Multimodel inference for biomarker development: an application to schizophrenia

Precision Medicine Approaches to Mental Health Care

Developing a more comprehensive understanding of the physiological underpinnings of mental illness, precision medicine has the potential to revolutionize psychiatric care. With recent breakthroughs in next-generation multi-omics technologies and data analytics, it is becoming more feasible to leverage multimodal biomarkers, from genetic variants to neuroimaging biomarkers, to objectify diagnostics and treatment decisions in psychiatry and improve patient outcomes. Ongoing work in precision psychiatry will parallel progress in precision oncology and cardiology to develop an expanded suite of blood- and neuroimaging-based diagnostic tests, empower monitoring of treatment efficacy over time, and reduce patient exposure to ineffective treatments. The emerging model of precision psychiatry has the potential to mitigate some of psychiatry's most pressing issues, including improving disease classification, lengthy treatment duration, and suboptimal treatment outcomes. This narrative-style review summarizes some of the emerging breakthroughs and recurring challenges in the application of precision medicine approaches to mental health care.

Development and Validation of Predictive Model for a Diagnosis of First Episode Psychosis Using the Multinational EU-GEI Case-control Study and Modern Statistical Learning Methods

Background and Hypothesis

It is argued that availability of diagnostic models will facilitate a more rapid identification of individuals who are at a higher risk of first episode psychosis (FEP). Therefore, we developed, evaluated, and validated a diagnostic risk estimation model to classify individual with FEP and controls across six countries.

Study Design

We used data from a large multi-center study encompassing 2627 phenotypically well-defined participants (aged 18-64 years) recruited from six countries spanning 17 research sites, as part of the European Network of National Schizophrenia Networks Studying Gene-Environment Interactions study. To build the diagnostic model and identify which of important factors for estimating an individual risk of FEP, we applied a binary logistic model with regularization by the least absolute shrinkage and selection operator. The model was validated employing the internal-external cross-validation approach. The model performance was assessed with the area under the receiver operating characteristic curve (AUROC), calibration, sensitivity, and specificity.

Study Results

Having included preselected 22 predictor variables, the model was able to discriminate adults with FEP and controls with high accuracy across all six countries (rangesAUROC = 0.84-0.86). Specificity (range = 73.9-78.0%) and sensitivity (range = 75.6-79.3%) were equally good, cumulatively indicating an excellent model accuracy; though, calibration slope for the diagnostic model showed a presence of some overfitting when applied specifically to participants from France, the UK, and The Netherlands.

Conclusions

The new FEP model achieved a good discrimination and good calibration across six countries with different ethnic contributions supporting its robustness and good generalizability.

Diagnostic model development for schizophrenia based on peripheral blood mononuclear cell subtype-specific expression of metabolic markers

A significant proportion of the personal and economic burden of schizophrenia can be attributed to the late diagnosis or misdiagnosis of the disorder. A novel, objective diagnostic approaches could facilitate the early detection and treatment of schizophrenia and improve patient outcomes. In the present study, we aimed to identify robust schizophrenia-specific blood biomarkers, with the goal of developing an accurate diagnostic model. The levels of selected serum and peripheral blood mononuclear cell (PBMC) markers relevant to metabolic and immune function were measured in healthy controls (n = 26) and recent-onset schizophrenia patients (n = 36) using multiplexed immunoassays and flow cytometry. Analysis of covariance revealed significant upregulation of insulin receptor (IR) and fatty acid translocase (CD36) levels in T helper cells (F = 10.75, P = 0.002, Q = 0.024 and F = 21.58, P = 2.8 × 10^-5, Q = 0.0004, respectively), as well as downregulation of glucose transporter 1 (GLUT1) expression in monocytes (F = 21.46, P = 2.9 × 10^-5, Q = 0.0004). The most robust predictors, monocyte GLUT1 and T helper cell CD36, were used to develop a diagnostic model, which showed a leave-one-out cross-validated area under the receiver operating characteristic curve (AUC) of 0.78 (95% CI: 0.66-0.92). The diagnostic model was validated in two independent datasets. The model was able to distinguish first-onset, drug-naïve schizophrenia patients (n = 34) from healthy controls (n = 39) with an AUC of 0.75 (95% CI: 0.64-0.86), and also differentiated schizophrenia patients (n = 22) from patients with other neuropsychiatric conditions, including bipolar disorder, major depressive disorder and autism spectrum disorder (n = 68), with an AUC of 0.83 (95% CI: 0.75-0.92). These findings indicate that PBMC-derived biomarkers have the potential to support an accurate and objective differential diagnosis of schizophrenia.

Learnings from user feedback of a novel digital mental health assessment

Digital mental health interventions (DMHI) have the potential to address barriers to face-to-face mental healthcare. In particular, digital mental health assessments offer the opportunity to increase access, reduce strain on services, and improve identification. Despite the potential of DMHIs there remains a high drop-out rate. Therefore, investigating user feedback may elucidate how to best design and deliver an engaging digital mental health assessment. The current study aimed to understand 1304 user perspectives of (1) a newly developed digital mental health assessment to determine which features users consider to be positive or negative and (2) the Composite International Diagnostic Interview (CIDI) employed in a previous large-scale pilot study. A thematic analysis method was employed to identify themes in feedback to three question prompts related to: (1) the questions included in the digital assessment, (2) the homepage design and reminders, and (3) the assessment results report. The largest proportion of the positive and negative feedback received regarding the questions included in the assessment (n = 706), focused on the quality of the assessment (n = 183, 25.92% and n = 284, 40.23%, respectively). Feedback for the homepage and reminders (n = 671) was overwhelmingly positive, with the largest two themes identified being positive usability (i.e., ease of use; n = 500, 74.52%) and functionality (i.e., reminders; n = 278, 41.43%). The most frequently identified negative theme in results report feedback (n = 794) was related to the report content (n = 309, 38.92%), with users stating it was lacking in-depth information. Nevertheless, the most frequent positive theme regarding the results report feedback was related to wellbeing outcomes (n = 145, 18.26%), with users stating the results report, albeit brief, encouraged them to seek professional support. Interestingly, despite some negative feedback, most users reported that completing the digital mental health assessment has been worthwhile (n = 1,017, 77.99%). Based on these findings, we offer recommendations to address potential barriers to user engagement with a digital mental health assessment. In summary, we recommend undertaking extensive co-design activities during the development of digital assessment tools, flexibility in answering modalities within digital assessment, customizable additional features such as reminders, transparency of diagnostic decision making, and an actionable results report with personalized mental health resources.

A highly predictive autoantibody-based biomarker panel for prognosis in early-stage NSCLC with potential therapeutic implications

BACKGROUND

Lung cancer is the leading cause of cancer-related death worldwide. Surgical resection remains the definitive curative treatment for early-stage disease offering an overall 5-year survival rate of 62%. Despite careful case selection, a significant proportion of early-stage cancers relapse aggressively within the first year post-operatively. Identification of these patients is key to accurate prognostication and understanding the biology that drives early relapse might open up potential novel adjuvant therapies.

METHODS

We performed an unsupervised interrogation of >1600 serum-based autoantibody biomarkers using an iterative machine-learning algorithm.

RESULTS

We identified a 13 biomarker signature that was highly predictive for survivorship in post-operative early-stage lung cancer; this outperforms currently used autoantibody biomarkers in solid cancers. Our results demonstrate significantly poor survivorship in high expressers of this biomarker signature with an overall 5-year survival rate of 7.6%.

CONCLUSIONS

We anticipate that the data will lead to the development of an off-the-shelf prognostic panel and further that the oncogenic relevance of the proteins recognised in the panel may be a starting point for a new adjuvant therapy.

Screening of key biomarkers and immune infiltration in Pulmonary Arterial Hypertension via integrated bioinformatics analysis

This study aimed to screen key biomarkers and investigate immune infiltration in pulmonary arterial hypertension (PAH) based on integrated bioinformatics analysis. The Gene Expression Omnibus (GEO) database was used to download three mRNA expression profiles comprising 91 PAH lung specimens and 49 normal lung specimens. Three mRNA expression datasets were combined, and differentially expressed genes (DEGs) were obtained. Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses and the protein-protein interaction (PPI) network of DEGs were performed using the STRING and DAVID databases, respectively. The diagnostic value of hub gene expression in PAH was also analyzed. Finally, the infiltration of immune cells in PAH was analyzed using the CIBERSORT algorithm. Total 182 DEGs (117 upregulated and 65 downregulated) were identified, and 15 hub genes were screened. These 15 hub genes were significantly associated with immune system functions such as myeloid leukocyte migration, neutrophil migration, cell chemotaxis, Toll-like receptor signaling pathway, and NF-κB signaling pathway. A 7-gene-based model was constructed and had a better diagnostic value in identifying PAH tissues compared with normal controls. The immune infiltration profiles of the PAH and normal control samples were significantly different. High proportions of resting NK cells, activated mast cells, monocytes, and neutrophils were found in PAH samples, while high proportions of resting T cells CD4 memory and Macrophages M1 cell were found in normal control samples. Functional enrichment of DEGs and immune infiltration analysis between PAH and normal control samples might help to understand the pathogenesis of PAH.

Quantitative Proteomic Approach for Discriminating Major Depressive Disorder and Bipolar Disorder by Multiple Reaction Monitoring-Mass Spectrometry

Because major depressive disorder (MDD) and bipolar disorder (BD) manifest with similar symptoms, misdiagnosis is a persistent issue, necessitating their differentiation through objective methods. This study was aimed to differentiate between these disorders using a targeted proteomic approach. Multiple reaction monitoring-mass spectrometry (MRM-MS) analysis was performed to quantify protein targets regarding the two disorders in plasma samples of 270 individuals (90 MDD, 90 BD, and 90 healthy controls (HCs)). In the training set (72 MDD and 72 BD), a generalizable model comprising nine proteins was developed. The model was evaluated in the test set (18 MDD and 18 BD). The model demonstrated a good performance (area under the curve (AUC) >0.8) in discriminating MDD from BD in the training (AUC = 0.84) and test sets (AUC = 0.81) and in distinguishing MDD from BD without current hypomanic/manic/mixed symptoms (90 MDD and 75 BD) (AUC = 0.83). Subsequently, the model demonstrated excellent performance for drug-free MDD versus BD (11 MDD and 10 BD) (AUC = 0.96) and good performance for MDD versus HC (AUC = 0.87) and BD versus HC (AUC = 0.86). Furthermore, the nine proteins were associated with neuro, oxidative/nitrosative stress, and immunity/inflammation-related biological functions. This proof-of-concept study introduces a potential model for distinguishing between the two disorders.

A machine learning algorithm to differentiate bipolar disorder from major depressive disorder using an online mental health questionnaire and blood biomarker data

The vast personal and economic burden of mood disorders is largely caused by their under- and misdiagnosis, which is associated with ineffective treatment and worsening of outcomes. Here, we aimed to develop a diagnostic algorithm, based on an online questionnaire and blood biomarker data, to reduce the misdiagnosis of bipolar disorder (BD) as major depressive disorder (MDD). Individuals with depressive symptoms (Patient Health Questionnaire-9 score ≥5) aged 18-45 years were recruited online. After completing a purpose-built online mental health questionnaire, eligible participants provided dried blood spot samples for biomarker analysis and underwent the World Health Organization World Mental Health Composite International Diagnostic Interview via telephone, to establish their mental health diagnosis. Extreme Gradient Boosting and nested cross-validation were used to train and validate diagnostic models differentiating BD from MDD in participants who self-reported a current MDD diagnosis. Mean test area under the receiver operating characteristic curve (AUROC) for separating participants with BD diagnosed as MDD (N = 126) from those with correct MDD diagnosis (N = 187) was 0.92 (95% CI: 0.86-0.97). Core predictors included elevated mood, grandiosity, talkativeness, recklessness and risky behaviour. Additional validation in participants with no previous mood disorder diagnosis showed AUROCs of 0.89 (0.86-0.91) and 0.90 (0.87-0.91) for separating newly diagnosed BD (N = 98) from MDD (N = 112) and subclinical low mood (N = 120), respectively. Validation in participants with a previous diagnosis of BD (N = 45) demonstrated sensitivity of 0.86 (0.57-0.96). The diagnostic algorithm accurately identified patients with BD in various clinical scenarios, and could help expedite accurate clinical diagnosis and treatment of BD.

Diagnostic prediction model development using data from dried blood spot proteomics and a digital mental health assessment to identify major depressive disorder among individuals presenting with low mood

With less than half of patients with major depressive disorder (MDD) correctly diagnosed within the primary care setting, there is a clinical need to develop an objective and readily accessible test to enable earlier and more accurate diagnosis. The aim of this study was to develop diagnostic prediction models to identify MDD patients among individuals presenting with subclinical low mood, based on data from dried blood spot (DBS) proteomics (194 peptides representing 115 proteins) and a novel digital mental health assessment (102 sociodemographic, clinical and personality characteristics). To this end, we investigated 130 low mood controls, 53 currently depressed individuals with an existing MDD diagnosis (established current MDD), 40 currently depressed individuals with a new MDD diagnosis (new current MDD), and 72 currently not depressed individuals with an existing MDD diagnosis (established non-current MDD). A repeated nested cross-validation approach was used to evaluate variation in model selection and ensure model reproducibility. Prediction models that were trained to differentiate between established current MDD patients and low mood controls (AUC = 0.94 ± 0.01) demonstrated a good predictive performance when extrapolated to differentiate between new current MDD patients and low mood controls (AUC = 0.80 ± 0.01), as well as between established non-current MDD patients and low mood controls (AUC = 0.79 ± 0.01). Importantly, we identified DBS proteins A1AG1, A2GL, AL1A1, APOE and CFAH as important predictors of MDD, indicative of immune system dysregulation; as well as poor self-rated mental health, BMI, reduced daily experiences of positive emotions, and tender-mindedness. Despite the need for further validation, our preliminary findings demonstrate the potential of such prediction models to be used as a diagnostic aid for detecting MDD in clinical practice.

Implementing Precision Psychiatry: A Systematic Review of Individualized Prediction Models for Clinical Practice

BACKGROUND

The impact of precision psychiatry for clinical practice has not been systematically appraised. This study aims to provide a comprehensive review of validated prediction models to estimate the individual risk of being affected with a condition (diagnostic), developing outcomes (prognostic), or responding to treatments (predictive) in mental disorders.

METHODS

PRISMA/RIGHT/CHARMS-compliant systematic review of the Web of Science, Cochrane Central Register of Reviews, and Ovid/PsycINFO databases from inception until July 21, 2019 (PROSPERO CRD42019155713) to identify diagnostic/prognostic/predictive prediction studies that reported individualized estimates in psychiatry and that were internally or externally validated or implemented. Random effect meta-regression analyses addressed the impact of several factors on the accuracy of prediction models.

FINDINGS

Literature search identified 584 prediction modeling studies, of which 89 were included. 10.4% of the total studies included prediction models internally validated (n = 61), 4.6% models externally validated (n = 27), and 0.2% (n = 1) models considered for implementation. Across validated prediction modeling studies (n = 88), 18.2% were diagnostic, 68.2% prognostic, and 13.6% predictive. The most frequently investigated condition was psychosis (36.4%), and the most frequently employed predictors clinical (69.5%). Unimodal compared to multimodal models (β = .29, P = .03) and diagnostic compared to prognostic (β = .84, p < .0001) and predictive (β = .87, P = .002) models were associated with increased accuracy.

INTERPRETATION

To date, several validated prediction models are available to support the diagnosis and prognosis of psychiatric conditions, in particular, psychosis, or to predict treatment response. Advancements of knowledge are limited by the lack of implementation research in real-world clinical practice. A new generation of implementation research is required to address this translational gap.

A Combined Digital and Biomarker Diagnostic Aid for Mood Disorders (the Delta Trial): Protocol for an Observational Study

Background

Mood disorders affect hundreds of millions of people worldwide, imposing a substantial medical and economic burden. Existing diagnostic methods for mood disorders often result in a delay until accurate diagnosis, exacerbating the challenges of these disorders. Advances in digital tools for psychiatry and understanding the biological basis of mood disorders offer the potential for novel diagnostic methods that facilitate early and accurate diagnosis of patients.

Objective

The Delta Trial was launched to develop an algorithm-based diagnostic aid combining symptom data and proteomic biomarkers to reduce the misdiagnosis of bipolar disorder (BD) as a major depressive disorder (MDD) and achieve more accurate and earlier MDD diagnosis.

Methods

Participants for this ethically approved trial were recruited through the internet, mainly through Facebook advertising. Participants were then screened for eligibility, consented to participate, and completed an adaptive digital questionnaire that was designed and created for the trial on a purpose-built digital platform. A subset of these participants was selected to provide dried blood spot (DBS) samples and undertake a World Health Organization World Mental Health Composite International Diagnostic Interview (CIDI). Inclusion and exclusion criteria were chosen to maximize the safety of a trial population that was both relevant to the trial objectives and generalizable. To provide statistical power and validation sets for the primary and secondary objectives, 840 participants were required to complete the digital questionnaire, submit DBS samples, and undertake a CIDI.

Results

The Delta Trial is now complete. More than 3200 participants completed the digital questionnaire, 924 of whom also submitted DBS samples and a CIDI, whereas a total of 1780 participants completed a 6-month follow-up questionnaire and 1542 completed a 12-month follow-up questionnaire. The analysis of the trial data is now underway.

Conclusions

If a diagnostic aid is able to improve the diagnosis of BD and MDD, it may enable earlier treatment for patients with mood disorders.

International Registered Report Identifier (IRRID)

DERR1-10.2196/18453

Bacterial DNA induces the formation of heat-resistant disease-associated proteins in human plasma

Our study demonstrated for the first time that bacterial extracellular DNA (eDNA) can change the thermal behavior of specific human plasma proteins, leading to an elevation of the heat-resistant protein fraction, as well as to de novo acquisition of heat-resistance. In fact, the majority of these proteins were not known to be heat-resistant nor do they possess any prion-like domain. Proteins found to become heat-resistant following DNA exposure were named "Tetz-proteins". Interestingly, plasma proteins that become heat-resistant following treatment with bacterial eDNA are known to be associated with cancer. In pancreatic cancer, the proportion of proteins exhibiting eDNA-induced changes in thermal behavior was found to be particularly elevated. Therefore, we analyzed the heat-resistant proteome in the plasma of healthy subjects and in patients with pancreatic cancer and found that exposure to bacterial eDNA made the proteome of healthy subjects more similar to that of cancer patients. These findings open a discussion on the possible novel role of eDNA in disease development following its interaction with specific proteins, including those involved in multifactorial diseases such as cancer.

Integrating proteomic, sociodemographic and clinical data to predict future depression diagnosis in subthreshold symptomatic individuals

Individuals with subthreshold depression have an increased risk of developing major depressive disorder (MDD). The aim of this study was to develop a prediction model to predict the probability of MDD onset in subthreshold individuals, based on their proteomic, sociodemographic and clinical data. To this end, we analysed 198 features (146 peptides representing 77 serum proteins (measured using MRM-MS), 22 sociodemographic factors and 30 clinical features) in 86 first-episode MDD patients (training set patient group), 37 subthreshold individuals who developed MDD within two or four years (extrapolation test set patient group), and 86 subthreshold individuals who did not develop MDD within four years (shared reference group). To ensure the development of a robust and reproducible model, we applied feature extraction and model averaging across a set of 100 models obtained from repeated application of group LASSO regression with ten-fold cross-validation on the training set. This resulted in a 12-feature prediction model consisting of six serum proteins (AACT, APOE, APOH, FETUA, HBA and PHLD), three sociodemographic factors (body mass index, childhood trauma and education level) and three depressive symptoms (sadness, fatigue and leaden paralysis). Importantly, the model demonstrated a fair performance in predicting future MDD diagnosis of subthreshold individuals in the extrapolation test set (AUC = 0.75), which involved going beyond the scope of the model. These findings suggest that it may be possible to detect disease indications in subthreshold individuals up to four years prior to diagnosis, which has important clinical implications regarding the identification and treatment of high-risk individuals.