Markers for early detection of cancer: statistical guidelines for nested case-control studies

Logistic regression and other statistical tools in diagnostic biomarker studies

A biomarker is a measured indicator of a variety of processes, and is often used as a clinical tool for the diagnosis of diseases. While the developmental process of biomarkers from lab to clinic is complex, initial exploratory stages often focus on characterizing the potential of biomarkers through utilizing various statistical methods that can be used to assess their discriminatory performance, establish an appropriate cut-off that transforms continuous data to apt binary responses of confirming or excluding a diagnosis, or establish a robust association when tested against confounders. This review aims to provide a gentle introduction to the most common tools found in diagnostic biomarker studies used to assess the performance of biomarkers with an emphasis on logistic regression.

Prediagnostic evaluation of multicancer detection tests: design and analysis considerations

There is growing interest in multicancer detection tests, which identify molecular signals in the blood that indicate a potential preclinical cancer. A key stage in evaluating these tests is a prediagnostic performance study, in which investigators store specimens from asymptomatic individuals and later test stored specimens from patients with cancer and a random sample of controls to determine predictive performance. Performance metrics include rates of cancer-specific true-positive and false-positive findings and a cancer-specific positive predictive value, with the latter compared with a decision-analytic threshold. The sample size trade-off method, which trades imprecise targeting of the true-positive rate for precise targeting of a zero-false-positive rate can substantially reduce sample size while increasing the lower bound of the positive predictive value. For a 1-year follow-up, with ovarian cancer as the rarest cancer considered, the sample size trade-off method yields a sample size of 163 000 compared with a sample size of 720 000, based on standard calculations. These design and analysis recommendations should be considered in planning a specimen repository and in the prediagnostic evaluation of multicancer detection tests.

Early Detection of Ovarian Cancer

The risk of death from ovarian cancer is highly associated with the clinical stage at diagnosis. Efforts to implement screening for ovarian cancer have been largely unsuccessful, due to the low prevalence of the disease in the general population and the heterogeneity of the various cancer types that fall under the ovarian cancer designation. A practical test for early detection will require both high sensitivity and high specificity to balance reducing the number of cancer deaths with minimizing surgical interventions for false positive screens. The technology must be cost-effective to deliver at scale, widely accessible, and relatively noninvasive. Most importantly, a successful early detection test must be effective not only at diagnosing ovarian cancer but also in reducing ovarian cancer deaths. Stepwise or multimodal approaches among the various areas under investigation will likely be required to make early detection a reality.

MS-based proteomics of body fluids: The end of the beginning

Accurate biomarkers are a crucial and necessary precondition for precision medicine, yet existing ones are often unspecific and new ones have been very slow to enter the clinic. Mass spectrometry (MS)-based proteomics excels by its untargeted nature, specificity of identification and quantification making it an ideal technology for biomarker discovery and routine measurement. It has unique attributes compared to affinity binder technologies, such as OLINK Proximity Extension Assay and SOMAscan. In a previous review we described technological and conceptual limitations that had held back success (Geyer et al., 2017). We proposed a 'rectangular strategy' to better separate true biomarkers by minimizing cohort-specific effects. Today, this has converged with advances in MS-based proteomics technology, such as increased sample throughput, depth of identification and quantification. As a result, biomarker discovery studies have become more successful, producing biomarker candidates that withstand independent verification and, in some cases, already outperform state-of-the-art clinical assays. We summarize developments over the last years, including the benefits of large and independent cohorts, which are necessary for clinical acceptance. They are also required for machine learning or deep learning. Shorter gradients, new scan modes and multiplexing are about to drastically increase throughput, cross-study integration, and quantification, including proxies for absolute levels. We have found that multi-protein panels are inherently more robust than current single analyte tests and better capture the complexity of human phenotypes. Routine MS measurement in the clinic is fast becoming a viable option. The full set of proteins in a body fluid (global proteome) is the most important reference and the best process control. Additionally, it increasingly has all the information that could be obtained from targeted analysis although the latter may be the most straightforward way to enter into regular use. Many challenges remain, not least of a regulatory and ethical nature, but the outlook for MS-based clinical applications has never been brighter.

Assess predictive values of a binary diagnostic test under a nested case-control design

Predictive values of a binary diagnostic test are often evaluated under a random sample design. When the disease is rare, however, such a design might not be as efficient as a nested case-control design where the cases are oversampled from a large existing cohort. Under a nested case-control design, direct proportion estimators of predictive values are biased because cases are oversampled. Consistent estimates of predictive values can be easily obtained by inverse probability weighting (IPW) method. The only difficulty with these IPW estimators has been the absence of expressions for their variances. To fill this gap, in the current paper, we obtain the asymptotic variance formulas for the IPW estimators of predictive values. Unlike their counterparts from weighted logistic regression, our variance formulas take into account the variance of the estimated weights in the IPW estimators of predictive values. We further use the proposed variance formulas to examine the gain in efficiency under a nested case-control design compared with a simple random sampling design. Our results clearly show that when the disease is rare, a nested case-control design can achieve a substantial amount of variance reduction by oversampling cases, compared with a random sample design. Finally, we compare via simulation the accuracy of the proposed variance formulas with the existing methods and illustrate the proposed method by a real data example evaluating the accuracy of D-dimer test.

Distinguishing Alzheimer's Disease Patients and Biochemical Phenotype Analysis Using a Novel Serum Profiling Platform: Potential Involvement of the VWF/ADAMTS13 Axis

It is important to develop minimally invasive biomarker platforms to help in the identification and monitoring of patients with Alzheimer's disease (AD). Assisting in the understanding of biochemical mechanisms as well as identifying potential novel biomarkers and therapeutic targets would be an added benefit of such platforms. This study utilizes a simplified and novel serum profiling platform, using mass spectrometry (MS), to help distinguish AD patient groups (mild and moderate) and controls, as well as to aid in understanding of biochemical phenotypes and possible disease development. A comparison of discriminating sera mass peaks between AD patients and control individuals was performed using leave one [serum sample] out cross validation (LOOCV) combined with a novel peak classification valuation (PCV) procedure. LOOCV/PCV was able to distinguish significant sera mass peak differences between a group of mild AD patients and control individuals with a p value of 10^-13. This value became non-significant (p = 0.09) when the same sera samples were randomly allocated between the two groups and reanalyzed by LOOCV/PCV. This is indicative of physiological group differences in the original true-pathology binary group comparison. Similarities and differences between AD patients and traumatic brain injury (TBI) patients were also discernable using this novel LOOCV/PCV platform. MS/MS peptide analysis was performed on serum mass peaks comparing mild AD patients with control individuals. Bioinformatics analysis suggested that cell pathways/biochemical phenotypes affected in AD include those involving neuronal cell death, vasculature, neurogenesis, and AD/dementia/amyloidosis. Inflammation, autoimmunity, autophagy, and blood-brain barrier pathways also appear to be relevant to AD. An impaired VWF/ADAMTS13 vasculature axis with connections to F8 (factor VIII) and LRP1 and NOTCH1 was indicated and is proposed to be important in AD development.

Evaluating the predictive performance of malaria antibodies and FCGR3B gene polymorphisms on Plasmodium falciparum infection outcome: a prospective cohort study

Background

Malaria antigen-specific antibodies and polymorphisms in host receptors involved in antibody functionality have been associated with different outcomes of Plasmodium falciparum infections. Thus, to identify key prospective malaria antigens for vaccine development, there is the need to evaluate the associations between malaria antibodies and antibody dependent host factors with more rigorous statistical methods. In this study, different statistical models were used to evaluate the predictive performance of malaria-specific antibodies and host gene polymorphisms on P. falciparum infection in a longitudinal cohort study involving Ghanaian children.

Methods

Models with different functional forms were built using known predictors (age, sickle cell status, blood group status, parasite density, and mosquito bed net use) and malaria antigen-specific immunoglobulin (Ig) G and IgG subclasses and FCGR3B polymorphisms shown to mediate antibody-dependent cellular functions. Malaria antigens studied were Merozoite surface proteins (MSP-1 and MSP-3), Glutamate Rich Protein (GLURP)-R0, R2, and the Apical Membrane Antigen (AMA-1). The models were evaluated through visualization and assessment of differences between the Area Under the Receiver Operating Characteristic Curve and Brier Score estimated by suitable internal cross-validation designs.

Results

This study found that the FCGR3B-c.233C>A genotype and IgG against AMA1 were relatively better compared to the other antibodies and FCGR3B genotypes studied in classifying or predicting malaria risk among children.

Conclusions

The data supports the P. falciparum, AMA1 as an important malaria vaccine antigen, while FCGR3B-c.233C>A under the additive and dominant models of inheritance could be an important modifier of the effect of malaria protective antibodies.

Simple Methods for Evaluating 4 Types of Biomarkers: Surrogate Endpoint, Prognostic, Predictive, and Cancer Screening

We review simple methods for evaluating 4 types of biomarkers. First, we discuss the evaluation of surrogate endpoint biomarkers (to shorten a randomized trial) using 2 statistical and 3 biological criteria. Second, we discuss the evaluation of prognostic biomarkers (to predict the risk of disease) by comparing data collection costs with the anticipated net benefit of risk prediction. Third, we discuss the evaluation of predictive markers (to search for a promising subgroup in a randomized trial) using a multivariate subpopulation treatment effect pattern plot involving a risk difference or responders-only benefit function. Fourth, we discuss the evaluation of cancer screening biomarkers (to predict cancer in asymptomatic persons) using methodology to substantially reduce the sample size with stored specimens.

Distinguishing patients with idiopathic epilepsy from solitary cysticercus granuloma epilepsy and biochemical phenotype assessment using a serum biomolecule profiling platform

A major source of epilepsy is Neurocysticercosis (NCC), caused by Taenia solium infection. Solitary cysticercus granuloma (SCG), a sub-group of NCC induced epilepsy, is the most common form of NCC in India. Current diagnostic criteria for SCG epilepsy require brain imaging which may not be available in communities where the disease is endemic. Identification of serum changes and potential biomolecules that could distinguish SCG epilepsy from idiopathic generalized epilepsy (IE), without the initial need for imaging, could assist in disease identification, understanding, and treatment. The objective here was to investigate, using mass spectrometry (MS), sera biomolecule differences between patients with SCG epilepsy or IE to help distinguish these disorders based on physiological differences, to understand underlying phenotypes and mechanisms, and to lay ground work for future therapeutic and biomarker analyses. Sera were obtained from patients with SCG or IE (N = 29 each group). Serum mass peak profiling was performed with electrospray ionization (ESI) MS, and mass peak area means in the two groups were compared using leave one [serum sample] out cross validation (LOOCV). Serum LOOCV analysis identified significant differences between SCG and IE patient groups (p = 10-20), which became non-significant (p = 0.074) when the samples were randomly allocated to the groups and reanalyzed. Tandem MS/MS peptide analysis of serum mass peaks from SCG or IE patients was performed to help identify potential peptide/protein biochemical and phenotypic changes involving these two forms of epilepsy. Bioinformatic analysis of these peptide/protein changes suggested neurological, inflammatory, seizure, blood brain barrier, cognition, ion channel, cell death, and behavior related biochemical systems were being altered in these disease states. This study provides groundwork for aiding in distinguishing SCG and IE patients in minimally invasive, lower-cost manners, for improving understanding of underlying epilepsy mechanisms, and for further identifying discriminatory biomarkers and potential therapeutic targets.

Serum discrimination and phenotype assessment of coronary artery disease patents with and without type 2 diabetes prior to coronary artery bypass graft surgery

Diabetes Mellitus (DM) accelerates coronary artery disease (CAD) and atherosclerosis, the causes of most heart attacks. The biomolecules involved in these inter-related disease processes are not well understood. This study analyzes biomolecules in the sera of patients with CAD, with and without type (T) 2DM, who are about to undergo coronary artery bypass graft (CABG) surgery. The goal is to develop methodology to help identify and monitor CAD patients with and without T2DM, in order to better understand these phenotypes and to glean relationships through analysis of serum biomolecules. Aorta, fat, muscle, and vein tissues from CAD T2DM patients display diabetic-related histologic changes (e.g., lipid accumulation, fibrosis, loss of cellularity) when compared to non-diabetic CAD patients. The patient discriminatory methodology utilized is serum biomolecule mass profiling. This mass spectrometry (MS) approach is able to distinguish the sera of a group of CAD patients from controls (p value 10⁻¹⁵), with the CAD group containing both T2DM and non-diabetic patients. This result indicates the T2DM phenotype does not interfere appreciably with the CAD determination versus control individuals. Sera from a group of T2DM CAD patients however are distinguishable from non-T2DM CAD patients (p value 10⁻⁸), indicating it may be possible to examine the T2DM phenotype within the CAD disease state with this MS methodology. The same serum samples used in the CAD T2DM versus non-T2DM binary group comparison were subjected to MS/MS peptide structure analysis to help identify potential biochemical and phenotypic changes associated with CAD and T2DM. Such peptide/protein identifications could lead to improved understanding of underlying mechanisms, additional biomarkers for discriminating and monitoring these disease conditions, and potential therapeutic targets. Bioinformatics/systems biology analysis of the peptide/protein changes associated with CAD and T2DM suggested cell pathways/systems affected include atherosclerosis, DM, fibrosis, lipogenesis, loss of cellularity (apoptosis), and inflammation.

Distinguishing and Biochemical Phenotype Analysis of Epilepsy Patients Using a Novel Serum Profiling Platform

Diagnosis of non-symptomatic epilepsy includes a history of two or more seizures and brain imaging to rule out structural changes like trauma, tumor, infection. Such analysis can be problematic. It is important to develop capabilities to help identify non-symptomatic epilepsy in order to better monitor and understand the condition. This understanding could lead to improved diagnostics and therapeutics. Serum mass peak profiling was performed using electrospray ionization mass spectrometry (ESI-MS). A comparison of sera mass peaks between epilepsy and control groups was performed via leave one [serum sample] out cross-validation (LOOCV). MS/MS peptide analysis was performed on serum mass peaks to compare epilepsy patient and control groups. LOOCV identified significant differences between the epilepsy patient group and control group (p = 10⁻²²). This value became non-significant (p = 0.10) when the samples were randomly allocated between the groups and reanalyzed by LOOCV. LOOCV was thus able to distinguish a non-symptomatic epilepsy patient group from a control group based on physiological differences and underlying phenotype. MS/MS was able to identify potential peptide/protein changes involved in this epilepsy versus control comparison, with 70% of the top 100 proteins indicating overall neurologic function. Specifically, peptide/protein sera changes suggested neuro-inflammatory, seizure, ion-channel, synapse, and autoimmune pathways changing between epilepsy patients and controls.

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.

Distinguishing and phenotype monitoring of traumatic brain injury and post-concussion syndrome including chronic migraine in serum of Iraq and Afghanistan war veterans

Traumatic Brain Injury (TBI) and persistent post-concussion syndrome (PCS) including chronic migraine (CM) are major health issues for civilians and the military. It is important to understand underlying biochemical mechanisms of these conditions, and be able to monitor them in an accurate and minimally invasive manner. This study describes the initial use of a novel serum analytical platform to help distinguish TBI patients, including those with post-traumatic headache (PTH), and to help identify phenotypes at play in these disorders. The hypothesis is that physiological responses to disease states like TBI and PTH and related bodily stresses are reflected in biomolecules in the blood in disease-specific manner. Leave one out (serum sample) cross validations (LOOCV) and sample randomizations were utilized to distinguished serum samples from the following TBI patient groups: TBI +PTSD + CM + severe depression (TBI "most affected" group) vs healthy controls, TBI "most affected" vs TBI, TBI vs controls, TBI + CM vs controls, and TBI + CM vs TBI. Inter-group discriminatory p values were ≤ 10-10, and sample group randomizations resulted in p non-significant values. Peptide/protein identifications of discriminatory mass peaks from the TBI "most affected" vs controls and from the TBI plus vs TBI minus CM groups yielded information of the cellular/molecular effects of these disorders (immune responses, amyloidosis/Alzheimer's disease/dementia, neuronal development). More specific biochemical disease effects appear to involve blood brain barrier, depression, migraine headache, autoimmunity, and autophagy pathways. This study demonstrated the ability for the first time of a novel, accurate, biomarker platform to monitor these conditions in serum, and help identify biochemical relationships leading to better understanding of these disorders and to potential therapeutic approaches.

Cancer Screening Markers: A Simple Strategy to Substantially Reduce the Sample Size for Validation

BACKGROUND

Studies to validate a cancer prediction model based on cancer screening markers collected in stored specimens from asymptomatic persons typically require large specimen collection sample sizes. A standard sample size calculation targets a true-positive rate (TPR) of 0.8 with a 2.5% lower bound of 0.7 at a false-positive rate (FPR) of 0.01 with a 5% upper bound of 0.03. If the probability of developing cancer during the study is P = 0.01, the specimen collection sample size based on the standard calculation is 7600.

METHODS

The strategy to reduce the specimen collection sample size is to decrease both the lower bound of TPR and the upper bound of FPR while keeping a positive lower bound on the anticipated clinical utility.

RESULTS

The new sample size calculation targets a TPR of 0.4 with a 2.5% lower bound of 0.10 and an FPR of 0.0 with a 5% upper bound of 0.005. With P = 0.01, the specimen collection sample size based on the new calculation is 1800 instead of 7600.

LIMITATIONS

The new sample size calculation requires a minimum benefit/cost ratio (number of false positives traded for a true positive). With P = 0.01, the minimum cost-benefit ratio is 5, which is plausible in many studies.

CONCLUSION

In validation studies for cancer screening markers, the strategy can substantially reduce the specimen collection sample size, substantially reducing costs and making some otherwise infeasible studies now feasible.

Distinguishing neurocysticercosis epilepsy from epilepsy of unknown etiology using a minimal serum mass profiling platform

Neurocysticercosis is associated with epilepsy in pig-raising communities with poor sanitation. Current internationally recognized diagnostic guidelines for neurocysticercosis rely on brain imaging, a technology that is frequently not available or not accessible in areas endemic for neurocysticercosis. Minimally invasive and low-cost aids for diagnosing neurocysticercosis epilepsy could improve treatment of neurocysticercosis. The goal of this study was to test the extent to which patients with neurocysticercosis epilepsy, epilepsy of unknown etiology, idiopathic headaches and among different types of neurocysticercosis lesions could be distinguished from each other based on serum mass profiling. For this, we collected sera from patients with neurocysticercosis-associated epilepsy, epilepsy of unknown etiology, recovered neurocysticercosis, and idiopathic headaches then performed binary group comparisons among them using electrospray ionization mass spectrometry. A leave one [serum sample] out cross validation procedure was employed to analyze spectral data. Sera from neurocysticercosis patients was distinguished from epilepsy of unknown etiology patients with a p-value of 10^-28. This distinction was lost when samples were randomized to either group (p-value = 0.22). Similarly, binary comparisons of patients with neurocysticercosis who has different types of lesions showed that different forms of this disease were also distinguishable from one another. These results suggest neurocysticercosis epilepsy can be distinguished from epilepsy of unknown etiology based on biomolecular differences in sera detected by mass profiling.

Validation of Biomarkers for Early Detection of Pancreatic Cancer: Summary of The Alliance of Pancreatic Cancer Consortia for Biomarkers for Early Detection Workshop

Pancreatic cancer is the third leading cause of cancer death in the United States, and the 5-year relative survival for patients diagnosed with pancreatic cancer is less than 10%. Early intervention is the key to a better survival outcome. Currently, there are no biomarkers that can reliably detect pancreatic cancer at an early stage or identify precursors that are destined to progress to malignancy. The National Cancer Institute in partnership with the Kenner Family Research Fund and the Pancreatic Cancer Action Network convened a Data Jamboree on Biomarkers workshop on December 5, 2016, to discuss and evaluate existing or newly developed biomarkers and imaging methods for early detection of pancreatic cancer. The primary goal of this workshop was to determine if there are any promising biomarkers for early detection of pancreatic cancer that are ready for clinical validation. The Alliance of Pancreatic Cancer Consortia for Biomarkers for Early Detection, formed under the auspices of this workshop, will provide the common platform and the resources necessary for validation. Although none of the biomarkers evaluated seemed ready for a large-scale biomarker validation trial, a number of them had sufficiently high sensitivity and specificity to warrant additional research, especially if combined with other biomarkers to form a panel.

Development of a sequential workflow based on LC-PRM for the verification of endometrial cancer protein biomarkers in uterine aspirate samples

About 30% of endometrial cancer (EC) patients are diagnosed at an advanced stage of the disease, which is associated with a drastic decrease in the 5-year survival rate. The identification of biomarkers in uterine aspirate samples, which are collected by a minimally invasive procedure, would improve early diagnosis of EC. We present a sequential workflow to select from a list of potential EC biomarkers, those which are the most promising to enter a validation study. After the elimination of confounding contributions by residual blood proteins, 52 potential biomarkers were analyzed in uterine aspirates from 20 EC patients and 18 non-EC controls by a high-resolution accurate mass spectrometer operated in parallel reaction monitoring mode. The differential abundance of 26 biomarkers was observed, and among them ten proteins showed a high sensitivity and specificity (AUC > 0.9). The study demonstrates that uterine aspirates are valuable samples for EC protein biomarkers screening. It also illustrates the importance of a biomarker verification phase to fill the gap between discovery and validation studies and highlights the benefits of high resolution mass spectrometry for this purpose. The proteins verified in this study have an increased likelihood to become a clinical assay after a subsequent validation phase.

Systematic review and critique of circulating miRNAs as biomarkers of stage I-II non-small cell lung cancer

Selected circulating microRNAs (miRNAs) have been suggested for non-invasive screening of non-small cell lung cancer (NSCLC), however the numerous proposed miRNA signatures are inconsistent. Aiming to identify miRNAs suitable specifically for stage I-II NSCLC screening in serum/plasma samples, we searched the databases "Pubmed", "Medline", "Scopus", "Embase" and "WOS" and systematically reviewed the publications reporting quantitative data on the efficacy [sensitivity, specificity and/or area under the curve (AUC)] of circulating miRNAs as biomarkers of NSCLC stage I and/or II. The 20 studies fulfilling the search criteria included 1110 NSCLC patients and 1009 controls, and were of medium quality according to Quality Assessment of Diagnostic Accuracy Studies checklist. In these studies, the patient cohorts as well as the control groups were heterogeneous for demographics and clinicopathological characteristics; moreover, numerous pre-analytical and analytical variables likely influenced miRNA determinations, and potential bias of hemolysis was often underestimated. We identified four circulating miRNAs scarcely influenced by hemolysis, each featuring high sensitivity (> 80%) and AUC (> 0.80) as biomarkers of stage I-II NSCLC: miR-223, miR-20a, miR-448 and miR-145; four other miRNAs showed high specificity (> 90%): miR-628-3p, miR-29c, miR-210 and miR-1244. In a model of two-step screening for stage I-II NSCLC using first the above panel of serum miRNAs with high sensitivity and high AUC, and subsequently the panel with high specificity, the estimated overall sensitivity is 91.6% and overall specificity is 93.4%. These and other circulating miRNAs suggested for stage I-II NSCLC screening require validation in multiple independent studies before they can be proposed for clinical application.

Evaluating Molecular Biomarkers for the Early Detection of Lung Cancer: When Is a Biomarker Ready for Clinical Use? An Official American Thoracic Society Policy Statement

BACKGROUND

Molecular biomarkers have the potential to improve the current state of early lung cancer detection. The goal of this project was to develop a policy statement that provides guidance about the level of evidence required to determine that a molecular biomarker, used to support early lung cancer detection, is appropriate for clinical use.

METHODS

An ad hoc project steering committee was formed, to include individuals with expertise in the early detection of lung cancer and molecular biomarker development, from inside and outside of the Assembly on Thoracic Oncology. Key questions, generated from the results of a survey of the project steering committee, were discussed at an in-person meeting. Results of the discussion were summarized in a policy statement that was circulated to the steering committee and revised multiple times to achieve consensus.

RESULTS

With a focus on the clinical applications of lung cancer screening and lung nodule evaluation, the policy statement outlines categories of results that should be reported in the early phases of molecular biomarker development, discusses the level of evidence that would support study of the clinical utility, describes the outcomes that should be proven to consider a molecular biomarker clinically useful, and suggests study designs capable of assessing these outcomes.

CONCLUSIONS

The application of molecular biomarkers to assist with the early detection of lung cancer has the potential to substantially improve our ability to select patients for lung cancer screening, and to assist with the characterization of indeterminate lung nodules. We have described relevant considerations and have suggested standards to apply when determining whether a molecular biomarker for the early detection of lung cancer is ready for clinical use.

Serum Monitoring and Phenotype Identification of Stage I Non-Small Cell Lung Cancer Patients

A stage I non-small cell lung cancer (NSCLC) serum profiling platform is presented which is highly efficient and accurate. Test sensitivity (0.95) for stage I NSCLC is the highest reported so far. Test metrics are reported for discriminating stage I adenocarcinoma vs squamous cell carcinoma subtypes. Blinded analysis identified 23 out of 24 stage I NSCLC and control serum samples. Group-discriminating mass peaks were targeted for tandem mass spectrometry peptide/protein identification, and yielded a lung cancer phenotype. Bioinformatic analysis revealed a novel lymphocyte adhesion pathway involved with early-stage lung cancer.

Prognostic value of routine laboratory variables in prediction of breast cancer recurrence

The prognostic value of routine laboratory variables in breast cancer has been largely overlooked. Based on laboratory tests commonly performed in clinical practice, we aimed to develop a new model to predict disease free survival (DFS) after surgical removal of primary breast cancer. In a cohort of 1,596 breast cancer patients, we analyzed the associations of 33 laboratory variables with patient DFS. Based on 3 significant laboratory variables (hemoglobin, alkaline phosphatase, and international normalized ratio), together with important demographic and clinical variables, we developed a prognostic model, achieving the area under the curve of 0.79. We categorized patients into 3 risk groups according to the prognostic index developed from the final model. Compared with the patients in the low-risk group, those in the medium- and high-risk group had a significantly increased risk of recurrence with a hazard ratio (HR) of 1.75 (95% confidence interval [CI] 1.30–2.38) and 4.66 (95% CI 3.54–6.14), respectively. The results from the training set were validated in the testing set. Overall, our prognostic model incorporating readily available routine laboratory tests is powerful in identifying breast cancer patients who are at high risk of recurrence. Further study is warranted to validate its clinical application.

Emerging Role of MicroRNAs as Liquid Biopsy Biomarkers in Gastrointestinal Cancers

Cancer has emerged as a leading cause of mortality worldwide, claiming more than 8 million lives annually. Gastrointestinal cancers account for about 35% of these mortalities. Recent advances in diagnostic and treatment strategies have reduced mortality among patients with gastrointestinal cancer, yet a significant number of patients still develop late-stage cancer, where treatment options are inadequate. Emerging interests in "liquid biopsies" have encouraged investigators to identify and develop clinically relevant noninvasive genomic and epigenomic signatures that can be exploited as biomarkers capable of detecting premalignant and early-stage cancers. In this context, microRNAs (miRNA), which are small, noncoding RNAs that are frequently dysregulated in cancers, have emerged as promising entities for such diagnostic purposes. Even though the future looks promising, current approaches for detecting miRNAs in blood and other biofluids remain inadequate. This review summarizes existing efforts to exploit circulating miRNAs as cancer biomarkers and evaluates their potential and challenges as liquid biopsy-based biomarkers for gastrointestinal cancers. Clin Cancer Res; 23(10); 2391-9. ©2017 AACR.

A modified risk set approach to biomarker evaluation studies

There is tremendous scientific and medical interest in the use of biomarkers to better facilitate medical decision making. In this article, we present a simple framework for assessing the predictive ability of a biomarker. The methodology requires use of techniques from a subfield of survival analysis termed semicompeting risks; results are presented to make the article self-contained. As we show in the article, one natural interpretation of semicompeting risks model is in terms of modifying the classical risk set approach to survival analysis that is more germane to medical decision making. A crucial parameter for evaluating biomarkers is the predictive hazard ratio, which is different from the usual hazard ratio from Cox regression models for right-censored data. This quantity will be defined; its estimation, inference and adjustment for covariates will be discussed. Aspects of causal inference related to these procedures will also be described. The methodology is illustrated with an evaluation of serum albumin in terms of predicting death in patients with primary biliary cirrhosis.

Biomarker discovery: success as a function of risk mitigation

Protein biomarker discovery is a fascinating enterprise; however, success in terms of products for in vitro diagnostic use is sparse. New developments in mass spectrometry-based quantitative proteomics as discovery technology have opened up new avenues for this endeavor. In addition to choice of technology, sample properties, study design and validation strategy are potent pillars required for project success. The challenge for successful biomarker discovery can be described by a series of risks that need to be mitigated. This article intends to describe the major risks along with possible solutions.

Cross-validation and Peeling Strategies for Survival Bump Hunting using Recursive Peeling Methods

We introduce a framework to build a survival/risk bump hunting model with a censored time-to-event response. Our Survival Bump Hunting (SBH) method is based on a recursive peeling procedure that uses a specific survival peeling criterion derived from non/semi-parametric statistics such as the hazards-ratio, the log-rank test or the Nelson--Aalen estimator. To optimize the tuning parameter of the model and validate it, we introduce an objective function based on survival or prediction-error statistics, such as the log-rank test and the concordance error rate. We also describe two alternative cross-validation techniques adapted to the joint task of decision-rule making by recursive peeling and survival estimation. Numerical analyses show the importance of replicated cross-validation and the differences between criteria and techniques in both low and high-dimensional settings. Although several non-parametric survival models exist, none addresses the problem of directly identifying local extrema. We show how SBH efficiently estimates extreme survival/risk subgroups unlike other models. This provides an insight into the behavior of commonly used models and suggests alternatives to be adopted in practice. Finally, our SBH framework was applied to a clinical dataset. In it, we identified subsets of patients characterized by clinical and demographic covariates with a distinct extreme survival outcome, for which tailored medical interventions could be made. An R package PRIMsrc (Patient Rule Induction Method in Survival, Regression and Classification settings) is available on CRAN (Comprehensive R Archive Network) and GitHub.

Diagnosing Lung Cancers through Examination of Micro-RNA Biomarkers in Blood, Plasma, Serum and Sputum: A Review and Summary of Current Literature

Lung cancer is the leading cause of cancer related morbidity and mortality worldwide. Currently, the vast majority of lung cancers are diagnosed at a late stage, when patients become symptomatic leading to dismal, less than 15% five-year survival rates. Evidence has demonstrated that screening computed tomography scans can be used to detect lung cancer, but these scans have high false positive rates. Therefore, there is a continued need for the development of minimally-invasive methods to screen the high risk population and diagnose lung cancer at an earlier, curable stage. One such promising area is the use micro-RNAs. These are short, non-coding RNA molecules that have been shown in previous research to be dysregulated in cancers. This review will focus on the potential use of miRNA levels in various biological fluids (whole blood, plasma, serum, and sputum) and demonstrate their potential utility as screening and diagnostic biomarkers for lung cancer. Current research will be analyzed and compared, and future directions in establishing the use of miRNAs for detecting lung cancer will be discussed.

R package PRIMsrc: Bump Hunting by Patient Rule Induction Method for Survival, Regression and Classification

PRIMsrc is a novel implementation of a non-parametric bump hunting procedure, based on the Patient Rule Induction Method (PRIM), offering a unified treatment of outcome variables, including censored time-to-event (Survival), continuous (Regression) and discrete (Classification) responses. To fit the model, it uses a recursive peeling procedure with specific peeling criteria and stopping rules depending on the response. To validate the model, it provides an objective function based on prediction-error or other specific statistic, as well as two alternative cross-validation techniques, adapted to the task of decision-rule making and estimation in the three types of settings. PRIMsrc comes as an open source R package, including at this point: (i) a main function for fitting a Survival Bump Hunting model with various options allowing cross-validated model selection to control model size (#covariates) and model complexity (#peeling steps) and generation of cross-validated end-point estimates; (ii) parallel computing; (iii) various S3-generic and specific plotting functions for data visualization, diagnostic, prediction, summary and display of results. It is available on CRAN and GitHub.

Cerebrospinal fluid peptides as potential Parkinson disease biomarkers: a staged pipeline for discovery and validation

Finding robust biomarkers for Parkinson disease (PD) is currently hampered by inherent technical limitations associated with imaging or antibody-based protein assays. To circumvent the challenges, we adapted a staged pipeline, starting from our previous proteomic profiling followed by high-throughput targeted mass spectrometry (MS), to identify peptides in human cerebrospinal fluid (CSF) for PD diagnosis and disease severity correlation. In this multicenter study consisting of training and validation sets, a total of 178 subjects were randomly selected from a retrospective cohort, matching age and sex between PD patients, healthy controls, and neurological controls with Alzheimer disease (AD). From ∼14,000 unique peptides displaying differences between PD and healthy control in proteomic investigations, 126 peptides were selected based on relevance and observability in CSF using bioinformatic analysis and MS screening, and then quantified by highly accurate and sensitive selected reaction monitoring (SRM) in the CSF of 30 PD patients versus 30 healthy controls (training set), followed by diagnostic (receiver operating characteristics) and disease severity correlation analyses. The most promising candidates were further tested in an independent cohort of 40 PD patients, 38 AD patients, and 40 healthy controls (validation set). A panel of five peptides (derived from SPP1, LRP1, CSF1R, EPHA4, and TIMP1) was identified to provide an area under curve (AUC) of 0.873 (sensitivity = 76.7%, specificity = 80.0%) for PD versus healthy controls in the training set. The performance was essentially confirmed in the validation set (AUC = 0.853, sensitivity = 82.5%, specificity = 82.5%). Additionally, this panel could also differentiate the PD and AD groups (AUC = 0.990, sensitivity = 95.0%, specificity = 97.4%). Furthermore, a combination of two peptides belonging to proteins TIMP1 and APLP1 significantly correlated with disease severity as determined by the Unified Parkinson's Disease Rating Scale motor scores in both the training (r = 0.381, p = 0.038)j and the validation (r = 0.339, p = 0.032) sets. The novel panel of CSF peptides, if validated in independent cohorts, could be used to assist in clinical diagnosis of PD and has the potential to help monitoring or predicting disease progression.

Network-based identification of biomarkers coexpressed with multiple pathways

Unraveling complex molecular interactions and networks and incorporating clinical information in modeling will present a paradigm shift in molecular medicine. Embedding biological relevance via modeling molecular networks and pathways has become increasingly important for biomarker identification in cancer susceptibility and metastasis studies. Here, we give a comprehensive overview of computational methods used for biomarker identification, and provide a performance comparison of several network models used in studies of cancer susceptibility, disease progression, and prognostication. Specifically, we evaluated implication networks, Boolean networks, Bayesian networks, and Pearson’s correlation networks in constructing gene coexpression networks for identifying lung cancer diagnostic and prognostic biomarkers. The results show that implication networks, implemented in Genet package, identified sets of biomarkers that generated an accurate prediction of lung cancer risk and metastases; meanwhile, implication networks revealed more biologically relevant molecular interactions than Boolean networks, Bayesian networks, and Pearson’s correlation networks when evaluated with MSigDB database.

Integrated DNA Copy Number and Gene Expression Regulatory Network Analysis of Non-small Cell Lung Cancer Metastasis

Integrative analysis of multi-level molecular profiles can distinguish interactions that cannot be revealed based on one kind of data in the analysis of cancer susceptibility and metastasis. DNA copy number variations (CNVs) are common in cancer cells, and their role in cell behaviors and relationship to gene expression (GE) is poorly understood. An integrative analysis of CNV and genome-wide mRNA expression can discover copy number alterations and their possible regulatory effects on GE. This study presents a novel framework to identify important genes and construct potential regulatory networks based on these genes. Using this approach, DNA copy number aberrations and their effects on GE in lung cancer progression were revealed. Specifically, this approach contains the following steps: (1) select a pool of candidate driver genes, which have significant CNV in lung cancer patient tumors or have a significant association with the clinical outcome at the transcriptional level; (2) rank important driver genes in lung cancer patients with good prognosis and poor prognosis, respectively, and use top-ranked driver genes to construct regulatory networks with the COpy Number and EXpression In Cancer (CONEXIC) method; (3) identify experimentally confirmed molecular interactions in the constructed regulatory networks using Ingenuity Pathway Analysis (IPA); and (4) visualize the refined regulatory networks with the software package Genatomy. The constructed CNV/mRNA regulatory networks provide important insights into potential CNV-regulated transcriptional mechanisms in lung cancer metastasis.

Cross-Validation of Survival Bump Hunting by Recursive Peeling Methods

We introduce a survival/risk bump hunting framework to build a bump hunting model with a possibly censored time-to-event type of response and to validate model estimates. First, we describe the use of adequate survival peeling criteria to build a survival/risk bump hunting model based on recursive peeling methods. Our method called "Patient Recursive Survival Peeling" is a rule-induction method that makes use of specific peeling criteria such as hazard ratio or log-rank statistics. Second, to validate our model estimates and improve survival prediction accuracy, we describe a resampling-based validation technique specifically designed for the joint task of decision rule making by recursive peeling (i.e. decision-box) and survival estimation. This alternative technique, called "combined" cross-validation is done by combining test samples over the cross-validation loops, a design allowing for bump hunting by recursive peeling in a survival setting. We provide empirical results showing the importance of cross-validation and replication.

Biomarker research with prospective study designs for the early detection of cancer

This article describes the principles of marker research with prospective studies along with examples for diagnostic tumor markers. A plethora of biomarkers have been claimed as useful for the early detection of cancer. However, disappointingly few biomarkers were approved for the detection of unrecognized disease, and even approved markers may lack a sound validation phase. Prospective studies aimed at the early detection of cancer are costly and long-lasting and therefore the bottleneck in marker research. They enroll a large number of clinically asymptomatic subjects and follow-up on incident cases. As invasive procedures cannot be applied to collect tissue samples from the target organ, biomarkers can only be determined in easily accessible body fluids. Marker levels increase during cancer development, with samples collected closer to the occurrence of symptoms or a clinical diagnosis being more informative than earlier samples. Only prospective designs allow the serial collection of pre-diagnostic samples. Their storage in a biobank upgrades cohort studies to serve for both, marker discovery and validation. Population-based cohort studies, which may collect a wealth of data, are commonly conducted with just one baseline investigation lacking serial samples. However, they can provide valuable information about factors that influence the marker level. Screening programs can be employed to archive serial samples but require significant efforts to collect samples and auxiliary data for marker research. Randomized controlled trials have the highest level of evidence in assessing a biomarker's benefit against usual care and present the most stringent design for the validation of promising markers as well as for the discovery of new markers. In summary, all kinds of prospective studies can benefit from a biobank as they can serve as a platform for biomarker research. This article is part of a Special Issue entitled: Biomarkers: A Proteomic Challenge.

Better cancer biomarker discovery through better study design

BACKGROUND

High-throughput laboratory technologies coupled with sophisticated bioinformatics algorithms have tremendous potential for discovering novel biomarkers, or profiles of biomarkers, that could serve as predictors of disease risk, response to treatment or prognosis. We discuss methodological issues in wedding high-throughput approaches for biomarker discovery with the case-control study designs typically used in biomarker discovery studies, especially focusing on nested case-control designs.

METHODS

We review principles for nested case-control study design in relation to biomarker discovery studies and describe how the efficiency of biomarker discovery can be effected by study design choices. We develop a simulated prostate cancer cohort data set and a series of biomarker discovery case-control studies nested within the cohort to illustrate how study design choices can influence biomarker discovery process.

RESULT

Common elements of nested case-control design, incidence density sampling and matching of controls to cases are not typically factored correctly into biomarker discovery analyses, inducing bias in the discovery process. We illustrate how incidence density sampling and matching of controls to cases reduce the apparent specificity of truly valid biomarkers 'discovered' in a nested case-control study. We also propose and demonstrate a new case-control matching protocol, we call 'antimatching', that improves the efficiency of biomarker discovery studies.

CONCLUSIONS

For a valid, but as yet undiscovered, biomarker(s) disjunctions between correctly designed epidemiologic studies and the practice of biomarker discovery reduce the likelihood that true biomarker(s) will be discovered and increases the false-positive discovery rate.

Comparison of approaches to estimate confidence intervals of post-test probabilities of diagnostic test results in a nested case-control study

Background

Nested case–control studies become increasingly popular as they can be very efficient for quantifying the diagnostic accuracy of costly or invasive tests or (bio)markers. However, they do not allow for direct estimation of the test’s predictive values or post-test probabilities, let alone for their confidence intervals (CIs). Correct estimates of the predictive values itself can easily be obtained using a simple correction by the (inverse) sampling fractions of the cases and controls. But using this correction to estimate the corresponding standard error (SE), falsely increases the number of patients that are actually studied, yielding too small CIs. We compared different approaches for estimating the SE and thus CI of predictive values or post-test probabilities of diagnostic test results in a nested case–control study.

Methods

We created datasets based on a large, previously published diagnostic study on 2 different tests (D-dimer test and calf difference test) with a nested case–control design. We compared six different approaches; the approaches were: 1. the standard formula for the SE of a proportion, 2. adaptation of the standard formula with the sampling fraction, 3. A bootstrap procedure, 4. A approach, which uses the sensitivity, the specificity and the prevalence, 5. Weighted logistic regression, and 6. Approach 4 on the log odds scale. The approaches were compared with respect to coverage of the CI and CI-width.

Results

The bootstrap procedure (approach 3) showed good coverage and relatively small CI widths. Approaches 4 and 6 showed some undercoverage, particularly for the D-dimer test with frequent positive results (positive results around 70%). Approaches 1, 2 and 5 showed clear overcoverage at low prevalences of 0.05 and 0.1 in the cohorts for all case–control ratios.

Conclusion

The results from our study suggest that a bootstrap procedure is necessary to assess the confidence interval for the predictive values or post-test probabilities of diagnostic tests results in studies using a nested case–control design.

Population-based study reveals new risk-stratification biomarker panel for Barrett's esophagus

BACKGROUND & AIMS

The risk of progression of Barrett's esophagus (BE) to esophageal adenocarcinoma (EAC) is low and difficult to calculate. Accurate tools to determine risk are needed to optimize surveillance and intervention. We assessed the ability of candidate biomarkers to predict which cases of BE will progress to EAC or high-grade dysplasia and identified those that can be measured in formalin-fixed tissues.

METHODS

We analyzed data from a nested case-control study performed using the population-based Northern Ireland BE Register (1993-2005). Cases who progressed to EAC (n = 89) or high-grade dysplasia ≥ 6 months after diagnosis with BE were matched to controls (nonprogressors, n = 291), for age, sex, and year of BE diagnosis. Established biomarkers (abnormal DNA content, p53, and cyclin A expression) and new biomarkers (levels of sialyl Lewis(a), Lewis(x), and Aspergillus oryzae lectin [AOL] and binding of wheat germ agglutinin) were assessed in paraffin-embedded tissue samples from patients with a first diagnosis of BE. Conditional logistic regression analysis was applied to assess odds of progression for patients with dysplastic and nondysplastic BE, based on biomarker status.

RESULTS

Low-grade dysplasia and all biomarkers tested, other than Lewis(x), were associated with risk of EAC or high-grade dysplasia. In backward selection, a panel comprising low-grade dysplasia, abnormal DNA ploidy, and AOL most accurately identified progressors and nonprogressors. The adjusted odds ratio for progression of patients with BE with low-grade dysplasia was 3.74 (95% confidence interval, 2.43-5.79) for each additional biomarker and the risk increased by 2.99 for each additional factor (95% confidence interval, 1.72-5.20) in patients without dysplasia.

CONCLUSIONS

Low-grade dysplasia, abnormal DNA ploidy, and AOL can be used to identify patients with BE most likely to develop EAC or high-grade dysplasia.

DNA methylation biomarkers offer improved diagnostic efficiency in lung cancer

The exceptional high mortality of lung cancer can be instigated to a high degree by late diagnosis. Despite the plethora of studies on potential molecular biomarkers for lung cancer diagnosis, very few have reached clinical implementation. In this study, we developed a panel of DNA methylation biomarkers and validated their diagnostic efficiency in bronchial washings from a large retrospective cohort. Candidate targets from previous high-throughput approaches were examined by pyrosequencing in an independent set of 48 lung tumor/normal paired. Ten promoters were selected and quantitative methylation-specific PCR (qMSP) assays were developed and used to screen 655 bronchial washings from the Liverpool Lung Project (LLP) subjects divided into training (194 cases and 214 controls) and validation (139 cases and 109 controls) sets. Three statistical models were used to select the optimal panel of markers and to evaluate the performance of the discriminatory algorithms. The final logit regression model incorporated hypermethylation at p16, TERT, WT1, and RASSF1. The performance of this 4-gene methylation signature in the validation set showed 82% sensitivity and 91% specificity. In comparison, cytology alone in this set provided 43% sensitivity at 100% specificity. The diagnostic efficiency of the panel did not show any biases with age, gender, smoking, and the presence of a nonlung neoplasm. However, sensitivity was predictably higher in central (squamous and small cell) than peripheral (adenocarcinomas) tumors, as well as in stage 2 or greater tumors. These findings clearly show the impact of DNA methylation-based assays in the diagnosis of cytologically occult lung neoplasms. A prospective trial is currently imminent in the LLP study to provide data on the enhancement of diagnostic accuracy in a clinical setting, including by additional markers.

A novel five gene signature derived from stem-like side population cells predicts overall and recurrence-free survival in NSCLC

Gene expression profiling has been used to characterize prognosis in various cancers. Earlier studies had shown that side population cells isolated from Non-Small Cell Lung Cancer (NSCLC) cell lines exhibit cancer stem cell properties. In this study we apply a systems biology approach to gene expression profiling data from cancer stem like cells isolated from lung cancer cell lines to identify novel gene signatures that could predict prognosis. Microarray data from side population (SP) and main population (MP) cells isolated from 4 NSCLC lines (A549, H1650, H460, H1975) were used to examine gene expression profiles associated with stem like properties. Differentially expressed genes that were over or under-expressed at least two fold commonly in all 4 cell lines were identified. We found 354 were upregulated and 126 were downregulated in SP cells compared to MP cells; of these, 89 up and 62 downregulated genes (average 2 fold changes) were used for Principle Component Analysis (PCA) and MetaCore™ pathway analysis. The pathway analysis demonstrated representation of 4 up regulated genes (TOP2A, AURKB, BRRN1, CDK1) in chromosome condensation pathway and 1 down regulated gene FUS in chromosomal translocation. Microarray data was validated using qRT-PCR on the 5 selected genes and all showed robust correlation between microarray and qRT-PCR. Further, we analyzed two independent gene expression datasets that included 360 lung adenocarcinoma patients from NCI Director's Challenge Set for overall survival and 63 samples from Sungkyunkwan University (SKKU) for recurrence free survival. Kaplan-Meier and log-rank test analysis predicted poor survival of patients in both data sets. Our results suggest that genes involved in chromosome condensation are likely related with stem-like properties and might predict survival in lung adenocarcinoma. Our findings highlight a gene signature for effective identification of lung adenocarcinoma patients with poor prognosis and designing more aggressive therapies for such patients.

Potential of computer-aided diagnosis to improve CT lung cancer screening

The development of low-dose spiral computed tomography (CT) has rekindled hope that effective lung cancer screening might yet be found. Screening is justified when there is evidence that it will extend lives at reasonable cost and acceptable levels of risk. A screening test should detect all extant cancers while avoiding unnecessary workups. Thus optimal screening modalities have both high sensitivity and specificity. Due to the present state of technology, radiologists must opt to increase sensitivity and rely on follow-up diagnostic procedures to rule out the incurred false positives. There is evidence in published reports that computer-aided diagnosis technology may help radiologists alter the benefit-cost calculus of CT sensitivity and specificity in lung cancer screening protocols. This review will provide insight into the current discussion of the effectiveness of lung cancer screening and assesses the potential of state-of-the-art computer-aided design developments.

Imaging mass spectrometry in biomarker discovery and validation

Biomarker discovery and validation involves the consideration of many issues and challenges in order to be effectively used for translation from bench to bedside. Imaging mass spectrometry (IMS) is a new technology to assess spatial molecular arrangements in tissue sections, going far beyond microscopy in providing hundreds of different molecular images from a single scan without the need of target-specific reagents. The possibility to correlate distribution maps of multiple analytes with histological and clinical features makes it an ideal tool to discover diagnostic and prognostic markers of diseases. Some recently published studies that show the usefulness and advantages of this technology in the field of cancer research are highlighted.

The state of molecular biomarkers for the early detection of lung cancer

Using biomarkers to select the most at-risk population, to detect the disease while measurable and yet not clinically apparent has been the goal of many investigations. Recent advances in molecular strategies and analytic platforms, including genomics, epigenomics, proteomics, and metabolomics, have identified increasing numbers of potential biomarkers in the blood, urine, exhaled breath condensate, bronchial specimens, saliva, and sputum, but none have yet moved to the clinical setting. Therefore, there is a recognized gap between the promise and the product delivery in the cancer biomarker field. In this review, we define clinical contexts where risk and diagnostic biomarkers may have use in the management of lung cancer, identify the most relevant candidate biomarkers of early detection, provide their state of development, and finally discuss critical aspects of study design in molecular biomarkers for early detection of lung cancer.

Markers of field cancerization: proposed clinical applications in prostate biopsies

Field cancerization denotes the occurrence of genetic, epigenetic, and biochemical aberrations in structurally intact cells in histologically normal tissues adjacent to cancerous lesions. This paper tabulates markers of prostate field cancerization known to date and discusses their potential clinical value in the analysis of prostate biopsies, including diagnosis, monitoring progression during active surveillance, and assessing efficacy of presurgical neoadjuvant and focal therapeutic interventions.

Signaling pathway-based identification of extensive prognostic gene signatures for lung adenocarcinoma

Tumor recurrence is the major cause of death in lung cancer treatment. To date, there is no clinically applied gene expression-based model to predict the risk for tumor recurrence in non-small cell lung cancer (NSCLC). We sought to embed crosstalk with major signaling pathways into biomarker identification. Three approaches were used to identify prognostic gene signatures from 442 lung adenocarcinoma samples. Candidate genes co-expressed with 6 or 7 major NSCLC signaling hallmarks were identified from genome-wide coexpression networks specifically associated with different prognostic groups. From these candidate genes, the first approach selected genes significantly associated with disease-specific survival using univariate Cox model. The second approach used random forests to refine the gene signatures; and the third approach used Relief algorithm to form the final gene sets. A total of 21 gene signatures were identified using these three approaches. These gene signatures generated significant prognostic stratifications (log-rank P<0.05 in Kaplan-Meier analyses; hazard ratio >1, P<0.05) in all tumors, stage I only, and in stage I patients not receiving chemotherapy in all training and test sets. In multivariate analyses with age, gender, race, smoking history, cancer stage, and tumor differentiation, a 10-gene signature had a hazard ratio of 3.23 (95% CI: [1.48, 7.06]), which was a more significant prognostic factor than other clinical factors, except cancer stage (III vs. I; with no significant difference). All identified 21 gene signatures outperformed other lung cancer signatures evaluated in the Director's Challenge Study. This study is an important step toward personalized prognosis of tumor recurrence and patient selection for adjuvant chemotherapy, with significant impact on down-stream clinical applications.

Assessing risk prediction models in case-control studies using semiparametric and nonparametric methods

The predictiveness curve is a graphical tool that characterizes the population distribution of Risk(Y)=P(D=1|Y), where D denotes a binary outcome such as occurrence of an event within a specified time period and Y denotes predictors. A wider distribution of Risk(Y) indicates better performance of a risk model in the sense that making treatment recommendations is easier for more subjects. Decisions are more straightforward when a subject's risk is deemed to be high or low. Methods have been developed to estimate predictiveness curves from cohort studies. However, early phase studies to evaluate novel risk prediction markers typically employ case-control designs. Here, we present semiparametric and nonparametric methods for evaluating a continuous risk prediction marker that accommodates case-control data. Small sample properties are investigated through simulation studies. The semiparametric methods are substantially more efficient than their nonparametric counterparts under a correctly specified model. We generalize them to settings where multiple prediction markers are involved. Applications to prostate cancer risk prediction markers illustrate methods for comparing the risk prediction capacities of markers and for evaluating the increment in performance gained by adding a marker to a baseline risk model. We propose a modified Hosmer-Lemeshow test for case-control study data to assess calibration of the risk model that is a natural complement to this graphical tool.

Hybrid models identified a 12-gene signature for lung cancer prognosis and chemoresponse prediction

Background

Lung cancer remains the leading cause of cancer-related deaths worldwide. The recurrence rate ranges from 35–50% among early stage non-small cell lung cancer patients. To date, there is no fully-validated and clinically applied prognostic gene signature for personalized treatment.

Methodology/Principal Findings

From genome-wide mRNA expression profiles generated on 256 lung adenocarcinoma patients, a 12-gene signature was identified using combinatorial gene selection methods, and a risk score algorithm was developed with Naïve Bayes. The 12-gene model generates significant patient stratification in the training cohort HLM & UM (n = 256; log-rank P = 6.96e-7) and two independent validation sets, MSK (n = 104; log-rank P = 9.88e-4) and DFCI (n = 82; log-rank P = 2.57e-4), using Kaplan-Meier analyses. This gene signature also stratifies stage I and IB lung adenocarcinoma patients into two distinct survival groups (log-rank P<0.04). The 12-gene risk score is more significant (hazard ratio = 4.19, 95% CI: [2.08, 8.46]) than other commonly used clinical factors except tumor stage (III vs. I) in multivariate Cox analyses. The 12-gene model is more accurate than previously published lung cancer gene signatures on the same datasets. Furthermore, this signature accurately predicts chemoresistance/chemosensitivity to Cisplatin, Carboplatin, Paclitaxel, Etoposide, Erlotinib, and Gefitinib in NCI-60 cancer cell lines (P<0.017). The identified 12 genes exhibit curated interactions with major lung cancer signaling hallmarks in functional pathway analysis. The expression patterns of the signature genes have been confirmed in RT-PCR analyses of independent tumor samples.

Conclusions/Significance

The results demonstrate the clinical utility of the identified gene signature in prognostic categorization. With this 12-gene risk score algorithm, early stage patients at high risk for tumor recurrence could be identified for adjuvant chemotherapy; whereas stage I and II patients at low risk could be spared the toxic side effects of chemotherapeutic drugs.

Semiparametric methods for evaluating the covariate-specific predictiveness of continuous markers in matched case-control studies

To assess the value of a continuous marker in predicting the risk of a disease, a graphical tool called the predictiveness curve has been proposed. It characterizes the marker's predictiveness, or capacity to risk stratify the population by displaying the distribution of risk endowed by the marker. Methods for making inference about the curve and for comparing curves in a general population have been developed. However, knowledge about a marker's performance in the general population only is not enough. Since a marker's effect on the risk model and its distribution can both differ across subpopulations, its predictiveness may vary when applied to different subpopulations. Moreover, information about the predictiveness of a marker conditional on baseline covariates is valuable for individual decision making about having the marker measured or not. Therefore, to fully realize the usefulness of a risk prediction marker, it is important to study its performance conditional on covariates. In this article, we propose semiparametric methods for estimating covariate-specific predictiveness curves for a continuous marker. Unmatched and matched case-control study designs are accommodated. We illustrate application of the methodology by evaluating serum creatinine as a predictor of risk of renal artery stenosis.