Investigation of metabolomic blood biomarkers for detection of adenocarcinoma lung cancer

A Comparative Study of Network-Based Machine Learning Approaches for Binary Classification in Metabolomics

Background/Objectives: Metabolomics has recently emerged as a key tool in the biological sciences, offering insights into metabolic pathways and processes. Over the last decade, network-based machine learning approaches have gained significant popularity and application across various fields. While several studies have utilized metabolomics profiles for sample classification, many network-based machine learning approaches remain unexplored for metabolomic-based classification tasks. This study aims to compare the performance of various network-based machine learning approaches, including recently developed methods, in metabolomics-based classification. Methods: A standard data preprocessing procedure was applied to 17 metabolomic datasets, and Bayesian neural network (BNN), convolutional neural network (CNN), feedforward neural network (FNN), Kolmogorov-Arnold network (KAN), and spiking neural network (SNN) were evaluated on each dataset. The datasets varied widely in size, mass spectrometry method, and response variable. Results: With respect to AUC on test data, BNN, CNN, FNN, KAN, and SNN were the top-performing models in 4, 1, 5, 3, and 4 of the 17 datasets, respectively. Regarding F1-score, the top-performing models were BNN (3 datasets), CNN (3 datasets), FNN (4 datasets), KAN (4 datasets), and SNN (3 datasets). For accuracy, BNN, CNN, FNN, KAN, and SNN performed best in 4, 1, 4, 4, and 4 datasets, respectively. Conclusions: No network-based modeling approach consistently outperformed others across the metrics of AUC, F1-score, or accuracy. Our results indicate that while no single network-based modeling approach is superior for metabolomics-based classification tasks, BNN, KAN, and SNN may be underappreciated and underutilized relative to the more commonly used CNN and FNN.

Monitoring central nervous system tumour metabolism using cerebrospinal fluid

Central nervous system (CNS) tumours are the most common cancer cause of death in under 40s in the UK, largely because they persist and recur and sometimes metastasise during treatment. Therefore, longitudinal monitoring of patients during and following treatment must be undertaken to understand the course of the disease and alter treatment plans reactively. This monitoring must be specific, sensitive, rapid, low cost, simple, and accepted by the patient. Cerebrospinal fluid (CSF) examination obtained following lumbar puncture, already a routine part of treatment in paediatric cases, could be better utilised with improved biomarkers. In this review, we discuss the potential for metabolites in the CSF to be used as biomarkers of CNS tumour remission, progression, response to drugs, recurrence and metastasis. We confer the clinical benefits and risks of this approach and conclude that there are many potential advantages over other tests and the required instrumentation is already present in UK hospitals. On the other hand, the approach needs more research investment to find more metabolite biomarkers, better understand their relation to the tumour, and validate those biomarkers in a standardised assay in order for the assay to become a clinical reality.

Combining Metabolomics and Machine Learning to Identify Diagnostic and Prognostic Biomarkers in Patients with Non-Small Cell Lung Cancer Pre- and Post-Radiation Therapy

Lung cancer is the leading cause of cancer-related deaths globally, with non-small cell lung cancer (NSCLC) accounting for over 85% of cases and poor prognosis in advanced stages. This study explored shifts in circulating metabolite levels in NSCLC patients versus healthy controls and examined the effects of conventionally fractionated radiation therapy (CFRT) and stereotactic body radiation therapy (SBRT). We enrolled 91 NSCLC patients (38 CFRT and 53 SBRT) and 40 healthy controls. Plasma metabolite levels were assessed using semi-targeted metabolomics, revealing 32 elevated and 18 reduced metabolites in patients. Key discriminatory metabolites included ethylmalonic acid, maltose, 3-phosphoglyceric acid, taurine, glutamic acid, glycocolic acid, and d-arabinose, with a combined Receiver Operating Characteristics curve indicating perfect discrimination between patients and controls. CFRT and SBRT affected different metabolites, but both changes suggested a partial normalization of energy and amino acid metabolism pathways. In conclusion, metabolomics identified distinct metabolic signatures in NSCLC patients with potential as diagnostic biomarkers. The differing metabolic responses to CFRT and SBRT reflect their unique therapeutic impacts, underscoring the utility of this technique in enhancing NSCLC diagnosis and treatment monitoring.

Metabolomics-based search for lung cancer markers among patients with different smoking status

Tobacco smoking is the main etiological factor of lung cancer (LC), which can also cause metabolome disruption. This study aimed to investigate whether the observed metabolic shift in LC patients was also associated with their smoking status. Untargeted metabolomics profiling was applied for the initial screening of changes in serum metabolic profile between LC and chronic obstructive pulmonary disease (COPD) patients, selected as a non-cancer group. Differences in metabolite profiles between current and former smokers were also tested. Then, targeted metabolomics methods were applied to verify and validate the proposed LC biomarkers. For untargeted metabolomics, a single extraction-dual separation workflow was applied. The samples were analyzed using a liquid chromatograph-high resolution quadrupole time-of-flight mass spectrometer. Next, the selected metabolites were quantified using liquid chromatography-triple-quadrupole mass spectrometry. The acquired data confirmed that patients' stratification based on smoking status impacted the discriminating ability of the identified LC marker candidates. Analyzing a validation set of samples enabled us to determine if the putative LC markers were truly robust. It demonstrated significant differences in the case of four metabolites: allantoin, glutamic acid, succinic acid, and sphingosine-1-phosphate. Our research showed that studying the influence of strong environmental factors, such as tobacco smoking, should be considered in cancer marker research since it reduces the risk of false positives and improves understanding of the metabolite shifts in cancer patients.

An amphiphilic dansyl based multianalyte sensor for the detection of Hg2+, PPi, and TNP: A three-in-one chemical sensor

A fluorescent dansyl-based amphiphilic probe, 5-(dimethylamino)-N-hexadecylnaphthalene-1-sulfonamide (DLC), was synthesized and characterized to detect multiple analytes at different sensing environments. In acetonitrile, DLC detects nitro explosives such as trinitrophenol (TNP) and 2,4-dinitrophenol (2,4-DNP) by an emission "on-off" response method, and the detection limits (LOD) were estimated to be as low as 4.3 µM and 17.4 µM, respectively. Amphiphilic long chains of the probe were embedded into lipid bilayers to form nanoscale vesicles DLC.Ves. Nanovesicular probe DLC.Ves was found to be highly selective for Hg2+ among other metal ions and for pyrophosphate (PPi) ions among various anions. DLC.Ves could detect Hg2+ with a turn "on-off" emission and PPi with ratiometric change in emission at 525 nm. It is proposed that DLC.Ves could detect Hg2+ via the Hg2+-induced aggregation quenching mechanism and PPi through the Hydrogen bonding. The LODs are estimated as 6.41 µM and 70.9 µM for Hg2+ and PPi, respectively. 1H NMR, SEM, and fluorescence lifetime measurements confirmed the binding mechanism. Thus, it is believed that the simple fluorescent probe DLC could be a prominent sensor to detect multiple analytes depending on the sensing medium.

Metabolomics and lipidomics in non-small cell lung cancer

Due to its insidious nature, lung cancer remains a leading cause of cancer-related deaths worldwide. Therefore, there is an urgent need to identify sensitive/specific biomarkers for early diagnosis and monitoring. The current study was designed to provide a current metabolic profile of non-small cell lung cancer (NSCLC) by systematically reviewing and summarizing various metabolomic/ lipidomic studies based on NSCLC blood samples, attempting to find biomarkers in human blood that can predict or diagnose NSCLC, and investigating the involvement of key metabolites in the pathogenesis of NSCLC. We searched all articles on lung cancer published in Elsevier, PubMed, Web of Science and the Cochrane Library between January 2012 and December 2022. After critical selection, a total of 31 studies (including 2768 NSCLC patients and 9873 healthy individuals) met the inclusion criteria, and 22 were classified as "high quality". Forty-six metabolites related to NSCLC were repeatedly identified, involving glucose metabolism, amino acid metabolism, lipid metabolism and nucleotide metabolism. Pyruvic acid, carnitine, phenylalanine, isoleucine, kynurenine and 3-hydroxybutyrate showed upward trends in all studies, citric acid, glycine, threonine, cystine, alanine, histidine, inosine, betaine and arachidic acid showed downward trends in all studies. This review summarizes the existing metabolomic/lipidomic studies related to the identification of blood biomarkers in NSCLC, examines the role of key metabolites in the pathogenesis of NSCLC, and provides an important reference for the clinical diagnosis and treatment of NSCLC. Due to the limited size and design heterogeneity of the existing studies, there is an urgent need for standardization of future studies, while validating existing findings with more studies.

The diagnostic performance of micro-RNA and metabolites in lung cancer: A meta-analysis

BACKGROUND

The diagnosis of lung cancer is based on the microscopic exam of tissue or liquid. During the recent decade, many biomarkers have been pointed to have a potential diagnostic role. These biomarkers may be assessed in blood, pleural effusion or sputum and they could avoid biopsies or other risky procedures. The authors aimed to assess the diagnostic performances of biomarkers focusing on micro-RNA and metabolites.

METHODS

This meta-analysis was conducted under the PRISMA guidelines during a nine-year-period (2013-2022). the Meta-Disc software 5.4 (free version) was used. Q test and I2 statistics were carried out to explore the heterogeneity among studies. Meta-regression was performed in case of significant heterogeneity. Publication bias was assessed using the funnel plot test and the Egger's test (free version JASP).

RESULTS

According to our inclusion criteria, 165 studies from 79 articles were included. The pooled SEN, SPE and dOR accounted, respectively, for 0.76, 0.79 and 13.927. The AUC was estimated to 0.859 suggesting a good diagnostic accuracy. The heterogeneity in the pooled SEN and SPE was statistically significant. The meta-regression analysis focusing on the technique used, the sample, the number of biomarkers, the biomarker subtype, the tumor stage and the ethnicity revealed the biomarker number (p  =  0.009) and the tumor stage (p  =  0.0241) as potential sources of heterogeneity.

CONCLUSION

Even if this meta-analysis highlighted the potential diagnostic utility of biomarkers, more prospective studies should be performed, especially to assess the biomarkers' diagnostic potential in early-stage lung cancers.

Two-Stage Deep-Learning Classifier for Diagnostics of Lung Cancer Using Metabolites

We developed a machine-learning system for the selective diagnostics of adenocarcinoma (AD), squamous cell carcinoma (SQ), and small-cell carcinoma lung (SC) cancers based on their metabolomic profiles. The system is organized as two-stage binary classifiers. The best accuracy for classification is 92%. We used the biomarkers sets that contain mostly metabolites related to cancer development. Compared to traditional methods, which exclude hierarchical classification, our method splits a challenging multiclass task into smaller tasks. This allows a two-stage classifier, which is more accurate in the scenario of lung cancer classification. Compared to traditional methods, such a "divide and conquer strategy" gives much more accurate and explainable results. Such methods, including our algorithm, allow for the systematic tracking of each computational step.

Recent Analytical Advances for Decoding Metabolic Reprogramming in Lung Cancer

Lung cancer is the leading cause of cancer-related death worldwide. Metabolic reprogramming is a fundamental trait associated with lung cancer development that fuels tumor proliferation and survival. Monitoring such metabolic pathways and their intermediate metabolites can provide new avenues concerning treatment strategies, and the identification of prognostic biomarkers that could be utilized to monitor drug responses in clinical practice. In this review, recent trends in the analytical techniques used for metabolome mapping of lung cancer are capitalized. These techniques include nuclear magnetic resonance (NMR), gas chromatography-mass spectrometry (GC-MS), liquid chromatography-mass spectrometry (LC-MS), and imaging mass spectrometry (MSI). The advantages and limitations of the application of each technique for monitoring the metabolite class or type are also highlighted. Moreover, their potential applications in the analysis of many biological samples will be evaluated.

An omics data analysis method based on feature linear relationship and graph convolutional network

Biological networks are known to be highly modular, and the dysfunction of network modules may cause diseases. Defining the key modules from the omics data and establishing the classification model is helpful in promoting the research of disease diagnosis and prognosis. However, for applying modules in downstream analysis such as disease states discrimination, most methods only utilize the node information, and ignore the node interactions or topological information, which may lead to false positives and limit the model performance. In this study, we propose an omics data analysis method based on feature linear relationship and graph convolutional network (LCNet). In LCNet, we adopt a way of applying the difference of feature linear relationships during disease development to characterize physiological and pathological changes and construct the differential linear relation network, which is simple and interpretable from the perspective of feature linear relationship. A greedy strategy is developed for searching the highly interactive modules with a strong discrimination ability. To fully utilize the information of the detected modules, the personalized sub-graphs for each sample based on the modules are defined, and the graph convolutional network (GCN) classifiers are trained to predict the sample labels. The experimental results on public datasets show the superiority of LCNet in classification performance. For Breast Cancer metabolic data, the identified metabolites by LCNet involve important pathways. Thus, LCNet can identify the module biomarkers by feature linear relationship and a greedy strategy, and label samples by personalized sub-graphs and GCN. It provides a new manner of utilizing node (molecule) information and topological information in the defined modules for better disease classification.

The Importance of the Immune System and Molecular Cell Signaling Pathways in the Pathogenesis and Progression of Lung Cancer

Lung cancer is a disease that in recent years has become one of the greatest threats to modern society. Every year there are more and more new cases and the percentage of deaths caused by this type of cancer increases. Despite many studies, scientists are still looking for answers regarding the mechanisms of lung cancer development and progression, with particular emphasis on the role of the immune system. The aim of this literature review was to present the importance of disorders of the immune system and the accompanying changes at the level of cell signaling in the pathogenesis of lung cancer. The collected results showed that in the process of immunopathogenesis of almost all subtypes of lung cancer, changes in the tumor microenvironment, deregulation of immune checkpoints and abnormalities in cell signaling pathways are involved, which contribute to the multistage and multifaceted carcinogenesis of this type of cancer. We, therefore, suggest that in future studies, researchers should focus on a detailed analysis of tumor microenvironmental immune checkpoints, and to validate their validity, perform genetic polymorphism analyses in a wide range of patients and healthy individuals to determine the genetic susceptibility to lung cancer development. In addition, further research related to the analysis of the tumor microenvironment; immune system disorders, with a particular emphasis on immunological checkpoints and genetic differences may contribute to the development of new personalized therapies that improve the prognosis of patients.

Interpretable machine learning with tree-based shapley additive explanations: Application to metabolomics datasets for binary classification

Machine learning (ML) models are used in clinical metabolomics studies most notably for biomarker discoveries, to identify metabolites that discriminate between a case and control group. To improve understanding of the underlying biomedical problem and to bolster confidence in these discoveries, model interpretability is germane. In metabolomics, partial least square discriminant analysis (PLS-DA) and its variants are widely used, partly due to the model's interpretability with the Variable Influence in Projection (VIP) scores, a global interpretable method. Herein, Tree-based Shapley Additive explanations (SHAP), an interpretable ML method grounded in game theory, was used to explain ML models with local explanation properties. In this study, ML experiments (binary classification) were conducted for three published metabolomics datasets using PLS-DA, random forests, gradient boosting, and extreme gradient boosting (XGBoost). Using one of the datasets, PLS-DA model was explained using VIP scores, while one of the best-performing models, a random forest model, was interpreted using Tree SHAP. The results show that SHAP has a more explanation depth than PLS-DA's VIP, making it a powerful method for rationalizing machine learning predictions from metabolomics studies.

Triazol-Methanaminium-Pillar[5]arene-Functionalized Single Nanochannel for Quantitative Analysis of Pyrophosphate in Water

Inorganic pyrophosphate (PPi) is an important biological functional anion and plays crucial roles in life science, environmental science, medicine, and chemical process. Quantification of PPi in water has far-reaching significance for life exploration, disease diagnosis, and water pollution control. The label-free quantitative detection of PPi anions with a nanofluidic sensing device based on a conical single nanochannel is demonstrated. The channel surface is functionalized with a synthetic PPi receptor, triazol-methanaminium-functionalized pillar[5]arene (TAMAP5), using carbodiimide coupling chemistry. Due to the specific binding between TAMAP5 and PPi, the functionalized nanochannel can discriminate PPi from other inorganic anions with high selectivity through ionic current recording, even in the presence of various interfering anions. The current response exhibits a linear correlation with PPi concentration in the range from 1 × 10^-7 to 1 × 10^-4 M with a limit of detection of 6.8 × 10^-7 M. A spike-and-recovery analysis of PPi in East Lake water samples indicates that the proposed nanofluidic sensor has the ability to quantitate micromolar concentrations of PPi in environmental water samples.

A Comprehensive Search of Non-Canonical Proteins in Non-Small Cell Lung Cancer and Their Impact on the Immune Response

There is substantial interest in mining neoantigens for cancer applications. Non-canonical proteins resulting from frameshift mutations have been identified as neoantigens in cancer. We investigated the landscape of non-canonical proteins in non-small cell lung cancer (NSCLC) and their induced immune response in the form of autoantibodies. A database of cryptoproteins was computationally constructed and comprised all alternate open reading frames (altORFs) and ORFs identified in pseudogenes, noncoding RNAs, and untranslated regions of mRNAs that did not align with known canonical proteins. Proteomic profiles of seventeen lung adenocarcinoma (LUAD) cell lines were searched to evaluate the occurrence of cryptoproteins. To assess the immunogenicity, immunoglobulin (Ig)-bound cryptoproteins in plasmas were profiled by mass spectrometry. The specimen set consisted of plasmas from 30 newly diagnosed NSCLC cases, pre-diagnostic plasmas from 51 NSCLC cases, and 102 control plasmas. An analysis of LUAD cell lines identified 420 cryptoproteins. Plasma Ig-bound analyses revealed 90 cryptoproteins uniquely found in cases and 14 cryptoproteins that had a fold-change >2 compared to controls. In pre-diagnostic samples, 17 Ig-bound cryptoproteins yielded an odds ratio ≥2. Eight Ig-bound cryptoproteins were elevated in both pre-diagnostic and newly diagnosed cases compared to controls. Cryptoproteins represent a class of neoantigens that induce an autoantibody response in NSCLC.

Imputation of Missing Values for Multi-Biospecimen Metabolomics Studies: Bias and Effects on Statistical Validity

The analysis of high-throughput metabolomics mass spectrometry data across multiple biological sample types (biospecimens) poses challenges due to missing data. During differential abundance analysis, dropping samples with missing values can lead to severe loss of data as well as biased results in group comparisons and effect size estimates. However, the imputation of missing data (the process of replacing missing data with estimated values such as a mean) may compromise the inherent intra-subject correlation of a metabolite across multiple biospecimens from the same subject, which in turn may compromise the efficacy of the statistical analysis of differential metabolites in biomarker discovery. We investigated imputation strategies when considering multiple biospecimens from the same subject. We compared a novel, but simple, approach that consists of combining the two biospecimen data matrices (rows and columns of subjects and metabolites) and imputes the two biospecimen data matrices together to an approach that imputes each biospecimen data matrix separately. We then compared the bias in the estimation of the intra-subject multi-specimen correlation and its effects on the validity of statistical significance tests between two approaches. The combined approach to multi-biospecimen studies has not been evaluated previously even though it is intuitive and easy to implement. We examine these two approaches for five imputation methods: random forest, k nearest neighbor, expectation-maximization with bootstrap, quantile regression, and half the minimum observed value. Combining the biospecimen data matrices for imputation did not greatly increase efficacy in conserving the correlation structure or improving accuracy in the statistical conclusions for most of the methods examined. Random forest tended to outperform the other methods in all performance metrics, except specificity.

Metabolomic Fingerprinting for the Detection of Early-Stage Lung Cancer: From the Genome to the Metabolome

The five-year survival rate of lung cancer patients is very low, mainly because most newly diagnosed patients present with locally advanced or metastatic disease. Therefore, early diagnosis is key to the successful treatment and management of lung cancer. Unfortunately, early detection methods of lung cancer are not ideal. In this brief review, we described early detection methods such as chest X-rays followed by bronchoscopy, sputum analysis followed by cytological analysis, and low-dose computed tomography (LDCT). In addition, we discussed the potential of metabolomic fingerprinting, compared to that of other biomarkers, including molecular targets, as a low-cost, high-throughput blood-based test that is both feasible and affordable for early-stage lung cancer screening of at-risk populations. Accordingly, we proposed a paradigm shift to metabolomics as an alternative to molecular and proteomic-based markers in lung cancer screening, which will enable blood-based routine testing and be accessible to those patients at the highest risk for lung cancer.

A Cost-Effectiveness Analysis of Lung Cancer Screening With Low-Dose Computed Tomography and a Diagnostic Biomarker

Background

The Lung Computed Tomography Screening Reporting and Data System (Lung-RADS) reduces the false-positive rate of lung cancer screening but introduces prolonged periods of uncertainty for indeterminate findings. We assess the cost-effectiveness of a screening program that assesses indeterminate findings earlier via a hypothetical diagnostic biomarker introduced in place of Lung-RADS 3 and 4A guidelines.

Methods

We evaluated the performance of the US Preventive Services Task Force (USPSTF) recommendations on lung cancer screening with and without a hypothetical noninvasive diagnostic biomarker using a validated microsimulation model. The diagnostic biomarker assesses the malignancy of indeterminate nodules, replacing Lung-RADS 3 and 4A guidelines, and is characterized by a varying sensitivity profile that depends on nodules' size, specificity, and cost. We tested the robustness of our findings through univariate sensitivity analyses.

Results

A lung cancer screening program per the USPSTF guidelines that incorporates a diagnostic biomarker with at least medium sensitivity profile and 90% specificity, that costs $250 or less, is cost-effective with an incremental cost-effectiveness ratio lower than $100 000 per quality-adjusted life year, and improves lung cancer-specific mortality reduction while requiring fewer screening exams than the USPSTF guidelines with Lung-RADS. A screening program with a biomarker costing $750 or more is not cost-effective. The health benefits accrued and costs associated with the screening program are sensitive to the disutility of indeterminate findings and specificity of the biomarker, respectively.

Conclusions

Lung cancer screening that incorporates a diagnostic biomarker, in place of Lung-RADS 3 and 4A guidelines, could improve the cost-effectiveness of the screening program and warrants further investigation.

Metabolomic Profiling in Lung Cancer: A Systematic Review

Lung cancer continues to be a significant burden worldwide and remains the leading cause of cancer-associated mortality. Two considerable challenges posed by this disease are the diagnosis of 61% of patients in advanced stages and the reduced five-year survival rate of around 4%. Noninvasively collected samples are gaining significant interest as new areas of knowledge are being sought and opened up. Metabolomics is one of these growing areas. In recent years, the use of metabolomics as a resource for the study of lung cancer has been growing. We conducted a systematic review of the literature from the past 10 years in order to identify some metabolites associated with lung cancer. More than 150 metabolites have been associated with lung cancer-altered metabolism. These were detected in different biological samples by different metabolomic analytical platforms. Some of the published results have been consistent, showing the presence/alteration of specific metabolites. However, there is a clear variability due to lack of a full clinical characterization of patients or standardized patients selection. In addition, few published studies have focused on the added value of the metabolomic profile as a means of predicting treatment response for lung cancer. This review reinforces the need for consistent and systematized studies, which will help make it possible to identify metabolic biomarkers and metabolic pathways responsible for the mechanisms that promote tumor progression, relapse and eventually resistance to therapy.

Metabolomics and health: from nutritional crops and plant-based pharmaceuticals to profiling of human biofluids

During the past decade metabolomics has emerged as one of the fastest developing branches of “-omics” technologies. Metabolomics involves documentation, identification, and quantification of metabolites through modern analytical platforms in various biological systems. Advanced analytical tools, such as gas chromatography–mass spectrometry (GC/MS), liquid chromatography–mass spectroscopy (LC/MS), and non-destructive nuclear magnetic resonance (NMR) spectroscopy, have facilitated metabolite profiling of complex biological matrices. Metabolomics, along with transcriptomics, has an influential role in discovering connections between genetic regulation, metabolite phenotyping and biomarkers identification. Comprehensive metabolite profiling allows integration of the summarized data towards manipulation of biosynthetic pathways, determination of nutritional quality markers, improvement in crop yield, selection of desired metabolites/genes, and their heritability in modern breeding. Along with that, metabolomics is invaluable in predicting the biological activity of medicinal plants, assisting the bioactivity-guided fractionation process and bioactive leads discovery, as well as serving as a tool for quality control and authentication of commercial plant-derived natural products. Metabolomic analysis of human biofluids is implemented in clinical practice to discriminate between physiological and pathological state in humans, to aid early disease biomarker discovery and predict individual response to drug therapy. Thus, metabolomics could be utilized to preserve human health by improving the nutritional quality of crops and accelerating plant-derived bioactive leads discovery through disease diagnostics, or through increasing the therapeutic efficacy of drugs via more personalized approach. Here, we attempt to explore the potential value of metabolite profiling comprising the above-mentioned applications of metabolomics in crop improvement, medicinal plants utilization, and, in the prognosis, diagnosis and management of complex diseases.

The Ability of Metabolomics to Discriminate Non-Small-Cell Lung Cancer Subtypes Depends on the Stage of the Disease and the Type of Material Studied

Identification of the NSCLC subtype at an early stage is still quite sophisticated. Metabolomics analysis of tissue and plasma of NSCLC patients may indicate new, and yet unknown, metabolic pathways active in the NSCLC. Our research characterized the metabolomics profile of tissue and plasma of patients with early and advanced NSCLC stage. Samples were subjected to thorough metabolomics analyses using liquid chromatography-mass spectrometry (LC-MS) technique. Tissue and/or plasma samples from 137 NSCLC patients were analyzed. Based on the early stage tissue analysis, more than 200 metabolites differentiating adenocarcinoma (ADC) and squamous cell lung carcinoma (SCC) subtypes as well as normal tissue, were identified. Most of the identified metabolites were amino acids, fatty acids, carnitines, lysoglycerophospholipids, sphingomyelins, plasmalogens and glycerophospholipids. Moreover, metabolites related to N-acyl ethanolamine (NAE) biosynthesis, namely glycerophospho (N-acyl) ethanolamines (GP-NAE), which discriminated early-stage SCC from ADC, have also been identified. On the other hand, the analysis of plasma of chronic obstructive pulmonary disease (COPD) and NSCLC patients allowed exclusion of the metabolites related to the inflammatory state in lungs and the identification of compounds (lysoglycerophospholipids, glycerophospholipids and sphingomyelins) truly characteristic to cancer. Our results, among already known, showed novel, thus far not described, metabolites discriminating NSCLC subtypes, especially in the early stage of cancer. Moreover, the presented results also indicated the activity of new metabolic pathways in NSCLC. Further investigations on the role of NAE biosynthesis pathways in the early stage of NSCLC may reveal new prognostic and diagnostic targets.

Combined Metabolomics with Transcriptomics Reveals Important Serum Biomarkers Correlated with Lung Cancer Proliferation through a Calcium Signaling Pathway

Lung cancer (LC) is one of the most malignant cancers in the world, but currently, it lacks effective noninvasive biomarkers to assist its early diagnosis. Our study aims to discover potential serum diagnostic biomarkers for LC. In our study, untargeted serum metabolomics of a discovery cohort and targeted analysis of a test cohort were performed based on gas chromatography-mass spectrometry. Both univariate and multivariate statistical analyses were employed to screen for differential metabolites between LC and healthy control (HC), followed by the selection of candidate biomarkers through multiple algorithms. The results showed that 15 metabolites were significantly dysregulated between LC and HC, and a panel, comprising cholesterol, oleic acid, myo-inositol, 2-hydroxybutyric acid, and 4-hydroxybutyric acid, was demonstrated to have excellent differentiating capability for LC based on multiple classification modelings. In addition, the molecular interaction analysis combined with transcriptomics revealed a close correlation between the candidate biomarkers and LC proliferation via a Ca²⁺ signaling pathway. Our study discovered that cholesterol, oleic acid, myo-inositol, 2-hydroxybutyric acid, and 4-hydroxybutyric acid in combination could be a promising diagnostic biomarker for LC, and most importantly, our results will shed some light on the pathophysiological mechanism underlying LC to understand it deeply. The data that support the findings of this study are openly available in MetaboLights at https://www.ebi.ac.uk/metabolights/, reference number MTBLS1517.

Evaluation of Single Sample Network Inference Methods for Metabolomics-Based Systems Medicine

Networks and network analyses are fundamental tools of systems biology. Networks are built by inferring pair-wise relationships among biological entities from a large number of samples such that subject-specific information is lost. The possibility of constructing these sample (individual)-specific networks from single molecular profiles might offer new insights in systems and personalized medicine and as a consequence is attracting more and more research interest. In this study, we evaluated and compared LIONESS (Linear Interpolation to Obtain Network Estimates for Single Samples) and ssPCC (single sample network based on Pearson correlation) in the metabolomics context of metabolite–metabolite association networks. We illustrated and explored the characteristics of these two methods on (i) simulated data, (ii) data generated from a dynamic metabolic model to simulate real-life observed metabolite concentration profiles, and (iii) 22 metabolomic data sets and (iv) we applied single sample network inference to a study case pertaining to the investigation of necrotizing soft tissue infections to show how these methods can be applied in metabolomics. We also proposed some adaptations of the methods that can be used for data exploration. Overall, despite some limitations, we found single sample networks to be a promising tool for the analysis of metabolomics data.

9G TestTM Cancer/Lung: A Desirable Companion to LDCT for Lung Cancer Screening

Simple Summary

Lung cancer is the most common cause of cancer-related deaths globally. Patients diagnosed at early-stage (0–I) have a higher survival rate than the metastasized stages (III–IV). Thus, there is great potential to reduce mortality by diagnosing lung cancer at stage 0~I through community screening. LDCT is a promising method, but it has a high false-positive rate. Therefore, a biomarker test that can be used in combination with LDCT for lung cancer screening to reduce false-positive rates is highly awaited. The present study evaluated the applicability of 9G test^TM Cancer/Lung test to detect stage 0~IV lung cancer. 9G test^TM Cancer/Lung test detects stage I, stage II, stage III, and stage IV cancers with the sensitivities of 77.5%, 78.1%, 67.4%, and 33.3%, respectively, at the specificity of 97.3%. These results indicate that the 9G test^TM Cancer/Lung can be used in conjunction with LDCT to screen lung cancer.

Abstract

A complimentary biomarker test that can be used in combination with LDCT for lung cancer screening is highly desirable to improve the diagnostic capacity of LDCT and reduce the false-positive rates. Most importantly, the stage I lung cancer detection rate can be dramatically increased by the simultaneous use of a biomarker test with LDCT. The present study was conducted to evaluate 9G test^TM Cancer/Lung’s sensitivity and specificity in detecting Stage 0~IV lung cancer. The obtained results indicate that the 9G test^TM Cancer/Lung can detect lung cancer with overall sensitivity and specificity of 75.0% (69.1~80.3) and 97.3% (95.0~98.8), respectively. The detection of stage I, stage II, stage III, and stage IV cancers with sensitivities of 77.5%, 78.1%, 67.4%, and 33.3%, respectively, at the specificity of 97.3% have never been reported before. The receiver operating characteristic curve analysis allowed us to determine the population-weighted AUC of 0.93 (95% CI, 0.91–0.95). These results indicate that the 9G test^TM Cancer/Lung can be used in conjunction with LDCT to screen lung cancer. Furthermore, obtained results indicate that the use of 9G test^TM Cancer/Lung with LDCT for lung cancer screening can increase stage I cancer detection, which is crucial to improve the currently low 5-year survival rates.

Predicting human health from biofluid-based metabolomics using machine learning

Biofluid-based metabolomics has the potential to provide highly accurate, minimally invasive diagnostics. Metabolomics studies using mass spectrometry typically reduce the high-dimensional data to only a small number of statistically significant features, that are often chemically identified—where each feature corresponds to a mass-to-charge ratio, retention time, and intensity. This practice may remove a substantial amount of predictive signal. To test the utility of the complete feature set, we train machine learning models for health state-prediction in 35 human metabolomics studies, representing 148 individual data sets. Models trained with all features outperform those using only significant features and frequently provide high predictive performance across nine health state categories, despite disparate experimental and disease contexts. Using only non-significant features it is still often possible to train models and achieve high predictive performance, suggesting useful predictive signal. This work highlights the potential for health state diagnostics using all metabolomics features with data-driven analysis.

Association between Metabolites and the Risk of Lung Cancer: A Systematic Literature Review and Meta-Analysis of Observational Studies

Globally, lung cancer is the most prevalent cancer type. However, screening and early detection is challenging. Previous studies have identified metabolites as promising lung cancer biomarkers. This systematic literature review and meta-analysis aimed to identify metabolites associated with lung cancer risk in observational studies. The literature search was performed in PubMed and EMBASE databases, up to 31 December 2019, for observational studies on the association between metabolites and lung cancer risk. Heterogeneity was assessed using the I2 statistic and Cochran's Q test. Meta-analyses were performed using either a fixed-effects or random-effects model, depending on study heterogeneity. Fifty-three studies with 297 metabolites were included. Most identified metabolites (252 metabolites) were reported in individual studies. Meta-analyses were conducted on 45 metabolites. Five metabolites (cotinine, creatinine riboside, N-acetylneuraminic acid, proline and r-1,t-2,3,c-4-tetrahydroxy-1,2,3,4-tetrahydrophenanthrene) and five metabolite groups (total 3-hydroxycotinine, total cotinine, total nicotine, total 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol (sum of concentrations of the metabolite and its glucuronides), and total nicotine equivalent (sum of total 3-hydroxycotinine, total cotinine and total nicotine)) were associated with higher lung cancer risk, while three others (folate, methionine and tryptophan) were associated with lower lung cancer risk. Significant heterogeneity was detected across most studies. These significant metabolites should be further evaluated as potential biomarkers for lung cancer.

The Expression Pattern of ZIC5 and its Prognostic Value in Lung Cancer

Background: Lung cancer (LC) represents a leading cause of cancer-related deaths. The aim of this study was to analyze the application value of ZIC5 in the prognosis evaluation of LC. Materials and Methods: The mRNA and protein levels of ZIC5 in LC tissues and adjacent normal controls were detected by quantitative real-time polymerase chain reaction and Western blot analysis, respectively. Chi-square test was performed to estimate the association between ZIC5 expression and clinical features. Furthermore, overall survival (OS) curves were plotted using Kaplan-Meier method with log rank test. The prognosis analysis was analyzed using Cox proportional hazards model. Results: Compared with that of noncancerous tissues, the expression of ZIC5 in LC tissues was significantly increased at both protein and mRNA levels (p < 0.001). Moreover, ZIC5 expression showed a positive correlation with differentiation (p = 0.001) and tumor size (p = 0.016). Survival analysis suggested that high expression of ZIC5 predicted shorter OS (log rank test, p = 0.007). ZIC5 might be an independent prognostic biomarker for LC (hazard ratio = 2.892; 95% confidence interval, 1.297-6.449; p = 0.009). Conclusions: ZIC5 expression is upregulated in LC, and it may be employed as a prognostic biomarker for patients with LC.

Early diagnosis and screening in lung cancer

Lung cancer is the third most diagnosed cancer, but the first cause of cancer-related deaths worldwide. This rather high death rate is due mainly to the fact that most patients are diagnosed with advanced-stage cancer, for which the conventional treatment does not work. The most used screening method for lung cancer is a low-dose CT scan, but it is recommended for specific age populations and it also started different debates on its advantages for lung cancer diagnosis. Over the year, several new techniques have been developed that are less invasive, have lower side effect, and can be implemented at all types of populations. This article aimed to present the advantages and disadvantages of using several methods for lung cancer diagnosis, including analysis of volatile organic compounds, exhaled breath condensate analysis and specific genomic approaches.

Using MetaboAnalyst 4.0 for Comprehensive and Integrative Metabolomics Data Analysis

MetaboAnalyst (https://www.metaboanalyst.ca) is an easy-to-use web-based tool suite for comprehensive metabolomic data analysis, interpretation, and integration with other omics data. Since its first release in 2009, MetaboAnalyst has evolved significantly to meet the ever-expanding bioinformatics demands from the rapidly growing metabolomics community. In addition to providing a variety of data processing and normalization procedures, MetaboAnalyst supports a wide array of functions for statistical, functional, as well as data visualization tasks. Some of the most widely used approaches include PCA (principal component analysis), PLS-DA (partial least squares discriminant analysis), clustering analysis and visualization, MSEA (metabolite set enrichment analysis), MetPA (metabolic pathway analysis), biomarker selection via ROC (receiver operating characteristic) curve analysis, as well as time series and power analysis. The current version of MetaboAnalyst (4.0) features a complete overhaul of the user interface and significantly expanded underlying knowledge bases (compound database, pathway libraries, and metabolite sets). Three new modules have been added to support pathway activity prediction directly from mass peaks, biomarker meta-analysis, and network-based multi-omics data integration. To enable more transparent and reproducible analysis of metabolomic data, we have released a companion R package (MetaboAnalystR) to complement the web-based application. This article provides an overview of the main functional modules and the general workflow of MetaboAnalyst 4.0, followed by 12 detailed protocols: © 2019 by John Wiley & Sons, Inc. Basic Protocol 1: Data uploading, processing, and normalization Basic Protocol 2: Identification of significant variables Basic Protocol 3: Multivariate exploratory data analysis Basic Protocol 4: Functional interpretation of metabolomic data Basic Protocol 5: Biomarker analysis based on receiver operating characteristic (ROC) curves Basic Protocol 6: Time-series and two-factor data analysis Basic Protocol 7: Sample size estimation and power analysis Basic Protocol 8: Joint pathway analysis Basic Protocol 9: MS peaks to pathway activities Basic Protocol 10: Biomarker meta-analysis Basic Protocol 11: Knowledge-based network exploration of multi-omics data Basic Protocol 12: MetaboAnalystR introduction.

Advances in lung cancer biomarkers: The role of (metal-) metabolites and selenoproteins

Lung cancer (LC) is the second most common cause of death in men after prostate cancer, and the third most recurrent type of tumor in women after breast and colon cancers. Unfortunately, when LC symptoms begin to appear, the disease is already in an advanced stage and the survival rate only reaches 2%. Thus, there is an urgent need for early diagnosis of LC using specific biomarkers, as well as effective therapies and strategies against LC. On the other hand, the influence of metals on more than 50% of proteins is responsible for their catalytic properties or structure, and their presence in molecules is determined in many cases by the genome. Research has shown that redox metal dysregulation could be the basis for the onset and progression of LC disease. Moreover, metals can interact between them through antagonistic, synergistic and competitive mechanisms, and for this reason metals ratios and correlations in LC should be explored. One of the most studied antagonists against the toxic action of metals is selenium, which plays key roles in medicine, especially related to selenoproteins. The study of potential biomarkers able to diagnose the disease in early stage is conditioned by the development of new analytical methodologies. In this sense, omic methodologies like metallomics, proteomics and metabolomics can greatly assist in the discovery of biomarkers for LC early diagnosis.

On the Use of Correlation and MI as a Measure of Metabolite-Metabolite Association for Network Differential Connectivity Analysis

Metabolite differential connectivity analysis has been successful in investigating potential molecular mechanisms underlying different conditions in biological systems. Correlation and Mutual Information (MI) are two of the most common measures to quantify association and for building metabolite-metabolite association networks and to calculate differential connectivity. In this study, we investigated the performance of correlation and MI to identify significantly differentially connected metabolites. These association measures were compared on (i) 23 publicly available metabolomic data sets and 7 data sets from other fields, (ii) simulated data with known correlation structures, and (iii) data generated using a dynamic metabolic model to simulate real-life observed metabolite concentration profiles. In all cases, we found more differentially connected metabolites when using correlation indices as a measure for association than MI. We also observed that different MI estimation algorithms resulted in difference in performance when applied to data generated using a dynamic model. We concluded that there is no significant benefit in using MI as a replacement for standard Pearson's or Spearman's correlation when the application is to quantify and detect differentially connected metabolites.

A comparative evaluation of the generalised predictive ability of eight machine learning algorithms across ten clinical metabolomics data sets for binary classification

INTRODUCTION

Metabolomics is increasingly being used in the clinical setting for disease diagnosis, prognosis and risk prediction. Machine learning algorithms are particularly important in the construction of multivariate metabolite prediction. Historically, partial least squares (PLS) regression has been the gold standard for binary classification. Nonlinear machine learning methods such as random forests (RF), kernel support vector machines (SVM) and artificial neural networks (ANN) may be more suited to modelling possible nonlinear metabolite covariance, and thus provide better predictive models.

OBJECTIVES

We hypothesise that for binary classification using metabolomics data, non-linear machine learning methods will provide superior generalised predictive ability when compared to linear alternatives, in particular when compared with the current gold standard PLS discriminant analysis.

METHODS

We compared the general predictive performance of eight archetypal machine learning algorithms across ten publicly available clinical metabolomics data sets. The algorithms were implemented in the Python programming language. All code and results have been made publicly available as Jupyter notebooks.

RESULTS

There was only marginal improvement in predictive ability for SVM and ANN over PLS across all data sets. RF performance was comparatively poor. The use of out-of-bag bootstrap confidence intervals provided a measure of uncertainty of model prediction such that the quality of metabolomics data was observed to be a bigger influence on generalised performance than model choice.

CONCLUSION

The size of the data set, and choice of performance metric, had a greater influence on generalised predictive performance than the choice of machine learning algorithm.

Metabolic profiles of serum samples from ground glass opacity represent potential diagnostic biomarkers for lung cancer

Background

Lung cancer is a leading cause of cancer deaths worldwide. Low-dose computed tomography (LDCT) screening trials indicated that LDCT is effective for the early detection of lung cancer, but the findings were accompanied by high false positive rates. Therefore, the detection of lung cancer needs complementary blood biomarker tests to reduce false positive rates.

Methods

In order to evaluate the potential of metabolite biomarkers for diagnosing lung cancer and increasing the effectiveness of clinical interventions, serum samples from subjects participating in a low-dose CT-scan screening were analyzed by using untargeted liquid chromatography-hybrid quadrupole time-of-flight mass spectrometry (LC-Q-TOF-MS). Samples were acquired from 34 lung patients with ground glass opacity diagnosed lung cancer and 39 healthy controls.

Results

In total, we identified 9 metabolites in electron spray ionization (ESI)(+) mode and 7 metabolites in ESI(-) mode. L-(+)-gulose, phosphatidylethanolamine (PE)(22:2(13Z,16Z)/15:0), cysteinyl-glutamine, S-japonin, threoninyl-glutamine, chlorate, 3-oxoadipic acid, dukunolide A, and malonic semialdehyde levels were observed to be elevated in serum samples of lung cancer cases when compared to those of healthy controls. By contrast, 1-(2-furanylmethyl)-1H-pyrrole, 2,4-dihydroxybenzoic acid, monoethyl carbonate, guanidinosuccinic acid, pseudouridine, DIMBOA-Glc, and 4-feruloyl-1,5-quinolactone levels were lower in serum samples of lung cancer cases compared with those of healthy controls.

Conclusions

This study demonstrates evidence of early metabolic alterations that can possibly distinguish malignant ground glass opacity from benign ground glass opacity. Further studies in larger pools of samples are warranted.

Integration of metabolomic and transcriptomic profiles to identify biomarkers in serum of lung cancer

We used blood serum samples collected from 31 lung cancer (LC) patients and 29 healthy volunteers in this study. Levels of serum metabolites were qualitative quantified with gas chromatography-mass spectrometry (GC-MS), and the data were analyzed by partial least-squares discrimination analysis (PLS-DA). Based on the Kyoto Encyclopedia of Genes and Genomes database, we performed pathway-based analysis utilizing metabolites presented at differential abundance between the LC serum samples and the normal healthy serum samples for systematical investigation on the metabolic alterations associated with LC pathogenesis. Finally, we analyzed the significantly enriched pathways as well as their relevant differentially expressed messenger RNAs, and drawn a correlation network plot to identify the serum metabolic biomarkers and the significantly altered metabolic pathways for LC. GC-MS analysis showed that 23 of the 169 metabolites identified were significantly different. PLS-DA model revealed that 13 of these metabolites were with variable importance > 1, and particularly five were with area under curve > 0.9. Pathway-based analysis demonstrated that five of eight enriched metabolic pathways were statistically significant with false discovery rate < 0.05. Lastly, the correlation networks between these pathways and their related genes suggested that 29 genes had correlation degree > 10, which were mainly engaged in the purine metabolism. In conclusion, we identified indole-3-lactate, erythritol, adenosine-5-phosphate, paracetamol and threitol as serum metabolic biomarkers for LC through metabolomics analysis. Besides, we identified the purine metabolism as the significantly altered metabolic pathway in LC with the help of transcriptomics analysis.

Cheminformatics approach to exploring and modeling trait-associated metabolite profiles

Developing predictive and transparent approaches to the analysis of metabolite profiles across patient cohorts is of critical importance for understanding the events that trigger or modulate traits of interest (e.g., disease progression, drug metabolism, chemical risk assessment). However, metabolites’ chemical structures are still rarely used in the statistical modeling workflows that establish these trait-metabolite relationships. Herein, we present a novel cheminformatics-based approach capable of identifying predictive, interpretable, and reproducible trait-metabolite relationships. As a proof-of-concept, we utilize a previously published case study consisting of metabolite profiles from non-small-cell lung cancer (NSCLC) adenocarcinoma patients and healthy controls. By characterizing each structurally annotated metabolite using both computed molecular descriptors and patient metabolite concentration profiles, we show that these complementary features enhance the identification and understanding of key metabolites associated with cancer. Ultimately, we built multi-metabolite classification models for assessing patients’ cancer status using specific groups of metabolites identified based on high structural similarity through chemical clustering. We subsequently performed a metabolic pathway enrichment analysis to identify potential mechanistic relationships between metabolites and NSCLC adenocarcinoma. This cheminformatics-inspired approach relies on the metabolites’ structural features and chemical properties to provide critical information about metabolite-trait associations. This method could ultimately facilitate biological understanding and advance research based on metabolomics data, especially with respect to the identification of novel biomarkers.

RNA sequencing and swarm intelligence-enhanced classification algorithm development for blood-based disease diagnostics using spliced blood platelet RNA

Blood-based diagnostics tests, using individual or panels of biomarkers, may revolutionize disease diagnostics and enable minimally invasive therapy monitoring. However, selection of the most relevant biomarkers from liquid biosources remains an immense challenge. We recently presented the thromboSeq pipeline, which enables RNA sequencing and cancer classification via self-learning and swarm intelligence-enhanced bioinformatics algorithms using blood platelet RNA. Here, we provide the wet-lab protocol for the generation of platelet RNA-sequencing libraries and the dry-lab protocol for the development of swarm intelligence-enhanced machine-learning-based classification algorithms. The wet-lab protocol includes platelet RNA isolation, mRNA amplification, and preparation for next-generation sequencing. The dry-lab protocol describes the automated FASTQ file pre-processing to quantified gene counts, quality controls, data normalization and correction, and swarm intelligence-enhanced support vector machine (SVM) algorithm development. This protocol enables platelet RNA profiling from 500 pg of platelet RNA and allows automated and optimized biomarker panel selection. The wet-lab protocol can be performed in 5 d before sequencing, and the algorithm development can be completed in 2 d, depending on computational resources. The protocol requires basic molecular biology skills and a basic understanding of Linux and R. In all, with this protocol, we aim to enable the scientific community to test platelet RNA for diagnostic algorithm development.

A postural unloading task to assess fast corrective responses in the upper limb following stroke

Background

Robotic technologies to measure human behavior are emerging as a new approach to assess brain function. Recently, we developed a robot-based postural Load Task to assess corrective responses to mechanical disturbances to the arm and found impairments in many participants with stroke compared to a healthy cohort (Bourke et al, J NeuroEngineering Rehabil 12: 7, 2015). However, a striking feature was the large range and skewed distribution of healthy performance. This likely reflects the use of different strategies across the healthy control sample, making it difficult to identify impairments. Here, we developed an intuitive “Unload Task”. We hypothesized this task would reduce healthy performance variability and improve the detection of impairment following stroke.

Methods

Performance on the Load and Unload Task in the KINARM exoskeleton robot was directly compared for healthy control (n = 107) and stroke (n = 31) participants. The goal was to keep a cursor representing the hand inside a virtual target and return “quickly and accurately” if the robot applied (or removed) an unexpected load to the arm (0.5–1.5 Nm). Several kinematic parameters quantified performance. Impairment was defined as performance outside the 95% of controls (corrected for age, sex and handedness). Task Scores were calculated using standardized parameter scores reflecting overall task performance.

Results

The distribution of healthy control performance was smaller and less skewed for the Unload Task compared to the Load Task. Fewer task outliers (outside 99.9 percentile for controls) were removed from the Unload Task (3.7%) compared to the Load Task (7.4%) when developing normative models of performance. Critically, more participants with stroke failed the Unload Task based on Task Score with their affected arm (68%) compared to the Load Task (23%). More impairments were found at the parameter level for the Unload (median = 52%) compared to Load Task (median = 29%).

Conclusions

The Unload Task provides an improved approach to assess fast corrective responses of the arm. We found that corrective responses are impaired in persons living with stroke, often equally in both arms. Impairments in generating rapid motor corrections may place individuals at greater risk of falls when they move and interact in the environment.

Study of the metabolomic relationship between lung cancer and chronic obstructive pulmonary disease based on direct infusion mass spectrometry

The high prevalence of lung cancer (LC) has triggered the search of biomarkers for early diagnosis of this disease. For this purpose the study of metabolic changes related to the development of lung cancer could provide interesting information about its early diagnosis. In this sense, chronic obstructive pulmonary disease (COPD), a disease associated with tumor development, is a comorbidity that increases the risk of onset and progression of lung neoplasia and has also to be considered in the study of pathology related to lung cancer. This work develop a metabolomic approach based on direct infusion mass spectrometry using a hybrid triple quadrupole-time of flight mass spectrometer (DI-ESI-QqQ-TOF-MS) in order to identify altered metabolites from serum of LC and COPD patients and evaluate its relationship and implication in the progression of LC. This methodology has been applied to 30 serum samples from LC, 30 healthy patients used as controls (HC) and 30 serum samples from COPD to found altered metabolites from both LC and COPD diseases. In addition, some metabolic differences and similarities were found in Pulmonary Emphysema and Chronic Bronchitis patients. On the other hand, altered metabolites were studied in different stages of LC (II, III and IV) to evaluate the perturbation of them throughout the progression of disease. The sample treatment consisted of the extraction of polar and non-polar metabolites from serum that was later infused into the mass spectrometer using an electrospray ionization source in positive and negative mode. Partial least squares discriminant analysis (PLS-DA) allowed a classification between LC, HC and COPD groups in all acquisition modes. A total of 35 altered and common metabolites between LC and COPD, including amino acids, fatty acids, lysophospholipids, phospholipids and triacylglycerides were identified, being alanine, aspartate and glutamate metabolism the most altered. Finally, ROC curves were applied to the dataset and metabolites with AUC value higher than 0.70 were considered as relevant in the progression of LC.

Metabolome-based biomarkers: their potential role in the early detection of lung cancer

Lung cancer is the leading cause of cancer-related death worldwide, and a major problem affecting its mortality is the late diagnosis of the majority of cases, where treatment options are limited and overall prognosis is very bad. Currently, a low-dose computed tomography (LD-CT) screening in the high-risk group is the only available diagnostic strategy that could reduce mortality due to this malignancy. However, the LD-CT screening test suffers from a high false positive rate. Hence, complementation of LD-CT examination with blood-based biomarkers is a rational approach to increase efficacy and reduce the cost of early lung cancer screening programs. Several molecular signatures that discriminate between patients with early lung cancer and healthy individuals have been proposed in recent years, which are based on components of serum/plasma metabolome. However, none of these signatures has been validated by independent studies based on material collected during real lung cancer screening. Therefore, the validation of the real diagnostic value of these otherwise promising candidates remains a critical step in this challenging field of cancer diagnostics.

Blood based biomarkers beyond genomics for lung cancer screening

While there is considerable interest at the present time in the development of so-called liquid biopsy approaches for cancer detection based notably on circulating tumor DNA, there are other types of potential biomarkers that show promise for lung cancer screening and early detection. Here we review approaches and some of the promising markers based on proteomics, metabolomics and the immune response to tumor antigens in the form of autoantibodies.

Simultaneous quantification of serum monounsaturated and polyunsaturated phosphatidylcholines as potential biomarkers for diagnosing non-small cell lung cancer

Non-small cell lung cancer (NSCLC) is one of the most common malignancies worldwide. In this study, we investigated Ultrahigh Performance Liquid Chromatography-Quadrupole Time-of-Flight Mass Spectrometry and Gas Chromatography Time-of-Flight/Mass Spectrometry-based non-targeted metabolomic profiles of serum samples obtained from early-stage NSCLC patients and healthy controls (HC). Metabolic pathways and the biological relevance of potential biomarkers were extensively studied to gain insights into dysregulated metabolism in NSCLC. The identified biomarker candidates were further externally validated via a targeted metabolomics analysis. The global metabolomics profiles could clearly distinguish NSCLC patients from HC. Phosphatidylcholine (PC) levels were found to be dysregulated in glycerophospholipid (GPL) metabolism, which was the top altered pathway in early-stage NSCLC. Compared with those in HC, significant increases in the levels of saturated and monounsaturated PCs such as PC (15:0/18:1), PC (18:0/16:0) and PC (18:0/20:1) were observed in NSCLC. Additionally, relative to those in HC, the levels of 9 polyunsaturated PCs, namely, PC (17:2/2:0), PC (18:4/3:0), and PC (15:0/18:2), and so on were significantly decreased in NSCLC patients. A panel of 12 altered PCs had good diagnostic performance in differentiating early-stage NSCLC patients from HC, and these PCs may thus be used as serum biomarkers for the early diagnosis of NSCLC.

High circulating activin A level is associated with tumor progression and predicts poor prognosis in lung adenocarcinoma

Activin A (ActA)/follistatin (FST) signaling has been shown to be deregulated in different tumor types including lung adenocarcinoma (LADC). Here, we report that serum ActA protein levels are significantly elevated in LADC patients (n=64) as compared to controls (n=46, p=0.015). ActA levels also correlated with more advanced disease stage (p<0.0001) and T (p=0.0035) and N (p=0.0002) factors. M1 patients had significantly higher ActA levels than M0 patients (p<0.001). High serum ActA level was associated with poor overall survival (p<0.0001) and was confirmed as an independent prognostic factor (p=0.004). Serum FST levels were increased only in female LADC patients (vs. female controls, p=0.031). Two out of five LADC cell lines secreted biologically active ActA, while FST was produced in all of them. Transcripts of both type I and II ActA receptors were detected in all five LADC cell lines. In conclusion, our study does not only suggest that measuring blood ActA levels in LADC patients might improve the prediction of prognosis, but also indicates that this parameter might be a novel non-invasive biomarker for identifying LADC patients with organ metastases.

Next-generation metabolomics in lung cancer diagnosis, treatment and precision medicine: mini review

Lung cancer is the leading cause of cancer-related death. Next-generation metabolomics is becoming a powerful emerging technology for studying the systems biology and chemistry of health and disease. This mini review summarized the main platforms of next-generation metabolomics and its main applications in lung cancer including early diagnosis, pathogenesis, classifications and precision medicine. The period covers between 2009 and August, 2017. The major issues and future directions of metabolomics in lung cancer research and clinical applications were also discussed.

Evaluation of Classifier Performance for Multiclass Phenotype Discrimination in Untargeted Metabolomics

Statistical classification is a critical component of utilizing metabolomics data for examining the molecular determinants of phenotypes. Despite this, a comprehensive and rigorous evaluation of the accuracy of classification techniques for phenotype discrimination given metabolomics data has not been conducted. We conducted such an evaluation using both simulated and real metabolomics datasets, comparing Partial Least Squares-Discriminant Analysis (PLS-DA), Sparse PLS-DA, Random Forests, Support Vector Machines (SVM), Artificial Neural Network, k-Nearest Neighbors (k-NN), and Naïve Bayes classification techniques for discrimination. We evaluated the techniques on simulated data generated to mimic global untargeted metabolomics data by incorporating realistic block-wise correlation and partial correlation structures for mimicking the correlations and metabolite clustering generated by biological processes. Over the simulation studies, covariance structures, means, and effect sizes were stochastically varied to provide consistent estimates of classifier performance over a wide range of possible scenarios. The effects of the presence of non-normal error distributions, the introduction of biological and technical outliers, unbalanced phenotype allocation, missing values due to abundances below a limit of detection, and the effect of prior-significance filtering (dimension reduction) were evaluated via simulation. In each simulation, classifier parameters, such as the number of hidden nodes in a Neural Network, were optimized by cross-validation to minimize the probability of detecting spurious results due to poorly tuned classifiers. Classifier performance was then evaluated using real metabolomics datasets of varying sample medium, sample size, and experimental design. We report that in the most realistic simulation studies that incorporated non-normal error distributions, unbalanced phenotype allocation, outliers, missing values, and dimension reduction, classifier performance (least to greatest error) was ranked as follows: SVM, Random Forest, Naïve Bayes, sPLS-DA, Neural Networks, PLS-DA and k-NN classifiers. When non-normal error distributions were introduced, the performance of PLS-DA and k-NN classifiers deteriorated further relative to the remaining techniques. Over the real datasets, a trend of better performance of SVM and Random Forest classifier performance was observed.

Integrated Metabolomics and Proteomics Highlight Altered Nicotinamide- and Polyamine Pathways in Lung Adenocarcinoma

Lung cancer is the leading cause of cancer mortality in the United States with non-small cell lung cancer (NSCLC) adenocarcinoma being the most common histological type. Early perturbations in cellular metabolism are a hallmark of cancer, but the extent of these changes in early stage lung adenocarcinoma remains largely unknown. In the current study, an integrated metabolomics and proteomics approach was utilized to characterize the biochemical and molecular alterations between malignant and matched control tissue from 27 subjects diagnosed with early stage lung adenocarcinoma. Differential analysis identified 71 metabolites and 1102 proteins that delineated tumor from control tissue. Integrated results indicated four major metabolic changes in early stage adenocarcinoma: (1) increased glycosylation and glutaminolysis; (2) elevated Nrf2 activation; (3) increase in nicotinic and nicotinamide salvaging pathways; and (4) elevated polyamine biosynthesis linked to differential regulation of the SAM/nicotinamide methyl-donor pathway. Genomic data from publicly available databases were included to strengthen proteomic findings. Our findings provide insight into the biochemical and molecular biological reprogramming that may accompanies early stage lung tumorigenesis and highlight potential therapeutic targets.