Establishment and Analysis of a Combined Diagnostic Model of Polycystic Ovary Syndrome with Random Forest and Artificial Neural Network.Biomed Res Int. 2020;2020:2613091. [PMID: 32884937 DOI: 10.1155/2020/2613091]

Establishment and Analysis of an Artificial Neural Network Model for Early Detection of Polycystic Ovary Syndrome Using Machine Learning Techniques

Background

To identify novel gene combinations and to develop an early diagnostic model for Polycystic Ovary Syndrome (PCOS) through the integration of artificial neural networks (ANN) and random forest (RF) methods.

Methods

We retrieved and processed gene expression datasets for PCOS from the Gene Expression Omnibus (GEO) database. Differential expression analysis of genes (DEGs) within the training set was performed using the "limma" R package. Enrichment analyses on DEGs using gene ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG), and immune cell infiltration. The identification of critical genes from DEGs was then performed using random forests, followed by the developing of new diagnostic models for PCOS using artificial neural networks.

Results

We identified 130 up-regulated genes and 132 down-regulated genes in PCOS compared to normal samples. Gene Ontology analysis revealed significant enrichment in myofibrils and highlighted crucial biological functions related to myofilament sliding, myofibril, and actin-binding. Compared with normal tissues, the types of immune cells expressed in PCOS samples are different. A random forest algorithm identified 10 significant genes proposed as potential PCOS-specific biomarkers. Using these genes, an artificial neural network diagnostic model accurately distinguished PCOS from normal samples. The diagnostic model underwent validation using the independent validation set, and the resulting area under the receiver operating characteristic curve (AUC) values was consistent with the anticipated outcomes.

Conclusion

Utilizing unique gene combinations, this research created a diagnostic model by merging random forest techniques with artificial neural networks. The AUC indicated a notably superior performance of the diagnostic model.

Machine learning classification of polycystic ovary syndrome based on radial pulse wave analysis

BACKGROUND

Patients with Polycystic ovary syndrome (PCOS) experienced endocrine disorders that may present vascular function changes. This study aimed to classify and predict PCOS by radial pulse wave parameters using machine learning (ML) methods and to provide evidence for objectifying pulse diagnosis in traditional Chinese medicine (TCM).

METHODS

A case-control study with 459 subjects divided into a PCOS group and a healthy (non-PCOS) group. The pulse wave parameters were measured and analyzed between the two groups. Seven supervised ML classification models were applied, including K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Decision Trees, Random Forest, Logistic Regression, Voting, and Long Short Term Memory networks (LSTM). Parameters that were significantly different were selected as input features and stratified k-fold cross-validations training was applied to the models.

RESULTS

There were 316 subjects in the PCOS group and 143 subjects in the healthy group. Compared to the healthy group, the pulse wave parameters h3/h1 and w/t from both left and right sides were increased while h4, t4, t, As, h4/h1 from both sides and right t1 were decreased in the PCOS group (P < 0.01). Among the ML models evaluated, both the Voting and LSTM with ensemble learning capabilities, demonstrated competitive performance. These models achieved the highest results across all evaluation metrics. Specifically, they both attained a testing accuracy of 72.174% and an F1 score of 0.818, their respective AUC values were 0.715 for the Voting and 0.722 for the LSTM.

CONCLUSION

Radial pulse wave signal could identify most PCOS patients accurately (with a good F1 score) and is valuable for early detection and monitoring of PCOS with acceptable overall accuracy. This technique can stimulate the development of individualized PCOS risk assessment using mobile detection technology, furthermore, gives physicians an intuitive understanding of the objective pulse diagnosis of TCM.

TRIAL REGISTRATION

Not applicable.

Revealing a potential necroptosis-related axis (RP11-138A9.1/hsa-miR-98-5p/ZBP1) in periodontitis by construction of the ceRNA network

BACKGROUND AND OBJECTIVES

Periodontitis, a prevalent chronic inflammatory condition, poses a significant risk of tooth loosening and subsequent tooth loss. Within the realm of programmed cell death, a recently recognized process known as necroptosis has garnered attention for its involvement in numerous inflammatory diseases. Nevertheless, its correlation with periodontitis is indistinct. Our study aimed to identify necroptosis-related lncRNAs and crucial lncRNA-miRNA-mRNA regulatory axes in periodontitis to further understand the pathogenesis of periodontitis.

MATERIALS AND METHODS

Gene expression profiles in gingival tissues were acquired from the Gene Expression Omnibus (GEO) database. Selecting hub necroptosis-related lncRNA and extracting the key lncRNA-miRNA-mRNA axes based on the ceRNA network by adding novel machine-learning models based on conventional analysis and combining qRT-PCR validation. Then, an artificial neural network (ANN) model was constructed for lncRNA in regulatory axes, and the accuracy of the model was validated by receiver operating characteristic (ROC) curve analysis. The clinical effect of the model was evaluated by decision curve analysis (DCA). Weighted correlation network analysis (WGCNA) and single-sample gene set enrichment analysis (ssGSEA) was performed to explore how these lncRNAs work in periodontitis.

RESULTS

Seven hub necroptosis-related lncRNAs and three lncRNA-miRNA-mRNA regulatory axes (RP11-138A9.1/hsa-miR-98-5p/ZBP1 axis, RP11-96D1.11/hsa-miR-185-5p/EZH2 axis, and RP4-773 N10.4/hsa-miR-21-5p/TLR3 axis) were predicted. WGCNA revealed that RP11-138A9.1 was significantly correlated with the "purple module". Functional enrichment analysis and ssGSEA demonstrated that the RP11-138A9.1/hsa-miR-98-5p/ZBP1 axis is closely related to the inflammation and immune processes in periodontitis.

CONCLUSION

Our study predicted a crucial necroptosis-related regulatory axis (RP11-138A9.1/hsa-miR-98-5p/ZBP1) based on the ceRNA network, which may aid in elucidating the role and mechanism of necroptosis in periodontitis.

Application of machine learning and artificial intelligence in the diagnosis and classification of polycystic ovarian syndrome: a systematic review

Introduction

Polycystic Ovarian Syndrome (PCOS) is the most common endocrinopathy in women of reproductive age and remains widely underdiagnosed leading to significant morbidity. Artificial intelligence (AI) and machine learning (ML) hold promise in improving diagnostics. Thus, we performed a systematic review of literature to identify the utility of AI/ML in the diagnosis or classification of PCOS.

Methods

We applied a search strategy using the following databases MEDLINE, Embase, the Cochrane Central Register of Controlled Trials, the Web of Science, and the IEEE Xplore Digital Library using relevant keywords. Eligible studies were identified, and results were extracted for their synthesis from inception until January 1, 2022.

Results

135 studies were screened and ultimately, 31 studies were included in this study. Data sources used by the AI/ML interventions included clinical data, electronic health records, and genetic and proteomic data. Ten studies (32%) employed standardized criteria (NIH, Rotterdam, or Revised International PCOS classification), while 17 (55%) used clinical information with/without imaging. The most common AI techniques employed were support vector machine (42% studies), K-nearest neighbor (26%), and regression models (23%) were the commonest AI/ML. Receiver operating curves (ROC) were employed to compare AI/ML with clinical diagnosis. Area under the ROC ranged from 73% to 100% (n=7 studies), diagnostic accuracy from 89% to 100% (n=4 studies), sensitivity from 41% to 100% (n=10 studies), specificity from 75% to 100% (n=10 studies), positive predictive value (PPV) from 68% to 95% (n=4 studies), and negative predictive value (NPV) from 94% to 99% (n=2 studies).

Conclusion

Artificial intelligence and machine learning provide a high diagnostic and classification performance in detecting PCOS, thereby providing an avenue for early diagnosis of this disorder. However, AI-based studies should use standardized PCOS diagnostic criteria to enhance the clinical applicability of AI/ML in PCOS and improve adherence to methodological and reporting guidelines for maximum diagnostic utility.

Systematic review registration

https://www.crd.york.ac.uk/prospero/, identifier CRD42022295287.

Identifying oxidative stress-related biomarkers in idiopathic pulmonary fibrosis in the context of predictive, preventive, and personalized medicine using integrative omics approaches and machine-learning strategies

Background

Idiopathic pulmonary fibrosis (IPF) is a rare interstitial lung disease with a poor prognosis that currently lacks effective treatment methods. Preventing the acute exacerbation of IPF, identifying the molecular subtypes of patients, providing personalized treatment, and developing individualized drugs are guidelines for predictive, preventive, and personalized medicine (PPPM / 3PM) to promote the development of IPF. Oxidative stress (OS) is an important pathological process of IPF. However, the relationship between the expression levels of oxidative stress-related genes (OSRGs) and clinical indices in patients with IPF is unclear; therefore, it is still a challenge to identify potential beneficiaries of antioxidant therapy. Because PPPM aims to recognize and manage diseases by integrating multiple methods, patient stratification and analysis based on OSRGs and identifying biomarkers can help achieve the above goals.

Methods

Transcriptome data from 250 IPF patients were divided into training and validation sets. Core OSRGs were identified in the training set and subsequently clustered to identify oxidative stress-related subtypes. The oxidative stress scores, clinical characteristics, and expression levels of senescence-associated secretory phenotypes (SASPs) of different subtypes were compared to identify patients who were sensitive to antioxidant therapy to conduct differential gene functional enrichment analysis and predict potential therapeutic drugs. Diagnostic markers between subtypes were obtained by integrating multiple machine learning methods, their expression levels were tested in rat models with different degrees of pulmonary fibrosis and validation sets, and nomogram models were constructed. CIBERSORT, single-cell RNA sequencing, and immunofluorescence staining were used to explore the effects of OSRGs on the immune microenvironment.

Results

Core OSRGs classified IPF into two subtypes. Patients classified into subtypes with low oxidative stress levels had better clinical scores, less severe fibrosis, and lower expression of SASP-related molecules. A reliable nomogram model based on five diagnostic markers was constructed, and these markers' expression stability was verified in animal experiments. The number of neutrophils in the immune microenvironment was significantly different between the two subtypes and was closely related to the degree of fibrosis.

Conclusion

Within the framework of PPPM, this work comprehensively explored the role of OSRGs and their mediated cellular senescence and immune processes in the progress of IPF and assessed their capabilities aspredictors of high oxidative stress and disease progression,targets of the vicious loop between regulated pulmonary fibrosis and OS for targeted secondary and tertiary prevention, andreferences for personalized antioxidant and antifibrotic therapies.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13167-023-00334-4.

An artificial neural network model to diagnose non-obstructive azoospermia based on RNA-binding protein-related genes

Non-obstructive azoospermia (NOA) is a severe form of male infertility, but its pathological mechanisms and diagnostic biomarkers remain obscure. Since the dysregulation of RNA-binding proteins (RBPs) had nonnegligible effects on spermatogenesis, we aimed to investigate the functions and diagnosis values of RBPs in NOA. 58 testicular samples (control = 11, NOA = 47) from Gene Expression Omnibus (GEO) were set as the training cohort. Three public datasets, containing GSE45885 (control = 4, NOA = 27), GSE45887 (control = 4, NOA = 16), and GSE145467 (control = 10, NOA = 10), and 44 clinical samples from the local hospital (control = 27, NOA = 17) were used for validation. Through a series of bioinformatical analyses and machine learning algorithms, including genomic difference detection, protein-protein interaction network analysis, LASSO, SVM-RFE, and Boruta, DDX20 and NCBP2 were determined as significant predictors of NOA. Single-cell RNA sequencing of 432 testicular cell samples from NOA patients indicated that DDX20 and NCBP2 were associated with spermatogenesis (false discovery rate < 0.05). Based on the transcriptome expressions of DDX20 and NCBP2, we constructed multiple diagnosis models using logistic regression, random forest, and artificial neural network (ANN). The ANN model exhibited the most reliable predictive performance in the training cohort (AUC = 0.840), GSE45885 (AUC = 0.731), GSE45887 (AUC = 0.781), GSE145467 (AUC = 0.850), and local cohort (AUC = 0.623). Totally, an ANN diagnosis model based on RBP DDX20 and NCBP2 was developed and externally validated in NOA, functioning as a promising tool in clinical practice.

A united model for diagnosing pulmonary tuberculosis with random forest and artificial neural network

Background: Pulmonary tuberculosis (PTB) is a chronic infectious disease and is the most common type of TB. Although the sputum smear test is a gold standard for diagnosing PTB, the method has numerous limitations, including low sensitivity, low specificity, and insufficient samples.

Methods: The present study aimed to identify specific biomarkers of PTB and construct a model for diagnosing PTB by combining random forest (RF) and artificial neural network (ANN) algorithms. Two publicly available cohorts of TB, namely, the GSE83456 (training) and GSE42834 (validation) cohorts, were retrieved from the Gene Expression Omnibus (GEO) database. A total of 45 and 61 differentially expressed genes (DEGs) were identified between the PTB and control samples, respectively, by screening the GSE83456 cohort. An RF classifier was used for identifying specific biomarkers, following which an ANN-based classification model was constructed for identifying PTB samples. The accuracy of the ANN model was validated using the receiver operating characteristic (ROC) curve. The proportion of 22 types of immunocytes in the PTB samples was measured using the CIBERSORT algorithm, and the correlations between the immunocytes were determined.

Results: Differential analysis revealed that 11 and 22 DEGs were upregulated and downregulated, respectively, and 11 biomarkers specific to PTB were identified by the RF classifier. The weights of these biomarkers were determined and an ANN-based classification model was subsequently constructed. The model exhibited outstanding performance, as revealed by the area under the curve (AUC), which was 1.000 for the training cohort. The AUC of the validation cohort was 0.946, which further confirmed the accuracy of the model.

Conclusion: Altogether, the present study successfully identified specific genetic biomarkers of PTB and constructed a highly accurate model for the diagnosis of PTB based on blood samples. The model developed herein can serve as a reliable reference for the early detection of PTB and provide novel perspectives into the pathogenesis of PTB.

Multi-omics and machine learning for the prevention and management of female reproductive health

Females typically carry most of the burden of reproduction in mammals. In humans, this burden is exacerbated further, as the evolutionary advantage of a large and complex human brain came at a great cost of women's reproductive health. Pregnancy thus became a highly demanding phase in a woman's life cycle both physically and emotionally and therefore needs monitoring to assure an optimal outcome. Moreover, an increasing societal trend towards reproductive complications partly due to the increasing maternal age and global obesity pandemic demands closer monitoring of female reproductive health. This review first provides an overview of female reproductive biology and further explores utilization of large-scale data analysis and -omics techniques (genomics, transcriptomics, proteomics, and metabolomics) towards diagnosis, prognosis, and management of female reproductive disorders. In addition, we explore machine learning approaches for predictive models towards prevention and management. Furthermore, mobile apps and wearable devices provide a promise of continuous monitoring of health. These complementary technologies can be combined towards monitoring female (fertility-related) health and detection of any early complications to provide intervention solutions. In summary, technological advances (e.g., omics and wearables) have shown a promise towards diagnosis, prognosis, and management of female reproductive disorders. Systematic integration of these technologies is needed urgently in female reproductive healthcare to be further implemented in the national healthcare systems for societal benefit.

Development and analysis of a comprehensive diagnostic model for aortic valve calcification using machine learning methods and artificial neural networks

BACKGROUND

Although advanced surgical and interventional treatments are available for advanced aortic valve calcification (AVC) with severe clinical symptoms, early diagnosis, and intervention is critical in order to reduce calcification progression and improve patient prognosis. The aim of this study was to develop therapeutic targets for improving outcomes for patients with AVC.

MATERIALS AND METHODS

We used the public expression profiles of individuals with AVC (GSE12644 and GSE51472) to identify potential diagnostic markers. First, the R software was used to identify differentially expressed genes (DEGs) and perform functional enrichment analysis. Next, we combined bioinformatics techniques with machine learning methodologies such as random forest algorithms and support vector machines to screen for and identify diagnostic markers of AVC. Subsequently, artificial neural networks were employed to filter and model the diagnostic characteristics for AVC incidence. The diagnostic values were determined using the receiver operating characteristic (ROC) curves. Furthermore, CIBERSORT immune infiltration analysis was used to determine the expression of different immune cells in the AVC. Finally, the CMap database was used to predict candidate small compounds as prospective AVC therapeutics.

RESULTS

A total of 78 strong DEGs were identified. The leukocyte migration and pid integrin 1 pathways were highly enriched for AVC-specific DEGs. CXCL16, GPM6A, BEX2, S100A9, and SCARA5 genes were all regarded diagnostic markers for AVC. The model was effectively constructed using a molecular diagnostic score system with significant diagnostic value (AUC = 0.987) and verified using the independent dataset GSE83453 (AUC = 0.986). Immune cell infiltration research revealed that B cell naive, B cell memory, plasma cells, NK cell activated, monocytes, and macrophage M0 may be involved in the development of AVC. Additionally, all diagnostic characteristics may have varying degrees of correlation with immune cells. The most promising small molecule medicines for reversing AVC gene expression are Doxazosin and Terfenadine.

CONCLUSION

It was identified that CXCL16, GPM6A, BEX2, S100A9, and SCARA5 are potentially beneficial for diagnosing and treating AVC. A diagnostic model was constructed based on a molecular prognostic score system using machine learning. The aforementioned immune cell infiltration may have a significant influence on the development and incidence of AVC.

Development of a joint diagnostic model of thyroid papillary carcinoma with artificial neural network and random forest

Objective: Papillary thyroid carcinoma (PTC) accounts for 80% of thyroid malignancy, and the occurrence of PTC is increasing rapidly. The present study was conducted with the purpose of identifying novel and important gene panels and developing an early diagnostic model for PTC by combining artificial neural network (ANN) and random forest (RF).

Methods and results: Samples were searched from the Gene Expression Omnibus (GEO) database, and gene expression datasets (GSE27155, GSE60542, and GSE33630) were collected and processed. GSE27155 and GSE60542 were merged into the training set, and GSE33630 was defined as the validation set. Differentially expressed genes (DEGs) in the training set were obtained by “limma” of R software. Then, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis as well as immune cell infiltration analysis were conducted based on DEGs. Important genes were identified from the DEGs by random forest. Finally, an artificial neural network was used to develop a diagnostic model. Also, the diagnostic model was validated by the validation set, and the area under the receiver operating characteristic curve (AUC) value was satisfactory.

Conclusion: A diagnostic model was established by a joint of random forest and artificial neural network based on a novel gene panel. The AUC showed that the diagnostic model had significantly excellent performance.

Identification and immune characteristics of molecular subtypes related to fatty acid metabolism in idiopathic pulmonary fibrosis

Background

Although fatty acid metabolism has been confirmed to be involved in the pathological process of idiopathic pulmonary fibrosis (IPF), systematic analyses on the immune process mediated by fatty acid metabolism-related genes (FAMRGs) in IPF remain lacking.

Methods

The gene expression data of 315 patients with IPF were obtained from Gene Expression Omnibus database and were divided into the training and verification sets. The core FAMRGs of the training set were identified through weighted gene co-expression network analysis. Then, the fatty acid metabolism-related subtypes in IPF were identified on the basis of k-means unsupervised clustering. The scores of fatty acid metabolism and the expression of the fibrosis biomarkers in different subtypes were compared, and functional enrichment analysis was carried out on the differentially expressed genes between subtypes. A random forest model was used to select important FAMRGs as diagnostic markers for distinguishing between subtypes, and a line chart model was constructed and verified by using other datasets and rat models with different degrees of pulmonary fibrosis. The difference in immune cell infiltration among subtypes was evaluated with CIBERSORT, and the correlation between core diagnostic markers and immune cells were analyzed.

Results

Twenty-four core FAMRGs were differentially expressed between the training set and normal samples, and IPF was divided into two subtypes. Significant differences were observed between the two subtypes in biological processes, such as linoleic acid metabolism, cilium movement, and natural killer (NK) cell activation. The subtype with high fatty acid metabolism had more severe pulmonary fibrosis than the other subtype. A reliable construction line chart model based on six diagnostic markers was constructed, and ABCA3 and CYP24A1 were identified as core diagnostic markers. Significant differences in immune cell infiltration were found between the two subtypes, and ABCA3 and CYP24A1 were closely related to NK cells.

Conclusion

Fatty acid metabolism and the immune process that it mediates play an important role in the occurrence and development of IPF. The analysis of the role of FAMRGs in IPF may provide a new potential therapeutic target for IPF.

Development and validation of immune-based biomarkers and deep learning models for Alzheimer’s disease

Background: Alzheimer’s disease (AD) is the most common form of dementia in old age and poses a severe threat to the health and life of the elderly. However, traditional diagnostic methods and the ATN diagnostic framework have limitations in clinical practice. Developing novel biomarkers and diagnostic models is necessary to complement existing diagnostic procedures.

Methods: The AD expression profile dataset GSE63060 was downloaded from the NCBI GEO public database for preprocessing. AD-related differentially expressed genes were screened using a weighted co-expression network and differential expression analysis, and functional enrichment analysis was performed. Subsequently, we screened hub genes by random forest, analyzed the correlation between hub genes and immune cells using ssGSEA, and finally built an AD diagnostic model using an artificial neural network and validated it.

Results: Based on the random forest algorithm, we screened a total of seven hub genes from AD-related DEGs, based on which we confirmed that hub genes play an essential role in the immune microenvironment and successfully established a novel diagnostic model for AD using artificial neural networks, and validated its effectiveness in the publicly available datasets GSE63060 and GSE97760.

Conclusion: Our study establishes a reliable model for screening and diagnosing AD that provides a theoretical basis for adding diagnostic biomarkers for the AD gene.

Establishment and Analysis of a Combined Diagnostic Model of Alzheimer's Disease With Random Forest and Artificial Neural Network

Alzheimer's disease (AD) is a neurodegenerative condition that causes cognitive decline over time. Because existing diagnostic approaches for AD are limited, improving upon previously established diagnostic models based on genetic biomarkers is necessary. Firstly, four AD gene expression datasets were collected from the Gene Expression Omnibus (GEO) database. Two datasets were used to establish diagnostic models, and the other two datasets were used to verify the model effect. We merged GSE5281 with GSE44771 as the training dataset and found 120 DEGs. Then, we used random forest (RF) to screen 6 key genes (KLF15, MAFF, ITPKB, SST, DDIT4, and NRXN3) as being critical for separating AD and normal samples. The weights of these key genes were measured, and a diagnostic model was created using an artificial neural network (ANN). The area under the curve (AUC) of the model is 0.953, while the accuracy is 0.914. In the final step, two validation datasets were utilized to assess AUC performance. In GSE109887, our model had an AUC of 0.854, and in GSE132903, it had an AUC of 0.810. To summarize, we successfully identified key gene biomarkers and developed a new AD diagnostic model.

A Joint Model of Random Forest and Artificial Neural Network for the Diagnosis of Endometriosis

Endometriosis (EM), an estrogen-dependent inflammatory disease with unknown etiology, affects thousands of childbearing-age couples, and its early diagnosis is still very difficult. With the rapid development of sequencing technology in recent years, the accumulation of many sequencing data makes it possible to screen important diagnostic biomarkers from some EM-related genes. In this study, we utilized public datasets in the Gene Expression Omnibus (GEO) and Array-Express database and identified seven important differentially expressed genes (DEGs) (COMT, NAA16, CCDC22, EIF3E, AHI1, DMXL2, and CISD3) through the random forest classifier. Among these DEGs, AHI1, DMXL2, and CISD3 have never been reported to be associated with the pathogenesis of EMs. Our study indicated that these three genes might participate in the pathogenesis of EMs through oxidative stress, epithelial–mesenchymal transition (EMT) with the activation of the Notch signaling pathway, and mitochondrial homeostasis, respectively. Then, we put these seven DEGs into an artificial neural network to construct a novel diagnostic model for EMs and verified its diagnostic efficacy in two public datasets. Furthermore, these seven DEGs were included in 15 hub genes identified from the constructed protein–protein interaction (PPI) network, which confirmed the reliability of the diagnostic model. We hope the diagnostic model can provide novel sights into the understanding of the pathogenesis of EMs and contribute to the clinical diagnosis and treatment of EMs.

Development of a random forest model for hypotension prediction after anesthesia induction for cardiac surgery

BACKGROUND

Hypotension after the induction of anesthesia is known to be associated with various adverse events. The involvement of a series of factors makes the prediction of hypotension during anesthesia quite challenging.

AIM

To explore the ability and effectiveness of a random forest (RF) model in the prediction of post-induction hypotension (PIH) in patients undergoing cardiac surgery.

METHODS

Patient information was obtained from the electronic health records of the Second Affiliated Hospital of Hainan Medical University. The study included patients, ≥ 18 years of age, who underwent cardiac surgery from December 2007 to January 2018. An RF algorithm, which is a supervised machine learning technique, was employed to predict PIH. Model performance was assessed by the area under the curve (AUC) of the receiver operating characteristic. Mean decrease in the Gini index was used to rank various features based on their importance.

RESULTS

Of the 3030 patients included in the study, 1578 (52.1%) experienced hypotension after the induction of anesthesia. The RF model performed effectively, with an AUC of 0.843 (0.808-0.877) and identified mean blood pressure as the most important predictor of PIH after anesthesia. Age and body mass index also had a significant impact.

CONCLUSION

The generated RF model had high discrimination ability for the identification of individuals at high risk for a hypotensive event during cardiac surgery. The study results highlighted that machine learning tools confer unique advantages for the prediction of adverse post-anesthesia events.

Tumor Nonimmune-Microenvironment-Related Gene Expression Signature Predicts Brain Metastasis in Lung Adenocarcinoma Patients after Surgery: A Machine Learning Approach Using Gene Expression Profiling

Simple Summary

It is important to be able to predict brain metastasis in lung adenocarcinoma patients; however, research in this area is still lacking. Much of the previous work on tumor microenvironments in lung adenocarcinoma with brain metastasis concerns the tumor immune microenvironment. The importance of the tumor nonimmune microenvironment (extracellular matrix (ECM), epithelial–mesenchymal transition (EMT) feature, and angiogenesis) has been overlooked with regard to brain metastasis. We evaluated tumor nonimmune-microenvironment-related gene expression signatures that could predict brain metastasis after the surgical resection of lung adenocarcinoma using a machine learning approach. We identified a tumor nonimmune-microenvironment-related 17-gene expression signature, and this signature showed high brain metastasis predictive power in four machine learning classifiers. The immunohistochemical expression of the top three genes of the 17-gene expression signature yielded similar results to NanoString tests. Our tumor nonimmune-microenvironment-related gene expression signatures are important biological markers that can predict brain metastasis and provide patient-specific treatment options.

Abstract

Using a machine learning approach with a gene expression profile, we discovered a tumor nonimmune-microenvironment-related gene expression signature, including extracellular matrix (ECM) remodeling, epithelial–mesenchymal transition (EMT), and angiogenesis, that could predict brain metastasis (BM) after the surgical resection of 64 lung adenocarcinomas (LUAD). Gene expression profiling identified a tumor nonimmune-microenvironment-related 17-gene expression signature that significantly correlated with BM. Of the 17 genes, 11 were ECM-remodeling-related genes. The 17-gene expression signature showed high BM predictive power in four machine learning classifiers (areas under the receiver operating characteristic curve = 0.845 for naïve Bayes, 0.849 for support vector machine, 0.858 for random forest, and 0.839 for neural network). Subgroup analysis revealed that the BM predictive power of the 17-gene signature was higher in the early-stage LUAD than in the late-stage LUAD. Pathway enrichment analysis showed that the upregulated differentially expressed genes were mainly enriched in the ECM–receptor interaction pathway. The immunohistochemical expression of the top three genes of the 17-gene expression signature yielded similar results to NanoString tests. The tumor nonimmune-microenvironment-related gene expression signatures found in this study are important biological markers that can predict BM and provide patient-specific treatment options.

Deep Learning Algorithm for Automated Detection of Polycystic Ovary Syndrome Using Scleral Images

The high prevalence of polycystic ovary syndrome (PCOS) among reproductive-aged women has attracted more and more attention. As a common disorder that is likely to threaten women's health physically and mentally, the detection of PCOS is a growing public health concern worldwide. In this paper, we proposed an automated deep learning algorithm for the auxiliary detection of PCOS, which explores the potential of scleral changes in PCOS detection. The algorithm was applied to the dataset that contains the full-eye images of 721 Chinese women, among which 388 are PCOS patients. Inputs of the proposed algorithm are scleral images segmented from full-eye images using an improved U-Net, and then a Resnet model was applied to extract deep features from scleral images. Finally, a multi-instance model was developed to achieve classification. Various performance indices such as AUC, classification accuracy, precision, recall, precision, and F1-score were adopted to assess the performance of our algorithm. Results show that our method achieves an average AUC of 0.979 and a classification accuracy of 0.929, which indicates the great potential of deep learning in the detection of PCOS.