Why your new cancer biomarker may never work: recurrent patterns and remarkable diversity in biomarker failures

Target Product Profiles for medical tests: a systematic review of current methods

BACKGROUND

A Target Product Profile (TPP) outlines the necessary characteristics of an innovative product to address an unmet clinical need. TPPs could be used to better guide manufacturers in the development of 'fit for purpose' tests, thus increasing the likelihood that novel tests will progress from bench to bedside. However, there is currently no guidance on how to produce a TPP specifically for medical tests.

METHODS

A systematic review was conducted to summarise the methods currently used to develop TPPs for medical tests, the sources used to inform these recommendations and the test characteristics for which targets are made. Database and website searches were conducted in November 2018. TPPs written in English for any medical test were included. Based on an existing framework, test characteristics were clustered into commonly recognised themes.

RESULTS

Forty-four TPPs were identified, all of which focused on diagnostic tests for infectious diseases. Three core decision-making phases for developing TPPs were identified: scoping, drafting and consensus-building. Consultations with experts and the literature mostly informed the scoping and drafting of TPPs. All TPPs provided information on unmet clinical need and desirable analytical performance, and the majority specified clinical validity characteristics. Few TPPs described specifications for clinical utility, and none included cost-effectiveness.

CONCLUSIONS

We have identified a commonly used framework that could be beneficial for anyone interested in drafting a TPP for a medical test. Currently, key outcomes such as utility and cost-effectiveness are largely overlooked within TPPs though and we foresee this as an area for further improvement.

Rapid Prostate Cancer Noninvasive Biomarker Screening Using Segmented Flow Mass Spectrometry-Based Untargeted Metabolomics

Spectrometric methods with rapid biomarker detection capacity through untargeted metabolomics are becoming essential in the clinical cancer research. Liquid chromatography-mass spectrometry (LC-MS) is a rapidly developing metabolomic-based biomarker technique due to its high sensitivity, reproducibility, and separation efficiency. However, its translation to clinical diagnostics is often limited due to long data acquisition times (∼20 min/sample) and laborious sample extraction procedures when employed for large-scale metabolomics studies. Here, we developed a segmented flow approach coupled with high-resolution mass spectrometry (SF-HRMS) for untargeted metabolomics, which has the capability to acquire data in less than 1.5 min/sample with robustness and reproducibility relative to LC-HRMS. The SF-HRMS results demonstrate the capability for screening metabolite-based urinary biomarkers associated with prostate cancer (PCa). The study shows that SF-HRMS-based global metabolomics has the potential to evolve into a rapid biomarker screening tool for clinical research.

Biobanking in Molecular Biomarker Research for the Early Detection of Cancer

Although population-wide screening programs for several cancer types have been implemented in multiple countries, screening procedures are invasive, time-consuming and often perceived as a burden for patients. Molecular biomarkers measurable in non-invasively collected samples (liquid biopsies) could facilitate screening, as they could have incremental value on early diagnosis of cancer, but could also predict prognosis or monitor treatment response. Although the shift towards biomarkers from liquid biopsies for early cancer detection was initiated some time ago, there are many challenges that hamper the development of such biomarkers. One of these challenges is large-scale validation that requires large prospectively collected biobanks with liquid biopsies. Establishing those biobanks involves several considerations, such as standardization of sample collection, processing and storage within and between biobanks. In this perspective, we will elaborate on several issues that need to be contemplated in biobanking, both in general and for certain specimen types specifically, to be able to facilitate biomarker validation for early detection of cancer.

CAncer bioMarker Prediction Pipeline (CAMPP)-A standardized framework for the analysis of quantitative biological data

With the improvement of -omics and next-generation sequencing (NGS) methodologies, along with the lowered cost of generating these types of data, the analysis of high-throughput biological data has become standard both for forming and testing biomedical hypotheses. Our knowledge of how to normalize datasets to remove latent undesirable variances has grown extensively, making for standardized data that are easily compared between studies. Here we present the CAncer bioMarker Prediction Pipeline (CAMPP), an open-source R-based wrapper (https://github.com/ELELAB/CAncer-bioMarker-Prediction-Pipeline -CAMPP) intended to aid bioinformatic software-users with data analyses. CAMPP is called from a terminal command line and is supported by a user-friendly manual. The pipeline may be run on a local computer and requires little or no knowledge of programming. To avoid issues relating to R-package updates, a renv .lock file is provided to ensure R-package stability. Data-management includes missing value imputation, data normalization, and distributional checks. CAMPP performs (I) k-means clustering, (II) differential expression/abundance analysis, (III) elastic-net regression, (IV) correlation and co-expression network analyses, (V) survival analysis, and (VI) protein-protein/miRNA-gene interaction networks. The pipeline returns tabular files and graphical representations of the results. We hope that CAMPP will assist in streamlining bioinformatic analysis of quantitative biological data, whilst ensuring an appropriate bio-statistical framework.

Biomarkers for Early Cancer Diagnosis: Prospects for Success through the Lens of Tumor Genetics

Thousands of candidate cancer biomarkers have been proposed, but so far, few are used in cancer screening. Failure to implement these biomarkers is attributed to technical and design flaws in the discovery and validation phases, but a major obstacle stems from cancer biology itself. Oncogenomics has revealed broad genetic heterogeneity among tumors of the same histology and same tissue (or organ) from different patients, while tumors of different tissue origins also share common genetic mutations. Moreover, there is wide intratumor genetic heterogeneity among cells within any single neoplasm. These findings seriously limit the prospects of finding a single biomarker with high specificity for early cancer detection. Current research focuses on developing biomarker panels, with data assessment by machine-learning algorithms. Whether such approaches will overcome the inherent limitations posed by tumor biology and lead to tests with true clinical value remains to be seen.

Noncoding RNAs orchestrate cell growth, death and drug resistance in renal cell carcinoma

Aim: We aimed to explore the roles of noncoding RNAs (ncRNAs) in renal cell carcinoma. Materials & methods: The altered expressions of miR-196a2, miR-499a, H19, MALAT1 and GAS5, as well as some target transcripts were identified by quantitative real-time reverse transcription polymerase chain reaction. Results: Up-regulation of miR-196a2, E2F3, HSPA4 and MALAT1 (median fold change: 5.69, 25.6, 4.15 and 19.6, respectively) and down-regulation of miR-499a, GAS5, PDCD4, ANXA1 and DFFA (median fold change: 0.28, 0.25, 0.12, 0.09 and 0.08, respectively) were reported compared with paired non-cancer tissue. PDCD4, DFFA and GAS5 down-regulation was associated with poor prognosis in terms of high grade, larger tumor, nodal invasion, capsular and pelvic infiltration. Conclusion: The identified ncRNAs could represent potential theranostic biomarkers for renal cell carcinoma.

Mass spectrometry-based metabolomics in health and medical science: a systematic review

Metabolomics is the study of the investigation of small molecules derived from cellular and organism metabolism, which reflects the outcomes of the complex network of biochemical reactions in living systems.

Biomarkers in respiratory diseases

The December issue of Breathe focuses on biomarkers in respiratory diseases: read the introductory editorial by Chief Editor @ClaudiaCDobler http://bit.ly/36nzAiW.

Systematic review and REMARK scoring of renal cell carcinoma prognostic circulating biomarker manuscripts

Background

No validated molecular biomarkers exist to help guide prognosis of renal cell carcinoma (RCC) patients. We seek to evaluate the quality of published prognostic circulating RCC biomarker manuscripts using the Reporting Recommendations for Tumor Marker Prognostic Studies (REMARK) guidelines.

Methods

The phrase “(renal cell carcinoma OR renal cancer OR kidney cancer OR kidney carcinoma) AND circulating AND (biomarkers OR cell free DNA OR tumor DNA OR methylated cell free DNA OR methylated tumor DNA)” was searched in Embase, Medline and PubMed March 2018. Relevant manuscripts were scored using 48 REMARK sub-criteria for a maximal score of 20 points.

Results

The search identified 535 publications: 33 were manuscripts of primary research and were analyzed. The mean REMARK score was 10.6 (range 6.42–14.2). All manuscripts stated their biomarker, study objectives and method of case selection. The lowest scoring criteria: time lapse between storage of blood/serum and marker assay (n = 2) and lack of flow diagram (n = 2). REMARK scores were significantly higher in publications stating adherence to REMARK guidelines (p = 0.0307) and reporting statistically significant results (p = 0.0318).

Conclusions

Most RCC prognostic biomarker manuscripts poorly adhere to the REMARK guidelines. Better designed studies and appropriate reporting are required to address this urgent unmet need.

CDK1, CCNB1, CDC20, BUB1, MAD2L1, MCM3, BUB1B, MCM2, and RFC4 May Be Potential Therapeutic Targets for Hepatocellular Carcinoma Using Integrated Bioinformatic Analysis

Hepatocellular carcinoma (HCC) is a malignant tumor with high mortality. The abnormal expression of genes is significantly related to the occurrence of HCC. The aim of this study was to explore the differentially expressed genes (DEGs) of HCC and to provide bioinformatics basis for the occurrence, prevention and treatment of HCC. The DEGs of HCC and normal tissues in GSE102079, GSE121248, GSE84402 and GSE60502 were obtained using R language. The GO function analysis and KEGG pathway enrichment analysis of DEGs were carried out using the DAVID database. Then, the protein–protein interaction (PPI) network was constructed using the STRING database. Hub genes were screened using Cytoscape software and verified using the GEPIA, UALCAN, and Oncomine database. We used HPA database to exhibit the differences in protein level of hub genes and used LinkedOmics to reveal the relationship between candidate genes and tumor clinical features. Finally, we obtained transcription factor (TF) of hub genes using NetworkAnalyst online tool. A total of 591 overlapping up-regulated genes were identified. These genes were related to cell cycle, DNA replication, pyrimidine metabolism, and p53 signaling pathway. Additionally, the GEPIA database showed that the CDK1, CCNB1, CDC20, BUB1, MAD2L1, MCM3, BUB1B, MCM2, and RFC4 were associated with the poor survival of HCC patients. UALCAN, Oncomine, and HPA databases and qRT-PCR confirmed that these genes were highly expressed in HCC tissues. LinkedOmics database indicated these genes were correlated with overall survival, pathologic stage, pathology T stage, race, and the age of onset. TF analysis showed that MYBL2, KDM5B, MYC, SOX2, and E2F4 were regulators to these nine hub genes. Overexpression of CDK1, CCNB1, CDC20, BUB1, MAD2L1, MCM3, BUB1B, MCM2, and RFC4 in tumor tissues predicted poor survival in HCC. They may be potential therapeutic targets for HCC.

False-positive pathology: improving reproducibility with the next generation of pathologists

The external validity of the scientific literature has recently come into question, popularly referred to as the "reproducibility crisis." It is now generally acknowledged that too many false positive or non-reproducible results are being published throughout the biomedical and social science literature due to misaligned incentives and poor methodology. Pathology is likely no exception to this problem, and may be especially prone to false positives due to common observational methodologies used in our research. Spurious findings in pathology contribute inefficiency to the scientific literature and detrimentally influence patient care. In particular, false positives in pathology affect patients through biomarker development, prognostic classification, and cancer overdiagnosis. We discuss possible sources of non-reproducible pathology studies and describe practical ways our field can improve research habits, especially among trainees.

Diagnostic Tumor Markers in Head and Neck Squamous Cell Carcinoma (HNSCC) in the Clinical Setting

Head and neck squamous cell carcinoma (HNSCC) represents a group of tumors arising in the oral cavity, oropharynx, and larynx. Although HNSCC is traditionally associated with tobacco and alcohol consumption, a growing proportion of head and neck tumors, mainly of the oropharynx, are associated with Human Papilloma Virus (HPV). Recurrent/metastatic disease is characterized by dismal prognosis and there is an unmet need for the development of biomarkers for detection of early disease, accurate prediction of prognosis, and appropriate selection of therapy. Based on the REMARK guidelines, a variety of diagnostic and prognostic biomarkers are being evaluated in clinical trials but their clinical significance is doubtful. Herein, we will focus on biomarkers in HNSCC used in the clinical setting and we will illustrate their clinical relevance.

Reproducibility and Transparency by Design

The reproducibility of bioinformatics analyses can be elevated to equal status with biological discovery. To achieve this, reproducibility must become part of the process, not an afterthought.

Diagnosis: From classification to prediction

Over the last two decades diagnostic labels have increasingly been sub-divided based on molecular and genetic 'signatures'. But this emphasis on disease sub-types defined in molecular terms, elides the central role of population-based predictive technologies in determining these new diagnoses. While molecular diagnostic sub-types might flow from the laboratory, the clinical validity of every putative diagnostic category must ultimately be tested against its predictive powers. In effect, the former logic of prognosis following diagnosis is reversed. This paper explores the emergence of this new method of diagnostic practice over the last half century.

The Search for Clinically Useful Biomarkers of Complex Disease: A Data Analysis Perspective

Unmet clinical diagnostic needs exist for many complex diseases, which (it is hoped) will be solved by the discovery of metabolomics biomarkers. However, at present, no diagnostic tests based on metabolomics have yet been introduced to the clinic. This review is presented as a research perspective on how data analysis methods in metabolomics biomarker discovery may contribute to the failure of biomarker studies and suggests how such failures might be mitigated. The study design and data pretreatment steps are reviewed briefly in this context, and the actual data analysis step is examined more closely.

Epigenetic IVD Tests for Personalized Precision Medicine in Cancer

Epigenetic alterations play a key role in the initiation and progression of cancer. Therefore, it is possible to use epigenetic marks as biomarkers for predictive and precision medicine in cancer. Precision medicine is poised to impact clinical practice, patients, and healthcare systems. The objective of this review is to provide an overview of the epigenetic testing landscape in cancer by examining commercially available epigenetic-based in vitro diagnostic tests for colon, breast, cervical, glioblastoma, lung cancers, and for cancers of unknown origin. We compile current commercial epigenetic tests based on epigenetic biomarkers (i.e., DNA methylation, miRNAs, and histones) that can actually be implemented into clinical practice.

Analysis of DNA methylation in cancer: location revisited

Changes in DNA methylation in cancer have been heralded as promising targets for the development of powerful diagnostic, prognostic, and predictive biomarkers. Despite the existence of more than 14,000 scientific publications describing DNA methylation-based biomarkers and their clinical associations in cancer, only 14 of these biomarkers have been translated into a commercially available clinical test. Methodological and experimental obstacles are both major causes of this disparity, but the genomic location of a DNA methylation-based biomarker is an intrinsic and essential property that also has an important and often overlooked role. Here, we examine the importance of the location of DNA methylation for the development of cancer biomarkers, and take a detailed look at the genomic location and other relevant characteristics of the various biomarkers with commercially available tests. We also emphasize the value of publicly available databases for the development of DNA methylation-based biomarkers and the importance of accurate reporting of the full methodological details of research findings.

The sensitivity and specificity of serum glycan-based biomarkers for cancer detection

Most of clinically used serum biomarkers for cancer detection were established in early 1980s when the Nobel Prize in physiology or medicine was awarded for the "discovery of the principle for the production of monoclonal antibodies." Using this "Nobel" technology, various monoclonal antibodies were obtained when different types of cancer cells were injected into mice and the ligands on the cancer cell surface were characterized. Both aberrant glycan structures and aberrant glycan-associated glycoproteins were revealed as a common feature of cancer cell surfaces through the specific interactions with the monoclonal antibodies. These results indicate that the biosynthesis of the environment-sensitive glycan structures goes awry in cancer cells, which is beyond genetic mutations. Later on, the glycan-related biomarkers were detected in the sera of cancer patients and then developed into serum biomarkers, such as CA125, CA153, CA195, CA199, CA242, CA27.29, CA50, and CA724, which are still in clinical use as of today. During the past 30 years, even with the advancement of different OMICS technologies not limited to genomics, epigenomics, proteomics, glycomics, lipidomics, and metabolomics, very few serum biomarkers have been introduced into clinical practice. The reason is that most of the newly discovered cancer biomarkers are inferior in terms of sensitivity and specificity to these biomarkers. We will summarize the reported sensitivity and specificity of currently used cancer biomarkers, especially the glycan-related biomarkers, in the forms of tables and radar plots and discuss the pros and cons of currently used cancer biomarkers.

The Possibility of Systematic Research Fraud Targeting Under-Studied Human Genes: Causes, Consequences, and Potential Solutions

A major reason for biomarker failure is the selection of candidate biomarkers based on inaccurate or incorrect published results. Incorrect research results leading to the selection of unproductive biomarker candidates are largely considered to stem from unintentional research errors. The additional possibility that biomarker research may be actively misdirected by research fraud has been given comparatively little consideration. This review discusses what we believe to be a new threat to biomarker research, namely, the possible systematic production of fraudulent gene knockdown studies that target under-studied human genes. We describe how fraudulent papers may be produced in series by paper mills using what we have described as a ‘theme and variations’ model, which could also be considered a form of salami slicing. We describe features of these single-gene knockdown publications that may allow them to evade detection by journal editors, peer reviewers, and readers. We then propose a number of approaches to facilitate their detection, including improved awareness of the features of publications constructed in series, broader requirements to post submitted manuscripts to preprint servers, and the use of semi-automated literature screening tools. These approaches may collectively improve the detection of fraudulent studies that might otherwise impede future biomarker research.

GradientScanSurv-An exhaustive association test method for gene expression data with censored survival outcome

Accurate assessment of the association between continuous variables such as gene expression and survival is a critical aspect of precision medicine. In this report, we provide a review of some of the available survival analysis and validation tools by referencing published studies that have utilized these tools. We have identified pitfalls associated with the assumptions inherent in those applications that have the potential to impact scientific research through their potential bias. In order to overcome these pitfalls, we have developed a novel method that enables the logrank test method to handle continuous variables that comprehensively evaluates survival association with derived aggregate statistics. This is accomplished by exhaustively considering all the cutpoints across the full expression gradient. Direct side-by-side comparisons, global ROC analysis, and evaluation of the ability to capture relevant biological themes based on current understanding of RAS biology all demonstrated that the new method shows better consistency between multiple datasets of the same disease, better reproducibility and robustness, and better detection power to uncover biological relevance within the selected datasets over the available survival analysis methods on univariate gene expression and penalized linear model-based methods.

Determining consistent prognostic biomarkers of overall survival and vascular invasion in hepatocellular carcinoma

Background: Potential prognostic biomarker candidates for hepatocellular carcinoma (HCC) are abundant, but their generalizability is unexplored. We cross-validated markers of overall survival (OS) and vascular invasion in independent datasets. Methods: The literature search yielded 318 genes related to survival and 52 related to vascular invasion. Validation was performed in three datasets (RNA-seq, n = 371; Affymetrix arrays, n = 91; Illumina gene chips, n = 135) by uni- and multivariate Cox regression and Mann-Whitney U-test, separately for Asian and Caucasian patients. Results: One hundred and eighty biomarkers remained significant in Asian and 128 in Caucasian subjects at p < 0.05. After multiple testing correction BIRC5 (p = 1.9 × 10^-10), CDC20 (p = 2.5 × 10^-9) and PLK1 (p = 3 × 10^-9) endured as best performing genes in Asian patients; however, none remained significant in the Caucasian cohort. In a multivariate analysis, significance was reached by stage (p = 0.0018) and expression of CENPH (p = 0.0038) and CDK4 (p = 0.038). KIF18A was the only gene predicting vascular invasion in the Affymetrix and Illumina cohorts (p = 0.003 and p = 0.025, respectively). Conclusion: Overall, about half of biomarker candidates failed to retain prognostic value and none were better than stage predicting OS. Impact: Our results help to eliminate biomarkers with limited capability to predict OS and/or vascular invasion.

Similarities in Blood Mononuclear Cell Membrane Phospholipid Profiles During Malignancy

Phospholipids (PLs), key elements of cellular membranes, are regulated reciprocally with membrane proteins and can act as sensors for alterations in physiological or pathological states of cells including initiation and development of cancer. On the other hand, peripheral blood mononuclear cells (MNCs) play an important role in antitumor immune response by reacting to cancerous modifications in distant organs. In the current study, we tested the hypothesis that tumor initiation and development are reflected in the alteration pattern of the MNC PL component. We analyzed MNC membrane PL fractions in samples from healthy individuals and from patients with diverse types of cancers to reveal possible alterations induced by malignancy. Compared to healthy controls, the cancer samples demonstrated shifts in several membrane PL profiles. In particular, when analyzing cancer data pooled together, there were significantly higher levels in lysophosphatidylcholine, phosphatidylcholine, and phosphatidylethanolamine fractions, and significantly lower quantities in phosphatidylinositol, phosphatidylserine, and phosphatidic acid fractions in cancer samples compared to controls. The levels of sphingomyelins and diphosphatidylglycerols were relatively unaffected. Most of the differences in PLs were sustained during the analysis of individual cancers such as breast cancer and chronic lymphocytic leukemia. Our findings suggest the presence of a common pattern of changes in MNC PLs during malignancy.

When Is Enough, Enough? When Are More Observational Epidemiologic Studies Needed to Resolve a Research Question: Illustrations Using Biomarker-Cancer Associations

BACKGROUND

Research reproducibility is vital for translation of epidemiologic findings. However, repeated studies of the same question may be undertaken without enhancing existing knowledge. To identify settings in which additional research is or is not warranted, we adapted research synthesis metrics to determine number of additional observational studies needed to change the inference from an existing meta-analysis.

METHODS

The fail-safe number (FSN) estimates number of additional studies of average weight and null effect needed to drive a statistically significant meta-analysis to null (P ≥ 0.05). We used conditional power to determine number of additional studies of average weight and equivalent heterogeneity to achieve 80% power in an updated meta-analysis to detect the observed summary estimate as statistically significant. We applied these metrics to a curated set of 98 meta-analyses on biomarkers and cancer risk.

RESULTS

Both metrics were influenced by number of studies, heterogeneity, and summary estimate size in the existing meta-analysis. For the meta-analysis on Helicobacter pylori and gastric cancer with 15 studies [OR = 2.29; 95% confidence interval (CI), 1.71-3.05], FSN was 805 studies, supporting futility of further study. For the meta-analysis on dehydroepiandrosterone sulfate and prostate cancer with 7 studies (OR = 1.29; 95% CI, 0.99-1.69), 5 more studies would be needed for 80% power, suggesting further study could change inferences.

CONCLUSIONS

Along with traditional assessments, these metrics could be used by stakeholders to decide whether additional studies addressing the same question are needed.

IMPACT

Systematic application of these metrics could lead to more judicious use of resources and acceleration from discovery to population-health impact.

Prediction of early breast cancer patient survival using ensembles of hypoxia signatures

Background

Biomarkers are a key component of precision medicine. However, full clinical integration of biomarkers has been met with challenges, partly attributed to analytical difficulties. It has been shown that biomarker reproducibility is susceptible to data preprocessing approaches. Here, we systematically evaluated machine-learning ensembles of preprocessing methods as a general strategy to improve biomarker performance for prediction of survival from early breast cancer.

Results

We risk stratified breast cancer patients into either low-risk or high-risk groups based on four published hypoxia signatures (Buffa, Winter, Hu, and Sorensen), using 24 different preprocessing approaches for microarray normalization. The 24 binary risk profiles determined for each hypoxia signature were combined using a random forest to evaluate the efficacy of a preprocessing ensemble classifier. We demonstrate that the best way of merging preprocessing methods varies from signature to signature, and that there is likely no ‘best’ preprocessing pipeline that is universal across datasets, highlighting the need to evaluate ensembles of preprocessing algorithms. Further, we developed novel signatures for each preprocessing method and the risk classifications from each were incorporated in a meta-random forest model. Interestingly, the classification of these biomarkers and its ensemble show striking consistency, demonstrating that similar intrinsic biological information are being faithfully represented. As such, these classification patterns further confirm that there is a subset of patients whose prognosis is consistently challenging to predict.

Conclusions

Performance of different prognostic signatures varies with pre-processing method. A simple classifier by unanimous voting of classifications is a reliable way of improving on single preprocessing methods. Future signatures will likely require integration of intrinsic and extrinsic clinico-pathological variables to better predict disease-related outcomes.

Practical guide for identifying unmet clinical needs for biomarkers

The development and evaluation of novel biomarkers and testing strategies requires a close examination of existing clinical pathways, including mapping of current pathways and identifying areas of unmet need. This approach enables early recognition of analytical and clinical performance criteria to guide evaluation studies, in a cyclical and iterative manner, all the time keeping the clinical pathway and patient health outcomes as the key drivers in the process. The Test Evaluation Working Group of the European Federation of Clinical Chemistry and Laboratory Medicine (EFLM TE-WG) https://www.eflm.eu/site/page/a/1158 has published a conceptual framework of the test evaluation cycle which is driven by the clinical pathway, inherent to which is the test purpose and role within the pathway that are defined by clinical need. To supplement this framework, the EFLM TE-WG has also published an interactive checklist for identifying unmet clinical needs for new biomarkers; a practical tool that laboratories, clinicians, researchers and industry can equally use in a consistent manner when new tests are developed and before they are released to the market. It is hoped that these practical tools will provide consistent and appropriate terminology in this diverse field and offer a platform that facilitates greater consultation and collaboration between all stakeholders. The checklist should assist the work of all colleagues involved in the discovery of novel biomarkers and implementation of new medical tests. The tool is aligned with the IOM recommendations and the FDA and CE regulating body's requirements.

Multivariate techniques enable a biochemical classification of children with autism spectrum disorder versus typically-developing peers: A comparison and validation study

Autism spectrum disorder (ASD) is a developmental disorder which is currently only diagnosed through behavioral testing. Impaired folate‐dependent one carbon metabolism (FOCM) and transsulfuration (TS) pathways have been implicated in ASD, and recently a study involving multivariate analysis based upon Fisher Discriminant Analysis returned very promising results for predicting an ASD diagnosis. This article takes another step toward the goal of developing a biochemical diagnostic for ASD by comparing five classification algorithms on existing data of FOCM/TS metabolites, and also validating the classification results with new data from an ASD cohort. The comparison results indicate a high sensitivity and specificity for the original data set and up to a 88% correct classification of the ASD cohort at an expected 5% misclassification rate for typically‐developing controls. These results form the foundation for the development of a biochemical test for ASD which promises to aid diagnosis of ASD and provide biochemical understanding of the disease, applicable to at least a subset of the ASD population.

Large-Scale Label-Free Quantitative Mapping of the Sputum Proteome

Analysis of induced sputum supernatant is a minimally invasive approach to study the epithelial lining fluid and, thereby, provide insight into normal lung biology and the pathobiology of lung diseases. We present here a novel proteomics approach to sputum analysis developed within the U-BIOPRED (unbiased biomarkers predictive of respiratory disease outcomes) international project. We present practical and analytical techniques to optimize the detection of robust biomarkers in proteomic studies. The normal sputum proteome was derived using data-independent HDMS^E applied to 40 healthy nonsmoking participants, which provides an essential baseline from which to compare modulation of protein expression in respiratory diseases. The "core" sputum proteome (proteins detected in ≥40% of participants) was composed of 284 proteins, and the extended proteome (proteins detected in ≥3 participants) contained 1666 proteins. Quality control procedures were developed to optimize the accuracy and consistency of measurement of sputum proteins and analyze the distribution of sputum proteins in the healthy population. The analysis showed that quantitation of proteins by HDMS^E is influenced by several factors, with some proteins being measured in all participants' samples and with low measurement variance between samples from the same patient. The measurement of some proteins is highly variable between repeat analyses, susceptible to sample processing effects, or difficult to accurately quantify by mass spectrometry. Other proteins show high interindividual variance. We also highlight that the sputum proteome of healthy individuals is related to sputum neutrophil levels, but not gender or allergic sensitization. We illustrate the importance of design and interpretation of disease biomarker studies considering such protein population and technical measurement variance.

Digitizing omics profiles by divergence from a baseline

Technological advances enable increasingly comprehensive profiling of the molecular landscapes of cells, and these data can inform the personalized treatment of complex diseases. Two major obstacles are the complexity of these data and the high degree of person-to-person heterogeneity. We develop a highly simplified, personalized data representation by comparing the profile of an individual to the range of landscapes in a baseline population, thereby mimicking basic clinical diagnostic testing for departures of selected variables from normal levels. Moreover, our method can be applied to any data modality and at any level of granularity, from single features to any subset of features treated as a single entity, for example the gene expression levels in a pathway. Experiments involve both healthy human tissues and various cancer subtypes.

Data collected from omics technologies have revealed pervasive heterogeneity and stochasticity of molecular states within and between phenotypes. A prominent example of such heterogeneity occurs between genome-wide mRNA, microRNA, and methylation profiles from one individual tumor to another, even within a cancer subtype. However, current methods in bioinformatics, such as detecting differentially expressed genes or CpG sites, are population-based and therefore do not effectively model intersample diversity. Here we introduce a unified theory to quantify sample-level heterogeneity that is applicable to a single omics profile. Specifically, we simplify an omics profile to a digital representation based on the omics profiles from a set of samples from a reference or baseline population (e.g., normal tissues). The state of any subprofile (e.g., expression vector for a subset of genes) is said to be “divergent” if it lies outside the estimated support of the baseline distribution and is consequently interpreted as “dysregulated” relative to that baseline. We focus on two cases: single features (e.g., individual genes) and distinguished subsets (e.g., regulatory pathways). Notably, since the divergence analysis is at the individual sample level, dysregulation can be analyzed probabilistically; for example, one can estimate the probability that a gene or pathway is divergent in some population. Finally, the reduction in complexity facilitates a more “personalized” and biologically interpretable analysis of variation, as illustrated by experiments involving tissue characterization, disease detection and progression, and disease–pathway associations.

Methylation-to-Expression Feature Models of Breast Cancer Accurately Predict Overall Survival, Distant-Recurrence Free Survival, and Pathologic Complete Response in Multiple Cohorts

Prognostic biomarkers serve a variety of purposes in cancer treatment and research, such as prediction of cancer progression, and treatment eligibility. Despite growing interest in multi-omic data integration for defining prognostic biomarkers, validated methods have been slow to emerge. Given that breast cancer has been the focus of intense research, it is amenable to studying the benefits of multi-omic prognostic models due to the availability of datasets. Thus, we examined the efficacy of our methylation-to-expression feature model (M2EFM) approach to combining molecular and clinical predictors to create risk scores for overall survival, distant metastasis, and chemosensitivity in breast cancer. Gene expression, DNA methylation, and clinical variables were integrated via M2EFM to build models of overall survival using 1028 breast tumor samples and applied to validation cohorts of 61 and 327 samples. Models of distant recurrence-free survival and pathologic complete response were built using 306 samples and validated on 182 samples. Despite different populations and assays, M2EFM models validated with good accuracy (C-index or AUC ≥ 0.7) for all outcomes and had the most consistent performance compared to other methods. Finally, we demonstrated that M2EFM identifies functionally relevant genes, which could be useful in translating an M2EFM biomarker to the clinic.

Cell-based reference samples designed with specific differences in microRNA biomarkers

BACKGROUND

We demonstrate the feasibility of creating a pair of reference samples to be used as surrogates for clinical samples measured in either a research or clinical laboratory setting. The reference sample paradigm presented and evaluated here is designed to assess the capability of a measurement process to detect true differences between two biological samples. Cell-based reference samples can be created with a biomarker signature pattern designed in silico. Clinical laboratories working in regulated applications are required to participate in proficiency testing programs; research laboratories doing discovery typically do not. These reference samples can be used in proficiency tests or as process controls that allow a laboratory to evaluate and optimize its measurement systems, monitor performance over time (process drift), assess changes in protocols, reagents, and/or personnel, maintain standard operating procedures, and most importantly, provide evidence for quality results.

RESULTS

The biomarkers of interest in this study are microRNAs (miRNAs), small non-coding RNAs involved in the regulation of gene expression. Multiple lung cancer associated cell lines were determined by reverse transcription (RT)-PCR to have sufficiently different miRNA profiles to serve as components in mixture designs as reference samples. In silico models based on the component profiles were used to predict miRNA abundance ratios between two different cell line mixtures, providing target values for profiles obtained from in vitro mixtures. Two reference sample types were tested: total RNA mixed after extraction from cell lines, and intact cells mixed prior to RNA extraction. MicroRNA profiling of a pair of samples composed of extracted RNA derived from these cell lines successfully replicated the target values. Mixtures of intact cells from these lines also approximated the target values, demonstrating potential utility as mimics for clinical specimens. Both designs demonstrated their utility as reference samples for inter- or intra-laboratory testing.

CONCLUSIONS

Cell-based reference samples can be created for performance assessment of a measurement process from biomolecule extraction through quantitation. Although this study focused on miRNA profiling with RT-PCR using cell lines associated with lung cancer, the paradigm demonstrated here should be extendable to genome-scale platforms and other biomolecular endpoints.

Protein Biomarkers for Early Detection of Pancreatic Ductal Adenocarcinoma: Progress and Challenges

Approximately 75% of patients with pancreatic ductal adenocarcinoma are diagnosed with advanced cancer, which cannot be safely resected. The most commonly used biomarker CA19-9 has inadequate sensitivity and specificity for early detection, which we define as Stage I/II cancers. Therefore, progress in next-generation biomarkers is greatly needed. Recent reports have validated a number of biomarkers, including combination assays of proteins and DNA mutations; however, the history of translating promising biomarkers to clinical utility suggests that several major hurdles require careful consideration by the medical community. The first set of challenges involves nominating and verifying biomarkers. Candidate biomarkers need to discriminate disease from benign controls with high sensitivity and specificity for an intended use, which we describe as a two-tiered strategy of identifying and screening high-risk patients. Community-wide efforts to share samples, data, and analysis methods have been beneficial and progress meeting this challenge has been achieved. The second set of challenges is assay optimization and validating biomarkers. After initial candidate validation, assays need to be refined into accurate, cost-effective, highly reproducible, and multiplexed targeted panels and then validated in large cohorts. To move the most promising candidates forward, ideally, biomarker panels, head-to-head comparisons, meta-analysis, and assessment in independent data sets might mitigate risk of failure. Much more investment is needed to overcome these challenges. The third challenge is achieving clinical translation. To moonshot an early detection test to the clinic requires a large clinical trial and organizational, regulatory, and entrepreneurial know-how. Additional factors, such as imaging technologies, will likely need to improve concomitant with molecular biomarker development. The magnitude of the clinical translational challenge is uncertain, but interdisciplinary cooperation within the PDAC community is poised to confront it.

Contribution of bladder cancer pathology assessment in planning clinical trials

Bladder cancer is a heterogeneous disease that demonstrates a wide spectrum of histologic features. The modern classification of bladder cancer is largely based on pathologic analysis, which assesses tumor grade, stage, type, size, and other features that are essential for understanding the biological behavior of bladder cancer. Bladder cancers with similar histologic features are likely to show comparable responses to a new therapeutic agent in clinical trial. Furthermore, pathologic analysis also evaluates the quality of tissue samples in clinical trial to ensure the integrity of various molecular tests. In spite of the emerging role of genomic and molecular studies, pathology remains the cornerstone in the diagnosis, prognosis, and treatment of bladder cancer. Herein, the pathologic considerations for bladder cancer clinical trial planning are reviewed.

Time for change: a new training programme for morpho-molecular pathologists?

The evolution of cellular pathology as a specialty has always been driven by technological developments and the clinical relevance of incorporating novel investigations into diagnostic practice. In recent years, the molecular characterisation of cancer has become of crucial relevance in patient treatment both for predictive testing and subclassification of certain tumours. Much of this has become possible due to the availability of next-generation sequencing technologies and the whole-genome sequencing of tumours is now being rolled out into clinical practice in England via the 100 000 Genome Project. The effective integration of cellular pathology reporting and genomic characterisation is crucial to ensure the morphological and genomic data are interpreted in the relevant context, though despite this, in many UK centres molecular testing is entirely detached from cellular pathology departments. The CM-Path initiative recognises there is a genomics knowledge and skills gap within cellular pathology that needs to be bridged through an upskilling of the current workforce and a redesign of pathology training. Bridging this gap will allow the development of an integrated ‘morphomolecular pathology’ specialty, which can maintain the relevance of cellular pathology at the centre of cancer patient management and allow the pathology community to continue to be a major influence in cancer discovery as well as playing a driving role in the delivery of precision medicine approaches. Here, several alternative models of pathology training, designed to address this challenge, are presented and appraised.

CANcer-specific Evaluation System (CANES): a high-accuracy platform, for preclinical single/multi-biomarker discovery

The recent creation of enormous, cancer-related “Big Data” public depositories represents a powerful means for understanding tumorigenesis. However, a consistently accurate system for clinically evaluating single/multi-biomarkers remains lacking, and it has been asserted that oft-failed clinical advancement of biomarkers occurs within the very early stages of biomarker assessment. To address these challenges, we developed a clinically testable, web-based tool, CANcer-specific single/multi-biomarker Evaluation System (CANES), to evaluate biomarker effectiveness, across 2,134 whole transcriptome datasets, from 94,147 biological samples (from 18 tumor types). For user-provided single/multi-biomarkers, CANES evaluates the performance of single/multi-biomarker candidates, based on four classification methods, support vector machine, random forest, neural networks, and classification and regression trees. In addition, CANES offers several advantages over earlier analysis tools, including: 1) survival analysis; 2) evaluation of mature miRNAs as markers for user-defined diagnostic or prognostic purposes; and 3) provision of a “pan-cancer” summary view, based on each single marker. We believe that such “landscape” evaluation of single/multi-biomarkers, for diagnostic therapeutic/prognostic decision-making, will be highly valuable for the discovery and “repurposing” of existing biomarkers (and their specific targeted therapies), leading to improved patient therapeutic stratification, a key component of targeted therapy success for the avoidance of therapy resistance.

Mass spectrometry-assisted gel-based proteomics in cancer biomarker discovery: approaches and application

There is a critical need for the discovery of novel biomarkers for early detection and targeted therapy of cancer, a major cause of deaths worldwide. In this respect, proteomic technologies, such as mass spectrometry (MS), enable the identification of pathologically significant proteins in various types of samples. MS is capable of high-throughput profiling of complex biological samples including blood, tissues, urine, milk, and cells. MS-assisted proteomics has contributed to the development of cancer biomarkers that may form the foundation for new clinical tests. It can also aid in elucidating the molecular mechanisms underlying cancer. In this review, we discuss MS principles and instrumentation as well as approaches in MS-based proteomics, which have been employed in the development of potential biomarkers. Furthermore, the challenges in validation of MS biomarkers for their use in clinical practice are also reviewed.

Cascaded Multi-view Canonical Correlation (CaMCCo) for Early Diagnosis of Alzheimer's Disease via Fusion of Clinical, Imaging and Omic Features

The introduction of mild cognitive impairment (MCI) as a diagnostic category adds to the challenges of diagnosing Alzheimer’s Disease (AD). No single marker has been proven to accurately categorize patients into their respective diagnostic groups. Thus, previous studies have attempted to develop fused predictors of AD and MCI. These studies have two main limitations. Most do not simultaneously consider all diagnostic categories and provide suboptimal fused representations using the same set of modalities for prediction of all classes. In this work, we present a combined framework, cascaded multiview canonical correlation (CaMCCo), for fusion and cascaded classification that incorporates all diagnostic categories and optimizes classification by selectively combining a subset of modalities at each level of the cascade. CaMCCo is evaluated on a data cohort comprising 149 patients for whom neurophysiological, neuroimaging, proteomic and genomic data were available. Results suggest that fusion of select modalities for each classification task outperforms (mean AUC = 0.92) fusion of all modalities (mean AUC = 0.54) and individual modalities (mean AUC = 0.90, 0.53, 0.71, 0.73, 0.62, 0.68). In addition, CaMCCo outperforms all other multi-class classification methods for MCI prediction (PPV: 0.80 vs. 0.67, 0.63).

Reaching the limits of prognostication in non-small cell lung cancer: an optimized biomarker panel fails to outperform clinical parameters

Numerous protein biomarkers have been analyzed to improve prognostication in non-small cell lung cancer, but have not yet demonstrated sufficient value to be introduced into clinical practice. Here, we aimed to develop and validate a prognostic model for surgically resected non-small cell lung cancer. A biomarker panel was selected based on (1) prognostic association in published literature, (2) prognostic association in gene expression data sets, (3) availability of reliable antibodies, and (4) representation of diverse biological processes. The five selected proteins (MKI67, EZH2, SLC2A1, CADM1, and NKX2-1 alias TTF1) were analyzed by immunohistochemistry on tissue microarrays including tissue from 326 non-small cell lung cancer patients. One score was obtained for each tumor and each protein. The scores were combined, with or without the inclusion of clinical parameters, and the best prognostic model was defined according to the corresponding concordance index (C-index). The best-performing model was subsequently validated in an independent cohort consisting of tissue from 345 non-small cell lung cancer patients. The model based only on protein expression did not perform better compared to clinicopathological parameters, whereas combining protein expression with clinicopathological data resulted in a slightly better prognostic performance (C-index: all non-small cell lung cancer 0.63 vs 0.64; adenocarcinoma: 0.66 vs 0.70, squamous cell carcinoma: 0.57 vs 0.56). However, this modest effect did not translate into a significantly improved accuracy of survival prediction. The combination of a prognostic biomarker panel with clinicopathological parameters did not improve survival prediction in non-small cell lung cancer, questioning the potential of immunohistochemistry-based assessment of protein biomarkers for prognostication in clinical practice.

Overcoming intratumoural heterogeneity for reproducible molecular risk stratification: a case study in advanced kidney cancer

BACKGROUND

Metastatic clear cell renal cell cancer (mccRCC) portends a poor prognosis and urgently requires better clinical tools for prognostication as well as for prediction of response to treatment. Considerable investment in molecular risk stratification has sought to overcome the performance ceiling encountered by methods restricted to traditional clinical parameters. However, replication of results has proven challenging, and intratumoural heterogeneity (ITH) may confound attempts at tissue-based stratification.

METHODS

We investigated the influence of confounding ITH on the performance of a novel molecular prognostic model, enabled by pathologist-guided multiregion sampling (n = 183) of geographically separated mccRCC cohorts from the SuMR trial (development, n = 22) and the SCOTRRCC study (validation, n = 22). Tumour protein levels quantified by reverse phase protein array (RPPA) were investigated alongside clinical variables. Regularised wrapper selection identified features for Cox multivariate analysis with overall survival as the primary endpoint.

RESULTS

The optimal subset of variables in the final stratification model consisted of N-cadherin, EPCAM, Age, mTOR (NEAT). Risk groups from NEAT had a markedly different prognosis in the validation cohort (log-rank p = 7.62 × 10^-7; hazard ratio (HR) 37.9, 95% confidence interval 4.1-353.8) and 2-year survival rates (accuracy = 82%, Matthews correlation coefficient = 0.62). Comparisons with established clinico-pathological scores suggest favourable performance for NEAT (Net reclassification improvement 7.1% vs International Metastatic Database Consortium score, 25.4% vs Memorial Sloan Kettering Cancer Center score). Limitations include the relatively small cohorts and associated wide confidence intervals on predictive performance. Our multiregion sampling approach enabled investigation of NEAT validation when limiting the number of samples analysed per tumour, which significantly degraded performance. Indeed, sample selection could change risk group assignment for 64% of patients, and prognostication with one sample per patient performed only slightly better than random expectation (median logHR = 0.109). Low grade tissue was associated with 3.5-fold greater variation in predicted risk than high grade (p = 0.044).

CONCLUSIONS

This case study in mccRCC quantitatively demonstrates the critical importance of tumour sampling for the success of molecular biomarker studies research where ITH is a factor. The NEAT model shows promise for mccRCC prognostication and warrants follow-up in larger cohorts. Our work evidences actionable parameters to guide sample collection (tumour coverage, size, grade) to inform the development of reproducible molecular risk stratification methods.

Did the reporting of prognostic studies of tumour markers improve since the introduction of REMARK guideline? A comparison of reporting in published articles

Although biomarkers are perceived as highly relevant for future clinical practice, few biomarkers reach clinical utility for several reasons. Among them, poor reporting of studies is one of the major problems. To aid improvement, reporting guidelines like REMARK for tumour marker prognostic (TMP) studies were introduced several years ago. The aims of this project were to assess whether reporting quality of TMP-studies improved in comparison to a previously conducted study assessing reporting quality of TMP-studies (PRE-study) and to assess whether articles citing REMARK (citing group) are better reported, in comparison to articles not citing REMARK (not-citing group). For the POST-study, recent articles citing and not citing REMARK (53 each) were identified in selected journals through systematic literature search and evaluated in same way as in the PRE-study. Ten of the 20 items of the REMARK checklist were evaluated and used to define an overall score of reporting quality. The observed overall scores were 53.4% (range: 10%-90%) for the PRE-study, 57.7% (range: 20%-100%) for the not-citing group and 58.1% (range: 30%-100%) for the citing group of the POST-study. While there is no difference between the two groups of the POST-study, the POST-study shows a slight but not relevant improvement in reporting relative to the PRE-study. Not all the articles of the citing group, cited REMARK appropriately. Irrespective of whether REMARK was cited, the overall score was slightly higher for articles published in journals requesting adherence to REMARK than for those published in journals not requesting it: 59.9% versus 51.9%, respectively. Several years after the introduction of REMARK, many key items of TMP-studies are still very poorly reported. A combined effort is needed from authors, editors, reviewers and methodologists to improve the current situation. Good reporting is not just nice to have but is essential for any research to be useful.

Cancer-cell intrinsic gene expression signatures overcome intratumoural heterogeneity bias in colorectal cancer patient classification

Stromal-derived intratumoural heterogeneity (ITH) has been shown to undermine molecular stratification of patients into appropriate prognostic/predictive subgroups. Here, using several clinically relevant colorectal cancer (CRC) gene expression signatures, we assessed the susceptibility of these signatures to the confounding effects of ITH using gene expression microarray data obtained from multiple tumour regions of a cohort of 24 patients, including central tumour, the tumour invasive front and lymph node metastasis. Sample clustering alongside correlative assessment revealed variation in the ability of each signature to cluster samples according to patient-of-origin rather than region-of-origin within the multi-region dataset. Signatures focused on cancer-cell intrinsic gene expression were found to produce more clinically useful, patient-centred classifiers, as exemplified by the CRC intrinsic signature (CRIS), which robustly clustered samples by patient-of-origin rather than region-of-origin. These findings highlight the potential of cancer-cell intrinsic signatures to reliably stratify CRC patients by minimising the confounding effects of stromal-derived ITH.

Tumour expression profiling is currently used for prognostic and predictive purposes without taking into account the intra patient heterogeneity. Here the authors show that cancer cell specific signatures overcome the tumour heterogeneity effect and result in better classification of colorectal cancer patients.

Standardising RNA profiling based biomarker application in cancer-The need for robust control of technical variables

Histopathology-based staging of colorectal cancer (CRC) has utility in assessing the prognosis of patient subtypes, but as yet cannot accurately predict individual patient's treatment response. Transcriptomics approaches, using array based or next generation sequencing (NGS) platforms, of formalin fixed paraffin embedded tissue can be harnessed to develop multi-gene biomarkers for predicting both prognosis and treatment response, leading to stratification of treatment. While transcriptomics can shape future biomarker development, currently <1% of published biomarkers become clinically validated tests, often due to poor study design or lack of independent validation. In this review of a large number of CRC transcriptional studies, we identify recurrent sources of technical variability that encompass collection, preservation and storage of malignant tissue, nucleic acid extraction, methods to quantitate RNA transcripts and data analysis pipelines. We propose a series of defined steps for removal of these confounding issues, to ultimately aid in the development of more robust clinical biomarkers.

A Model-Based Personalized Cancer Screening Strategy for Detecting Early-Stage Tumors Using Blood-Borne Biomarkers

An effective cancer blood biomarker screening strategy must distinguish aggressive from nonaggressive tumors at an early, intervenable time. However, for blood-based strategies to be useful, the quantity of biomarker shed into the blood and its relationship to tumor growth or progression must be validated. To study how blood biomarker levels correlate with early-stage viable tumor growth in a mouse model of human cancer, we monitored early tumor growth of engineered human ovarian cancer cells (A2780) implanted orthotopically into nude mice. Biomarker shedding was monitored by serial blood sampling, whereas tumor viability and volume were monitored by bioluminescence imaging and ultrasound imaging. From these metrics, we developed a mathematical model of cancer biomarker kinetics that accounts for biomarker shedding from tumor and healthy cells, biomarker entry into vasculature, biomarker elimination from plasma, and subject-specific tumor growth. We validated the model in a separate set of mice in which subject-specific tumor growth rates were accurately predicted. To illustrate clinical translation of this strategy, we allometrically scaled model parameters from mouse to human and used parameters for PSA shedding and prostate cancer. In this manner, we found that blood biomarker sampling data alone were capable of enabling the detection and discrimination of simulated aggressive (2-month tumor doubling time) and nonaggressive (18-month tumor doubling time) tumors as early as 7.2 months and 8.9 years before clinical imaging, respectively. Our model and screening strategy offers broad impact in their applicability to any solid cancer and associated biomarkers shed, thereby allowing a distinction between aggressive and nonaggressive tumors using blood biomarker sampling data alone. Cancer Res; 77(10); 2570-84. ©2017 AACR.

Characteristics and Validation Techniques for PCA-Based Gene-Expression Signatures

Background. Many gene-expression signatures exist for describing the biological state of profiled tumors. Principal Component Analysis (PCA) can be used to summarize a gene signature into a single score. Our hypothesis is that gene signatures can be validated when applied to new datasets, using inherent properties of PCA. Results. This validation is based on four key concepts. Coherence: elements of a gene signature should be correlated beyond chance. Uniqueness: the general direction of the data being examined can drive most of the observed signal. Robustness: if a gene signature is designed to measure a single biological effect, then this signal should be sufficiently strong and distinct compared to other signals within the signature. Transferability: the derived PCA gene signature score should describe the same biology in the target dataset as it does in the training dataset. Conclusions. The proposed validation procedure ensures that PCA-based gene signatures perform as expected when applied to datasets other than those that the signatures were trained upon. Complex signatures, describing multiple independent biological components, are also easily identified.