Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Proliferation Index: A Continuous Model to Predict Prognosis in Patients with Tumours of the Ewing's Sarcoma Family

  • Samantha Brownhill ,

    s.c.brownhill@leeds.ac.uk

    Affiliation Children's Cancer Research Group, Leeds Institute of Cancer and Pathology, St James's University Hospital, Leeds, United Kingdom

    ⨯
  • Dena Cohen,

    Affiliation Clinical Trials Research Unit, Leeds Institute of Cancer and Pathology, St James's University Hospital, Leeds, United Kingdom

    ⨯
  • Sue Burchill

    Affiliation Children's Cancer Research Group, Leeds Institute of Cancer and Pathology, St James's University Hospital, Leeds, United Kingdom

    ⨯

Abstract

The prognostic value of proliferation index (PI) and apoptotic index (AI), caspase-8, -9 and -10 expression have been investigated in primary Ewing's sarcoma family of tumours (ESFT). Proliferating cells, detected by immunohistochemistry for Ki-67, were identified in 91% (91/100) of tumours with a median PI of 14 (range 0–87). Apoptotic cells, identified using the TUNEL assay, were detected in 96% (76/79) of ESFT; the median AI was 3 (range 0–33). Caspase-8 protein expression was negative (0) in 14% (11/79), low (1) in 33% (26/79), medium (2) in 38% (30/79) and high (3) in 15% (12/79) of tumours, caspase-9 expression was low (1) in 66% (39/59) and high (3) in 34% (20/59), and caspase-10 protein was low (1) in 37% (23/62) and negative (0) in 63% (39/62) of primary ESFT. There was no apparent relationship between caspase-8, -9 and -10 expression, PI and AI.

PI was predictive of relapse-free survival (RFS; p = 0.011) and overall survival (OS; p = <0.001) in a continuous model, whereas AI did not predict outcome. Patients with tumours expressing low levels of caspase-9 protein had a trend towards a worse RFS than patients with tumours expressing higher levels of caspase-9 protein (p = 0.054, log rank test), although expression of caspases-8, -9 and/or -10 did not significantly predict RFS or OS. In a multivariate analysis model that included tumour site, tumour volume, the presence of metastatic disease at diagnosis, PI and AI, PI independently predicts OS (p = 0.003). Consistent with previous publications, patients with pelvic tumours had a significantly worse OS than patients with tumours at other sites (p = 0.028); patients with a pelvic tumour and a PI≥20 had a 6 fold-increased risk of death. These studies advocate the evaluation of PI in a risk model of outcome for patients with ESFT.

Citation: Brownhill S, Cohen D, Burchill S (2014) Proliferation Index: A Continuous Model to Predict Prognosis in Patients with Tumours of the Ewing's Sarcoma Family. PLoS ONE 9(8): e104106. https://doi.org/10.1371/journal.pone.0104106

Editor: David Loeb, Johns Hopkins University, United States of America

Received: February 19, 2014; Accepted: July 9, 2014; Published: August 26, 2014

Copyright: © 2014 Brownhill et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: This work was funded by Cancer Research UK (CA1010/A2862; http://www.cancerresearchuk.org) and the Bone Cancer Research Trust (BCRT; http://www.bcrt.org.uk). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

Ewing's sarcoma family of tumours (ESFT) can arise in bone or soft tissue sites at any age, but most frequently are diagnosed in bony sites in children and young adults [1]. Five year survival rates for patients diagnosed with localised disease are between 60 and 70%, although outcome for patients with metastasis is lower despite multimodality treatment incorporating combination chemotherapy, surgery and radiotherapy. Since improved patient outcome is anticipated by adapting therapy based on risk, a number of prognostic clinical factors have been described including the presence of metastatic disease at the time of diagnosis [2], tumour volume (greater than 200 ml) [3] and pelvic primary tumours [4], [5] at diagnosis in patients with localised disease. Despite these observations there is currently no universally accepted informative staging system in ESFT.

High proliferation index (PI), is reported to predict poor outcome in colon carcinoma [6], renal cell carcinoma [7], cervical carcinoma [8], neuroblastoma [9], bladder [10] and breast [11]–[13] cancer, as well as ESFT [14]–[16]. Although apoptotic index (AI) affects tumour viability and growth, the relationship between AI and prognosis is controversial and has not previously been investigated in ESFT. High AI is predictive of increased survival in patients with osteosarcoma [17] and gastric cancer [18], and low AI has been associated with high grade tumours of the ovary [19], kidney [20] and colon [6]. In contrast, low AI has been associated with a higher mean survival in childhood acute lymphoblastic leukaemia (ALL) [21] and patients with colorectal carcinoma [22].

Since tumour growth reflects the number of both proliferating and apoptotic cells, we have hypothesised that PI and AI may more reliably predict outcome than PI alone. This hypothesis is supported by studies in patients with adenocarcinoma of the uterine cervix, where the ratio of PI/AI was predictive of survival but not PI or AI alone [23]. The relationship between expression of the effector caspases, PI, AI and outcome is poorly investigated, caspase-8 being the most frequently studied. Methylation of the CASP8 gene has been associated with low levels of caspase-8 expression, reduced response to chemotherapeutic agents and poor outcome in a number of different cancer types including medulloblastoma [24] and neuroblastoma [25], [26], although this remains controversial at least in neuroblastoma [27]. Lack or low caspase-8 expression has also been reported in a number of other cancer types including osteosarcoma [28] and squamous cell carcinoma of the tongue [29]. In ESFT cell lines [30] and tumours [31], expression is variable [32], [33] and is reported to have no correlation with outcome [33].

Although caspase-9 and -10 are executors of apoptosis, little is known of their prognostic value in primary tumours. Expression of cleaved caspase-9 protein correlates with a longer OS in patients with Hodgkin's lymphoma [34], although the clinical significance of low caspase-9 expression in colon carcinoma [35], medulloblastoma [24] and gastric carcinoma [36] remains unclear. Expression of caspase-10 is also low in gastric carcinoma [36], rectal [37] and colorectal [38] cancers. The aims of this study were to examine PI, AI and expression of caspases-8, -9 and -10 in a panel of primary ESFT to evaluate and compare their prognostic power.

Materials and Methods

Clinical samples

Tumour samples were collected at diagnosis from 105 patients with ESFT, diagnosis was confirmed by reverse transcriptase polymerase chain reaction (RT-PCR) for the EWS-ETS gene rearrangements in 80/105 tumours. Of the remaining 25 tumour samples, CD99 expression was confirmed by immunohistochemistry (IHC) for MIC-2 in 25/25 tumours. It was not possible to analyse all tumours for all markers due to limited amounts of material; suitable material was available to analyse PI in 100/105, AI in 79/105, caspase-8 in 79/105, caspase-9 in 59/105 and caspase-10 in 62/105 of cases. The presence of metastases was detected by conventional imaging and examination of bone marrow aspirates by light microscopy.

Ethics statement

Ethical approval for this study was obtained from Trent Multi-centre Research Ethics Committee (MREC 98/4/023; MREC 98/0/44). Informed written consent was obtained from all participants.

Immunohistochemistry for Ki-67, caspases-8, -9 and -10

Immunohistochemistry and antigen retrieval were optimised using sections (5 µm) of frozen or formalin fixed-paraffin embedded (FF-PE) ESFT cell pellets; sections of TC-32 cell pellets were included as positive controls in subsequent assays. Sequential sections (5 µm) of frozen (n = 86) and FF-PE (n = 19) primary ESFT were analysed. FF-PE sections were hydrated and antigen retrieval was performed by boiling the sections in citric acid buffer for 10 minutes prior to analysis. Sections processed without primary antibody or with isotype matched antibody were included as negative controls. All sections were counterstained with haematoxylin and mounted in DePeX mounting medium.

Proliferating cells were detected by immunohistochemistry for Ki-67 using a mouse monoclonal antibody (1∶100; Dako, Cambridgeshire, UK) and 3 stage peroxidase [39]. PI = [number of Ki-67 positive cells ÷ number of cells scored]×100; 100 cells were scored in 3 different fields of view by two independent observers.

Caspase-8 and caspase-9 proteins were detected using the Rabbit Envision Kit (Dako) and caspase-8 (1 in 50; Santa Cruz, Heidelberg, Germany) or caspase-9 (1 in 100; Abcam, Cambridge, UK) antibodies. Prior to incubation with primary antibody overnight at 4°C, sections were incubated with normal goat serum (1 in 10; Dako) for 1 hour at room temperature, wash steps were performed with TBS plus 1% Tween 20. Caspase-10 protein was detected using caspase-10 primary antibody (1 in 150; Santa Cruz) and the Goat Vectastain ABC kit (Vector Laboratories, Peterborough, UK); sections were incubated with primary antibody for one hour at 37°C and antibody was visualised using DAB+ substrate (Dako).

Identification of apoptotic cells by terminal deoxynucleotidyl transferase (TdT) mediated dUTP nick-end labelling (TUNEL)

Apoptotic cells were detected by TUNEL using the ApopTag Peroxidase In Situ Apoptosis Detection Kit (Millipore, UK), according to manufacturer's instructions. AI = [number of TUNEL positive cells ÷ number of cells scored]×100; 100 cells were scored in three different fields of view by two independent observers.

Scoring of immunohistochemistry

Sections were visualised by light microscopy (Zeiss Axioplan microscope, Zeiss, UK) and scored using the following criteria. Caspase-8 expression was negative (0), low (1), medium (2) or high (3) based on intensity of staining compared to a caspase-8 positive (2, tumour 11) and negative (0, tumour 45) tissue. Caspase-9 expression was scored as negative (0), low (1), medium (2) or high (3) dependant on staining intensity compared to tissue with high (3, tumour 55) or low (1, tumour 57) expression, and caspase-10 as negative (0), low (1), medium (2) or high (3) compared to a caspase-10 expressing tumour (1, tumour 1) and a negative (0, tumour 57). Staining was independently evaluated by two observers.

Statistical analysis

All statistical analyses were carried out using SAS v9.2 (SAS Institute Inc., Cary, NC, USA). RFS was calculated from the date of diagnosis to relapse or death, or the date of last assessment without event. OS was defined as the time from diagnosis to death or the date last seen alive. Univariate analyses explored whether a number of covariate factors predict prognosis: age group (<14, ≥14) [40]; tumour site (pelvic, other) [4]; tumour volume (<200 ml, ≥200 ml) [3]; metastases at diagnosis (any, none) [2]; response to treatment (<90%, ≥90%) [41]; PI (as a continuous variable); AI (as a continuous variable); caspases-8, -9 and -10 (0/1, 2/3). RFS and OS curves were calculated using the Kaplan-Meier method by each covariate group, and the results for each compared using the log rank test. Caution was taken in statistical interpretation to minimise the effect of multiplicity, and the p-values were compared to 0.01 in order to control the type 1 error rate.

Cox proportional hazards regression was carried out on OS, incorporating the most significant covariates to assess which are prognostic in a multivariate setting. A stepwise procedure was used, with significance level for entering an explanatory variable into the model set to 0.05 and significance level for removing an explanatory variable from the model set to 0.1.

An investigation was carried out into the optimal cut-point choice for PI. PI was investigated as a continuous covariate within a Cox model; the log-rank value was calculated over the whole range of cut-points to visualise any patterns, and the cut-point log rank values were compared to those when the data is described as a continuous covariate. Fractional polynomials were calculated to investigate how the data could best be described as a continuous covariate, using a series of predefined transformations of predictor variables [42]. If the cut-point describes the data significantly better than a fitted model, it is likely that the cut-point effect is real; otherwise the survival may be related to the PI on a continuous scale.

Results

Frozen and FF-PE tumour distribution with age at diagnosis, tumour site and PI confirmed that frozen and FF-PE tumours could be considered as a single sample set (Table S1). Patient and tumour details, PI, AI and protein expression profiles are summarised in Table S2.

Prognostic power of clinical parameters

The prognostic power of parameters previously reported to be of clinical value was investigated in this patient cohort. Consistent with previous studies a tumour volume ≥200 ml was predictive of a poor OS and RFS (p = 0.019 and 0.013 respectively, log rank test, Figure 1A) compared to a tumour volume <200 ml. Furthermore, patients with pelvic tumours had a worse OS and RFS (p = 0.019 and 0.009 respectively, log rank test, Figure 1B) than patients with tumours presenting in other sites. Age of the patient (<14 years compared to ≥14 years) and response to treatment (<90% necrosis of tumour post-treatment compared to ≥90% necrosis) did not predict outcome in this cohort. The presence of metastatic disease at diagnosis detected by conventional imaging and examination of bone marrow aspirates by light microscopy for tumour cells did not predict outcome, although there was a non-significant trend towards patients with metastasis at diagnosis having a worse OS (p = 0.099, log rank test). Site of metastasis was available for 20/32 (63%) patients with metastasis; patients with bone metastasis at diagnosis had a significantly worse OS and RFS (p = <0.001 and <0.001 respectively, log rank test) than patients with no metastatic disease or metastasis at other sites.

thumbnail
Download:
Figure 1. A. Kaplan Meier survival plots comparing the overall and relapse-free survival of patients with a tumour volume ≥200 ml with that of patients with a tumour volume <200 ml, p = 0.019 and 0.0133 respectively; log rank test. Crosses indicate censored events.

B. Kaplan Meier survival plots to compare the overall and relapse-free survival of patients with pelvic tumours to that of patients with tumours at other sites, p = 0.0185 and 0.0088 respectively; log rank test. Crosses indicate censored events.

https://doi.org/10.1371/journal.pone.0104106.g001

Proliferation and apoptosis in primary ESFT

Proliferating cells were identified in 91% (91/100) of tumours with a median PI of 14 (range = 0–87) (Figure 2A; Table S2). Fractional polynomial analysis showed that none of the transformations were significantly better than the simple linear model, suggesting that PI is linearly related to OS. The log-rank values over the entire range of cut-points show that there is no optimal cut-point, and since the log-rank values were not substantially higher than those for the linear model supports the conclusion that PI is a truly continuous variable.

thumbnail
Download:
Figure 2. A. Immunohistochemistry for Ki-67 performed on primary ESFT (5 µm).

Ki-67 protein was visualised with the DAB substrate, proliferating cells are identified by Ki-67 positive nuclei (see arrows). Immunohistochemistry performed in the absence of primary antibody was included to control for non-specific binding of the secondary antibody (No primary AB). Magnification ×400. B. TUNEL on primary ESFT (5 µm). Apoptotic cells were identified by TUNEL positive nuclei (see arrows). The TUNEL assay was performed in the absence TdT enzyme to control for non-specific staining (No enzyme). Magnification ×400.

https://doi.org/10.1371/journal.pone.0104106.g002

A Cox regression for OS incorporating PI as a continuous covariate demonstrates that those patients with tumours that have a high PI at diagnosis will have a worse OS than those patients with a lower PI (χ2 = 11.5, p<0.001, HR = 1.04, 95% CI = 1.02–1.06). PI was also included in a multivariate analysis for OS along with the covariates tumour site, tumour volume, metastases at diagnosis and AI. In this analysis PI remained strongly significant (χ2 = 9.11, p = 0.003, HR = 1.04, 95% CI = 1.01–1.06), along with tumour site (χ2 = 4.85, p = 0.028, HR = 2.81, 95% CI = 1.12–7.04), where patients with pelvic tumours had a worse outcome. With these two variables in the model, none of the other factors significantly added to the prediction of OS. Patients with a pelvic tumour and a high PI have a worse OS than those with either factor alone; patients with a pelvic tumour and a PI≥20 have a HR of 5.8. A Cox regression for RFS incorporating PI as a continuous covariate demonstrates that those patients with tumours that have a high PI at diagnosis will have a worse RFS than those patients with a lower PI (χ2 = 6.5, p = 0.011, HR = 1.03, 95% CI = 1.01–1.06). Apoptotic cells were identified in 96% (76/79) of ESFT with a median AI of 3 (range = 0–33) (Figure 2B; Table S2); AI does not predict OS (p = 0.137, Cox regression) or RFS (p = 0.596, Cox regression).

Caspase-8, -9 and -10 protein expression in primary ESFT

Caspase-8 protein was heterogeneously expressed in the cytoplasm (Figure 3A; Table S2); negative (0) in 11/79 (14%), low (1) in 26/79 (33%), medium (2) in 30/79 (38%) and high (3) in 12/79 (15%) primary ESFT. Caspase-8 protein expression did not predict OS or RFS (p = 0.250 and 0.235 respectively, log rank test). Caspase-10 protein was expressed at a low level (1) in the cytoplasm of 23/62 (37%) of tumours and not expressed (0) in 39/62 (63%; Figure 3B; Table S2) and did not predict OS (p = 0.732, log rank test) or RFS (p = 0.693, log rank test).

thumbnail
Download:
Figure 3. Immunohistochemistry on primary ESFT (5 µm) for caspases-8, -9 and -10.

Immunohistochemistry in the absence of primary antibody was to control for non-specific binding of the secondary antibody (No primary AB). Magnification of images = ×400. Caspase expression was scored as negative (0), low (1), medium (2) or high (3). A. Caspase-8 protein expression. B. Caspase-10 protein expression. C. Caspase-9 protein expression.

https://doi.org/10.1371/journal.pone.0104106.g003

Caspase-9 protein was detected in the cytoplasm and nucleus of ESFT (Figure 3C; Table S2), expression was low (1) in 39/59 (66%) and high (3) in 20/59 (34%) of tumours. Expression of caspase-9 was not predictive of OS (p = 0.275, log rank test), however, patients with tumours expressing low levels of caspase-9 protein had a strong trend towards a worse RFS than patients with tumours expressing high levels of caspase-9 protein (p = 0.054, log rank test, Figure S1). Nuclear or cytoplasmic expression did not appear to contribute to the prognostic value of low and high expression across the tumour. This is worthy of investigation in a validation tumour cohort.

Discussion

In this study we demonstrate for the first time that high numbers of proliferating cells in primary ESFT are predictive of a worse RFS and OS in a continuous model. In a multivariate analysis, high PI was predictive of poor OS independent of tumour site, tumour volume, metastasis at diagnosis or AI. The prognostic value of PI is consistent with results from univariate [14], [16] and multivariate [15] discontinuous studies in ESFT. However there is considerable variation in the level above which a tumour is scored as having a high number of proliferating cells; in the published literature high PI has being defined as ≥50% [16] or >5% [15]. This wide discrepancy can be explained by the variable having a truly continuous distribution. One group has reported no association between PI and outcome using an arbitrary cut-point of 20 (based on the literature at the time, p = 0.71) or a statistically defined cut-point of 8.3 (generated by analysis of the ROC curve, p = 0.06) [14]; this most likely reflects the continuous relationship between PI and outcome and the distribution of patients in this small group (n = 37). We have detected proliferating cells in 91% of ESFT, higher than previous studies where they reported Ki-67 staining in 57% [14] and 34% [15]. This is likely to be due to the fact that these authors only reported Ki-67 expression once the PI had reached the cut points of 8.3% and 5% respectively, whereas we have reported tumours to be positive with a PI of 1% or greater. Since a higher PI is predictive of worse OS in a continuous model, patient risk should optimally be defined by incorporating the PI value rather than an arbitrary cut-point into a predictive model. Alternatively the introduction of PI as a biomarker of risk into the clinic may be enabled by defining a cut-point that identifies patients at ultra-high risk, for whom current treatment offers no benefit and might be offered novel investigational treatments. For example, we have identified that patients with a PI value of ≥20 (33% of patients in this study) have a two-fold increased risk of death (HR = 2.0) or relapse (HR = 1.9), whereas patients with a pelvic tumour and a PI≥20 have a 6 fold-increased risk of death (HR = 5.8) suggesting PI and tumour site should be included in a risk model for patients with ESFT. The quantification of PI by immunohistochemistry for Ki-67 protein is a robust reproducible assay that can be performed reliably on paraffin embedded tumour at low cost, and so may readily be introduced into pathology practice. The utility of this marker, and others where quantification is important, may benefit from an observer independent assessment using digital imaging to allow high-throughput analysis and to overcome subjectivity and heterogeneity of expression. A recent study found that Ki-67 quantification by digital image analysis and manual counting were highly concordant [43]. It is now important to evaluate the prognostic value of PI as a continuous variable with other prognostic biomarkers to identify the most clinically informative algorithm of risk for patients with ESFT.

The number of apoptotic cells was low in the majority of primary ESFT, and AI did not predict prognosis in the univariate or multivariate models. This is the first study to describe this relationship in ESFT. AI has been correlated with survival in other cancers with contradictory findings, however analysis has been based values above and below cut points rather than as a continuous variable [17]–[19]. An increase in AI in tumours post-treatment may be a useful surrogate tumour marker of response to chemotherapy. For example, AI has been shown to correlate with pathological response in breast cancer biopsies collected after therapy [44], [45]. Some studies have suggested that immunohistochemistry for cleaved caspase-3 might replace measures of apoptosis such as TUNEL [46], [47], and absence or low numbers of cleaved caspase-3 positive cells have been associated with a worse prognosis in patients with nasopharyngeal carcinoma [48] and glioma [49]. However the transient expression of cleaved caspases and homology between the cleavage sites [50] may limit the clinical utility of IHC for cleaved caspases as predictive biomarkers. In contrast we have examined the hypothesis that expression of full-length caspases-8, -9 and -10 may be surrogate markers of tumour apoptotic potential, response to treatment and outcome. We have found considerable heterogeneity in the expression of these effectors of cell death, with no correlation between expression of the different caspases. Expression of caspase 8 and 10 was not significantly associated with OS or RFS. There was however a non-significant trend towards patients with tumours expressing low levels of caspase-9 protein having a worse RFS, which may reflect the importance of caspase-9 as an effector molecule of apoptotic cell death following treatment with current therapeutics. Low levels of caspase-9 protein in primary tumour could therefore identify patients that are most likely to benefit from non-caspase-9 dependent treatments, such as death receptor (DR)-mediated apoptosis through caspase-8 cleavage and induction of the non-mitochondrial apoptotic pathway [51], [52]. Consistent with this hypothesis ESFT are reported to be sensitive to TRAIL-induced apoptosis in preclinical models of ESFT growth [53]–[55]. Heterogeneous expression was not explained by differential methylation of the CASP8 or CASP9 gene promoters (Figure S2; [56]).

This study highlights the value of PI as a predictive biomarker of RFS and OS in patients with ESFT when measured in primary tumour. PI remained predictive of OS in a multivariate analysis, indicating that PI provides additional prognostic information independent of clinical markers used in current practice. This observation was authentic in a continuous statistical model, demonstrating that survival worsens as PI increases and therefore that risk can be defined most accurately based on PI as a continuous measure rather than choosing an arbitrary cut-point. The introduction of this assay into the pathological examination of ESFT at diagnosis may improve the ability to identify those patients destined to do badly that might benefit from alternative investigational treatment.

Supporting Information

Figure S1.

Kaplan Meier survival plot to compare the relapse-free survival of patients with tumours that had low caspase-9 expression to that of patients with tumours that had high caspase-9 expression, p = 0.0538; log rank test. Crosses identify censored events.

https://doi.org/10.1371/journal.pone.0104106.s001

(TIF)

Figure S2.

Methylation of the CASP9 promoter in ESFT. PCR products in lanes labelled U and M indicate the presence of unmethylated and methylated CASP9 promoter regions respectively. DNA extracted from peripheral blood (PB) from healthy volunteers was included as an unmethylated control and CpG Methylase treated peripheral blood DNA (Meth PB) was used as a methylation control.

https://doi.org/10.1371/journal.pone.0104106.s002

(PDF)

Table S1.

Comparison between frozen and FF-PE tumours and their distribution by age at diagnosis, tumour site and PI.

https://doi.org/10.1371/journal.pone.0104106.s003

(DOC)

Table S2.

Summary of patient and tumour information, PI, AI and protein expression. Details of the 105 ESFT samples analysed with clinical information, PI and AI and expression of caspases-8, -9 and -10.

https://doi.org/10.1371/journal.pone.0104106.s004

(XLS)

Acknowledgments

Thank you to Andrea Berry (AB) and Dr Catherine Cullinane (Paediatric Pathology, SJUH) for help scoring immunohistochemistry. Thanks also to Mr R Grimer, Royal Orthopaedic Hospital, Birmingham and the CCLG Tumour Bank and Data Centre for providing tumour samples and clinical outcome data.

Author Contributions

Conceived and designed the experiments: S.Burchill. Performed the experiments: S.Brownhill. Analyzed the data: S.Brownhill DC S.Burchill. Contributed reagents/materials/analysis tools: S.Burchill. Wrote the paper: S.Brownhill DC S.Burchill. Developed the study concept, secured grant funding: S.Burchill. Contributed to the interpretation and opinions expressed in this manuscript, and approved the final version of the manuscript: S.Brownhill DC S.Burchill.

References

  1. 1. Burchill SA (2003) Ewing's sarcoma: diagnostic, prognostic, and therapeutic implications of molecular abnormalities. Journal of clinical pathology 56: 96–102.
  2. 2. Terrier P, Llombart-Bosch A, Contesso G (1996) Small round blue cell tumors in bone: prognostic factors correlated to Ewing's sarcoma and neuroectodermal tumors. Semin Diagn Pathol 13: 250–257.
  3. 3. Ahrens S, Hoffmann C, Jabar S, Braun-Munzinger G, Paulussen M, et al. (1999) Evaluation of prognostic factors in a tumor volume-adapted treatment strategy for localized Ewing sarcoma of bone: the CESS 86 experience. Cooperative Ewing Sarcoma Study. Med Pediatr Oncol 32: 186–195.
  4. 4. Craft AW, Cotterill SJ, Bullimore JA, Pearson D (1997) Long-term results from the first UKCCSG Ewing's Tumour Study (ET-1). United Kingdom Children's Cancer Study Group (UKCCSG) and the Medical Research Council Bone Sarcoma Working Party. Eur J Cancer 33: 1061–1069.
  5. 5. Nesbit ME Jr, Gehan EA, Burgert EO Jr, Vietti TJ, Cangir A, et al. (1990) Multimodal therapy for the management of primary, nonmetastatic Ewing's sarcoma of bone: a long-term follow-up of the First Intergroup study. J Clin Oncol 8: 1664–1674.
  6. 6. Sinicrope FA, Hart J, Hsu HA, Lemoine M, Michelassi F, et al. (1999) Apoptotic and mitotic indices predict survival rates in lymph node-negative colon carcinomas. Clin Cancer Res 5: 1793–1804.
  7. 7. Rioux-Leclercq N, Turlin B, Bansard J, Patard J, Manunta A, et al. (2000) Value of immunohistochemical Ki-67 and p53 determinations as predictive factors of outcome in renal cell carcinoma. Urology 55: 501–505.
  8. 8. Gasinska A, Urbanski K, Gruchala A, Biesaga B, Kojs Z (2002) A ratio of apoptosis to mitosis, proliferation pattern and prediction of radiotherapy response in cervical carcinoma. Neoplasma 49: 379–386.
  9. 9. Mejia C, Navarro S, Pellin A, Ruiz A, Castel V, et al. (2003) Prognostic significance of cell proliferation in human neuroblastoma: comparison with other prognostic factors. Oncol Rep 10: 243–247.
  10. 10. Lopez-Beltran A, Luque RJ, Alvarez-Kindelan J, Quintero A, Merlo F, et al. (2004) Prognostic factors in stage T1 grade 3 bladder cancer survival: the role of G1-S modulators (p53, p21Waf1, p27kip1, Cyclin D1, and Cyclin D3) and proliferation index (ki67-MIB1). Eur Urol 45: 606–612.
  11. 11. Ali HR, Dawson SJ, Blows FM, Provenzano E, Leung S, et al. (2012) A Ki67/BCL2 index based on immunohistochemistry is highly prognostic in ER-positive breast cancer. The Journal of pathology 226: 97–107.
  12. 12. Ibrahim T, Farolfi A, Scarpi E, Mercatali L, Medri L, et al. (2012) Hormonal Receptor, Human Epidermal Growth Factor Receptor-2, and Ki67 Discordance between Primary Breast Cancer and Paired Metastases: Clinical Impact. Oncology 84: 150–157.
  13. 13. van Diest PJ, van der Wall E, Baak JP (2004) Prognostic value of proliferation in invasive breast cancer: a review. J Clin Pathol 57: 675–681.
  14. 14. Amir G, Issakov J, Meller I, Sucher E, Peyser A, et al. (2002) Expression of p53 gene product and cell proliferation marker Ki-67 in Ewing's sarcoma: correlation with clinical outcome. Hum Pathol 33: 170–174.
  15. 15. Lopez-Guerrero JA, Machado I, Scotlandi K, Noguera R, Pellin A, et al. (2011) Clinicopathological significance of cell cycle regulation markers in a large series of genetically confirmed Ewing's sarcoma family of tumors. International journal of cancer Journal international du cancer 128: 1139–1150.
  16. 16. Sollazzo MR, Benassi MS, Magagnoli G, Gamberi G, Molendini L, et al. (1999) Increased c-myc oncogene expression in Ewing's sarcoma: correlation with Ki67 proliferation index. Tumori 85: 167–173.
  17. 17. Wu X, Cai ZD, Lou LM, Zhu YB (2012) Expressions of p53, c-MYC, BCL-2 and apoptotic index in human osteosarcoma and their correlations with prognosis of patients. Cancer epidemiology 36: 212–216.
  18. 18. Jia Y, Dong B, Tang L, Liu Y, Du H, et al. (2012) Apoptosis index correlates with chemotherapy efficacy and predicts the survival of patients with gastric cancer. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine 33: 1151–1158.
  19. 19. Yamasaki F, Tokunaga O, Sugimori H (1997) Apoptotic index in ovarian carcinoma: correlation with clinicopathologic factors and prognosis. Gynecol Oncol 66: 439–448.
  20. 20. Tannapfel A, Hahn HA, Katalinic A, Fietkau RJ, Kuhn R, et al. (1997) Incidence of apoptosis, cell proliferation and P53 expression in renal cell carcinomas. Anticancer Res 17: 1155–1162.
  21. 21. Stammler G, Sauerbrey A, Zintl F, Volm M (1997) Apoptotic index, Fas and bcl-2 in initial and relapsed childhood acute lymphoblastic leukaemia. Apoptosis 2: 377–383.
  22. 22. Bendardaf R, Ristamaki R, Kujari H, Laine J, Lamlum H, et al. (2003) Apoptotic index and bcl-2 expression as prognostic factors in colorectal carcinoma. Oncology 64: 435–442.
  23. 23. Leung TW, Xue WC, Cheung AN, Khoo US, Ngan HY (2004) Proliferation to apoptosis ratio as a prognostic marker in adenocarcinoma of uterine cervix. Gynecol Oncol 92: 866–872.
  24. 24. Pingoud-Meier C, Lang D, Janss AJ, Rorke LB, Phillips PC, et al. (2003) Loss of caspase-8 protein expression correlates with unfavorable survival outcome in childhood medulloblastoma. Clin Cancer Res 9: 6401–6409.
  25. 25. Teitz T, Wei T, Valentine MB, Vanin EF, Grenet J, et al. (2000) Caspase 8 is deleted or silenced preferentially in childhood neuroblastomas with amplification of MYCN. Nat Med 6: 529–535.
  26. 26. Yang Q, Kiernan CM, Tian Y, Salwen HR, Chlenski A, et al. (2007) Methylation of CASP8, DCR2, and HIN-1 in neuroblastoma is associated with poor outcome. Clin Cancer Res 13: 3191–3197.
  27. 27. Fulda S, Poremba C, Berwanger B, Hacker S, Eilers M, et al. (2006) Loss of caspase-8 expression does not correlate with MYCN amplification, aggressive disease, or prognosis in neuroblastoma. Cancer Res 66: 10016–10023.
  28. 28. Kaseta MK, Gomatos IP, Khaldi L, Tzagarakis GP, Alevizos L, et al. (2007) Prognostic value of bax, cytochrome C, and caspase-8 protein expression in primary osteosarcoma. Hybridoma (Larchmt) 26: 355–362.
  29. 29. Andressakis D, Lazaris AC, Tsiambas E, Kavantzas N, Rapidis A, et al. (2008) Evaluation of caspase-3 and caspase-8 deregulation in tongue squamous cell carcinoma, based on immunohistochemistry and computerised image analysis. J Laryngol Otol 122: 1213–1218.
  30. 30. Takita J, Yang HW, Bessho F, Hanada R, Yamamoto K, et al. (2000) Absent or reduced expression of the caspase 8 gene occurs frequently in neuroblastoma, but not commonly in Ewing sarcoma or rhabdomyosarcoma. Med Pediatr Oncol 35: 541–543.
  31. 31. Harada K, Toyooka S, Shivapurkar N, Maitra A, Reddy JL, et al. (2002) Deregulation of caspase 8 and 10 expression in pediatric tumors and cell lines. Cancer Res 62: 5897–5901.
  32. 32. de Hooge AS, Berghuis D, Santos SJ, Mooiman E, Romeo S, et al. (2007) Expression of cellular FLICE inhibitory protein, caspase-8, and protease inhibitor-9 in Ewing sarcoma and implications for susceptibility to cytotoxic pathways. Clin Cancer Res 13: 206–214.
  33. 33. Lissat A, Vraetz T, Tsokos M, Klein R, Braun M, et al. (2007) Interferon-gamma sensitizes resistant Ewing's sarcoma cells to tumor necrosis factor apoptosis-inducing ligand-induced apoptosis by up-regulation of caspase-8 without altering chemosensitivity. Am J Pathol 170: 1917–1930.
  34. 34. Santon A, Garcia-Cosio M, Cristobal E, Pascual A, Muriel A, et al. (2011) Expression of heat shock proteins in classical Hodgkin lymphoma: correlation with apoptotic pathways and prognostic significance. Histopathology 58: 1072–1080.
  35. 35. Strater J, Herter I, Merkel G, Hinz U, Weitz J, et al. (2010) Expression and prognostic significance of APAF-1, caspase-8 and caspase-9 in stage II/III colon carcinoma: caspase-8 and caspase-9 is associated with poor prognosis. Int J Cancer 127: 873–880.
  36. 36. Liu LX, Liu ZH, Jiang HC, Qu X, Zhang WH, et al. (2002) Profiling of differentially expressed genes in human gastric carcinoma by cDNA expression array. World J Gastroenterol 8: 580–585.
  37. 37. Xu B, Zhou ZG, Li Y, Wang L, Yang L, et al. (2008) Clinicopathological significance of caspase-8 and caspase-10 expression in rectal cancer. Oncology 74: 229–236.
  38. 38. Shen XG, Wang C, Li Y, Zhou B, Xu B, et al. (2011) Downregulation of caspase-10 predicting poor survival after resection of stage II colorectal cancer. International journal of colorectal disease 26: 1519–1524.
  39. 39. Dalal S, Berry AM, Cullinane CJ, Mangham DC, Grimer R, et al. (2005) Vascular endothelial growth factor: a therapeutic target for tumors of the Ewing's sarcoma family. Clinical cancer research : an official journal of the American Association for Cancer Research 11: 2364–2378.
  40. 40. Rosito P, Mancini AF, Rondelli R, Abate ME, Pession A, et al. (1999) Italian Cooperative Study for the treatment of children and young adults with localized Ewing sarcoma of bone: a preliminary report of 6 years of experience. Cancer 86: 421–428.
  41. 41. Bacci G, Ferrari S, Bertoni F, Rimondini S, Longhi A, et al. (2000) Prognostic factors in nonmetastatic Ewing's sarcoma of bone treated with adjuvant chemotherapy: analysis of 359 patients at the Istituto Ortopedico Rizzoli. J Clin Oncol 18: 4–11.
  42. 42. Royston P, Sauerbrei W (2004) A new approach to modelling interactions between treatment and continuous covariates in clinical trials by using fractional polynomials. Statistics in medicine 23: 2509–2525.
  43. 43. Tang LH, Gonen M, Hedvat C, Modlin IM, Klimstra DS (2012) Objective quantification of the Ki67 proliferative index in neuroendocrine tumors of the gastroenteropancreatic system: a comparison of digital image analysis with manual methods. The American journal of surgical pathology 36: 1761–1770.
  44. 44. Davis DW, Buchholz TA, Hess KR, Sahin AA, Valero V, et al. (2003) Automated quantification of apoptosis after neoadjuvant chemotherapy for breast cancer: early assessment predicts clinical response. Clin Cancer Res 9: 955–960.
  45. 45. Stearns V, Singh B, Tsangaris T, Crawford JG, Novielli A, et al. (2003) A prospective randomized pilot study to evaluate predictors of response in serial core biopsies to single agent neoadjuvant doxorubicin or paclitaxel for patients with locally advanced breast cancer. Clin Cancer Res 9: 124–133.
  46. 46. Duan WR, Garner DS, Williams SD, Funckes-Shippy CL, Spath IS, et al. (2003) Comparison of immunohistochemistry for activated caspase-3 and cleaved cytokeratin 18 with the TUNEL method for quantification of apoptosis in histological sections of PC-3 subcutaneous xenografts. J Pathol 199: 221–228.
  47. 47. Gown AM, Willingham MC (2002) Improved detection of apoptotic cells in archival paraffin sections: immunohistochemistry using antibodies to cleaved caspase 3. J Histochem Cytochem 50: 449–454.
  48. 48. Oudejans JJ, Harijadi A, Cillessen SA, Busson P, Tan IB, et al. (2005) Absence of caspase 3 activation in neoplastic cells of nasopharyngeal carcinoma biopsies predicts rapid fatal outcome. Mod Pathol 18: 877–885.
  49. 49. Kobayashi T, Masumoto J, Tada T, Nomiyama T, Hongo K, et al. (2007) Prognostic significance of the immunohistochemical staining of cleaved caspase-3, an activated form of caspase-3, in gliomas. Clin Cancer Res 13: 3868–3874.
  50. 50. McStay GP, Salvesen GS, Green DR (2008) Overlapping cleavage motif selectivity of caspases: implications for analysis of apoptotic pathways. Cell Death Differ 15: 322–331.
  51. 51. Ashkenazi A (2008) Directing cancer cells to self-destruct with pro-apoptotic receptor agonists. Nat Rev Drug Discov 7: 1001–1012.
  52. 52. Debatin KM, Krammer PH (2004) Death receptors in chemotherapy and cancer. Oncogene 23: 2950–2966.
  53. 53. Van Valen F, Fulda S, Truckenbrod B, Eckervogt V, Sonnemann J, et al. (2000) Apoptotic responsiveness of the Ewing's sarcoma family of tumours to tumour necrosis factor-related apoptosis-inducing ligand (TRAIL). International journal of cancer Journal international du cancer 88: 252–259.
  54. 54. van Valen F, Harrer H, Hotfilder M, Dirksen U, Pap T, et al. (2012) A Novel Role of IGF1 in Apo2L/TRAIL-Mediated Apoptosis of Ewing Tumor Cells. Sarcoma 2012: 782970.
  55. 55. White DE, Burchill SA (2010) Fenretinide-dependent upregulation of death receptors through ASK1 and p38alpha enhances death receptor ligand-induced cell death in Ewing's sarcoma family of tumours. Br J Cancer
  56. 56. Martinez R, Setien F, Voelter C, Casado S, Quesada MP, et al. (2007) CpG island promoter hypermethylation of the pro-apoptotic gene caspase-8 is a common hallmark of relapsed glioblastoma multiforme. Carcinogenesis 28: 1264–1268.