were scored 1, absence of these criteria was scored 0. The sum of these criteria defined a score. The low- (scores 0–3) and high-grade (scores 4–7) groups included 17 and 14 tumors, respectively. Two chordomas were excluded from the grading system due to missing data. The classification into 2 groups has proved to be significantly correlated with PFS and OS (Table; Fig. 3). Figure 3. View largeDownload slide Prognostic value of the proposed histopathological and immunohistochemical grading system. Grades significantly correlated to PFS and OS: mean 5 years PFS = 52.9 months for low-grade and 14.3 months for high-grade, p = 0.03; mean 5 years OS = 64.7 months for low-grade and 35.7 months for high-grade, p = 0.002). OS, overall survival; PFS, progression-free survival. Figure 3. View largeDownload slide Prognostic value of the proposed histopathological and immunohistochemical grading system. Grades significantly correlated to PFS and OS: mean 5 years PFS = 52.9 months for low-grade and 14.3 months for high-grade, p = 0.03; mean 5 years OS = 64.7 months for low-grade and 35.7 months for high-grade, p = 0.002). OS, overall survival; PFS, progression-free survival. DISCUSSION Our results agree with those previously published, as classical chordoma is the most frequent subtype in children, followed by pure chondroid chordoma. Poorly differentiated chordomas (46/222, 20.7%), including dedifferentiated chordomas (12/222, 5.4%), and “atypical” forms (34/222, 15.3%), including epithelioid or solid forms (12 cases), have also been described (3–35) but this histopathological subtype is not included in the 2014 WHO classification of tumors of soft tissue and bone (3, 5, 12, 36, 37). In our series, a poorly differentiated component represented a significant prognostic factor in terms of PFS and OS as previously assumed (2, 36) (Table). Among other histopathological parameters, we observed that nuclear pleomorphism, necrosis, and apoptosis were significantly correlated to OS and PFS (Table). The poorer prognostic value of nuclear pleomorphism and necrosis had previously been described in adults (2). However, apoptosis correlated with a shorter survival in our series, contrary to previous studies (2). Ki67 LI remains a controversial prognostic factor in the literature (3). Our results along with those obtained in a previous study indicate that Ki67 LI is higher in pediatric chordomas than in adult cases (2). We also observed a higher p53 LI with no prognostic significance, as opposed to other adult and pediatric studies (2). Finally, loss of E-cadherin expression significantly correlated with a better OS (p = 0.03, log-rank; Table), according to a previous smaller series (2). Loss of BAF47 (SMARCB1) immunoexpression represents a powerful prognostic factor in pediatric chordomas. Originally described in atypical teratoid and rhabdoid tumors, alterations of SMARCB1 gene were also described in chordomas and correlated with more aggressive behavior (38–40). The mean age of previously reported patients with BAF47-deficient chordomas was 6.9 years (1–17 years) (4, 38, 39, 41, 42). Chordomas were predominantly located within the skull base, and 19/26 cases (73.1%) were poorly differentiated with sarcoma-like features (4, 38–42). This loss of expression is due to mutations (38), or large deletions of the SMARCB1 gene (40, 42), as observed in 22.6% of cases in our series. Treatments were extremely variable and progression was observed in 13/15 cases (38, 40–42). The loss of BAF47 immunoreactivity correlated with a poorer prognosis and death occurred within the first 3 years after the initial presentation in 14/20 children (4, 38, 40–42). This poorer prognosis was confirmed in our series (PFS, p < 0.001, log-rank; Table), and has led to considering BAF47 immunoexpression as a major prognostic factor for pediatric chordomas. The prognostic value of the histopathological grading system suggested by our team in the adult series has been confirmed by our pediatric series: patients with a low-grade tumor have a significantly better outcome than patients with a high-grade tumor, both in PFS and OS (Table). Loss of BAF47 immunoexpression was an independent prognostic factor in term of PFS, and it led us to create a prognostic algorithm based on its immunoexpression and data obtained from our grading system (Fig. 4). Considering the significant impact of radical surgery and histopathological grading on outcome, we infer that patients with low-grade tumors for whom a complete resection has been performed have a low risk of recurrence. For these patients, we wonder if radiotherapy is necessary as the first line of treatment and if it could be postponed and subsequently offered only for recurrent tumors. Figure 4. View largeDownload slide Proposed prognostic algorithm of pediatric chordomas. The first step evaluates the status of the expression of BAF47, the worst prognosis being associated with a loss of this immunoexpression. The second step includes the results of the histopathological and immunohistochemical grading system. Figure 4. View largeDownload slide Proposed prognostic algorithm of pediatric chordomas. The first step evaluates the status of the expression of BAF47, the worst prognosis being associated with a loss of this immunoexpression. The second step includes the results of the histopathological and immunohistochemical grading system. We further investigated the expression of molecules that may influence the growth and survival of cancer cells and could lead to targeted therapies. Hypoxia-driven VEGF, a major inducer of tumor angiogenesis, was highly expressed in this series (94.9% of chordomas), as well as in the adult series (2). While EGFR immunolabeling is highly variable in the literature (8%–81%) (2), it was intensely expressed in 47.4% of tumors in our study. These results are in accordance with a previous but small pediatric series (12). Previous studies have reported partial response to both anti-VEGF (43) and anti-EGFR drugs (2) in adults, and anti-EGFR therapy has been shown to inhibit growth in a patient-derived chordoma xenograft (2); therefore, these drugs could potentially be useful in the treatment of chordomas. We also found pSTAT3 expression in 87.5% of primary pediatric chordomas, that is, higher than in adult studies (2). Inhibitors of the STAT3 pathway increased the rate of apoptosis in chemoresistant cancer cell lines and decreased the growth rate of chordoma cell lines (2), so this biomarker could be considered an interesting therapeutic target. Loss of PTEN expression has previously been observed in chordoma samples and is correlated with tumor invasion (44). In our study, a loss of expression in the majority of chordomas (73.9% of all cases), was demonstrated, concordant with other studies in which PTEN deletion and loss of the PTEN gene were reported (45, 46). Treatment of chordomas is based on multidisciplinary management, and all authors agree on the fact that surgery is the first step of this treatment (47–50). Surgery has 2 aims: to reduce the tumor volume and, if possible, to obtain complete resection, and to keep any residue away from vascular or nervous anatomical structures for adjuvant radiotherapy. Limited data exist on pediatric series due to the small size of cohorts but Ridenour et al observed a tendency for a better OS in case of total resection compared to subtotal resection (n = 35, p = 0.374) (3). In our series, this difference was significant in terms of OS and PFS (p = 0.04, log-rank; Table). Complete resection of these tumors remains challenging and varies from 0% to 36.4% in the main pediatric studies (22.5% in our series) (3, 47, 49, 51). High-dose radiotherapy is considered to be the standard treatment after surgery (30, 52–54). Limited data are available for children but in a review on intracranial chordomas, the benefit of radiotherapy was clearly established (OS, p = 0.004) (36). Proton therapy has become the technique of choice (55, 56) because it allows for higher doses of radiation on tumor tissue while sparing adjacent healthy tissue (5, 57, 58). In our cohort, patients who were irradiated had a better outcome in terms of PFS and OS (p < 0.001, log-rank, Table), but patients who were not irradiated died rapidly or presented a rapid growth of the tumor. Thus, it is difficult to conclude on the direct consequences of the absence of irradiation of our patients. Considering the high potential for local recurrence of chordomas, the benefits of reirradiation may also have to be evaluated in the case of progressive disease, as is recommended for other pediatric tumors (59). In conclusion, our results provide evidence that the aforementioned clinical and histopathological parameters are correlated with prognosis. Our grading system combined with the quality of surgical resection could help in categorizing patients and postponing radiotherapy in low-risk cases. The immunoexpression of biomarkers implicated in targeted therapy (EGFR, VEGF, and STAT3 pathways) could help in the selection of therapeutic strategies among those highly expressed. Table. Univariate and Multivariate Analysis for Prognostic Factors     Univariate Analysis       n  5-Year PFS (%)  p Value  5-Year OS (%)  p Value  Location  SB  29  56.2  0.01  0.9  0.42    NSB  11  27.3    60.6    Extent of surgery                STR  33  40.7  0.04  61.3  0.04    TR  7  85.7    100    Treatment                No RxT  6  0  <0.001  22.2  <0.001    Photon Rxt  7  71.4    71.4      PT ± Photon Rxt  26  55.1    79.3    Nuclear pleomorphism                Absent  13  65.8  0.04  91.7  0.04    Present  18  31.1    47.6    Histopathological subtype                Well differentiated  24  50.7  0.01  77.2  <0.01    Poorly differentiated  9  22.2    25.4    Apoptosis                Absent  20  61.7  <0.01  83.9  <0.01    Present  11  18.2    34.1    Necrosis                Absent  19  65.7  <0.01  88.5  <0.01    Present  12  12.5    31.3    Grade                Low  17  52.9  0.03  64.7  0.002    High  14  14.3    35.7    pSTAT3 expression                Absent or weak  7  14.3  0.02  21.4  0.01    High  23  53.4    77.3    E-cadherin expression                < 50%  28  46.6  0.21  73.4  0.03    ≥ 50%  5  20    20    BAF47 expression                Lost  7  14.3  <0.01  51.4  0.27    Preserved  24  50.9    68.7        Multivariate analysis  PFS    Grade  <0.01            Extent of surgery  0.01            Nuclear pleomorphism  0.02            BAF47 expression  0.03        OS    Grade  <0.01            E-cadherin expression  0.02            Univariate Analysis       n  5-Year PFS (%)  p Value  5-Year OS (%)  p Value  Location  SB  29  56.2  0.01  0.9  0.42    NSB  11  27.3    60.6    Extent of surgery                STR  33  40.7  0.04  61.3  0.04    TR  7  85.7    100    Treatment                No RxT  6  0  <0.001  22.2  <0.001    Photon Rxt  7  71.4    71.4      PT ± Photon Rxt  26  55.1    79.3    Nuclear pleomorphism                Absent  13  65.8  0.04  91.7  0.04    Present  18  31.1    47.6    Histopathological subtype                Well differentiated  24  50.7  0.01  77.2  <0.01    Poorly differentiated  9  22.2    25.4    Apoptosis                Absent  20  61.7  <0.01  83.9  <0.01    Present  11  18.2    34.1    Necrosis                Absent  19  65.7  <0.01  88.5  <0.01    Present  12  12.5    31.3    Grade                Low  17  52.9  0.03  64.7  0.002    High  14  14.3    35.7    pSTAT3 expression                Absent or weak  7  14.3  0.02  21.4  0.01    High  23  53.4    77.3    E-cadherin expression                < 50%  28  46.6  0.21  73.4  0.03    ≥ 50%  5  20    20    BAF47 expression                Lost  7  14.3  <0.01  51.4  0.27    Preserved  24  50.9    68.7        Multivariate analysis  PFS    Grade  <0.01            Extent of surgery  0.01            Nuclear pleomorphism  0.02            BAF47 expression  0.03        OS    Grade  <0.01            E-cadherin expression  0.02        NSB, non-skull base; OS, overall survival; PFS, progression-free survival; Photon RxT, Photon beam radiotherapy; PT, Proton beam radiotherapy; RxT, radiotherapy; SB, skull base; STR, subtotal resection; TR, total resection. ACKNOWLEDGMENTS The authors are grateful to all laboratory technicians for their technical assistance. They thank Albane Gareton for assistance in translation. Part of this study has been presented at the 24th ESPN Congress in Rome, Italy, and at the ISPN 2016 44th Annual Meeting in Kobe, Japan. REFERENCES 1 Beccaria K, Sainte-Rose C, Zerah M, et al.   Paediatric chordomas. Orphanet J Rare Dis  2015; 10: 116 http://dx.doi.org/10.1186/s13023-015-0340-8 Google Scholar CrossRef Search ADS PubMed  2 Tauziède-Espariat A, Bresson D, Polivka M, et al.   Prognostic and therapeutic markers in chordomas: A study of 287 tumors. J Neuropathol Exp Neurol  2016; 75: 111– 20 Google Scholar CrossRef Search ADS PubMed  3 Ridenour RVIII, Ahrens WA, Folpe AL, et al.   Clinical and histopathologic features of chordomas in children and young adults. Pediatr Dev Pathol  2010; 13: 9– 17 http://dx.doi.org/10.2350/09-01-0584.1 Google Scholar CrossRef Search ADS PubMed  4 Yadav R, Sharma MC, Malgulwar PB, et al.   Prognostic value of MIB-1, p53, epidermal growth factor receptor, and INI1 in childhood chordomas. Neuro-oncology  2014; 16: 372– 81 http://dx.doi.org/10.1093/neuonc/not228 Google Scholar CrossRef Search ADS PubMed  5 Hoch BL, Nielsen GP, Liebsch NJ, et al.   Base of skull chordomas in children and adolescents: A clinicopathologic study of 73 cases. Am J Surg Pathol  2006; 30: 811– 8 http://dx.doi.org/10.1097/01.pas.0000209828.39477.ab Google Scholar CrossRef Search ADS PubMed  6 Moore KA, Bohnstedt BN, Shah SU, et al.   Intracranial chordoma presenting as acute hemorrhage in a child: Case report and literature review. Surg Neurol Int  2015; 6: 63 http://dx.doi.org/10.4103/2152-7806.155445 Google Scholar CrossRef Search ADS PubMed  7 Hashim H, Rosman AK, Abdul Aziz A, et al.   Atypical clival chordoma in an adolescent without imaging evidence of bone involvement. Malays J Med Sci  2014; 21: 78– 82 Google Scholar PubMed  8 Dhall G, Traverso M, Finlay JL, et al.   The role of chemotherapy in pediatric clival chordomas. J Neurooncol  2011; 103: 657– 62 http://dx.doi.org/10.1007/s11060-010-0441-0 Google Scholar CrossRef Search ADS PubMed  9 Schechter MM, Liebeskind AL, Azar-Kia B. Intracranial chordomas. Neuroradiol  1974; 8: 67– 82 http://dx.doi.org/10.1007/BF00345038 Google Scholar CrossRef Search ADS   10 Fink FM, Ausserer B, Schröcksnadel W, et al.   Clivus chordoma in a 9-year-old child: Case report and review of the literature. Pediatr Hematol Oncol  1987; 4: 91– 100 Google Scholar CrossRef Search ADS PubMed  11 Inagaki H, Anno Y, Hori T, et al.   [Clival chordoma in an infant; case report and review of the literature]. No Shinkei Geka  1992; 20: 809– 13 Google Scholar PubMed  12 Yadav YR, Kak VK, Khosla VK, et al.   Cranial chordoma in the first decade. Clin Neurol Neurosurg  1992; 94: 241– 6 http://dx.doi.org/10.1016/0303-8467(92)90096-L Google Scholar CrossRef Search ADS PubMed  13 Schamschula RG, Soo MY. Clival chordomas. Australas Radiol  1993; 37: 259– 64 http://dx.doi.org/10.1111/j.1440-1673.1993.tb00069.x Google Scholar CrossRef Search ADS PubMed  14 Auger M, Raney B, Callender D, et al.   Metastatic intracranial chordoma in a child with massive pulmonary tumor emboli. Pediatr Pathol  1994; 14: 763– 70 http://dx.doi.org/10.3109/15513819409037673 Google Scholar CrossRef Search ADS PubMed  15 Scimeca PG, James-Herry AG, Black KS, et al.   Chemotherapeutic treatment of malignant chordoma in children. J Pediatr Hematol Oncol  1996; 18: 237– 40 http://dx.doi.org/10.1097/00043426-199605000-00032 Google Scholar CrossRef Search ADS PubMed  16 Soo MY. Chordoma: Review of clinicoradiological features and factors affecting survival. Australas Radiol  2001; 45: 427– 34 http://dx.doi.org/10.1046/j.1440-1673.2001.00950.x Google Scholar CrossRef Search ADS PubMed  17 Lountzis NI, Hogarty MD, Kim HJ, et al.   Cutaneous metastatic chordoma with concomitant tuberous sclerosis. J Am Acad Dermatol  2006; 55: S6– S10 Google Scholar CrossRef Search ADS PubMed  18 Yoneoka Y, Tsumanuma I, Fukuda M, et al.   Cranial base chordoma—long term outcome and review of the literature. Acta Neurochir (Wien)  2008; 150: 773– 8 Google Scholar CrossRef Search ADS PubMed  19 Chang SW, Gore PA, Nakaji P, et al.   Juvenile intradural chordoma: Case report. Neurosurgery  2008; 62:E525–6; discussion: E527 20 Al-Rahawan MM, Siebert JD, Mitchell CS, et al.   Durable complete response to chemotherapy in an infant with a clival chordoma. Pediatr Blood Cancer  2011; 59: 323– 5 Google Scholar CrossRef Search ADS PubMed  21 Richards AT, Stricke L, Spitz L. Sacrococcygeal chordomas in children. J Pediatr Surg  1973; 8: 911– 4 http://dx.doi.org/10.1016/0022-3468(73)90010-9 Google Scholar CrossRef Search ADS PubMed  22 Garofalo E, Minerva A, Baltieri G. [Sacrococcygeal chordoma in a 19-month-old boy]. Minerva Pediatr  1976; 28: 1909– 14 Google Scholar PubMed  23 Nix WL, Steuber CP, Hawkins EP, et al.   Sacrococcygeal chordoma in a neonate with multiple anomalies. J Pediat  1978; 93: 995– 8 http://dx.doi.org/10.1016/S0022-3476(78)81235-9 Google Scholar CrossRef Search ADS PubMed  24 Kozlowski K, Barylak A, Campbell J, et al.   Primary sacral bone tumours in children (report of 16 cases with a short literature review). Australas Radiol  1990; 34: 142– 9 http://dx.doi.org/10.1111/j.1440-1673.1990.tb02830.x Google Scholar CrossRef Search ADS PubMed  25 Cable DG, Moir C. Pediatric sacrococcygeal chordomas: A rare tumor to be differentiated from sacrococcygeal teratoma. J Pediatr Surg  1997; 32: 759– 61 http://dx.doi.org/10.1016/S0022-3468(97)90028-2 Google Scholar CrossRef Search ADS PubMed  26 Shinmura Y, Miura K, Yajima S, et al.   Sacrococcygeal chordoma in infancy showing an aggressive clinical course: An autopsy case report. Pathol Int  2003; 53: 473– 7 http://dx.doi.org/10.1046/j.1440-1827.2003.01496.x Google Scholar CrossRef Search ADS PubMed  27 Al-Adra D, Bennett A, Gill R, et al.   Pediatric metastatic sacrococcygeal chordoma treated with surgery. Eur J Pediatr Surg  2011; 21: 196– 8 http://dx.doi.org/10.1055/s-0031-1271635 Google Scholar CrossRef Search ADS PubMed  28 Windeyer BW. Chordoma. Proc R Soc Med  1959; 52: 1088– 100 Google Scholar PubMed  29 Proyard G, DONY H. [Chordoma of the cervical spine]. Acta Chir Belg  1965; 64: 132– 9 Google Scholar PubMed  30 Higinbotham NL, Phillips RF, Farr HW, et al.   Chordoma. Thirty-five-year study at Memorial Hospital. Cancer  1967; 20: 1841– 50 http://dx.doi.org/10.1002/1097-0142(196711)20:11<1841::AID-CNCR2820201107>3.0.CO;2-2 Google Scholar CrossRef Search ADS PubMed  31 Occhipinti E, Mastrostefano R, Pompili A, et al.   Spinal chordomas in infancy. Report of a case and analysis of the literature. Childs Brain  1981; 8: 198– 206 Google Scholar PubMed  32 Huang S-M, Chen C-C, Chiu P-C, et al.   Unusual presentation of posterior mediastinal chordoma in a 2-year-old boy. J Pediatr Hematol Oncol  2003; 25: 743– 6 http://dx.doi.org/10.1097/00043426-200309000-00014 Google Scholar CrossRef Search ADS PubMed  33 Killampalli VV, Power D, Stirling AJ. Preadolescent presentation of a lumbar chordoma: Results of vertebrectomy and fibula strut graft reconstruction at 8 years. Eur Spine J  2006; 15: 621– 5 http://dx.doi.org/10.1007/s00586-006-0138-4 Google Scholar CrossRef Search ADS PubMed  34 Wang Y, Xiao J, Wu Z, et al.   Primary chordomas of the cervical spine: A consecutive series of 14 surgically managed cases. J Neurosurg Spine  2012; 17: 292– 9 http://dx.doi.org/10.3171/2012.7.SPINE12175 Google Scholar CrossRef Search ADS PubMed  35 Hart ES. Cervical spine chordoma. Orthop Nurs  2012; 31: 355– 6 http://dx.doi.org/10.1097/NOR.0b013e318274272f Google Scholar CrossRef Search ADS PubMed  36 Borba LA, Al-Mefty O, Mrak RE, et al.   Cranial chordomas in children and adolescents. J Neurosurg  1996; 84: 584– 91 http://dx.doi.org/10.3171/jns.1996.84.4.0584 Google Scholar CrossRef Search ADS PubMed  37 Jo VY, Fletcher CDM. WHO classification of soft tissue tumours: An update based on the 2013 (4th) edition. Pathology  2014; 46: 95– 104 http://dx.doi.org/10.1097/PAT.0000000000000050 Google Scholar CrossRef Search ADS PubMed  38 Antonelli M, Raso A, Mascelli S, et al.   SMARCB1/INI1 Involvement in pediatric chordoma: A mutational and immunohistochemical analysis. Am J Surg Pathol  2017; 41: 56– 61 http://dx.doi.org/10.1097/PAS.0000000000000741 Google Scholar CrossRef Search ADS PubMed  39 Chavez JA, Nasir Ud D, Memon A, et al.   Anaplastic chordoma with loss of INI1 and Brachyury expression in a 2-year-old girl. Clin Neuropathol  2014; 33: 418– 20 http://dx.doi.org/10.5414/NP300724 Google Scholar CrossRef Search ADS PubMed  40 Hasselblatt M, Thomas C, Hovestadt V, et al.   Poorly differentiated chordoma with SMARCB1/INI1 loss: A distinct molecular entity with dismal prognosis. Acta Neuropathol  2016; 132: 149– 51 http://dx.doi.org/10.1007/s00401-016-1574-9 Google Scholar CrossRef Search ADS PubMed  41 Renard C, Pissaloux D, Decouvelaere AV, et al.   Non-rhabdoid pediatric SMARCB1-deficient tumors: Overlap between chordomas and malignant rhabdoid tumors? Cancer Genet  2014; 207: 384– 9 Google Scholar CrossRef Search ADS PubMed  42 Mobley BC, McKenney JK, Bangs CD, et al.   Loss of SMARCB1/INI1 expression in poorly differentiated chordomas. Acta Neuropathol  2010; 120: 745– 53 http://dx.doi.org/10.1007/s00401-010-0767-x Google Scholar CrossRef Search ADS PubMed  43 Asklund T, Sandström M, Shahidi S, et al.   Durable stabilization of three chordoma cases by bevacizumab and erlotinib. Acta Oncol  2014; 53: 980– 4 Google Scholar CrossRef Search ADS PubMed  44 Chen K, Mo J, Zhou M, et al.   Expression of PTEN and mTOR in sacral chordoma and association with poor prognosis. Med Oncol  2014; 31: 886– 5 http://dx.doi.org/10.1007/s12032-014-0886-7 Google Scholar CrossRef Search ADS PubMed  45 Choy E, MacConaill LE, Cote GM, et al.   Genotyping cancer-associated genes in chordoma identifies mutations in oncogenes and areas of chromosomal loss involving CDKN2A, PTEN, and SMARCB1. PLoS ONE  2014; 9: e101283 Google Scholar CrossRef Search ADS PubMed  46 Horbinski C, Oakley GJ, Cieply K, et al.   The prognostic value of Ki-67, p53, epidermal growth factor receptor, 1p36, 9p21, 10q23, and 17p13 in skull base chordomas. Arch Pathol Lab Med  2010; 134: 1170– 6 Google Scholar PubMed  47 Wold LE, Laws ER. Cranial chordomas in children and young adults. J Neurosurg  1983; 59: 1043– 7 http://dx.doi.org/10.3171/jns.1983.59.6.1043 Google Scholar CrossRef Search ADS PubMed  48 Heffelfinger MJ, Dahlin DC, MacCarty CS, et al.   Chordomas and cartilaginous tumors at the skull base. Cancer  1973; 32: 410– 20 http://dx.doi.org/10.1002/1097-0142(197308)32:2<410::AID-CNCR2820320219>3.0.CO;2-S Google Scholar CrossRef Search ADS PubMed  49 Benk V, Liebsch NJ, Munzenrider JE, et al.   Base of skull and cervical spine chordomas in children treated by high-dose irradiation. Int J Radiat Oncol Biol Phys  1995; 31: 577– 81 http://dx.doi.org/10.1016/0360-3016(94)00395-2 Google Scholar CrossRef Search ADS PubMed  50 Carpentier A, Polivka M, Blanquet A, et al.   Suboccipital and cervical chordomas: The value of aggressive treatment at first presentation of the disease. J Neurosurg  2002; 97: 1070– 7 http://dx.doi.org/10.3171/jns.2002.97.5.1070 Google Scholar CrossRef Search ADS PubMed  51 Coffin CM, Swanson PE, Wick MR, et al.   Chordoma in childhood and adolescence. A clinicopathologic analysis of 12 cases. Arch Pathol Lab Med  1993; 117: 927– 33 Google Scholar PubMed  52 Dahlin DC, MacCarty CS. Chordoma. Cancer  1952; 5: 1170– 8 Google Scholar CrossRef Search ADS PubMed  53 Tewfik HH, McGinnis WL, Nordstrom DG, et al.   Chordoma: Evaluation of clinical behavior and treatment modalities. Int J Radiat Oncol Biol Phys  1977; 2: 959– 62 http://dx.doi.org/10.1016/0360-3016(77)90194-8 Google Scholar CrossRef Search ADS PubMed  54 Reddy EK, Mansfield CM, Hartman GV. Chordoma. Int J Radiat Oncol Biol Phys  1981; 7: 1709– 11 http://dx.doi.org/10.1016/0360-3016(81)90197-8 Google Scholar CrossRef Search ADS PubMed  55 Hug EB, Sweeney RA, Nurre PM, et al.   Proton radiotherapy in management of pediatric base of skull tumors. Int J Radiat Oncol Biol Phys  2002; 52: 1017– 24 http://dx.doi.org/10.1016/S0360-3016(01)02725-0 Google Scholar CrossRef Search ADS PubMed  56 Igaki H, Tokuuye K, Okumura T, et al.   Clinical results of proton beam therapy for skull base chordoma. Int J Rad Oncol Biol Phys  2004; 60: 1120– 6 http://dx.doi.org/10.1016/j.ijrobp.2004.05.064 Google Scholar CrossRef Search ADS   57 Habrand JL, Bolle S, Datchary J, et al.   La protonthérapie en radiothérapie pédiatrique. Cancer Radiothér  2009; 13: 550– 5 Google Scholar CrossRef Search ADS PubMed  58 Habrand JL, Bondiau PY, Dupuis O, et al.   [Late effects of radiotherapy in children]. Cancer Radiother  1997; 1: 810– 6 Google Scholar CrossRef Search ADS PubMed  59 Bouffet E, Hawkins CE, Ballourah W, et al.   Survival benefit for pediatric patients with recurrent ependymoma treated with reirradiation. Int J Radiat Oncol Biol Phys  2012; 83: 1541– 8 http://dx.doi.org/10.1016/j.ijrobp.2011.10.039 Google Scholar CrossRef Search ADS PubMed  © 2018 American Association of Neuropathologists, Inc. All rights reserved.

Journal

Journal of Neuropathology & Experimental NeurologyOxford University Press

Published: Mar 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off