Skip to main content
NCBI home page
Cancer Reports logoLink to Cancer Reports
. 2019 Feb 19;3(4):e1166. doi: 10.1002/cnr2.1166

Development and validation of an immune prognostic signature for ovarian carcinoma

Seema Khadirnaikar 1,2, Pranjal Kumar 1, Sudhanshu Kumar Shukla 1,✉
PMCID: PMC7941549  PMID: 32794637

Abstract

Background

Ovarian cancer (OC) causes a significant proportion of cancer‐related deaths in women. Recently, immunotherapy has emerged as a substantial player in cancer treatment. Lymphocyte infiltration, an important indicator of immune activity and disease aggressiveness, can be identified by gene expression profiling of immune‐related genes of tumours which may prove useful in prognosis of patients.

Aims

The aim of this study is to identify and validate a novel immune gene‐based prognostic signature for OC.

Methods and results

Here, we extracted the expression of immune‐related genes and performed the Cox regression analysis and identified five genes with significant correlation with survival in training cohort of patients (n = 286). We utilised regression coefficient and expression level of five genes to calculate immune prognostic signature (IPS) score for OC patients. In univariate and multivariate Cox regression analysis with other clinicopathological factors, we showed that IPS is an independent predictor of survival (P value <0.01). More importantly, we utilised 404 patients from TCGA dataset as the validation cohort and validated the survival capability of IPS in the univariate and multivariate analysis (P value <0.001). Interestingly, KM analysis showed a significant difference in survival of patients with high and low IPS score in both datasets (training dataset P value <0.01, validation dataset P value <0.01). Further, we showed that all the five genes are differentially expressed and involved in immune modulation among other pathways. Interestingly, GSEA analysis showed that high IPS patients had low immune activity and activated EMT and other oncogenic pathways.

Conclusion

In summary, we have developed and validated robust immune‐related gene‐based prognostic signature to identify the OC patients with high immune activity who can be taken for immunotherapy.

Keywords: immune activity, immunotherapy, ovarian cancer, prognostic signature

Abbreviations

IPS

Immune prognostic signature

OC

Ovarian cancer

TCGA

The Cancer Genome Atlas

1. INTRODUCTION

Ovarian cancer (OC) is the second most common cause of cancer‐related deaths in women, with estimated 239 000 new cases and 152 000 deaths each year.1, 2 The highest rate of OC is reported in Europe.1 An overwhelming majority of OC gets diagnosed at a later stage/metastatic stage of the disease.3 Majority of patients undergo chemotherapy and surgical removal of the tumours, but eventually, disease relapses with a resistant tumour, and only 30% patients show 5‐year survival rate.4 The immune system has been shown to play an essential role in tumour development and progression.5 Also, immune infiltration has been proposed as one of the hallmarks of cancer.6 Recently, immunotherapy (PD‐L1 inhibitor) is emerging as a practical option for advance OC treatment.7 Immunotherapy which increases tumour‐infiltrating T cells in the tumours helps in the clearing of tumour cells effectively in OC because of immunoactive nature of tumours.8, 9, 10 Melanomas with high expression immune‐related genes have been shown to respond better to immunotherapy.11, 12 Even though immunotherapy has proved to be useful in cancer treatment, the response rate of patients is meagre.13 Thus, it is essential to identify the patients with high immunological activity for immunotherapy. The correlation between high lymphocyte infiltration and patient's survival has been well established in OC.14 Lymphocyte infiltration level can be identified by gene expression profiling of immune‐related genes which may prove useful in prognosis of patients. However, the molecular nature of tumour‐immune system interaction is poorly exploited for their prognostic potential in OC. Some studies have developed gene expression‐based prognostic signatures for prognosis and stratification of OC patients.15 Unfortunately, none of them has utilised the immune‐related gene's expression to predict survival and immune filtration which can be used for immunotherapy decisions.

In this work, we have utilised two datasets with 690 patients to develop and validate an immune prognostic signature (IPS) score for OC. We have used various statistical techniques to show the robustness of IPS in prognostication of OC patients. We have also demonstrated that five component genes of IPS are differentially expressed in OC patients compared with normal and are involved in immune modulation. Most importantly, we showed that patients with high IPS have less immune infiltration and high oncogenic pathway activity which could be the cause of their poor survival.

2. MATERIALS AND METHODS

2.1. Patient samples

We have used 690 patient's data for this study. Expression and clinical data for GSE9899 were downloaded using Oncomine and used as training cohort.16 Clinical data from TCGA was downloaded using Oncomine, whereas expression data were downloaded from the TCGA website. TCGA patient data were used as a validation cohort. The training dataset has similar age distribution (median age 59 years) and follow‐up time (median follow‐up time—28.5 months) compared with validation cohort (median age—59 years, median follow‐up time—29.4 months) (Table 1).

Table 1.

Clinicopathological data of training and validation cohort

Patients' Clinicopathological Data
Training Dataset Validation Dataset
Total number of patients 286 404
Median age Minimum 22.36 30
Median 58.95 59
Maximum 80.05 87
Gradea Grade 1 19 1
Grade 2 96 47
Grade 3 160 348
FIGO stage Stage I 22 Not available
Stage II 16
Stage III 155
Stage IV 13
Open in a new tab
a

Eleven patients in training cohort and eight patients in validation cohort do not have this information.

2.2. Cox regression analysis and development of IPS

Cox regression analysis was carried out using the survival package in R (https://CRAN.R‐project.org/package=survival). Five genes that were significantly associated with survival were identified by univariate Cox analysis, and regression coefficients thus obtained were used to calculate the IPS score for each patient using the following formula:

IPS = (0.251 × C1QTNF3 expression) + (0.308 × CD246 expression) + (0.502 × ADA expression) + (−0.539 × CASP8 expression) + (−0.395 × C6 expression).

2.3. Kaplan‐Meier and multivariate analysis

IPS score was divided at the 60th percentile to perform the Cox regression analysis using GraphPad Prism 7.0. Multivariate analysis was performed using SPSS version 25. For multivariate analysis, forward conditional method was used. Stage and grade were used as categorical variables in the analysis.

2.4. Differential expression analysis and GSEA analysis

We used MiPanda software (http://www.mipanda.org) for the differential expression analysis of the five component genes. Q values were obtained from MiPanda database. The correlation coefficient was calculated using the correl command in Excel 2016. The correlation heat map was plotted in the GraphPad version 7.0. For GO analysis, we used the GOrilla database, and the output was used as input for REViGO software.17 REViGO output was plotted in R.18 For GSEA analysis, differentially expressed genes were identified by performing t test between high and low IPS group. Significantly differentially expressed genes were then ranked and used for preranked GSEA analysis.19 A flow chart illustrating various analysis performed in this study is shown in Figure 1 .

Figure 1.

Figure 1

Open in a new tab

Flow chart of the overall analysis scheme

3. RESULTS

3.1. Patient characteristics

In this study, we have included a total of 690 patient's samples with median age [range] of 59 [22‐87] years. We used 286 patients with median age [range] of 58.5 [22‐80.5] years from the GSE9899 dataset as training cohort and 404 patients with median age [range] of 59 [30‐87] years from TCGA dataset as the validation cohort. GSE9899 dataset had stage and grade data available, whereas TCGA samples had only grade data available. The clinicopathological data of patients from both the cohorts are given in Table 1 . There was no significant difference in the age of patients in the two cohorts (Table 1).

3.2. Development of immune prognostic score

Firstly, to develop the IPS, we extracted the expression level of immune genes (n = 824, downloaded from http://www.innatedb.com) from training cohort. The immune gene expression data of samples were then subjected to Cox regression analysis using the survival package in R to identify the genes whose expression was significantly correlating with the survival. Top five genes (C1QTNF3, CD246, ADA, C6, and CASP8) that were significantly correlated with survival in the training cohort were used for calculation of IPS score using the formula described before (Section 2, Table S1). The proportional hazard assumptions in the Cox model for all the five covariates were validated and found to be time‐independent globally (Table S2). Expression pattern of all the five genes in the training cohort is shown in the heat map (Figure 2A). To carry out the Kaplan‐Meier analysis, we divided the samples into two groups (low risk and high risk) at multiple cut‐offs and measured the difference in survival by log‐rank test (Table S3). We also used the Cutoff Finder to find the optimal cut‐off and found the values to be near to 60th percentile.20 Hence, 60% of the samples were considered to be in the low‐risk group and 40% of the samples in the high‐risk group. The pattern of IPS distribution is given in Figure 2B. Patients with high IPS showed significantly more deaths compared with low IPS patients (Fisher's exact test, P value <0.01) (Figure 2C). Most importantly, high IPS patients showed significantly lower survival compared with low IPS patients (HR = 2.60, P value <0.01) (Figure 2D). Further, we divided patients into three groups (33% samples each) based on IPS score and showed the presence of an intermediate population with the significant difference in survival compared with high IPS and low IPS patients (P value <0.01, Figure S1).

Figure 2.

Figure 2

Open in a new tab

Development of IPS. A, Heat map of five genes identified as prognostic in Cox regression analysis in the training cohort. The colour gradient shows the expression of genes (black = higher expression, white = low expression). B, Scatter plot showing the distribution of IPS score with the number of patients (black dots indicate patients with high IPS, and grey dots indicate patients with low IPS). C, Scatter plot showing the distribution of patients with survival (black dots stated dead and grey alive patients). D, A Kaplan‐Meier plot showing the difference in survival of patients. E, Heat map of five genes identified as prognostic in Cox regression analysis in the validation cohort. The colour gradient shows the expression of genes (black = higher expression, white = low expression). F, Scatter plot showing the distribution of IPS score with the number of patients (black dots indicate patients with high IPS, and grey dots indicate patients with low IPS). G, Scatter plot showing the distribution of patients with their survival (black dots stated dead and grey alive patients). H, A Kaplan‐Meier plot showing the difference in survival of patients

3.3. Validation of prognostic ability of IPS in OC

We showed that IPS is a significant predictor of survival in training cohort. Further, we validated the prognostic ability of IPS in a cohort of 404 patient samples obtained from TCGA. We downloaded the RNA‐Seq and clinical data from the TCGA database and calculated the read counts and FPKM value using Tuxedo pipeline.21 We then extracted the FPKM values for five genes which constitute the IPS. Expression values were then used to calculate the IPS in the validation cohort using the same formula which was used in the training cohort (Section 2, Table S4). Similar to the training cohort, patients in the validation cohort were divided into low IPS and high IPS at 60th percentile. The expression pattern of five genes that constitute the IPS score is shown in the heat map (Figure 2E). Distribution of IPS in the validation cohort is shown in Figure 2F. Correlation of survival with IPS showed that patients belonging to low IPS group tend to survive better (Figure 2G), which was then validated in Kaplan‐Meier analysis (Figure 2H). KM analysis showed that patients with high IPS had significantly less survival compared with patients with low IPS (HR = 1.51, P value <0.01).

3.4. IPS is an independent predictor of survival

To show that IPS is an independent predictor of survival, we performed univariate and multivariate Cox regression analysis in training cohort with IPS and other clinical markers age, grade, and stage as covariates. In univariate analysis, we showed that only IPS and age were the significant predictors of prognosis of OC patients (IPS—HR: 2.74, CI: 2.03‐3.69, P value <0.01; age—HR: 1.02, CI: 1.01‐1.04, P value = 0.01) (Figure 3A). In multivariate analysis with IPS and age as covariates, we showed that IPS was an independent predictor of survival in training cohort (IPS—HR: 2.63, CI: 1.95‐3.55, P value <0.01; age—HR: 1.02, CI: 1.01‐1.04, P value = 0.04) (Figure 3A and 3B). Similarly, to validate that IPS was an independent predictor of survival in the validation cohort, we performed univariate and multivariate Cox regression analysis using age, grade, and IPS as covariates. Similar to training cohort, univariate Cox regression in validation cohort showed that only age and IPS were significant predictors of survival (IPS—HR: 1.08, CI: 1.03‐1.15, P value <0.01; age—HR: 1.02, CI: 1.01‐1.03, P value <0.01). Multivariate analysis with age and IPS showed that both age and IPS were independent predictors of survival in the validation cohort (IPS—HR: 1.06, CI: 1.01‐1.12, P value = 0.03; age—HR: 1.02, CI: 1.00‐1.03, P value = 0.02) (Figure 3A and 3C). Further, leave‐one‐out analysis confirmed that the proposed five‐gene model was a better prognosticator among other models (Table S5).

Figure 3.

Figure 3

Open in a new tab

Cox regression analysis of patients. A, The table is showing the univariate and multivariate analysis results in training and validation cohort. B, Forest plot is indicating the hazard ratios of univariate analysis of different variable with IPS in training cohort. C, Forest plot is showing the hazard ratios of univariate and multivariate analysis of varying variable with IPS in the validation cohort

3.5. Component genes of the IPS are differentially regulated and affect the immune system

First, to check the regulation of IPS component genes, we utilised MiPanda expression database. Interestingly, we found C1QTNF3, ADA, CD246, and CASP8 to be overexpressed and C6 to be underexpressed in OC compared with normal (Figure 4A). The expression pattern of these genes suggests their essential role in OC development and progression. Further, we found that the correlation among the expression of these genes was very poor (Figure 4B), suggesting that these genes are independently involved in prognosis and are not coregulated with each other.

Figure 4.

Figure 4

Open in a new tab

Characterisation of IPS components. A, Expression pattern of all the five genes in OC and normal as obtained from MiPanda website (q values are shown as obtained from MiPanda). B, Heatmap of correlation coefficient as obtained by correlating five genes with each other (grey indicates that there is no correlation between the variables). C, The REViGO output of GO terms enriched in the list of genes obtained as genes correlating with all the five genes (colour of bubble shows the P value of enrichment of GO term, and size indicates the number of GO terms enriched in each category)

To understand the functions regulated by these genes, we identified the list of genes coregulated with IPS component genes. This gene list was then used to determine the enriched GO terms using the GOrilla database, which was further used as input for REViGO analysis. REViGO analysis recognized immune system process as one of the top enriched pathways, suggesting the role of IPS component genes in immune regulation (Figure 4C). Other enriched pathways were the response to another organism, microtubule‐based process, anatomical structure development, a vascular process in the circulatory system, localization, regulation of biological quality, and transport (Figure 4C). Pathways regulated by correlating genes with C1QTNF3, CD246, ADA, C6, and CASP8 are given in Figure S2.

We have also checked the connection among the member of IPS, C1QTNF3, ADA, CD246, C6, and CASP8 using Metascape (http://metascape.org) software. Interestingly, the analysis showed that IPS component genes together regulate pathways involved in the regulation of defence response, in utero embryonic development, and inflammatory response (Figure S3).

3.6. The immune activity of patients with high IPS

We identified in the previous analysis that the patients with high IPS survived significantly lower than patients with low IPS. To understand the molecular mechanism behind the differential survival of these groups of patients, we performed a differential expression analysis between high and low IPS group of patients. The differentially expressed genes were then ranked in decreasing order of fold change and used as input in preranked GSEA analysis using C1 hallmark as gene set (Figure 5A). Interestingly, we found that patients with high IPS were significantly negatively enriched in interferon alpha and interferon gamma response, and significantly positive enriched with epithelial‐mesenchymal transition (EMT), hypoxia, and KRAS signalling (Figure 5A to 5F), suggesting the immune inactive and high oncogenic nature of high IPS patients.

Figure 5.

Figure 5

Open in a new tab

GSEA analysis of high and low IPS groups. A, Scatter plot showing the number of gene cohorts enriched in GSEA analysis (black dots indicate the positively enriched gene cohorts, and grey dots show nonsignificantly enriched gene sets). B and C, GSEA enrichment plot of negatively enriched gene cohorts in high IPS compare with low IPS. D to F, GSEA enrichment plot of positively enriched gene sets in high IPS compared with low IPS

4. DISCUSSION

OC is the leading cause of gynaecological cancer‐related deaths.12 Similar to many other cancers, late detection of OC is the primary hurdle in therapy.22 While because of the advancement in surgical technologies and chemotherapy, a significant difference has been made, the survival of OC patients has moderately changed.23 Majority of women affected with OC show relapse of disease, and more than 50% of these patients die within 5 years.22 Recently, immunotherapy has emerged as an important option in cancer treatment10 as it enhances the tumour‐suppressive nature of the immune system. FDA has recently approved immunotherapy for the treatment of melanoma, nonsmall cell lung cancer, renal cell carcinoma, bladder cancer, and classical Hodgkin lymphoma.10 Two most important factors responsible for immunotherapy response are (1) availability of immune cell in the tumour environment and (2) nature of the immune checkpoint pathway in antitumor immunity. OC patients are the suitable candidate for immunotherapy because of high activity of immune cells in tumour milieu.1, 8 However, there are no studies reported to identify the patients with poor prognosis based on the immune system activity in the tumour.

Here in this work, we have utilised expression data available on GEO and TCGA to develop and validate an IPS score. We have identified five genes, C1QTNF3, CD246, ADA, C6, and CASP8, as prognostically essential genes in OC. IPS, a score generated using the expression of these genes, is an independent prognostic factor in the training and validation cohorts. Patients divided into low and high IPS showed a significant difference in survival.

Detailed in silico study of the components of IPS showed that all five genes were differentially expressed in OC compared with normal tumours. Further, we explained that there was no correlation among the five genes, suggesting that all the genes predict survival independently and not selected because of coregulation. GO terms associated with coregulated genes indicated the related immune function of these genes among others. Interestingly, many of the five genes are already shown to be involved in cancer. CD246 is one of the most critical genes in NSCLC.24 CD246 is a tyrosine kinase and is a driver genetic aberration in NSCLC.24 CD246 inhibitors have been approved for the treatment of EML4‐CD246 fusion‐positive tumours.24 CD246 has also been proposed to be an important target for OC treatment.25 ADA is an enzyme involved in the hydrolysis of adenosine to inosine. Various mutation of this gene has been associated with different kinds of diseases. Genetic variation in ADA genes has been shown to be involved in uterine leiomyomas.26 ADA gene has been proposed to be an important prognostic marker in patients with malignant pleural effusion.27 C6 protein is an essential factor in complement cascade and is involved in membrane attack complex.28 C6 genotype rs9200 is shown to be associated with hepatocellular carcinoma recurrence.29 Also, C6 gene is found to be underexpressed in oesophageal carcinoma.30 CASP8 is a member of the caspase family and is involved in the apoptosis. CASP8 has been shown to be involved in triggering and sensing DNA damage in liver cancer, a nonapoptotic function.31 C1QTNF3 is not well‐studied protein in the context of cancer. However, C1QTNF3 is an essential gene in the type II diabetes.32

Most importantly, in this study, we have identified various pathways activated in different IPS groups. GSEA analysis using differentially expressed genes between high and low IPS groups showed that high IPS group had activated EMT, hypoxia, and KRAS signalling pathways suggesting the reason why this group has a more aggressive tumour with poor survival. Also, we found that immune‐related pathways interferon alpha and interferon gamma responses are negatively enriched, suggesting a poor enrichment of immune cells in these tumours. This finding indicates that this group of patients have less immune activity and may not be an ideal candidate for the immunotherapy.

Limitation of our study is the retrospective nature. Although we have included two completely different datasets as training and testing cohort, further validation of signature would be required to validate the usefulness of the study. Also, we have not performed molecular biology lab work to show the function of IPS component genes.

In summary, we have developed and validated a robust IPS for prognostication of ovarian tumours. We also showed that the component genes are associated with immune modulation. Importantly, we showed that the patients with poor survival have weak immune activation in the milieu.

FUNDING

No funding to disclose.

CONFLICT OF INTEREST

None declared.

AUTHORS' CONTRIBUTIONS

All authors had full access to the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Conceptualization, S.K.S.; Methodology, S.K. and S.K.S.; Investigation, S.K.S., S.K. and P.K.; Resources, S.K.S.; Writing ‐ Original Draft, P.K. and S.K.; Writing ‐ Review & Editing, S.K.S., S.K. and P.K.; Supervision, S.K.S.

Supporting information

Figure S1. KM plot showing the di_erence in survival when patients were divided into three groups.

Figure S2 A, Pathways enriched with C1QTNF3 correlating genes. B, Pathways regulated by ALK co‐regulated genes. C, Pathways regulated by ADA correlated genes. D, Pathways regulated by C6 co‐regulated genes

Figure S3: Metascape analysis of IPS member genes. The enriched terms were used for plotting the network layout. More specifically, each term is represented by a circle node, where its size is proportional to the number of input genes fall into that term, and its color represent its cluster identity

Table S1: Training set five genes expression and IPS

Table S2: Fitness of Cox proportional hazard assumptions

Table S3: Cut‐off to divide samples into low‐ and high‐risk groups in training cohort

Table S4: Validation cohort five genes expression and IPS

Table S5: Leave‐one‐out analysis in validation cohort (TCGA cohort)

Click here for additional data file. (2MB, pdf)

ACKNOWLEDGEMENTS

Authors acknowledge the Indian Institute of Technology Dharwad for computer facilities. Dr Shruti Bhargava is acknowledged for her critical comments. “The results <published or shown> here are in whole or part based upon data generated by the TCGA Research Network: http://cancergenome.nih.gov/.”

Khadirnaikar S, Kumar P, Shukla SK. Development and validation of an immune prognostic signature for ovarian carcinoma. Cancer Reports. 2020;3:e1166. 10.1002/cnr2.1166

REFERENCES

  • 1. Reid BM, Permuth JB, Sellers TA. Epidemiology of ovarian cancer: a review. Cancer Biol Med. 2017;14(1):9‐32. 10.20892/j.issn.2095-3941.2016.0084. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2017. CA Cancer J Clin. 2017;67(1):7‐30. 10.3322/caac.21387 [DOI] [PubMed] [Google Scholar]
  • 3. Coward JIG, Middleton K, Murphy F. New perspectives on targeted therapy in ovarian cancer. Int. J. Womens Health 2015:189–203. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Matulonis UA, Sood AK, Fallowfield L, Howitt BE, Sehouli J. Ovarian cancer. 2016. 10.1038/nrdp.2016.61 [DOI] [PMC free article] [PubMed]
  • 5. Signatures M. Review the continuum of cancer immunosurveillance: prognostic, predictive, and mechanistic signatures. Immunity. 2013;11‐26. [DOI] [PubMed] [Google Scholar]
  • 6. Hanahan D, Review WRA. Hallmarks of cancer: the next generation. Cell. 2011;144(5):646‐674. 10.1016/j.cell.2011.02.013 [DOI] [PubMed] [Google Scholar]
  • 7. Drerup JM, Liu Y, Padron AS, et al. HHS public access. 2016;16(1):1‐20. 10.1007/s11864-014-0317-1.Immunotherapy [DOI] [Google Scholar]
  • 8. Baert T, Vergote I, Coosemans A. Ovarian cancer and the immune system. Gynecol Oncol Rep. 2017;19:57‐58. 10.1016/j.gore.2017.01.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Arimappamagan A, Somasundaram K, Thennarasu K, et al. A fourteen gene GBM prognostic signature identifies association of immune response pathway and mesenchymal subtype with high risk group. PLoS ONE. 2013;8(4):e62042. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Farkona S, Diamandis EP, Blasutig IM. Cancer immunotherapy: the beginning of the end of cancer ? BMC Med. 2016;14(1):‐18. 10.1186/s12916-016-0623-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Sharma P, Hu‐lieskovan S, Wargo JA, Ribas A. Review primary, adaptive, and acquired resistance to cancer immunotherapy. Cell. 2017;168(4):707‐723. 10.1016/j.cell.2017.01.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Achkar T, Tarhini AA. The use of immunotherapy in the treatment of melanoma. J Hematol Oncol. 2017;10(1):1‐9. 10.1186/s13045-017-0458-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Hiniker SM, Maecker HT, Knox SJ. Predictors of clinical response to immunotherapy with or without radiotherapy. 2015:339–345. 10.1007/s13566-015-0219-2 [DOI] [PMC free article] [PubMed]
  • 14. Outcomes M. Dose‐response association of CD8 + tumor‐infiltrating lymphocytes and survival time in high‐grade serous ovarian cancer. JAMA Oncol. 2017;55905. 10.1001/jamaoncol.2017.3290 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Waldron L, Haibe‐kains B, Culhane AC, et al. Comparative meta‐analysis of prognostic gene signatures for late‐stage ovarian cancer. JNCI. 2014;27(5). 10.1093/jnci/dju049 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Tothill RW, Tinker A V, George J, et al. Novel molecular subtypes of serous and endometrioid ovarian cancer linked to clinical outcome. Clin Cancer Res 2008;14(16):5198–5209. 10.1158/1078-0432.CCR-08-0196 [DOI] [PubMed] [Google Scholar]
  • 17. Eden E, Navon R, Steinfeld I, Lipson D, Yakhini Z. Gorilla: a tool for discovery and visualization of enriched GO terms in ranked gene lists. BMC Bioinformatics. 2009;10(1):48. 10.1186/1471-2105-10-48 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Supek F, Bošnjak M, Škunca N, Šmuc T. Revigo summarizes and visualizes long lists of gene ontology terms. PLoS ONE. 2011;6(7):e21800. 10.1371/journal.pone.0021800 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Mootha VK, Lindgren CM, Eriksson KF, et al. PGC‐1α‐responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat Genet. 2003;34(3):267‐273. 10.1038/ng1180 [DOI] [PubMed] [Google Scholar]
  • 20. Cutoff Finder: a comprehensive and straightforward web application enabling rapid biomarker cutoff optimization https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0051862. Accessed January 26, 2019, 7, 12, e51862. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. Trapnell C, Roberts A, Goff L, et al. Differential gene and transcript expression analysis of RNA‐seq experiments with TopHat and Cufflinks. Nat Protoc. 2012;7(3):562‐578. 10.1038/nprot.2012.016 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Gubbels JAA, Claussen N, Kapur AK, Connor JP, Patankar MS. The detection, treatment, and biology of epithelial ovarian cancer. J Ovarian Res. 2010;1‐11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Wallace S, Kumar A, Mc M, et al. Gynecologic oncology efforts at maximal cytoreduction improve survival in ovarian cancer patients, even when complete gross resection is not feasible ☆. Gynecol Oncol. 2017;145(1):21‐26. 10.1016/j.ygyno.2017.01.029 [DOI] [PubMed] [Google Scholar]
  • 24. Hallberg B, Palmer RH. Mechanistic insight into ALK cancer biology. Nat Rev Cancer 10.1038/nrc3580. 13(10):685‐700. [DOI] [PubMed] [Google Scholar]
  • 25. Ren H, Tan Z, Zhu X, et al. Identification of anaplastic lymphoma kinase as a potential therapeutic target in ovarian cancer. Cancer Res. 2012;72(9):3312‐3324. 10.1158/0008-5472.CAN-11-3931 [DOI] [PubMed] [Google Scholar]
  • 26. Gloria‐bottini F, Saccucci P, Ammendola M, Neri A, Magrini A, Bottini E. European Journal of Obstetrics & Gynecology and reproductive biology genetic variability within adenosine deaminase gene and uterine leiomyomas. Euro J Obs Gyne. 2016;199:108‐109. 10.1016/j.ejogrb.2016.02.002 [DOI] [PubMed] [Google Scholar]
  • 27. Terra RM, Antonangelo L, Wasum A, Terra RM. Pleural fluid adenosine deaminase (ADA) predicts survival in patients with malignant pleural effusion. Lung. 2016. 10.1007/s00408-016-9891-2;194(4):681‐686. [DOI] [PubMed] [Google Scholar]
  • 28. Sarma JV, Ward PA. The complement system. Cell Tissue Res. 2012;343(1):227‐235. 10.1007/s00441-010-1034-0.The [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Wang Z, Liao J, Wu S, Li C, Fan J, Peng Z. Recipient C6 rs9200 genotype is associated with hepatocellular carcinoma recurrence after orthotopic liver transplantation in a Han Chinese population. 2016; (November 2015):157–161. 10.1038/cgt.2016.7 [DOI] [PubMed]
  • 30. Oka R, Sasagawa T, Ninomiya I, Miwa K, Tanii H, Saijoh K. Reduction in the local expression of complement component 6 (C6) and 7 (C7) mRNAs in oesophageal carcinoma. Eur. J. Cancer. 2001;37:1158‐1165. [DOI] [PubMed] [Google Scholar]
  • 31. Boege Y, Malehmir M, Healy ME, et al. A dual role of Caspase‐8 in triggering and sensing proliferation‐associated DNA damage, a key article a dual role of Caspase‐8 in triggering and sensing proliferation‐associated DNA damage, a key determinant of liver Cancer Development. Cancer Cell. 2017;32(3):342‐359. 10.1016/j.ccell.2017.08.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32. Fadaei R, Moradi N, Baratchian M, Aghajani H. Association of C1q/TNF‐related protein‐3 (CTRP3) and CTRP13 serum levels with coronary artery disease in subjects with and without type 2 diabetes mellitus. Plos One. 2016;3:1‐14. 10.1371/journal.pone.0168773 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figure S1. KM plot showing the di_erence in survival when patients were divided into three groups.

Figure S2 A, Pathways enriched with C1QTNF3 correlating genes. B, Pathways regulated by ALK co‐regulated genes. C, Pathways regulated by ADA correlated genes. D, Pathways regulated by C6 co‐regulated genes

Figure S3: Metascape analysis of IPS member genes. The enriched terms were used for plotting the network layout. More specifically, each term is represented by a circle node, where its size is proportional to the number of input genes fall into that term, and its color represent its cluster identity

Table S1: Training set five genes expression and IPS

Table S2: Fitness of Cox proportional hazard assumptions

Table S3: Cut‐off to divide samples into low‐ and high‐risk groups in training cohort

Table S4: Validation cohort five genes expression and IPS

Table S5: Leave‐one‐out analysis in validation cohort (TCGA cohort)

Click here for additional data file. (2MB, pdf)

Articles from Cancer Reports are provided here courtesy of Wiley

RESOURCES