netANOVA: novel graph clustering technique with significance assessment via hierarchical ANOVA.

Many problems in life sciences can be brought back to a comparison of graphs. Even though a multitude of such techniques exist, often, these assume prior knowledge about the partitioning or the number of clusters and fail to provide statistical significance of observed between-network heterogeneity. Addressing these issues, we developed an unsupervised workflow to identify groups of graphs from reliable network-based statistics. In particular, we first compute the similarity between networks via appropriate distance measures between graphs and use them in an unsupervised hierarchical algorithm to identify classes of similar networks. Then, to determine the optimal number of clusters, we recursively test for distances between two groups of networks. The test itself finds its inspiration in distance-wise ANOVA algorithms. Finally, we assess significance via the permutation of between-object distance matrices. Notably, the approach, which we will call netANOVA, is flexible since users can choose multiple options to adapt to specific contexts and network types. We demonstrate the benefits and pitfalls of our approach via extensive simulations and an application to two real-life datasets. NetANOVA achieved high performance in many simulation scenarios while controlling type I error. On non-synthetic data, comparison against state-of-the-art methods showed that netANOVA is often among the top performers. There are many application fields, including precision medicine, for which identifying disease subtypes via individual-level biological networks improves prevention programs, diagnosis and disease monitoring.

The prediabetes conundrum: striking the balance between risk and resources.

The current definition of prediabetes is controversial and subject to continuous debate. Nonetheless, prediabetes is a risk factor for type 2 diabetes, is highly prevalent and is associated with diabetic complications and mortality. Thereby, it has the potential to become a huge strain on healthcare systems in the future, necessitating action from legislators and healthcare providers. But how do we best reduce its associated burden on health? As a compromise between differing opinions in the literature and among the authors of this article, we suggest stratifying individuals with prediabetes according to estimated risk and only offering individual-level preventive interventions to those at high risk. At the same time, we argue to identify those with prediabetes and already established diabetes-related complications and treat them as we would treat individuals with established type 2 diabetes.

Statistical perspectives on using hepatocellular carcinoma risk models to inform surveillance decisions.

More than 50,000 people are diagnosed with hepatocellular carcinoma (HCC) every year in Europe. Many cases are known to specialist liver centres years before they present with HCC. Despite this, HCC is usually detected at a late stage, when prognosis is very poor. For more than two decades, clinical guidelines have recommended uniform surveillance for all patients with cirrhosis. However, studies continue to show that this broad-based approach is inefficient and poorly implemented in practice. A "personalised" approach, where the surveillance regimen is customised to the needs of the patient, is gaining growing support in the clinical community. The cornerstone of personalised surveillance is the HCC risk model - a mathematical equation predicting a patient's individualised probability of developing HCC within a specific time window. However, although numerous risk models have now been published, few are being used in routine care to inform HCC surveillance decisions. In this article, we discuss methodological issues stymieing the use of HCC risk models in routine practice - highlighting biases, evidence gaps and misconceptions that future research must address.

The Future of Precision Medicine to Predict Outcomes and Control Tissue Remodeling in Inflammatory Bowel Disease.

Inflammatory bowel disease is characterized by significant interindividual heterogeneity. With a wider selection of pharmacologic and nonpharmacologic interventions available and in advanced developmental stages, a priority for the coming decade is to determine accurate methods of predicting treatment response and disease course. Precision medicine strategies will allow tailoring of preventative and therapeutic decisions to individual patient needs. In this review, we consider the future of precision medicine in inflammatory bowel disease. We discuss the critical need to extend from research focused on short-term symptomatic response to integrative multi-omic systems biology strategies to identify and validate biomarkers that underpin precision approaches. Crucially, the international community has collective responsibility to provide well-phenotyped and -curated longitudinal datasets for scientific discovery and validation. Research must also study broader aspects of the immune response, including components of the extracellular matrix, to better understand biological pathways initiating and perpetuating tissue fibrosis and longer-term disease complications.

Adaptive enrichment designs with a continuous biomarker.

A popular design for clinical trials assessing targeted therapies is the two-stage adaptive enrichment design with recruitment in stage 2 limited to a biomarker-defined subgroup chosen based on data from stage 1. The data-dependent selection leads to statistical challenges if data from both stages are used to draw inference on treatment effects in the selected subgroup. If subgroups considered are nested, as when defined by a continuous biomarker, treatment effect estimates in different subgroups follow the same distribution as estimates in a group-sequential trial. This result is used to obtain tests controlling the familywise type I error rate (FWER) for six simple subgroup selection rules, one of which also controls the FWER for any selection rule. Two approaches are proposed: one based on multivariate normal distributions suitable if the number of possible subgroups, k, is small, and one based on Brownian motion approximations suitable for large k. The methods, applicable in the wide range of settings with asymptotically normal test statistics, are illustrated using survival data from a breast cancer trial.

Diagnostic sequencing to support genetically stratified medicine in a tertiary care setting.

PURPOSE: The goal of stratified medicine is to identify subgroups of patients with similar disease mechanisms and specific responses to treatments. To prepare for stratified clinical trials, genome-wide genetic analysis should occur across clinical areas to identify undiagnosed genetic diseases and new genetic causes of disease. METHODS: To advance genetically stratified medicine, we have developed and implemented broad exome sequencing infrastructure and research protocols at Columbia University Irving Medical Center/NewYork-Presbyterian Hospital. RESULTS: We enrolled 4889 adult and pediatric probands and identified a primary result in 572 probands. The cohort was phenotypically and demographically heterogeneous because enrollment occurred across multiple specialty clinics (eg, epilepsy, nephrology, fetal anomaly). New gene-disease associations and phenotypic expansions were discovered across clinical specialties. CONCLUSION: Our study processes have enabled the enrollment and exome sequencing/analysis of a phenotypically and demographically diverse cohort of patients within 1 tertiary care medical center. Because all genomic data are stored centrally with permission for longitudinal access to the electronic medical record, subjects can be recontacted with updated genetic diagnoses or for participation in future genotype-based clinical trials. This infrastructure has allowed for the promotion of genetically stratified clinical trial readiness within the Columbia University Irving Medical Center/NewYork-Presbyterian Hospital health care system.

IDNetwork: A deep illness-death network based on multi-state event history process for disease prognostication.

Multi-state models can capture the different patterns of disease evolution. In particular, the illness-death model is used to follow disease progression from a healthy state to an intermediate state of the disease and to a death-related final state. We aim to use those models in order to adapt treatment decisions according to the evolution of the disease. In state-of-the art methods, the risks of transition between the states are modeled via (semi-) Markov processes and transition-specific Cox proportional hazard (P.H.) models. The Cox P.H. model assumes that each variable makes a linear contribution to the model, but the relationship between covariates and risks can be more complex in clinical situations. To address this challenge, we propose a neural network architecture called illness-death network (IDNetwork) that relaxes the linear Cox P.H. assumption within an illness-death process. IDNetwork employs a multi-task architecture and uses a set of fully connected subnetworks in order to learn the probabilities of transition. Through simulations, we explore different configurations of the architecture and demonstrate the added value of our model. IDNetwork significantly improves the predictive performance compared to state-of-the-art methods on a simulated data set, on two clinical trials for patients with colon cancer and on a real-world data set in breast cancer.

Precision-Enhancing Risk Stratification Tools for Selecting Optimal Treatment Durations in Tuberculosis Clinical Trials.

Rationale: No evidence-based tools exist to enhance precision in the selection of patient-specific optimal treatment durations to study in tuberculosis clinical trials. Objectives: To develop risk stratification tools that assign patients with tuberculosis into risk groups of unfavorable outcome and inform selection of optimal treatment duration for each patient strata to study in clinical trials. Methods: Publicly available data from four phase 3 trials, each evaluating treatment duration shortening from 6 to 4 months, were used to develop parametric time-to-event models that describe unfavorable outcomes. Regimen, baseline, and on-treatment characteristics were evaluated as predictors of outcomes. Exact regression coefficients of predictors were used to assign risk groups and predict optimal treatment durations. Measurements and Main Results: The parametric model had an area under the receiver operating characteristic curve of 0.72. A six-item risk score (HIV status, smear grade, sex, cavitary disease status, body mass index, and Month 2 culture status) successfully grouped participants into low (1,060/3,791; 28%), moderate (1,740/3,791; 46%), and high (991/3,791; 26%) risk, requiring treatment durations of 4, 6, and greater than 6 months, respectively, to reach a target cure rate of 93% when receiving standard-dose rifamycin-containing regimens. With current one-duration-fits-all approaches, high-risk groups have a 3.7-fold (95% confidence interval, 2.7-5.1) and 2.4-fold (1.9-2.9) higher hazard risk of unfavorable outcomes compared with low- and moderate-risk groups, respectively. Four-month regimens were noninferior to the standard 6-month regimen in the low-risk group. Conclusions: Our model discrimination was modest but consistent with current models of unfavorable outcomes. Our results showed that stratified medicine approaches are feasible and may achieve high cure rates in all patients with tuberculosis. An interactive risk stratification tool is provided to facilitate decision-making in the regimen development pathway.

Using heterogeneity in disease to understand the relationship between health and personality.

The aim of this study was to compare the relationship between two health outcomes (pain and self-reported health) and personality while accounting for heterogeneity in arthritic disease. Traditionally health research has treated patients' disease experiences as homogeneous but stratified medicine suggests that doing so might over-generalise findings and miss important effects. We present a longitudinal analysis over 14 years, on a subsample of 443 arthritic respondents from the English Longitudinal Study of Ageing (ELSA). Using linear regressions, we modelled how the Big Five domains of personality (wave 5) moderated the relationship between past (at wave 1) and present health (at wave 7). Then, to model heterogeneity in arthritis experience we included assignment to 4 different sub-groups based on their experience of pain progression. The results showed that modelling heterogeneity led to the identification of specific stratified effects for personality (neuroticism, agreeableness, and extraversion) not observed when these data are treated as homogenous. Higher agreeableness was associated with worse pain for those in a sub-group reporting the greatest pain, and higher extraversion was protective against pain among those whose pain improved. The results highlight the importance of modelling heterogeneity of disease.

Predictors of Microvascular Reperfusion After Myocardial Infarction.

PURPOSE OF REVIEW: In acute ST-segment elevation myocardial infarction (STEMI), successful restoration of blood flow in the infarct-related coronary artery may not secure effective myocardial reperfusion. The mortality and morbidity associated with acute MI remain significant. Microvascular obstruction (MVO) represents failed microvascular reperfusion. MVO is under-recognized, independently associated with adverse cardiac prognosis and represents an unmet therapeutic need. RECENT FINDINGS: Multiple factors including clinical presentation, patient characteristics, biochemical markers, and imaging parameters are associated with MVO after MI. Impaired microvascular reperfusion is common following percutaneous coronary intervention (PCI). New knowledge about disease mechanisms underpins precision medicine with individualized risk assessment, investigation, and stratified therapy. To date, there are no evidence-based therapies to prevent or treat MVO post-MI. Identifying novel therapy for MVO is the next frontier.

Omics Data and Their Integrative Analysis to Support Stratified Medicine in Neurodegenerative Diseases.

Molecular and clinical heterogeneity is increasingly recognized as a common characteristic of neurodegenerative diseases (NDs), such as Alzheimer's disease, Parkinson's disease and amyotrophic lateral sclerosis. This heterogeneity makes difficult the development of early diagnosis and effective treatment approaches, as well as the design and testing of new drugs. As such, the stratification of patients into meaningful disease subgroups, with clinical and biological relevance, may improve disease management and the development of effective treatments. To this end, omics technologies-such as genomics, transcriptomics, proteomics and metabolomics-are contributing to offer a more comprehensive view of molecular pathways underlying the development of NDs, helping to differentiate subtypes of patients based on their specific molecular signatures. In this article, we discuss how omics technologies and their integration have provided new insights into the molecular heterogeneity underlying the most prevalent NDs, aiding to define early diagnosis and progression markers as well as therapeutic targets that can translate into stratified treatment approaches, bringing us closer to the goal of personalized medicine in neurology.

Accounting for grouped predictor variables or pathways in high-dimensional penalized Cox regression models.

BACKGROUND: The standard lasso penalty and its extensions are commonly used to develop a regularized regression model while selecting candidate predictor variables on a time-to-event outcome in high-dimensional data. However, these selection methods focus on a homogeneous set of variables and do not take into account the case of predictors belonging to functional groups; typically, genomic data can be grouped according to biological pathways or to different types of collected data. Another challenge is that the standard lasso penalisation is known to have a high false discovery rate. RESULTS: We evaluated different penalizations in a Cox model to select grouped variables in order to further penalize variables that, in addition to having a low effect, belong to a group with a low overall effect; and to favor the selection of variables that, in addition to having a large effect, belong to a group with a large overall effect. We considered the case of prespecified and disjoint groups and proposed diverse weights for the adaptive lasso method. In particular we proposed the product Max Single Wald by Single Wald weighting (MSW*SW) which takes into account the information of the group to which it belongs and of this biomarker. Through simulations, we compared the selection and prediction ability of our approach with the standard lasso, the composite Minimax Concave Penalty (cMCP), the group exponential lasso (gel), the Integrative L1-Penalized Regression with Penalty Factors (IPF-Lasso), and the Sparse Group Lasso (SGL) methods. In addition, we illustrated the methods using gene expression data of 614 breast cancer patients. CONCLUSIONS: The adaptive lasso with the MSW*SW weighting method incorporates both the information in the grouping structure and the individual variable. It outperformed the competitors by reducing the false discovery rate without severely increasing the false negative rate.

Confidence interval estimation for the changepoint of treatment stratification in the presence of a qualitative covariate-treatment interaction.

The goal in stratified medicine is to administer the "best" treatment to a patient. Not all patients might benefit from the same treatment; the choice of best treatment can depend on certain patient characteristics. In this article, it is assumed that a time-to-event outcome is considered as a patient-relevant outcome and a qualitative interaction between a continuous covariate and treatment exists, ie, that patients with different values of one specific covariate should be treated differently. We suggest and investigate different methods for confidence interval estimation for the covariate value, where the treatment recommendation should be changed based on data collected in a randomized clinical trial. An adaptation of Fieller's theorem, the delta method, and different bootstrap approaches (normal, percentile-based, wild bootstrap) are investigated and compared in a simulation study. Extensions to multivariable problems are presented and evaluated. We observed appropriate confidence interval coverage following Fieller's theorem irrespective of sample size but at the cost of very wide or even infinite confidence intervals. The delta method and the wild bootstrap approach provided the smallest intervals but inadequate coverage for small to moderate event numbers, also depending on the location of the true changepoint. For the percentile-based bootstrap, wide intervals were observed, and it was slightly conservative regarding coverage, whereas the normal bootstrap did not provide acceptable results for many scenarios. The described methods were also applied to data from a randomized clinical trial comparing two treatments for patients with symptomatic, severe carotid artery stenosis, considering patient's age as predictive marker.

Point and interval estimation in two-stage adaptive designs with time to event data and biomarker-driven subpopulation selection.

In personalized medicine, it is often desired to determine if all patients or only a subset of them benefit from a treatment. We consider estimation in two-stage adaptive designs that in stage 1 recruit patients from the full population. In stage 2, patient recruitment is restricted to the part of the population, which, based on stage 1 data, benefits from the experimental treatment. Existing estimators, which adjust for using stage 1 data for selecting the part of the population from which stage 2 patients are recruited, as well as for the confirmatory analysis after stage 2, do not consider time to event patient outcomes. In this work, for time to event data, we have derived a new asymptotically unbiased estimator for the log hazard ratio and a new interval estimator with good coverage probabilities and probabilities that the upper bounds are below the true values. The estimators are appropriate for several selection rules that are based on a single or multiple biomarkers, which can be categorical or continuous.

Qualifying and quantifying the precision medicine rhetoric.

BACKGROUND: With the rise of precision medicine efforts worldwide, our study objective was to describe and map the emerging precision medicine landscape. A Google search was conducted between June 19, 2017 to July 20, 2017 to examine how "precision medicine" and its analogous terminology were used to describe precision medicine efforts. Resulting web-pages were reviewed for geographic location, data type(s), program aim(s), sample size, duration, and the key search terms used and recorded in a database. Descriptive statistics were applied to quantify terminology used to describe specific precision medicine efforts. Qualitative data were analyzed for content and patterns. RESULTS: Of the 108 programs identified through our search, 84% collected only biospecimen(s) and, of those that collected at least two data types, 42% mentioned both Electronic Health Records (EHR) and biospecimen. Given the majority of efforts limited to biospecimen(s) use, genetic research seems to be prioritized in association with precision medicine. Roughly, 54% were found to collect two or more data types, which limits the output of information that may contribute to understanding of the interplay of genetic, lifestyle, and environmental factors. Over half were government-funded with roughly a third being industry-funded. Most initiatives were concentrated in the United States, Europe, and Asia. CONCLUSIONS: To our knowledge, this is the first study to map and qualify the global precision medicine landscape. Our findings reveal that precision medicine efforts range from large model cohort studies involving multidimensional, longitudinal data to biorepositories with a collection of blood samples. We present a spectrum where past, present, and future PM-like efforts can fall based on their scope and potential impact. If precision medicine is based on genes, lifestyle and environmental factors, we recommend programs claiming to be precision medicine initiatives to incorporate multidimensional data that can inform a holistic approach to healthcare.

Effects of Age and Sex of Response to Ursodeoxycholic Acid and Transplant-free Survival in Patients With Primary Biliary Cholangitis.

BACKGROUND & AIMS: Primary biliary cholangitis (PBC) predominantly affects middle-aged women; there are few data on disease phenotypes and outcomes of PBC in men and younger patients. We investigated whether differences in sex and/or age at the start of ursodeoxycholic acid (UDCA) treatment are associated with response to therapy, based on biochemical markers, or differences in transplant-free survival. METHODS: We performed a longitudinal retrospective study of 4355 adults in the Global PBC Study cohort, collected from 17 centers across Europe and North America. Patients received a diagnosis of PBC from 1961 through 2014. We evaluated the effects of sex and age on response to UDCA treatment (based on GLOBE score) and transplant-free survival using logistic regression and Cox regression analyses, respectively. RESULTS: Male patients were older at the start of treatment (58.3±12.1 years vs 54.3±11.6 years for women; P<.0001) and had higher levels of bilirubin and lower circulating platelet counts (P<.0001). Younger patients (45 years or younger) had increased serum levels of transaminases than older patients (older than 45 years). Patients older than 45 years at time of treatment initiation had increased odds of a biochemical response to UDCA therapy, based on GLOBE score, compared to younger patients. The greatest odds of response to UDCA were observed in patients older than 65 years (odds ratio compared to younger patients 45 years or younger, 5.48; 95% CI, 3.92-7.67; P<.0001). Risk of liver transplant or death (compared to a general population matched for age, sex, and birth year) decreased significantly with advancing age: hazard ratio for patients 35 years or younger, 14.59 (95% CI, 9.66-22.02) vs hazard ratio for patients older than 65 years, 1.39 (95% CI, 1.23-1.57) (P<.0001). On multivariable analysis, sex was not independently associated with response or transplant-free survival. CONCLUSION: In longitudinal analysis of 4355 adults in the Global PBC Study, we associated patient age, but not sex, with response to UDCA treatment and transplant-free survival. Younger age at time of treatment initiation is associated with increased risk of treatment failure, liver transplant, and death.

Patient-centered simulations to assess the usefulness of the 70-gene signature for adjuvant chemotherapy administration in early-stage breast cancer.

PURPOSE: From the MINDACT trial, Cardoso et al. did not demonstrate a significant efficacy for adjuvant chemotherapy (CT) for women with early-stage breast cancer presenting high clinical and low genomic risks. Our objective was to assess the usefulness of the 70-gene signature in this population by using an alternative endpoint: the number of Quality-Adjusted Life-Years (QALYs), i.e., a synthetic measure of quantity and quality of life. METHODS: Based on the results of the MINDACT trial, we simulated a randomized clinical trial consisting of 1497 women with early-stage breast cancer presenting high clinical and low genomic risks. The individual preferences for the different health states and corresponding decrements were obtained from the literature. RESULTS: The gain in terms of 5-year disease-free survival was 2.8% (95% CI from - 0.1 to 5.7%, from 90.4% for women without CT to 93.3% for women with CT). In contrast, due to the associated side effects, CT significantly reduced the number of QALYs by 62 days (95% CI from 55 to 70 days, from 4.13 years for women without CT to 3.96 years for women with CT). CONCLUSION: Our results support the conclusions published by Cardoso et al. by providing additional evidence that the 70-gene signature can be used to avoid overtreatment by CT for women with high clinical risk but low genomic risk.

Working together to deliver stratified medicine research effectively.

INTRODUCTION OR BACKGROUND: Stratified medicine is an important area of research across all clinical specialties, with far reaching impact in many spheres. Despite recently formulated global policy and research programmes, major challenges for delivering stratified medicine studies persist. Across the globe, clinical research infrastructures have been setup to facilitate high quality clinical research. SOURCES OF DATA: This article reviews the literature and summarizes views collated from a workshop held by the UK Pharmacogenetics and Stratified Medicine Network and the NIHR Clinical Research Network in November 2016. AREAS OF AGREEMENT: Stratified medicine is an important area of clinical research and health policy, benefitting from substantial international, cross-sector investment and has the potential to transform patient care. However there are significant challenges to the delivery of stratified medicine studies. AREAS OF CONTROVERSY: Complex methodology and lack of consistency of definition and agreement on key approaches to the design, regulation and delivery of research contribute to these challenges and would benefit from greater focus. GROWING POINTS: Effective partnership and development of consistent approaches to the key factors relating to stratified medicine research is required to help overcome these challenges. AREAS TIMELY FOR DEVELOPING RESEARCH: This paper examines the critical contribution clinical research networks can make to the delivery of national (and international) initiatives in the field of stratified medicine. Importantly, it examines the position of clinical research in stratified medicine at a time when pressures on the clinical and social services are mounting.

Meta-analysis of predictive models to assess the clinical validity and utility for patient-centered medical decision making: application to the CAncer of the Prostate Risk Assessment (CAPRA).

BACKGROUND: The Cancer of the Prostate Risk Assessment (CAPRA) score was designed and validated several times to predict the biochemical recurrence-free survival after a radical prostatectomy. Our objectives were, first, to study the clinical validity of the CAPRA score, and, second, to assess its clinical utility for stratified medicine from an original patient-centered approach. METHODS: We proposed a meta-analysis based on a literature search using MEDLINE. Observed and predicted biochemical-recurrence-free survivals were compared to assess the calibration of the CAPRA score. Discriminative capacities were evaluated by estimating the summary time-dependent ROC curve. The clinical utility of the CAPRA score was evaluated according to the following stratified decisions: active monitoring for low-risk patients, prostatectomy for intermediate-risk patients, or radio-hormonal therapy for high risk patients. For this purpose, we assessed CAPRA's clinical utility in terms of its ability to maximize time-dependent utility functions (i.e. Quality-Adjusted Life-Years - QALYs). RESULTS: We identified 683 manuscripts and finally retained 9 studies. We reported good discriminative capacities with an area under the SROCt curve at 0.73 [95%CI from 0.67 to 0.79], while graphical calibration seemed acceptable. Nevertheless, we also described that the CAPRA score was unable to discriminate between the three medical alternatives, i.e. it did not allow an increase in the number of life years in perfect health (QALYs) of patients with prostate cancer. CONCLUSIONS: We confirmed the prognostic capacities of the CAPRA score. In contrast, we were not able to demonstrate its clinical usefulness for stratified medicine from a patient-centered perspective. Our results also highlighted the confusion between clinical validity and utility. This distinction should be better considered in order to develop predictive tools useful in practice.

From hype to reality: data science enabling personalized medicine.

BACKGROUND: Personalized, precision, P4, or stratified medicine is understood as a medical approach in which patients are stratified based on their disease subtype, risk, prognosis, or treatment response using specialized diagnostic tests. The key idea is to base medical decisions on individual patient characteristics, including molecular and behavioral biomarkers, rather than on population averages. Personalized medicine is deeply connected to and dependent on data science, specifically machine learning (often named Artificial Intelligence in the mainstream media). While during recent years there has been a lot of enthusiasm about the potential of 'big data' and machine learning-based solutions, there exist only few examples that impact current clinical practice. The lack of impact on clinical practice can largely be attributed to insufficient performance of predictive models, difficulties to interpret complex model predictions, and lack of validation via prospective clinical trials that demonstrate a clear benefit compared to the standard of care. In this paper, we review the potential of state-of-the-art data science approaches for personalized medicine, discuss open challenges, and highlight directions that may help to overcome them in the future. CONCLUSIONS: There is a need for an interdisciplinary effort, including data scientists, physicians, patient advocates, regulatory agencies, and health insurance organizations. Partially unrealistic expectations and concerns about data science-based solutions need to be better managed. In parallel, computational methods must advance more to provide direct benefit to clinical practice.