Search

The discovery of novel therapeutic targets, defined as proteins which drugs can interact with to induce therapeutic benefits, typically represent the first and most important step of drug discovery. One solution for target discovery is target repositioning, a strategy which relies on the repurposing of known targets for new diseases, leading to new treatments, less side effects and potential drug synergies. Biological networks have emerged as powerful tools for integrating heterogeneous data and facilitating the prediction of biological or therapeutic properties. Consequently, they are widely employed to predict new therapeutic targets by characterizing potential candidates, often based on their interactions within a Protein-Protein Interaction (PPI) network, and their proximity to genes associated with the disease. However, over-reliance on PPI networks and the assumption that potential targets are necessarily near known genes can introduce biases that may limit the effectiveness of these methods. This study addresses these limitations in two ways. First, by exploiting a multi-layer network which incorporates additional information such as gene regulation, metabolite interactions, metabolic pathways, and several disease signatures such as Differentially Expressed Genes, mutated genes, Copy Number Alteration, and structural variants. Second, by extracting relevant features from the network using several approaches including proximity to disease-associated genes, but also unbiased approaches such as propagation-based methods, topological metrics, and module detection algorithms. Using prostate cancer as a case study, the best features were identified and utilized to train machine learning algorithms to predict 5 novel promising therapeutic targets for prostate cancer: IGF2R, C5AR, RAB7, SETD2 and NPBWR1.

IntroductionMagnetic resonance imaging (MRI) research has advanced our understanding of neurodegeneration in sporadic early-onset Alzheimer's disease (EOAD) but studies include small samples, mostly amnestic EOAD, and have not focused on developing an MRI biomarker.MethodsWe analyzed MRI scans to define the sporadic EOAD-signature atrophy in a small sample (n = 25) of Massachusetts General Hospital (MGH) EOAD patients, investigated its reproducibility in the large longitudinal early-onset Alzheimer's disease study (LEADS) sample (n = 211), and investigated the relationship of the magnitude of atrophy with cognitive impairment.ResultsThe EOAD-signature atrophy was replicated across the two cohorts, with prominent atrophy in the caudal lateral temporal cortex, inferior parietal lobule, and posterior cingulate and precuneus cortices, and with relative sparing of the medial temporal lobe. The magnitude of EOAD-signature atrophy was associated with the severity of cognitive impairment.DiscussionThe EOAD-signature atrophy is a reliable and clinically valid biomarker of AD-related neurodegeneration that could be used in clinical trials for EOAD.HighlightsWe developed an early-onset Alzheimer's disease (EOAD)-signature of atrophy based on magnetic resonance imaging (MRI) scans. EOAD signature was robustly reproducible across two independent patient cohorts. EOAD signature included prominent atrophy in parietal and posterior temporal cortex. The EOAD-signature atrophy was associated with the severity of cognitive impairment. EOAD signature is a reliable and clinically valid biomarker of neurodegeneration.

(1) Background: Barrett's esophagus is a major risk factor for esophageal adenocarcinoma. In this pilot study, we employed precision mass spectrometry to map global (phospho)protein perturbations in Barrett's esophagus lesions and adjacent normal tissue to glean insights into disease progression. (2) Methods: Biopsies were collected from two small but independent cohorts. Comparative analyses were performed between Barrett's esophagus samples and adjacent matched (normal) tissues from patients with known pathology, while specimens from healthy patients served as additional controls. (3) Results: We identified and quantified 6810 proteins and 6395 phosphosites in the discovery cohort, revealing hundreds of statistically significant differences in protein abundances and phosphorylation states. We identified a robust proteomic signature that accurately classified the disease status of samples from the independent patient cohorts. Pathway-level analysis of the phosphoproteomic profiles revealed the dysregulation of specific cellular processes, including DNA repair, in Barrett's esophagus relative to paired controls. Comparative analysis with previously published transcriptomic profiles provided independent evidence in support of these preliminary findings. (4) Conclusions: This pilot study establishes the feasibility of using unbiased quantitative phosphoproteomics to identify molecular perturbations associated with disease progression in Barrett's esophagus to define potentially clinically actionable targets warranting further assessment.

Abstract Background The S209F variant of Abelson Interactor Protein 3 (ABI3) increases risk for Alzheimer’s disease (AD), but little is known about its function in relation to AD pathogenesis. Methods Here, we use a mouse model that is deficient in Abi3 locus to study how the loss of function of Abi3 impacts two cardinal neuropathological hallmarks of AD—amyloid β plaques and tau pathology. Our study employs extensive neuropathological and transcriptomic characterization using transgenic mouse models and adeno-associated virus-mediated gene targeting strategies. Results Analysis of bulk RNAseq data confirmed age-progressive increase in Abi3 levels in rodent models of AD-type amyloidosis and upregulation in AD patients relative to healthy controls. Using RNAscope in situ hybridization, we localized the cellular distribution of Abi3 in mouse and human brains, finding that Abi3 is expressed in both microglial and non-microglial cells. Next, we evaluated Abi3 −/− mice and document that both Abi3 and its overlapping gene, Gngt2, are disrupted in these mice. Using multiple transcriptomic datasets, we show that expression of Abi3 and Gngt2 are tightly correlated in rodent models of AD and human brains, suggesting a tight co-expression relationship. RNAseq of the Abi3-Gngt2−/− mice revealed upregulation of Trem2, Plcg2, and Tyrobp, concomitant with induction of an AD-associated neurodegenerative signature, even in the absence of AD-typical neuropathology. In APP mice, loss of Abi3-Gngt2 resulted in a gene dose- and age-dependent reduction in Aβ deposition. Additionally, in Abi3-Gngt2 −/− mice, expression of a pro-aggregant form of human tau exacerbated tauopathy and astrocytosis. Further, using in vitro culture assays, we show that the AD-associated S209F mutation alters the extent of ABI3 phosphorylation. Conclusions These data provide an important experimental framework for understanding the role of Abi3-Gngt2 function and early inflammatory gliosis in AD. Our studies also demonstrate that inflammatory gliosis could have opposing effects on amyloid and tau pathology, highlighting the unpredictability of targeting immune pathways in AD.

Background: The S209F variant of Abelson Interactor Protein 3 (ABI3) increases risk for Alzheimer's disease (AD), but little is known about its function in relation to AD pathogenesis. Methods: Here, we use a mouse model that is deficient in Abi3 locus to study how the loss of function of Abi3 impacts two cardinal neuropathological hallmarks of AD—amyloid β plaques and tau pathology. Our study employs extensive neuropathological and transcriptomic characterization using transgenic mouse models and adeno-associated virus-mediated gene targeting strategies. Results: Analysis of bulk RNAseq data confirmed age-progressive increase in Abi3 levels in rodent models of AD-type amyloidosis and upregulation in AD patients relative to healthy controls. Using RNAscope in situ hybridization, we localized the cellular distribution of Abi3 in mouse and human brains, finding that Abi3 is expressed in both microglial and non-microglial cells. Next, we evaluated Abi3−/− mice and document that both Abi3 and its overlapping gene, Gngt2, are disrupted in these mice. Using multiple transcriptomic datasets, we show that expression of Abi3 and Gngt2 are tightly correlated in rodent models of AD and human brains, suggesting a tight co-expression relationship. RNAseq of the Abi3-Gngt2−/− mice revealed upregulation of Trem2, Plcg2, and Tyrobp, concomitant with induction of an AD-associated neurodegenerative signature, even in the absence of AD-typical neuropathology. In APP mice, loss of Abi3-Gngt2 resulted in a gene dose- and age-dependent reduction in Aβ deposition. Additionally, in Abi3-Gngt2−/− mice, expression of a pro-aggregant form of human tau exacerbated tauopathy and astrocytosis. Further, using in vitro culture assays, we show that the AD-associated S209F mutation alters the extent of ABI3 phosphorylation. Conclusions: These data provide an important experimental framework for understanding the role of Abi3-Gngt2 function and early inflammatory gliosis in AD. Our studies also demonstrate that inflammatory gliosis could have opposing effects on amyloid and tau pathology, highlighting the unpredictability of targeting immune pathways in AD. [ABSTRACT FROM AUTHOR]

Background: Lymph node metastasis (LNM) of lung cancer is an important factor associated with prognosis. Dysregulated microRNAs (miRNAs) are becoming a new powerful tool to characterize tumorigenesis and metastasis. We have developed and validated a miRNA disease signature to predict LNM in lung adenocarcinoma (LUAD). Method: LUAD miRNAs and clinical data from The Cancer Genome Atlas (TCGA) were obtained and divided randomly into training (n = 259) and validation (n = 83) cohorts. A miRNA signature was built using least absolute shrinkage and selection operator (LASSO) (λ = -1.268) and logistic regression model. The performance of the miRNA signature was evaluated using the area under curve (AUC) of receiver operating characteristic curve (ROC). We performed decision curve analysis (DCA) to assess the clinical usefulness of the signature. We also conducted a miRNA-regulatory network analysis to look for potential genes engaged in LNM in LUAD. Result: Thirteen miRNAs were selected to build our miRNA disease signature. The model showed good calibration in the training cohort, with an AUC of 0.782 (95% CI: 0.725-0.839). In the validation cohort, AUC was 0.691 (95% CI: 0.575-0.806). DCA demonstrated that the miRNA signature was clinically useful. Conclusion: The miRNA disease signature can be used as a noninvasive method to predict LNM in patients with lung adenocarcinoma objectively and the signature achieved high accuracy for prediction.

Introduction: With the shift of research focus to personalized medicine in Alzheimer's Dementia (AD), there is an urgent need for tools that are capable of quantifying a patient's risk using diagnostic biomarkers. The Medical Informatics Platform (MIP) is a distributed e-infrastructure federating large amounts of data coupled with machine-learning (ML) algorithms and statistical models to define the biological signature of the disease. The present study assessed (i) the accuracy of two ML algorithms, i.e., supervised Gradient Boosting (GB) and semi-unsupervised 3C strategy (Categorize, Cluster, Classify—CCC) implemented in the MIP and (ii) their contribution over the standard diagnostic workup.Methods: We examined individuals coming from the MIP installed across 3 Italian memory clinics, including subjects with Normal Cognition (CN, n = 432), Mild Cognitive Impairment (MCI, n = 456), and AD (n = 451). The GB classifier was applied to best discriminate the three diagnostic classes in 1,339 subjects, and the CCC strategy was used to refine the classical disease categories. Four dementia experts provided their diagnostic confidence (DC) of MCI conversion on an independent cohort of 38 patients. DC was based on clinical, neuropsychological, CSF, and structural MRI information and again with addition of the outcome from the MIP tools.Results: The GB algorithm provided a classification accuracy of 85% in a nested 10-fold cross-validation for CN vs. MCI vs. AD discrimination. Accuracy increased to 95% in the holdout validation, with the omission of each Italian clinical cohort out in turn. CCC identified five homogeneous clusters of subjects and 36 biomarkers that represented the disease fingerprint. In the DC assessment, CCC defined six clusters in the MCI population used to train the algorithm and 29 biomarkers to improve patients staging. GB and CCC showed a significant impact, evaluated as +5.99% of increment on physicians' DC. The influence of MIP on DC was rated from “slight” to “significant” in 80% of the cases.Discussion: GB provided fair results in classification of CN, MCI, and AD. CCC identified homogeneous and promising classes of subjects via its semi-unsupervised approach. We measured the effect of the MIP on the physician's DC. Our results pave the way for the establishment of a new paradigm for ML discrimination of patients who will or will not convert to AD, a clinical priority for neurology.

Introduction: With the shift of research focus to personalized medicine in Alzheimer's Dementia (AD), there is an urgent need for tools that are capable of quantifying a patient's risk using diagnostic biomarkers. The Medical Informatics Platform (MIP) is a distributed e-infrastructure federating large amounts of data coupled with machine-learning (ML) algorithms and statistical models to define the biological signature of the disease. The present study assessed (i) the accuracy of two ML algorithms, i.e., supervised Gradient Boosting (GB) and semi-unsupervised 3C strategy (Categorize, Cluster, Classify—CCC) implemented in the MIP and (ii) their contribution over the standard diagnostic workup. Methods: We examined individuals coming from the MIP installed across 3 Italian memory clinics, including subjects with Normal Cognition (CN, n = 432), Mild Cognitive Impairment (MCI, n = 456), and AD (n = 451). The GB classifier was applied to best discriminate the three diagnostic classes in 1,339 subjects, and the CCC strategy was used to refine the classical disease categories. Four dementia experts provided their diagnostic confidence (DC) of MCI conversion on an independent cohort of 38 patients. DC was based on clinical, neuropsychological, CSF, and structural MRI information and again with addition of the outcome from the MIP tools. Results: The GB algorithm provided a classification accuracy of 85% in a nested 10-fold cross-validation for CN vs. MCI vs. AD discrimination. Accuracy increased to 95% in the holdout validation, with the omission of each Italian clinical cohort out in turn. CCC identified five homogeneous clusters of subjects and 36 biomarkers that represented the disease fingerprint. In the DC assessment, CCC defined six clusters in the MCI population used to train the algorithm and 29 biomarkers to improve patients staging. GB and CCC showed a significant impact, evaluated as +5.99% of increment on physicians' DC. The influence of MIP on DC was rated from "slight" to "significant" in 80% of the cases. Discussion: GB provided fair results in classification of CN, MCI, and AD. CCC identified homogeneous and promising classes of subjects via its semi-unsupervised approach. We measured the effect of the MIP on the physician's DC. Our results pave the way for the establishment of a new paradigm for ML discrimination of patients who will or will not convert to AD, a clinical priority for neurology. [ABSTRACT FROM AUTHOR]

Hypertrophic Cardiomyopathy (HCM) is a common inherited heart disease with poor risk prediction due to incomplete penetrance and a lack of clear genotype–phenotype correlations. Advanced imaging techniques have shown altered myocardial energetics already in preclinical gene variant carriers. To determine whether disturbed myocardial energetics with the potential to serve as biomarkers are also reflected in the serum metabolome, we analyzed the serum metabolome of asymptomatic carriers in comparison to healthy controls and obstructive HCM patients (HOCM). We performed non-quantitative direct-infusion high-resolution mass spectrometry-based untargeted metabolomics on serum from fasted asymptomatic gene variant carriers, symptomatic HOCM patients and healthy controls (n = 31, 14 and 9, respectively). Biomarker panels that discriminated the groups were identified by performing multivariate modeling with gradient-boosting classifiers. For all three group-wise comparisons we identified a panel of 30 serum metabolites that best discriminated the groups. These metabolite panels performed equally well as advanced cardiac imaging modalities in distinguishing the groups. Seven metabolites were found to be predictive in two different comparisons and may play an important role in defining the disease stage. This study reveals unique metabolic signatures in serum of preclinical carriers and HOCM patients that may potentially be used for HCM risk stratification and precision therapeutics.

Identifying disease signatures in order to facilitate accurate diagnosis/treatment has been the focus of research efforts in the last decade. However, the term "disease signature" has not been properly defined, resulting in inconsistencies between studies, as well as limited ability to fully utilize the tools/information available in the evolving field of healthcare big data. Research was conducted according to the Preferred Reporting Items for Systematic Review and Meta-analyses (PRISMA) guidelines. The search (in PubMed, Cochrane, and Web of Science) was limited to English articles published up to 31/12/2016. The search string was "disease signature" OR "disease signatures" OR "disease fingerprint" OR "disease fingerprints" OR "subtype signature" OR "subtype signatures" OR "subgroup signature" OR "subgroup signatures." The full text of the articles was reviewed to determine the meaning of the phrase "disease signature" as well as the context of its use. Of 285 articles identified in the search, 129 were included in the final analysis. The term disease signature was first found in an article from 2001. In the last 10 years, the use of the term increased by approximately ninefold, which is double the general increase in the number of published articles. Only one article attempted to define the term. The two major medical fields where the term was used were oncology (31%) and neurology (20%); 71% of the identified articles used a single biomarker to define the term, 13% of the articles used a pair of biomarkers, and 16% used signatures with multiple biomarker; in 42% of the identified articles, genomic biomarkers were used for the signature, in 17% measurements of biochemical compounds in body fluids, and in 10%, changes in imaging studies were used for the signature. Our findings identified a lack of consistency in defining the term disease signature. We suggest a novel hierarchical multidimensional concept for this term that would combine both current approaches for identifying diseases (one focusing on undesired effects of the disease and the other on its causes). This model can improve disease signature definition consistency which will enable to generalize and classify diseases, resulting in more precise treatments and better outcomes. Ultimately, this model could lead to developing a statistical confidence in a disease signature that would allow physicians/patients to estimate the precision of the diagnosis, which, in turn, may have important implications on patients' prognosis and treatment. [ABSTRACT FROM AUTHOR]

• The molecular signature of severe COVID-19 in lung autopsies was stablished. • Six transcription factors were found as Master Regulators for COVID-19 severe form. • 52 drugs were proposed as reversers of the severe COVID-19 molecular signature. • A limited overlap between clinical and pre-clinical COVID-19 data was identified. Coronavirus disease 2019 (COVID-19) is an acute viral disease with millions of cases worldwide. Although the number of daily new cases and deaths has been dropping, there is still a need for therapeutic alternatives to deal with severe cases. A promising strategy to prospect new therapeutic candidates is to investigate the regulatory mechanisms involved in COVID-19 progression using integrated transcriptomics approaches. In this work, we aimed to identify COVID-19 Master Regulators (MRs) using a series of publicly available gene expression datasets of lung tissue from patients which developed the severe form of the disease. We were able to identify a set of six potential COVID-19 MRs related to its severe form, namely TAL1, TEAD4, EPAS1, ATOH8, ERG , and ARNTL2. In addition, using the Connectivity Map drug repositioning approach, we identified 52 different drugs which could be used to revert the disease signature, thus being candidates for the design of novel clinical treatments. Furthermore, we compared the identified signature and drugs with the ones obtained from the analysis of nasopharyngeal swab samples from infected patients and preclinical cell models. This comparison showed significant similarities between them, although also revealing some limitations on the overlap between clinical and preclinical data in COVID-19, highlighting the need for careful selection of the best model for each disease stage. [ABSTRACT FROM AUTHOR]

Extensive surface profiles of colorectal cancer (CRC) cells and tumor infiltrating lymphocytes (TIL) have been obtained from 45 surgical resection samples. Live cells were captured on an antibody microarray and stained with fluorescently-labeled antibodies. Minimal panels of 11 CRC antigens (CD13, CD24, CD26, CD49d, CD138, CD166, CA-125, CA19-9, EGFR, Galectin-4 and HLA-DR) and 11 T-cell antigens (CD10, CD11b, CD11c, CD25, CD31, CD95, CD151, CD181, Galectin-4, CA19-9, TSP-1) provide signatures for relapse and survival. Hierarchical clustering of profiles from CRC cells and TIL identified groups of patients for survival, systemic relapse and death. The groups from CRC and TIL profiles for systemic relapse showed 79.2% concordance, enabling prediction of relapse after surgery. The results demonstrate communication between CRC cells and TIL. [ABSTRACT FROM AUTHOR]

Objective: Developmental dyscalculia (DD) is a specific learning disability characterized by a deficiency in math learning. It is suggested that dyscalculia might be caused by a connectivity deficit in the brain. The main purpose of our study was to provide a method to assist the diagnosis of dyscalculia using effective connectivity modeled by discrete Dynamic Bayesian Network (dDBN). Methods: dDBN was modeled using the functional magnetic resonance imaging (fMRI) data obtained from twelve children with dyscalculia (8 female, mean age: 11.25) and fifteen typically developing children (9 female, mean age: 11.26) volunteers while they were performing a numerosity task. The approach consisted of three steps. Firstly, imaging data were analyzed by repeated-measures ANOVA with group (dyscalculic/control), number modality (symbolic/non-symbolic tasks), and difficulty (0.7/0.5 ratios) as factors, using SPM12 software. Significantly activated voxel clusters related to the main effect of group were selected as the region of interests (ROI). There were six ROIs: the left and right intraparietal sulci, right anterior cingulate cortex, left medial prefrontal cortex, left hippocampus, and occipital cortex. Secondly, dDBN was used to model the brain connectivity using the fMRI time series data of these ROIs. The effective connectivity is modeled for each task by using two methods, which were virtual-typical-subject and individual-structure methods. Finally, the Bayesian score was used to classify dyscalculic participants from control participants for each method separately. Results: A prefrontal-parietal and hippocampal network in the left hemisphere shows effective connectivity in dyscalculia, which might be the result of their brain's compensation mechanisms. In addition, when the connectivity maps of the symbolic task were used, the classification performance reached 70% accuracy. Conclusion: Hyper-connectivity was observed in children with DD. The significant connectivity differences are mainly reported in the left hemisphere for the symbolic task. This study was supported within the scope of TUBITAK project number 214S069. [ABSTRACT FROM AUTHOR]

Hypertrophic Cardiomyopathy (HCM) is a common inherited heart disease with poor risk prediction due to incomplete penetrance and a lack of clear genotype–phenotype correlations. Advanced imaging techniques have shown altered myocardial energetics already in preclinical gene variant carriers. To determine whether disturbed myocardial energetics with the potential to serve as biomarkers are also reflected in the serum metabolome, we analyzed the serum metabolome of asymptomatic carriers in comparison to healthy controls and obstructive HCM patients (HOCM). We performed non-quantitative direct-infusion high-resolution mass spectrometry-based untargeted metabolomics on serum from fasted asymptomatic gene variant carriers, symptomatic HOCM patients and healthy controls (n = 31, 14 and 9, respectively). Biomarker panels that discriminated the groups were identified by performing multivariate modeling with gradient-boosting classifiers. For all three group-wise comparisons we identified a panel of 30 serum metabolites that best discriminated the groups. These metabolite panels performed equally well as advanced cardiac imaging modalities in distinguishing the groups. Seven metabolites were found to be predictive in two different comparisons and may play an important role in defining the disease stage. This study reveals unique metabolic signatures in serum of preclinical carriers and HOCM patients that may potentially be used for HCM risk stratification and precision therapeutics. [ABSTRACT FROM AUTHOR]

Today, the introduction of metabolomics, like other omics sciences, into clinical practice as a personal omics test that realizes the perfect analytical capabilities of this science has become an important subject. The assembled data show that the metabolome of biosamples is a collection of highly informative and accurate signatures of virtually all diseases that are widespread in the population. However, we have not seen the emergence of personalized metabolomics in clinical practice. This article analyzes the causes of this problem. The complexity of personal metabolic data analysis and its incompatibility with widely accepted data treatment in metabolomics are shown. As a result, the impossibility of translating metabolic signatures accumulated in databases into a personal test is revealed. Problem-solving strategies that may radically change the situation and realize the analytical capabilities of metabolomics in medical laboratory practice are discussed. [ABSTRACT FROM AUTHOR]

NMR measures a relatively small portion of the human metabolome, which roughly corresponds to all free small molecules present in concentrations ≥1 μM. However, the method is intrinsically quantitative and highly reproducible and, whenever standardized pre-analytical and analytical procedures are used, it allows for fast fingerprinting and profiling of a variety of biosamples. While developing methods for accurate and automated identification and quantitation of detectable metabolites, we have proposed a Urine Shift Predictor that also provides the concentration of NMR-invisible inorganic ions. For biomedical applications, two peculiar features are associated to the fast, untargeted NMR-metabolomic fingerprinting of biofluids, i.e. the ability to: i) identify the individual phenotype that constitutes the metabolic signature of a person and monitor its behavior over time; ii) identify the signature of different diseases. The combination of the two in longitudinal cohort studies could become a tool for fast and untargeted screening of populations. • NMR permits fast fingerprinting and profiling of a large variety of biosamples. • It measures metabolites >1 μM but can predict concentrations of invisible ions. • Each healthy individual has a well-defined and stable metabolic phenotype. • 1H NMR spectra contain the signature of diseases. • Untargeted NMR metabolomics is suitable for population screening. [ABSTRACT FROM AUTHOR]

Background: Lymph node metastasis (LNM) of lung cancer is an important factor associated with prognosis. Dysregulated microRNAs (miRNAs) are becoming a new powerful tool to characterize tumorigenesis and metastasis. We have developed and validated a miRNA disease signature to predict LNM in lung adenocarcinoma (LUAD). Method: LUAD miRNAs and clinical data from The Cancer Genome Atlas (TCGA) were obtained and divided randomly into training (n = 259) and validation (n = 83) cohorts. A miRNA signature was built using least absolute shrinkage and selection operator (LASSO) (λ =-1.268) and logistic regression model. The performance of the miRNA signature was evaluated using the area under curve (AUC) of receiver operating characteristic curve (ROC). We performed decision curve analysis (DCA) to assess the clinical usefulness of the signature. We also conducted a miRNA-regulatory network analysis to look for potential genes engaged in LNM in LUAD. Result: Thirteen miRNAs were selected to build our miRNA disease signature. The model showed good calibration in the training cohort, with an AUC of 0.782 (95% CI: 0.725-0.839). In the validation cohort, AUC was 0.691 (95% CI: 0.575-0.806). DCA demonstrated that the miRNA signature was clinically useful. Conclusion: The miRNA disease signature can be used as a noninvasive method to predict LNM in patients with lung adenocarcinoma objectively and the signature achieved high accuracy for prediction.