Comparative meta-analysis of prognostic gene signatures for late-stage ovarian cancer

Urinary-based detection of MSL, HE4 and CA125 as an additional dimension for predictive and prognostic modelling in ovarian cancer

Objectives

We have recently described a predictive/prognostic model for ovarian cancer, exploiting commonly available clinico-pathological parameters and the ovarian serum biomarkers mesothelin (MSL), human epididymis protein 4 (HE4) and cancer-antigen 125 (CA125). Considering urine as a prototype non-invasive sample, we investigated whether serum levels of these biomarkers are mirrored in urine and compared their clinical relevance in matched serum vs. urine samples.

Methods

MSL, HE4 and CA125 were quantified in urinary (n=172) and matched serum samples (n=188) from ovarian cancer patients (n=192) using the Lumipulse® G chemiluminescent enzyme immunoassay (Fujirebio).

Results

While absolute concentrations of MSL or CA125 were higher in serum than in matched urine samples, HE4 concentrations were considerably higher in urine than in serum. Nonetheless, the levels of all three biomarkers strongly correlated between matched serum vs. urine samples and were unrelated to BRCA1/2 mutational status. Consequently, prediction of surgical outcome or relapse/death by MSL, HE4 or CA125 was similarly efficient among urinary- vs. serum-based detection. HE4 provided the highest capacity to predict surgical outcome or relapse/death among both body fluids (urine: AUC=0.854; serum: AUC=0.750, respectively). All clinically relevant findings regarding the investigated urinary biomarkers were equally reproducible among raw vs. creatinine-normalized datasets, suggesting that normalization may have subordinate priority for urine-based analysis of these biomarkers.

Conclusion

We report that the capacity of MSL, HE4 and CA125 to predict surgical outcome and relapse/death is equivalent between serum vs. urine-based detection. Urinary biomarkers, in particular HE4, may provide an additional dimension for prognostic modeling in ovarian cancer.

A predictive and prognostic model for surgical outcome and prognosis in ovarian cancer computed by clinico-pathological and serological parameters (CA125, HE4, mesothelin)

OBJECTIVES

Numerous prognostic models have been proposed for ovarian cancer, extending from single serological factors to complex gene-expression signatures. Nonetheless, these models have not been routinely translated into clinical practice. We constructed a robust and readily calculable model for predicting surgical outcome and prognosis of ovarian cancer patients by exploiting commonly available clinico-pathological factors and three selected serum parameters.

METHODS

Serum CA125, human epididymis protein 4 (HE4) and mesothelin (MSL) were quantified by Lumipulse® G chemiluminescent enzyme immunoassay (Fujirebio) in a total of 342 serum samples from 190 ovarian cancer patients, including 152 paired pre- and post-operative samples.

RESULTS

Detection of pre-operative HE4 and CA125 was the optimal marker combination for blood-based prediction of surgical outcome (AUC=0.86). We constructed a prognostic model, computed by serum levels of pre-operative CA125, post-operative HE4, post-operative MSL and surgical outcome. Prognostic performance of our model was superior to any of these parameters alone and was independent from BRCA1/2 mutational status. We subsequently transformed our model into a prognostic risk index, stratifying patients as "lower risk" or "higher risk". In "higher risk" patients, relapse or death was predicted with an AUC of 0.89 and they had a significantly shorter progression free survival (HR: 9.74; 95 % CI: 5.95-15.93; p<0.0001) and overall survival (HR: 5.62; 95 % CI: 3.16-9.99; p<0.0001) compared to "lower risk" patients.

CONCLUSIONS

We present a robust predictive/prognostic model for ovarian cancer, which could readily be implemented into routine diagnostics in order to identify ovarian cancer patients at high risk of recurrence.

Personalization of Therapy in High-Grade Serous Tubo-Ovarian Cancer-The Possibility or the Necessity?

High-grade serous tubo-ovarian cancer (HGSTOC) is the most lethal tumor of the female genital tract. The foregoing therapy consists of cytoreduction followed by standard platinum/taxane chemotherapy; alternatively, for primary unresectable tumors, neo-adjuvant platinum/taxane chemotherapy followed by delayed interval cytoreduction. In patients with suboptimal surgery or advanced disease, different forms of targeted therapy have been accepted or tested in clinical trials. Studies on HGSTOC discovered its genetic and proteomic heterogeneity, epigenetic regulation, and the role of the tumor microenvironment. These findings turned attention to the fact that there are several distinct primary tumor subtypes of HGSTOC and the unique biology of primary, metastatic, and recurrent tumors may result in a differential drug response. This results in both chemo-refractoriness of some primary tumors and, what is significantly more frequent and destructive, secondary chemo-resistance of metastatic and recurrent HGSTOC tumors. Treatment possibilities for platinum-resistant disease include several chemotherapeutics with moderate activity and different targeted drugs with difficult tolerable effects. Therefore, the question appears as to why different subtypes of ovarian cancer are predominantly treated based on the same therapeutic schemes and not in an individualized way, adjusted to the biology of a specific tumor subtype and temporal moment of the disease. The paper reviews the genomic, mutational, and epigenetic signatures of HGSTOC subtypes and the tumor microenvironment. The clinical trials on personalized therapy and the overall results of a new, comprehensive approach to personalized therapy for ovarian cancer have been presented and discussed.

Unveiling the novel immune and molecular signatures of ovarian cancer: insights and innovations from single-cell sequencing

Ovarian cancer is a highly heterogeneous and lethal malignancy with limited treatment options. Over the past decade, single-cell sequencing has emerged as an advanced biological technology capable of decoding the landscape of ovarian cancer at the single-cell resolution. It operates at the level of genes, transcriptomes, proteins, epigenomes, and metabolisms, providing detailed information that is distinct from bulk sequencing methods, which only offer average data for specific lesions. Single-cell sequencing technology provides detailed insights into the immune and molecular mechanisms underlying tumor occurrence, development, drug resistance, and immune escape. These insights can guide the development of innovative diagnostic markers, therapeutic strategies, and prognostic indicators. Overall, this review provides a comprehensive summary of the diverse applications of single-cell sequencing in ovarian cancer. It encompasses the identification and characterization of novel cell subpopulations, the elucidation of tumor heterogeneity, the investigation of the tumor microenvironment, the analysis of mechanisms underlying metastasis, and the integration of innovative approaches such as organoid models and multi-omics analysis.

Accounting for time dependency in meta-analyses of concordance probability estimates

Recent years have seen the development of many novel scoring tools for disease prognosis and prediction. To become accepted for use in clinical applications, these tools have to be validated on external data. In practice, validation is often hampered by logistical issues, resulting in multiple small-sized validation studies. It is therefore necessary to synthesize the results of these studies using techniques for meta-analysis. Here we consider strategies for meta analyzing the concordance probability for time-to-event data ("C-index"), which has become a popular tool to evaluate the discriminatory power of prediction models with a right-censored outcome. We show that standard meta-analysis of the C-index may lead to biased results, as the magnitude of the concordance probability depends on the length of the time interval used for evaluation (defined e.g., by the follow-up time, which might differ considerably between studies). To address this issue, we propose a set of methods for random-effects meta-regression that incorporate time directly as covariate in the model equation. In addition to analyzing nonlinear time trends via fractional polynomial, spline, and exponential decay models, we provide recommendations on suitable transformations of the C-index before meta-regression. Our results suggest that the C-index is best meta-analyzed using fractional polynomial meta-regression with logit-transformed C-index values. Classical random-effects meta-analysis (not considering time as covariate) is demonstrated to be a suitable alternative when follow-up times are small. Our findings have implications for the reporting of C-index values in future studies, which should include information on the length of the time interval underlying the calculations.

Machine learning-based integration develops an immune-related risk model for predicting prognosis of high-grade serous ovarian cancer and providing therapeutic strategies

Background

High-grade serous ovarian cancer (HGSOC) is a highly lethal gynecological cancer that requires accurate prognostic models and personalized treatment strategies. The tumor microenvironment (TME) is crucial for disease progression and treatment. Machine learning-based integration is a powerful tool for identifying predictive biomarkers and developing prognostic models. Hence, an immune-related risk model developed using machine learning-based integration could improve prognostic prediction and guide personalized treatment for HGSOC.

Methods

During the bioinformatic study in HGSOC, we performed (i) consensus clustering to identify immune subtypes based on signatures of immune and stromal cells, (ii) differentially expressed genes and univariate Cox regression analysis to derive TME- and prognosis-related genes, (iii) machine learning-based procedures constructed by ten independent machine learning algorithms to screen and construct a TME-related risk score (TMErisk), and (iv) evaluation of the effect of TMErisk on the deconstruction of TME, indication of genomic instability, and guidance of immunotherapy and chemotherapy.

Results

We identified two different immune microenvironment phenotypes and a robust and clinically practicable prognostic scoring system. TMErisk demonstrated superior performance over most clinical features and other published signatures in predicting HGSOC prognosis across cohorts. The low TMErisk group with a notably favorable prognosis was characterized by BRCA1 mutation, activation of immunity, and a better immune response. Conversely, the high TMErisk group was significantly associated with C-X-C motif chemokine ligands deletion and carcinogenic activation pathways. Additionally, low TMErisk group patients were more responsive to eleven candidate agents.

Conclusion

Our study developed a novel immune-related risk model that predicts the prognosis of ovarian cancer patients using machine learning-based integration. Additionally, the study not only depicts the diversity of cell components in the TME of HGSOC but also guides the development of potential therapeutic techniques for addressing tumor immunosuppression and enhancing the response to cancer therapy.

Sensitivity Analysis for Survival Prognostic Prediction with Gene Selection: A Copula Method for Dependent Censoring

Prognostic analysis for patient survival often employs gene expressions obtained from high-throughput screening for tumor tissues from patients. When dealing with survival data, a dependent censoring phenomenon arises, and thus the traditional Cox model may not correctly identify the effect of each gene. A copula-based gene selection model can effectively adjust for dependent censoring, yielding a multi-gene predictor for survival prognosis. However, methods to assess the impact of various types of dependent censoring on the multi-gene predictor have not been developed. In this article, we propose a sensitivity analysis method using the copula-graphic estimator under dependent censoring, and implement relevant methods in the R package “compound.Cox”. The purpose of the proposed method is to investigate the sensitivity of the multi-gene predictor to a variety of dependent censoring mechanisms. In order to make the proposed sensitivity analysis practical, we develop a web application. We apply the proposed method and the web application to a lung cancer dataset. We provide a template file so that developers can modify the template to establish their own web applications.

Molecular characteristics and clinical behaviour of epithelial ovarian cancers

Ovarian carcinoma (OC) is an umbrella term for multiple distinct diseases (histotypes), each with their own developmental origins, clinical behaviour and molecular profile. Accordingly, OC management is progressing away from a one-size-fits all approach, toward more molecularly-driven, histotype-specific management strategies. Our knowledge of driver events in high grade serous OC, the most common histotype, has led to major advances in treatments, including PARP inhibitor use. However, these agents are not suitable for all patients, most notably for many of those with rare OC histotypes. Identification of additional targeted therapeutic strategies will require a detailed understanding of the molecular landscape in each OC histotype. Until recently, tumour profiling studies in rare histotypes were sparse; however, significant advances have been made over the last decade. In particular, reports of genomic characterisation in endometrioid, clear cell, mucinous and low grade serous OC have significantly expanded our understanding of mutational events in these tumour types. Nonetheless, substantial knowledge gaps remain. This review summarises our current understanding of each histotype, highlighting recent advances in these unique diseases and outlining immediate research priorities for accelerating progress toward improving patient outcomes.

ncRNAs mediated RPS6KA2 inhibits ovarian cancer proliferation via p38/MAPK signaling pathway

Background

Ovarian cancer is the most lethal gynecology malignancy in the world, therefore, research on the molecular biological mechanism of ovarian cancer tumorigenesis and progression has received widespread attention.

Methods

We identified RPS6KA2 as the prognosis-related gene of ovarian cancer from TCGA, GSE26712 and GSE26193 database via bioinformatic analysis. qRT-PCR and western blot detected the differential expression of RPS6KA2 in normal ovaries and ovarian cancer tissues. The biological functions of RPS6KA2 were verified by in vitro and in vivo. GSEA analysis was used to select candidate signaling pathway of RPS6KA2 which was further verified by western blot. The possible binding sites of RPS6KA2 with miRNAs and circRNAs were predicted by bioinformatics analysis, and then a circRNA-miRNA-mRNA interaction network was constructed.

Results

We found the expression of RPS6KA2 was down-regulated in ovarian cancer tissues. Overexpression of RPS6KA2 could suppress cell proliferation, whereas knockdown of RPS6KA2 had the opposite effects on proliferation. GSEA analysis showed that the MARK signaling pathway was closely associated with RPS6KA2. Bioinformatics analysis and dual-luciferase reporter assay showed that RPS6KA2 was regulated with miR-19a-3p, miR-106a-5p and miR-519d-3p. Further analysis showed that circFAM169A was the common ceRNA of miR-19a-3p, miR-106a-5p and miR-519d-3p. Dual-luciferase reporter assay showed the relationship of circFAM169A and miR-106a-5p and miR-519d-3p. After network analysis, one circRNA-miRNA-mRNA axis (circFAM169A/miR-106a-5p, miR-519d-3p/RPS6KA2) was identified.

Conclusions

We demonstrated that circFAM169A/miR-106a-5p, miR-519d-3p mediated low expression of RPS6KA2 could affect the proliferation of ovarian cancer cells via p38/MAPK signaling pathway.

Integrating single-cell RNA sequencing and prognostic model revealed the carcinogenicity and clinical significance of FAM83D in ovarian cancer

Background

Ovarian cancer (OC) is a fatal gynecological tumor with high mortality and poor prognosis. Yet, its molecular mechanism is still not fully explored, and early prognostic markers are still missing. In this study, we assessed carcinogenicity and clinical significance of family with sequence similarity 83 member D (FAM83D) in ovarian cancer by integrating single-cell RNA sequencing (scRNA-seq) and a prognostic model.

Methods

A 10x scRNA-seq analysis was performed on cells from normal ovary and high-grade serous ovarian cancer (HGSOC) tissue. The prognostic model was constructed by Lasso-Cox regression analysis. The biological function of FAM83D on cell growth, invasion, migration, and drug sensitivity was examined in vitro in OC cell lines. Luciferase reporter assay was performed for binding analysis between FAM83D and microRNA-138-5p (miR-138-5p).

Results

Our integrative analysis identified a subset of malignant epithelial cells (C1) with epithelial-mesenchymal transition (EMT) and potential hyperproliferation gene signature. A FAM83D⁺ malignant epithelial subcluster (FAM83D⁺ MEC) was associated with cell cycle regulation, apoptosis, DNA repair, and EMT activation. FAM83D resulted as a viable prognostic marker in a prognostic model that efficiently predict the overall survival of OC patients. FAM83D downregulation in SKOV3 and A2780 cells increased cisplatin sensitivity, reducing OC cell proliferation, migration, and invasion. MiR-138-5p was identified to regulate FAM83D's carcinogenic effect in OC cells.

Conclusions

Our findings highlight the importance of miR-138 -5p/FAM83D/EMT signaling and may provide new insights into therapeutic strategies for OC.

Integration of survival data from multiple studies

We introduce a statistical procedure that integrates datasets from multiple biomedical studies to predict patients' survival, based on individual clinical and genomic profiles. The proposed procedure accounts for potential differences in the relation between predictors and outcomes across studies, due to distinct patient populations, treatments and technologies to measure outcomes and biomarkers. These differences are modeled explicitly with study-specific parameters. We use hierarchical regularization to shrink the study-specific parameters towards each other and to borrow information across studies. The estimation of the study-specific parameters utilizes a similarity matrix, which summarizes differences and similarities of the relations between covariates and outcomes across studies. We illustrate the method in a simulation study and using a collection of gene expression datasets in ovarian cancer. We show that the proposed model increases the accuracy of survival predictions compared to alternative meta-analytic methods.

The limits of molecular signatures for pancreatic ductal adenocarcinoma subtyping

Molecular signatures have been suggested as biomarkers to classify pancreatic ductal adenocarcinoma (PDAC) into two, three, four or five subtypes. Since the robustness of existing signatures is controversial, we performed a systematic evaluation of four established signatures for PDAC stratification across nine publicly available datasets. Clustering revealed inconsistency of subtypes across independent datasets and in some cases a different number of PDAC subgroups than in the original study, casting doubt on the actual number of existing subtypes. Next, we built sixteen classification models to investigate the ability of the signatures for tumor subtype prediction. The overall classification performance ranged from ∼35% to ∼90% accuracy, suggesting instability of the signatures. Notably, permuted subtypes and random gene sets achieved very similar performance. Cellular decomposition and functional pathway enrichment analysis revealed strong tissue-specificity of the predicted classes. Our study highlights severe limitations and inconsistencies that can be attributed to technical biases in sample preparation and tumor purity, suggesting that PDAC molecular signatures do not generalize across datasets. How stromal heterogeneity and immune compartment interplay in the diverging development of PDAC is still unclear. Therefore, a more mechanistic or a cross-platform multi-omic approach seems necessary to extract more robust and clinically exploitable insights.

Searching for robust associations with a multi-environment knockoff filter

This paper develops a method based on model-X knockoffs to find conditional associations that are consistent across environments, controlling the false discovery rate. The motivation for this problem is that large data sets may contain numerous associations that are statistically significant and yet misleading, as they are induced by confounders or sampling imperfections. However, associations replicated under different conditions may be more interesting. In fact, consistency sometimes provably leads to valid causal inferences even if conditional associations do not. While the proposed method is widely applicable, this paper highlights its relevance to genome-wide association studies, in which robustness across populations with diverse ancestries mitigates confounding due to unmeasured variants. The effectiveness of this approach is demonstrated by simulations and applications to the UK Biobank data.

Multiomic Characterization of High-Grade Serous Ovarian Carcinoma Enables High-Resolution Patient Stratification

PURPOSE

High-grade serous ovarian carcinoma (HGSOC) is the most common ovarian cancer type; most patients experience disease recurrence that accumulates chemoresistance, leading to treatment failure. Genomic and transcriptomic features have been associated with differential outcome and treatment response. However, the relationship between events at the gene sequence, copy number, and gene-expression levels remains poorly defined.

EXPERIMENTAL DESIGN

We perform multiomic characterization of a large HGSOC cohort (n = 362) with detailed clinical annotation to interrogate the relationship between patient subgroups defined by specific molecular events.

RESULTS

BRCA2-mutant (BRCA2m) and EMSY-overexpressing cases demonstrated prolonged survival [multivariable hazard ratios (HR) 0.40 and 0.51] and significantly higher first- and second-line chemotherapy response rate. CCNE1-gained (CCNE1g) cases demonstrated underrepresentation of FIGO stage IV cases, with shorter survival but no significant difference in treatment response. We demonstrate marked overlap between the TCGA- and Tothill-derived subtypes. IMR/C2 cases displayed higher BRCA1/2m frequency (25.5%, 32.5%) and significantly greater immune cell infiltration, whereas PRO/C5 cases had the highest CCNE1g rate (23.9%, 22.2%) and were uniformly low in immune cell infiltration. The survival benefit for cases with aberrations in homologous recombination repair (HRR) genes was apparent across all transcriptomic subtypes (HR range, 0.48-0.68). There was significant co-occurrence of RB loss and HRR gene aberrations; RB loss was further associated with favorable survival within HRR-aberrant cases (multivariable HR, 0.50).

CONCLUSIONS

These data paint a high-resolution picture of the molecular landscape in HGSOC, better defining patients who may benefit most from specific molecular therapeutics and highlighting those for whom novel treatment strategies are needed to improve outcomes.

Replicability in cancer omics data analysis: measures and empirical explorations

In biomedical research, the replicability of findings across studies is highly desired. In this study, we focus on cancer omics data, for which the examination of replicability has been mostly focused on important omics variables identified in different studies. In published literature, although there have been extensive attention and ad hoc discussions, there is insufficient quantitative research looking into replicability measures and their properties. The goal of this study is to fill this important knowledge gap. In particular, we consider three sensible replicability measures, for which we examine distributional properties and develop a way of making inference. Applying them to three The Cancer Genome Atlas (TCGA) datasets reveals in general low replicability and significant across-data variations. To further comprehend such findings, we resort to simulation, which confirms the validity of the findings with the TCGA data and further informs the dependence of replicability on signal level (or equivalently sample size). Overall, this study can advance our understanding of replicability for cancer omics and other studies that have identification as a key goal.

Cross-Platform Omics Prediction procedure: a statistical machine learning framework for wider implementation of precision medicine

In this modern era of precision medicine, molecular signatures identified from advanced omics technologies hold great promise to better guide clinical decisions. However, current approaches are often location-specific due to the inherent differences between platforms and across multiple centres, thus limiting the transferability of molecular signatures. We present Cross-Platform Omics Prediction (CPOP), a penalised regression model that can use omics data to predict patient outcomes in a platform-independent manner and across time and experiments. CPOP improves on the traditional prediction framework of using gene-based features by selecting ratio-based features with similar estimated effect sizes. These components gave CPOP the ability to have a stable performance across datasets of similar biology, minimising the effect of technical noise often generated by omics platforms. We present a comprehensive evaluation using melanoma transcriptomics data to demonstrate its potential to be used as a critical part of a clinical screening framework for precision medicine. Additional assessment of generalisation was demonstrated with ovarian cancer and inflammatory bowel disease studies.

Dynamic Risk Prediction via a Joint Frailty-Copula Model and IPD Meta-Analysis: Building Web Applications

Clinical risk prediction formulas for cancer patients can be improved by dynamically updating the formulas by intermediate events, such as tumor progression. The increased accessibility of individual patient data (IPD) from multiple studies has motivated the development of dynamic prediction formulas accounting for between-study heterogeneity. A joint frailty-copula model for overall survival and time to tumor progression has the potential to develop a dynamic prediction formula of death from heterogenous studies. However, the process of developing, validating, and publishing the prediction formula is complex, which has not been sufficiently described in the literature. In this article, we provide a tutorial in order to build a web-based application for dynamic risk prediction for cancer patients on the basis of the R packages joint.Cox and Shiny. We demonstrate the proposed methods using a dataset of breast cancer patients from multiple clinical studies. Following this tutorial, we demonstrate how one can publish web applications available online, which can be manipulated by any user through a smartphone or personal computer. After learning this tutorial, developers acquire the ability to build an online web application using their own datasets.

Identification of key gene signatures for the overall survival of ovarian cancer

BACKGROUND

The five-year overall survival (OS) of advanced-stage ovarian cancer remains nearly 25-35%, although several treatment strategies have evolved to get better outcomes. A considerable amount of heterogeneity and complexity has been seen in ovarian cancer. This study aimed to establish gene signatures that can be used in better prognosis through risk prediction outcome for the survival of ovarian cancer patients. Different studies' heterogeneity into a single platform is presented to explore the penetrating genes for poor or better survival. The integrative analysis of multiple data sets was done to determine the genes that influence poor or better survival. A total of 6 independent data sets was considered. The Cox Proportional Hazard model was used to obtain significant genes that had an impact on ovarian cancer patients. The gene signatures were prepared by splitting the over-expressed and under-expressed genes parallelly by the variable selection technique. The data visualisation techniques were prepared to predict the overall survival, and it could support the therapeutic regime.

RESULTS

We preferred to select 20 genes in each data set as upregulated and downregulated. Irrespective of the selection of multiple genes, not even a single gene was found common among data sets for the survival of ovarian cancer patients. However, the same analytical approach adopted. The chord plot was presented to make a comprehensive understanding of the outcome.

CONCLUSIONS

This study helps us to understand the results obtained from different studies. It shows the impact of the heterogeneity from one study to another. It shows the requirement of integrated studies to make a holistic view of the gene signature for ovarian cancer survival.

Integrin α2β1 Represents a Prognostic and Predictive Biomarker in Primary Ovarian Cancer

Currently, the same first-line chemotherapy is administered to almost all patients suffering from primary ovarian cancer. The high recurrence rate emphasizes the need for precise drug treatment in primary ovarian cancer. Being crucial in ovarian cancer progression and chemotherapeutic resistance, integrins became promising therapeutic targets. To evaluate its prognostic and predictive value, in the present study, the expression of integrin α2β1 was analyzed immunohistochemically and correlated with the survival data and other therapy-relevant biomarkers. The significant correlation of a high α2β1-expression with the estrogen receptor alpha (ERα; p = 0.035) and epithelial growth factor receptor (EGFR; p = 0.027) was observed. In addition, high α2β1-expression was significantly associated with a low number of tumor-infiltrating immune cells (CD3 intratumoral, p = 0.017; CD3 stromal, p = 0.035; PD-1 intratumoral, p = 0.002; PD-1 stromal, p = 0.049) and the lack of PD-L1 expression (p = 0.005). In Kaplan–Meier survival analysis, patients with a high expression of integrin α2β1 revealed a significant shorter progression-free survival (PFS, p = 0.035) and platinum-free interval (PFI, p = 0.034). In the multivariate Cox regression analysis, integrin α2β1 was confirmed as an independent prognostic factor for both PFS (p = 0.021) and PFI (p = 0.020). Dual expression of integrin α2β1 and the hepatocyte growth factor receptor (HGFR; PFS/PFI, p = 0.004) and CD44v6 (PFS, p = 0.000; PFI, p = 0.001; overall survival [OS], p = 0.025) impaired survival. Integrin α2β1 was established as a prognostic and predictive marker in primary ovarian cancer with the potential to stratify patients for chemotherapy and immunotherapy, and to design new targeted treatment strategies.

Gene expression profile association with poor prognosis in epithelial ovarian cancer patients

Ovarian cancer (OC) is the eighth most common type of cancer for women worldwide. The current diagnostic and prognostic routine available for OC management either lack specificity or are very costly. Gene expression profiling has shown to be a very effective tool in exploring new molecular markers for patients with OC, although association of such markers with patient survival and clinical outcome is still elusive. Here, we performed gene expression profiling of different subtypes of OC to evaluate its association with patient overall survival (OS) and aggressive forms of the disease. By global mRNA microarray profiling in a total of 196 epithelial OC patients (161 serous, 15 endometrioid, 11 mucinous, and 9 clear cell carcinomas), we found four candidates-HSPA1A, CD99, RAB3A and POM121L9P, which associated with OS and poor clinicopathological features. The overexpression of all combined was correlated with shorter OS and progression-free survival (PFS). Furthermore, the combination of at least two markers were further associated with advanced grade, chemotherapy resistance, and progressive disease. These results indicate that a panel comprised of a few predictors that associates with a more aggressive form of OC may be clinically relevant, presenting a better performance than one marker alone.

Reliability of microarray analysis for studying periodontitis: low consistency in 2 periodontitis cohort data sets from different platforms and an integrative meta-analysis

PURPOSE

The aim of this study was to compare the characteristic expression patterns of advanced periodontitis in 2 cohort data sets analyzed using different microarray platforms, and to identify differentially expressed genes (DEGs) through a meta-analysis of both data sets.

METHODS

Twenty-two patients for cohort 1 and 40 patients for cohort 2 were recruited with the same inclusion criteria. The 2 cohort groups were analyzed using different platforms: Illumina and Agilent. A meta-analysis was performed to increase reliability by removing statistical differences between platforms. An integrative meta-analysis based on an empirical Bayesian methodology (ComBat) was conducted. DEGs for the integrated data sets were identified using the limma package to adjust for age, sex, and platform and compared with the results for cohorts 1 and 2. Clustering and pathway analyses were also performed.

RESULTS

This study detected 557 and 246 DEGs in cohorts 1 and 2, respectively, with 146 and 42 significantly enriched gene ontology (GO) terms. Overlapping between cohorts 1 and 2 was present in 59 DEGs and 18 GO terms. However, only 6 genes from the top 30 enriched DEGs overlapped, and there were no overlapping GO terms in the top 30 enriched pathways. The integrative meta-analysis detected 34 DEGs, of which 10 overlapped in all the integrated data sets of cohorts 1 and 2.

CONCLUSIONS

The characteristic expression pattern differed between periodontitis and the healthy periodontium, but the consistency between the data sets from different cohorts and metadata was too low to suggest specific biomarkers for identifying periodontitis.

Tumor-intrinsic and -extrinsic (immune) gene signatures robustly predict overall survival and treatment response in high grade serous ovarian cancer patients

In the present study, we developed a transcriptomic signature capable of predicting prognosis and response to primary therapy in high grade serous ovarian cancer (HGSOC). Proportional hazard analysis was performed on individual genes in the TCGA RNAseq data set containing 229 HGSOC patients. Ridge regression analysis was performed to select genes and develop multigenic models. Survival analysis identified 120 genes whose expression levels were associated with overall survival (OS) (HR = 1.49-2.46 or HR = 0.48-0.63). Ridge regression modeling selected 38 of the 120 genes for development of the final Ridge regression models. The consensus model based on plurality voting by 68 individual Ridge regression models classified 102 (45%) as low, 23 (10%) as moderate and 104 patients (45%) as high risk. The median OS was 31 months (HR = 7.63, 95% CI = 4.85-12.0, P < 1.0-10) and 77 months (HR = ref) in the high and low risk groups, respectively. The gene signature had two components: intrinsic (proliferation, metastasis, autophagy) and extrinsic (immune evasion). Moderate/high risk patients had more partial and non-responses to primary therapy than low risk patients (odds ratio = 4.54, P < 0.001). We concluded that the overall survival and response to primary therapy in ovarian cancer is best assessed using a combination of gene signatures. A combination of genes which combines both tumor intrinsic and extrinsic functions has the best prediction. Validation studies are warranted in the future.

The impact of different sources of heterogeneity on loss of accuracy from genomic prediction models

Cross-study validation (CSV) of prediction models is an alternative to traditional cross-validation (CV) in domains where multiple comparable datasets are available. Although many studies have noted potential sources of heterogeneity in genomic studies, to our knowledge none have systematically investigated their intertwined impacts on prediction accuracy across studies. We employ a hybrid parametric/non-parametric bootstrap method to realistically simulate publicly available compendia of microarray, RNA-seq, and whole metagenome shotgun microbiome studies of health outcomes. Three types of heterogeneity between studies are manipulated and studied: (i) imbalances in the prevalence of clinical and pathological covariates, (ii) differences in gene covariance that could be caused by batch, platform, or tumor purity effects, and (iii) differences in the "true" model that associates gene expression and clinical factors to outcome. We assess model accuracy, while altering these factors. Lower accuracy is seen in CSV than in CV. Surprisingly, heterogeneity in known clinical covariates and differences in gene covariance structure have very limited contributions in the loss of accuracy when validating in new studies. However, forcing identical generative models greatly reduces the within/across study difference. These results, observed consistently for multiple disease outcomes and omics platforms, suggest that the most easily identifiable sources of study heterogeneity are not necessarily the primary ones that undermine the ability to accurately replicate the accuracy of omics prediction models in new studies. Unidentified heterogeneity, such as could arise from unmeasured confounding, may be more important.

HGNChelper: identification and correction of invalid gene symbols for human and mouse

Gene symbols are recognizable identifiers for gene names but are unstable and error-prone due to aliasing, manual entry, and unintentional conversion by spreadsheets to date format. Official gene symbol resources such as HUGO Gene Nomenclature Committee (HGNC) for human genes and the Mouse Genome Informatics project (MGI) for mouse genes provide authoritative sources of valid, aliased, and outdated symbols, but lack a programmatic interface and correction of symbols converted by spreadsheets. We present HGNChelper, an R package that identifies known aliases and outdated gene symbols based on the HGNC human and MGI mouse gene symbol databases, in addition to common mislabeling introduced by spreadsheets, and provides corrections where possible. HGNChelper identified invalid gene symbols in the most recent Molecular Signatures Database (mSigDB 7.0) and in platform annotation files of the Gene Expression Omnibus, with prevalence ranging from ~3% in recent platforms to 30-40% in the earliest platforms from 2002-03. HGNChelper is installable from CRAN.

HGNChelper: identification and correction of invalid gene symbols for human and mouse

Gene symbols are recognizable identifiers for gene names but are unstable and error-prone due to aliasing, manual entry, and unintentional conversion by spreadsheets to date format. Official gene symbol resources such as HUGO Gene Nomenclature Committee (HGNC) for human genes and the Mouse Genome Informatics project (MGI) for mouse genes provide authoritative sources of valid, aliased, and outdated symbols, but lack a programmatic interface and correction of symbols converted by spreadsheets. We present HGNChelper, an R package that identifies known aliases and outdated gene symbols based on the HGNC human and MGI mouse gene symbol databases, in addition to common mislabeling introduced by spreadsheets, and provides corrections where possible. HGNChelper identified invalid gene symbols in the most recent Molecular Signatures Database (MSigDB 7.0) and in platform annotation files of the Gene Expression Omnibus, with prevalence ranging from ~3% in recent platforms to 30-40% in the earliest platforms from 2002-03. HGNChelper is installable from CRAN.

Whole transcriptome signature for prognostic prediction (WTSPP): application of whole transcriptome signature for prognostic prediction in cancer

Developing prognostic biomarkers for specific cancer types that accurately predict patient survival is increasingly important in clinical research and practice. Despite the enormous potential of prognostic signatures, proposed models have found limited implementations in routine clinical practice. Herein, we propose a generic, RNA sequencing platform independent, statistical framework named whole transcriptome signature for prognostic prediction to generate prognostic gene signatures. Using ovarian cancer and lung adenocarcinoma as examples, we provide evidence that our prognostic signatures overperform previous reported signatures, capture prognostic features not explained by clinical variables, and expose biologically relevant prognostic pathways, including those involved in the immune system and cell cycle. Our approach demonstrates a robust method for developing prognostic gene expression signatures. In conclusion, our statistical framework can be generally applied to all cancer types for prognostic prediction and might be extended to other human diseases. The proposed method is implemented as an R package (PanCancerSig) and is freely available on GitHub ( https://github.com/Cheng-Lab-GitHub/PanCancer_Signature ).

Multiomic Analysis of Subtype Evolution and Heterogeneity in High-Grade Serous Ovarian Carcinoma

Multiple studies have identified transcriptome subtypes of high-grade serous ovarian carcinoma (HGSOC), but their interpretation and translation are complicated by tumor evolution and polyclonality accompanied by extensive accumulation of somatic aberrations, varying cell type admixtures, and different tissues of origin. In this study, we examined the chronology of HGSOC subtype evolution in the context of these factors using a novel integrative analysis of absolute copy-number analysis and gene expression in The Cancer Genome Atlas complemented by single-cell analysis of six independent tumors. Tumor purity, ploidy, and subclonality were reliably inferred from different genomic platforms, and these characteristics displayed marked differences between subtypes. Genomic lesions associated with HGSOC subtypes tended to be subclonal, implying subtype divergence at later stages of tumor evolution. Subclonality of recurrent HGSOC alterations was evident for proliferative tumors, characterized by extreme genomic instability, absence of immune infiltration, and greater patient age. In contrast, differentiated tumors were characterized by largely intact genome integrity, high immune infiltration, and younger patient age. Single-cell sequencing of 42,000 tumor cells revealed widespread heterogeneity in tumor cell type composition that drove bulk subtypes but demonstrated a lack of intrinsic subtypes among tumor epithelial cells. Our findings prompt the dismissal of discrete transcriptome subtypes for HGSOC and replacement by a more realistic model of continuous tumor development that includes mixtures of subclones, accumulation of somatic aberrations, infiltration of immune and stromal cells in proportions correlated with tumor stage and tissue of origin, and evolution between properties previously associated with discrete subtypes. SIGNIFICANCE: This study infers whether transcriptome-based groupings of tumors differentiate early in carcinogenesis and are, therefore, appropriate targets for therapy and demonstrates that this is not the case for HGSOC.

Mechanistically derived patient-level framework for precision medicine identifies a personalized immune prognostic signature in high-grade serous ovarian cancer

An accurate prognosis assessment for cancer patients could aid in guiding clinical decision-making. Reliance on traditional clinical features alone in a complex clinical environment is challenging and unsatisfactory in the era of precision medicine; thus, reliable prognostic biomarkers are urgently required to improve a patient staging system. In this study, we proposed a patient-level computational framework from mechanistic and translational perspectives to establish a personalized prognostic signature (named PLPPS) in high-grade serous ovarian carcinoma (HGSOC). The PLPPS composed of 68 immune genes achieved accurate prognostic risk stratification for 1190 patients in the meta-training cohort and was rigorously validated in multiple cross-platform independent cohorts comprising 792 HGSOC patients. Furthermore, the PLPPS was shown to be the better prognostic factor compared with clinical parameters in the univariate analysis and retained a significant independent association with prognosis after adjusting for clinical parameters in the multivariate analysis. In benchmark comparisons, the performance of PLPPS (hazard ratio (HR), 1.371; concordance index (C-index), 0.604 and area under the curve (AUC), 0.637) is comparable to or better than other published gene signatures (HR, 0.972 to 1.340; C-index, 0.495 to 0.592 and AUC, 0.48-0.624). With further validation in prospective clinical trials, we hope that the PLPPS might become a promising genomic tool to guide personalized management and decision-making of HGSOC in clinical practice.

MicroRNA-1271-5p inhibits the tumorigenesis of ovarian cancer through targeting E2F5 and negatively regulates the mTOR signaling pathway

BACKGROUND

MicroRNA-1271-5p (miR-1271-5p) has been reported to participate in the progression of many human cancers. However, the role of miR-1271-5p still remains unclear in ovarian cancer (OC). Therefore, we explored the effect of miR-1271-5p on the development of OC in present study.

METHODS

We measured the miR-1271-5p expression via the qRT-PCR assay. Then the function of miR-1271-5p was analyzed through MTT and Transwell assays. The relationship among miR-1271-5p and E2F5 was verified by dual luciferase assay. The protein expression levels were examined through western blot.

RESULTS

MiR-1271-5p was downregulated in OC tissues which predicted poor prognosis of OC patients. Moreover, E2F5 was a direct target of miR-1271-5p in OC. And miR-1271-5p suppressed cell proliferation, migration and invasion in OC through targeting E2F5. Furthermore, E2F5 was upregulated in OC tissues which predicted poor prognosis of OC patients. Besides that, miR-1271-5p suppressed EMT and mTOR pathway in OC.

CONCLUSIONS

MiR-1271-5p inhibited the tumorigenesis of OC through targeting E2F5 and negatively regulated the mTOR signaling pathway.

The Impact of Stroma Admixture on Molecular Subtypes and Prognostic Gene Signatures in Serous Ovarian Cancer

BACKGROUND

Recent efforts to improve outcomes for high-grade serous ovarian cancer, a leading cause of cancer death in women, have focused on identifying molecular subtypes and prognostic gene signatures, but existing subtypes have poor cross-study robustness. We tested the contribution of cell admixture in published ovarian cancer molecular subtypes and prognostic gene signatures.

METHODS

Gene signatures of tumor and stroma were developed using paired microdissected tissue from two independent studies. Stromal genes were investigated in two molecular subtype classifications and 61 published gene signatures. Prognostic performance of gene signatures of stromal admixture was evaluated in 2,527 ovarian tumors (16 studies). Computational simulations of increasing stromal cell proportion were performed by mixing gene-expression profiles of paired microdissected ovarian tumor and stroma.

RESULTS

Recently described ovarian cancer molecular subtypes are strongly associated with the cell admixture. Tumors were classified as different molecular subtypes in simulations where the percentage of stromal cells increased. Stromal gene expression in bulk tumors was associated with overall survival (hazard ratio, 1.17; 95% confidence interval, 1.11-1.23), and in one data set, increased stroma was associated with anatomic sampling location. Five published prognostic gene signatures were no longer prognostic in a multivariate model that adjusted for stromal content.

CONCLUSIONS

Cell admixture affects the interpretation and reproduction of ovarian cancer molecular subtypes and gene signatures derived from bulk tissue. Elucidating the role of stroma in the tumor microenvironment and in prognosis is important.

IMPACT

Single-cell analyses may be required to refine the molecular subtypes of high-grade serous ovarian cancer.

Modeling Between-Study Heterogeneity for Improved Replicability in Gene Signature Selection and Clinical Prediction

In the genomic era, the identification of gene signatures associated with disease is of significant interest. Such signatures are often used to predict clinical outcomes in new patients and aid clinical decision-making. However, recent studies have shown that gene signatures are often not replicable. This occurrence has practical implications regarding the generalizability and clinical applicability of such signatures. To improve replicability, we introduce a novel approach to select gene signatures from multiple datasets whose effects are consistently non-zero and account for between-study heterogeneity. We build our model upon some rank-based quantities, facilitating integration over different genomic datasets. A high dimensional penalized Generalized Linear Mixed Model (pGLMM) is used to select gene signatures and address data heterogeneity. We compare our method to some commonly used strategies that select gene signatures ignoring between-study heterogeneity. We provide asymptotic results justifying the performance of our method and demonstrate its advantage in the presence of heterogeneity through thorough simulation studies. Lastly, we motivate our method through a case study subtyping pancreatic cancer patients from four gene expression studies.

Novel Model for Comprehensive Assessment of Robust Prognostic Gene Signature in Ovarian Cancer Across Different Independent Datasets

Different analytical methods or models can often find completely different prognostic biomarkers for the same cancer. In the study of prognostic molecular biomarkers of ovarian cancer (OvCa), different studies have reported a variety of prognostic gene signatures. In the current study, based on geometric concepts, the linearity-clustering phase diagram with integrated P-value (LCP) method was used to comprehensively consider three indicators that are commonly employed to estimate the quality of a prognostic gene signature model. The three indicators, namely, concordance index, area under the curve, and level of the hazard ratio were determined via calculation of the prognostic index of various gene signatures from different datasets. As evaluation objects, we selected 13 gene signature models (Cox regression model) and 16 OvCa genomic datasets (including gene expression information and follow-up data) from published studies. The results of LCP showed that three models were universal and better than other models. In addition, combining the three models into one model showed the best performance in all datasets by LCP calculation. The combination gene signature model provides a more reliable model and could be validated in various datasets of OvCa. Thus, our method and findings can provide more accurate prognostic biomarkers and effective reference for the precise clinical treatment of OvCa.

Dual-platform affinity proteomics identifies links between the recurrence of ovarian carcinoma and proteins released into the tumor microenvironment

The peritoneal fluid (ascites), replete with abundant tumor-promoting factors and extracellular vesicles (EVs) reflecting the tumor secretome, plays an essential role in ovarian high-grade serous carcinoma (HGSC) metastasis and immune suppression. A comprehensive picture of mediators impacting HGSC progression is, however, not available.

Methods: Proteins in ascites from HGSC patients were quantified by the aptamer-based SOMAscan affinity proteomic platform. SOMAscan data were analyzed by bioinformatic methods to reveal clinically relevant links and functional connections, and were validated using the antibody-based proximity extension assay (PEA) Olink platform. Mass spectrometry was used to identify proteins in extracellular microvesicles released by HGSC cells.

Results: Consistent with the clinical features of HGSC, 779 proteins in ascites identified by SOMAscan clustered into groups associated either with metastasis and a short relapse-free survival (RFS), or with immune regulation and a favorable RFS. In total, 346 proteins were linked to OC recurrence in either direction. Reanalysis of 214 of these proteins by PEA revealed an excellent median Spearman inter-platform correlation of ρ=0.82 for the 46 positively RFS-associated proteins in both datasets. Intriguingly, many proteins strongly associated with clinical outcome were constituents of extracellular vesicles. These include proteins either linked to a poor RFS, such as HSPA1A, BCAM and DKK1, or associated with a favorable outcome, such as the protein kinase LCK. Finally, based on these data we defined two protein signatures that clearly classify short-term and long-term relapse-free survivors.

Conclusion: The ascites secretome points to metastasis-promoting events and an anti-tumor response as the major determinants of the clinical outcome of HGSC. Relevant proteins include both bone fide secreted and vesicle-encapsulated polypeptides, many of which have previously not been linked to HGSC recurrence. Besides a deeper understanding of the HGSC microenvironment our data provide novel potential tools for HGSC patient stratification. Furthermore, the first large-scale inter-platform validation of SOMAscan and PEA will be invaluable for other studies using these affinity proteomics platforms.

MetaGxData: Clinically Annotated Breast, Ovarian and Pancreatic Cancer Datasets and their Use in Generating a Multi-Cancer Gene Signature

A wealth of transcriptomic and clinical data on solid tumours are under-utilized due to unharmonized data storage and format. We have developed the MetaGxData package compendium, which includes manually-curated and standardized clinical, pathological, survival, and treatment metadata across breast, ovarian, and pancreatic cancer data. MetaGxData is the largest compendium of curated transcriptomic data for these cancer types to date, spanning 86 datasets and encompassing 15,249 samples. Open access to standardized metadata across cancer types promotes use of their transcriptomic and clinical data in a variety of cross-tumour analyses, including identification of common biomarkers, and assessing the validity of prognostic signatures. Here, we demonstrate that MetaGxData is a flexible framework that facilitates meta-analyses by using it to identify common prognostic genes in ovarian and breast cancer. Furthermore, we use the data compendium to create the first gene signature that is prognostic in a meta-analysis across 3 cancer types. These findings demonstrate the potential of MetaGxData to serve as an important resource in oncology research, and provide a foundation for future development of cancer-specific compendia.

Multi-study factor analysis

We introduce a novel class of factor analysis methodologies for the joint analysis of multiple studies. The goal is to separately identify and estimate (1) common factors shared across multiple studies, and (2) study-specific factors. We develop an Expectation Conditional-Maximization algorithm for parameter estimates and we provide a procedure for choosing the numbers of common and specific factors. We present simulations for evaluating the performance of the method and we illustrate it by applying it to gene expression data in ovarian cancer. In both, we clarify the benefits of a joint analysis compared to the standard factor analysis. We have provided a tool to accelerate the pace at which we can combine unsupervised analysis across multiple studies, and understand the cross-study reproducibility of signal in multivariate data. An R package (MSFA), is implemented and is available on GitHub.

Prediction of response to anti-cancer drugs becomes robust via network integration of molecular data

Despite the widening range of high-throughput platforms and exponential growth of generated data volume, the validation of biomarkers discovered from large-scale data remains a challenging field. In order to tackle cancer heterogeneity and comply with the data dimensionality, a number of network and pathway approaches were invented but rarely systematically applied to this task. We propose a new method, called NEAmarker, for finding sensitive and robust biomarkers at the pathway level. scores from network enrichment analysis transform the original space of altered genes into a lower-dimensional space of pathways. These dimensions are then correlated with phenotype variables. The method was first tested using in vitro data from three anti-cancer drug screens and then on clinical data of The Cancer Genome Atlas. It proved superior to the single-gene and alternative enrichment analyses in terms of (1) universal applicability to different data types with a possibility of cross-platform integration, (2) consistency of the discovered correlates between independent drug screens, and (3) ability to explain differential survival of treated patients. Our new screen of anti-cancer compounds validated the performance of multivariate models of drug sensitivity. The previously proposed methods of enrichment analysis could achieve comparable levels of performance in certain tests. However, only our method could discover predictors of both in vitro response and patient survival given administration of the same drug.

Development and validation of an immune prognostic signature for ovarian carcinoma

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.

Development and validation of an immune gene-set based Prognostic signature in ovarian cancer

BACKGROUND

Ovarian cancer (OV) is the most lethal gynecological cancer in women. We aim to develop a generalized, individualized immune prognostic signature that can stratify and predict overall survival for ovarian cancer.

METHODS

The gene expression profiles of ovarian cancer tumor tissue samples were collected from 17 public cohorts, including 2777 cases totally. Single sample gene set enrichment (ssGSEA) analysis was used for the immune genes from ImmPort database to develop an immune-based prognostic score for OV (IPSOV). The signature was trained and validated in six independent datasets (n = 519, 409, 606, 634, 415, 194).

FINDINGS

The IPSOV significantly stratified patients into low- and high-immune risk groups in the training set and in the 5 validation sets (HR range: 1.71 [95%CI: 1.32-2.19; P = 4.04 × 10-5] to 2.86 [95%CI: 1.72-4.74; P = 4.89 × 10-5]). Further, we compared IPSOV with nine reported ovarian cancer prognostic signatures as well as the clinical characteristics including stage, grade and debulking status. The IPSOV achieved the highest mean C-index (0.625) compared with the other signatures (0.516 to 0.602) and clinical characteristics (0.555 to 0.583). Further, we integrated IPSOV with stage, grade and debulking, which showed improved prognostic accuracy than clinical characteristics only.

INTERPRETATION

The proposed clinical-immune signature is a promising biomarker for estimating overall survival in ovarian cancer. Prospective studies are needed to further validate its analytical accuracy and test the clinical utility. FUND: This work was supported by National Key Research and Development Program of China, National Natural Science Foundation of China and Natural Science Foundation of the Jiangsu Higher Education Institutions of China.

compound.Cox: Univariate feature selection and compound covariate for predicting survival

BACKGROUND AND OBJECTIVE

Univariate feature selection is one of the simplest and most commonly used techniques to develop a multigene predictor for survival. Presently, there is no software tailored to perform univariate feature selection and predictor construction.

METHODS

We develop the compound.Cox R package that implements univariate significance tests (via the Wald tests or score tests) for feature selection. We provide a cross-validation algorithm to measure predictive capability of selected genes and a permutation algorithm to assess the false discovery rate. We also provide three algorithms for constructing a multigene predictor (compound covariate, compound shrinkage, and copula-based methods), which are tailored to the subset of genes obtained from univariate feature selection. We demonstrate our package using survival data on the lung cancer patients. We examine the predictive capability of the developed algorithms by the lung cancer data and simulated data.

RESULTS

The developed R package, compound.Cox, is available on the CRAN repository. The statistical tools in compound.Cox allow researchers to determine an optimal significance level of the tests, thus providing researchers an optimal subset of genes for prediction. The package also allows researchers to compute the false discovery rate and various prediction algorithms.

Identification of grade and origin specific cell populations in serous epithelial ovarian cancer by single cell RNA-seq

Here we investigated different cell populations within ovarian cancer using single-cell RNA seq: fourteen samples from nine patients with differing grades (high grade, low grade and benign) as well as different origin sites (primary and metastatic tumor site, ovarian in origin and fallopian in origin). We were able to identify sixteen distinct cell populations with specific cells correlated to high grade tumors, low grade tumors, benign and one population unique to a patient with a breast cancer relapse. Furthermore the proportion of these populations changes from primary to metastatic in a shift from mainly epithelial cells to leukocytes with few cancer epithelial cells in the metastases. Differential gene expression shows myeloid lineage cells are the primary cell group expressing soluble factors in primary samples while fibroblasts do so in metastatic samples. The leukocytes that were captured did not seem to be suppressed through known pro-tumor cytokines from any of the cell populations. Single cell RNA-seq is necessary to de-tangle cellular heterogeneity for better understanding of ovarian cancer progression.

Consensus on Molecular Subtypes of High-Grade Serous Ovarian Carcinoma

Purpose: The majority of ovarian carcinomas are of high-grade serous histology, which is associated with poor prognosis. Surgery and chemotherapy are the mainstay of treatment, and molecular characterization is necessary to lead the way to targeted therapeutic options. To this end, various computational methods for gene expression-based subtyping of high-grade serous ovarian carcinoma (HGSOC) have been proposed, but their overlap and robustness remain unknown.Experimental Design: We assess three major subtype classifiers by meta-analysis of publicly available expression data, and assess statistical criteria of subtype robustness and classifier concordance. We develop a consensus classifier that represents the subtype classifications of tumors based on the consensus of multiple methods, and outputs a confidence score. Using our compendium of expression data, we examine the possibility that a subset of tumors is unclassifiable based on currently proposed subtypes.Results: HGSOC subtyping classifiers exhibit moderate pairwise concordance across our data compendium (58.9%-70.9%; P < 10^-5) and are associated with overall survival in a meta-analysis across datasets (P < 10^-5). Current subtypes do not meet statistical criteria for robustness to reclustering across multiple datasets (prediction strength < 0.6). A new subtype classifier is trained on concordantly classified samples to yield a consensus classification of patient tumors that correlates with patient age, survival, tumor purity, and lymphocyte infiltration.Conclusions: A new consensus ovarian subtype classifier represents the consensus of methods and demonstrates the importance of classification approaches for cancer that do not require all tumors to be assigned to a distinct subtype. Clin Cancer Res; 24(20); 5037-47. ©2018 AACR.

Consensus on Molecular Subtypes of High-Grade Serous Ovarian Carcinoma

Purpose: The majority of ovarian carcinomas are of high-grade serous histology, which is associated with poor prognosis. Surgery and chemotherapy are the mainstay of treatment, and molecular characterization is necessary to lead the way to targeted therapeutic options. To this end, various computational methods for gene expression-based subtyping of high-grade serous ovarian carcinoma (HGSOC) have been proposed, but their overlap and robustness remain unknown.Experimental Design: We assess three major subtype classifiers by meta-analysis of publicly available expression data, and assess statistical criteria of subtype robustness and classifier concordance. We develop a consensus classifier that represents the subtype classifications of tumors based on the consensus of multiple methods, and outputs a confidence score. Using our compendium of expression data, we examine the possibility that a subset of tumors is unclassifiable based on currently proposed subtypes.Results: HGSOC subtyping classifiers exhibit moderate pairwise concordance across our data compendium (58.9%-70.9%; P < 10-5) and are associated with overall survival in a meta-analysis across datasets (P < 10-5). Current subtypes do not meet statistical criteria for robustness to reclustering across multiple datasets (prediction strength < 0.6). A new subtype classifier is trained on concordantly classified samples to yield a consensus classification of patient tumors that correlates with patient age, survival, tumor purity, and lymphocyte infiltration.Conclusions: A new consensus ovarian subtype classifier represents the consensus of methods and demonstrates the importance of classification approaches for cancer that do not require all tumors to be assigned to a distinct subtype. Clin Cancer Res; 24(20); 5037-47. ©2018 AACR.

Immunogenomic Analyses of Advanced Serous Ovarian Cancer Reveal Immune Score is a Strong Prognostic Factor and an Indicator of Chemosensitivity

Purpose: Ovarian cancer is one of the first human cancers for which in situ immune response was reported to be important for the clinical outcome. To elucidate the mechanistic relationship between immune repertoire and cancer genotype in ovarian cancer, the development of a well-defined immune score for ovarian cancer is required.Experimental Design: From a collection of 2,203 patient samples of advanced ovarian cancer from public available resources, we evaluated the prognostic values for a compendium of immune marker genes and proposed an immune score. The relationships between immune score, tumor-infiltrating immune cells, cancer genotypes, and their impact on patient outcome were characterized.Results: Loss of chemokine and IFNγ pathway genes is frequent in ovarian cancer and is significantly associated with low immune score and poor outcome. Chemotherapy can increase the immune score of tumors by inducing the expression of IFNγ inducible chemokines. High immune score is significantly associated with BRCA1/2 mutation status and the response to chemotherapy. Multivariate analysis revealed that immune score is a strong predictor of patient survival and the response to immunotherapy.Conclusions: Our results reveal the drivers of the immune repertoire of advanced ovarian cancer and demonstrate the importance of immune score as an independent prognostic signature and a potent indicator of intratumoral immune status. Clin Cancer Res; 24(15); 3560-71. ©2018 AACR.

Transcriptomic Characterization of Endometrioid, Clear Cell, and High-Grade Serous Epithelial Ovarian Carcinoma

Background: Endometrioid carcinoma (EC) and clear cell carcinoma (CC) histotypes of epithelial ovarian cancer are understudied compared with the more common high-grade serous carcinomas (HGSC). We therefore sought to characterize EC and CC transcriptomes in relation to HGSC.Methods: Following bioinformatics processing and gene abundance normalization, differential expression analysis of RNA sequence data collected on fresh-frozen tumors was completed with nonparametric statistical analysis methods (55 ECs, 19 CCs, 112 HGSCs). Association of gene expression with progression-free survival (PFS) was completed with Cox proportional hazards models. Eight additional multi-histotype expression array datasets (N = 852 patients) were used for replication.Results: In the discovery set, tumors generally clustered together by histotype. Thirty-two protein-coding genes were differentially expressed across histotype (P < 1 × 10^-10) and showed similar associations in replication datasets, including MAP2K6, KIAA1324, CDH1, ENTPD5, LAMB1, and DRAM1 Nine genes associated with PFS (P < 0.0001) showed similar associations in replication datasets. In particular, we observed shorter PFS time for CC and EC patients with high gene expression for CCNB2, CORO2A, CSNK1G1, FRMD8, LIN54, LINC00664, PDK1, and PEX6, whereas, the converse was observed for HGSC patients.Conclusions: The results suggest important histotype differences that may aid in the development of treatment options, particularly those for patients with EC or CC.Impact: We present replicated findings on transcriptomic differences and how they relate to clinical outcome for two of the rarer ovarian cancer histotypes of EC and CC, along with comparison with the common histotype of HGSC. Cancer Epidemiol Biomarkers Prev; 27(9); 1101-9. ©2018 AACR.

Integrative multi-platform meta-analysis of gene expression profiles in pancreatic ductal adenocarcinoma patients for identifying novel diagnostic biomarkers

Applying differentially expressed genes (DEGs) to identify feasible biomarkers in diseases can be a hard task when working with heterogeneous datasets. Expression data are strongly influenced by technology, sample preparation processes, and/or labeling methods. The proliferation of different microarray platforms for measuring gene expression increases the need to develop models able to compare their results, especially when different technologies can lead to signal values that vary greatly. Integrative meta-analysis can significantly improve the reliability and robustness of DEG detection. The objective of this work was to develop an integrative approach for identifying potential cancer biomarkers by integrating gene expression data from two different platforms. Pancreatic ductal adenocarcinoma (PDAC), where there is an urgent need to find new biomarkers due its late diagnosis, is an ideal candidate for testing this technology. Expression data from two different datasets, namely Affymetrix and Illumina (18 and 36 PDAC patients, respectively), as well as from 18 healthy controls, was used for this study. A meta-analysis based on an empirical Bayesian methodology (ComBat) was then proposed to integrate these datasets. DEGs were finally identified from the integrated data by using the statistical programming language R. After our integrative meta-analysis, 5 genes were commonly identified within the individual analyses of the independent datasets. Also, 28 novel genes that were not reported by the individual analyses (‘gained’ genes) were also discovered. Several of these gained genes have been already related to other gastroenterological tumors. The proposed integrative meta-analysis has revealed novel DEGs that may play an important role in PDAC and could be potential biomarkers for diagnosing the disease.

Data and Statistical Methods To Analyze the Human Microbiome

The Waldron lab for computational biostatistics bridges the areas of cancer genomics and microbiome studies for public health, developing methods to exploit publicly available data resources and to integrate -omics studies.

The Waldron lab for computational biostatistics bridges the areas of cancer genomics and microbiome studies for public health, developing methods to exploit publicly available data resources and to integrate -omics studies.

Training replicable predictors in multiple studies

This article considers replicability of the performance of predictors across studies. We suggest a general approach to investigating this issue, based on ensembles of prediction models trained on different studies. We quantify how the common practice of training on a single study accounts in part for the observed challenges in replicability of prediction performance. We also investigate whether ensembles of predictors trained on multiple studies can be combined, using unique criteria, to design robust ensemble learners trained upfront to incorporate replicability into different contexts and populations.

MARCKS contributes to stromal cancer-associated fibroblast activation and facilitates ovarian cancer metastasis

The Cancer Genome Atlas network has revealed that the 'mesenchymal' epithelial ovarian cancer (EOC) subtype represents the poorest outcome, indicating a crucial role of stromal cancer-associated fibroblasts (CAFs) in disease progression. The cooperative role of CAFs in EOC metastasis has long been recognized, but the mechanisms of stromal CAFs activation are still obscure. Therefore, we carried out an integrative analysis to identify the regulator genes that are responsible for CAFs activation in microdissected tumor stroma profiles. Here, we determined that myristoylated alanine-rich C-kinase substrate (MARCKS) was highly expressed in ovarian stroma, and was required for the differentiation and tumor promoting function of CAFs. Suppression of MARCKS resulted in the loss of CAF features, and diminished role of CAFs in supporting tumor cell growth in 3D organotypic cultures and in murine xenograft model. Mechanistically, we found that MARCKS maintained CAF activation through suppression of cellular senescence and activation of the AKT/Twist1 signaling. Moreover, high MARCKS expression was associated with poor patient survival in EOC. Collectively, our findings identify the potential of MARCKS inhibition as a novel stroma-oriented therapy in EOC.