BioMedInformatics, Volume 2, Issue 3 (September 2022) – 11 articles

Background: The application of machine learning (ML) tools (MLTs) to support clinical trials outputs in evidence-based health informatics can be an effective, useful, feasible, and acceptable way to advance medical research and provide precision medicine. Methods: In this study, the author used the rapid review approach and snowballing methods. The review was conducted in the following databases: PubMed, Scopus, COCHRANE LIBRARY, clinicaltrials.gov, Semantic Scholar, and the first six pages of Google Scholar from the 10 July–15 August 2022 period. Results: Here, 49 articles met the required criteria and were included in this review. Accordingly, 32 MLTs and platforms were identified in this study that applied the automatic extraction of knowledge from clinical trial outputs. Specifically, the initial use of automated tools resulted in modest to satisfactory time savings compared with the manual management. In addition, the evaluation of performance, functionality, usability, user interface, and system requirements also yielded positive results. Moreover, the evaluation of some tools in terms of acceptance, feasibility, precision, accuracy, efficiency, efficacy, and reliability was also positive. Conclusions: In summary, design based on the application of clinical trial results in ML is a promising approach to apply more reliable solutions. Future studies are needed to propose common standards for the assessment of MLTs and to clinically validate the performance in specific healthcare and technical domains. Full article

Recent developments in research on artificial intelligence (AI) in medicine deal with the analysis of image data such as Magnetic Resonance Imaging (MRI) scans to support the of decision-making of medical personnel. For this purpose, machine learning (ML) algorithms are often used, which do not explain the internal decision-making process at all. Thus, it is often difficult to validate or interpret the results of the applied AI methods. This manuscript aims to overcome this problem by using methods of explainable AI (XAI) to interpret the decision-making of an ML algorithm in the use case of predicting the survival rate of patients with brain tumors based on MRI scans. Therefore, we explore the analysis of brain images together with survival data to predict survival in gliomas with a focus on improving the interpretability of the results. Using the Brain Tumor Segmentation dataset BraTS 2020, we used a well-validated dataset for evaluation and relied on a convolutional neural network structure to improve the explainability of important features by adding Shapley overlays. The trained network models were used to evaluate SHapley Additive exPlanations (SHAP) directly and were not optimized for accuracy. The resulting overfitting of some network structures is therefore seen as a use case of the presented interpretation method. It is shown that the network structure can be validated by experts using visualizations, thus making the decision-making of the method interpretable. Our study highlights the feasibility of combining explainers with 3D voxels and also the fact that the interpretation of prediction results significantly supports the evaluation of results. The implementation in python is available on gitlab as “XAIforBrainImgSurv”. Full article

Despite the great progress in its early diagnosis and treatment, colon adenocarcinoma (COAD) is still poses important issues to clinical management. Therefore, the identification of novel biomarkers or therapeutic targets for this disease is important. Using UALCAN, the top 25 upregulated and downregulated genes in COAD were identified. Then, a Kaplan–Meier plotter was employed for these genes for survival analysis, revealing the correlation with overall survival rate only for MMP3 (Matrix Metallopeptidase 3) and TESC (Tescalcin). Despite this, the mRNA expression levels were not correlated with the tumor stages or nodal metastatic status. MMP3 and TESC are relevant targets in COAD that should be additionally validated as biomarkers for early diagnosis and prevention. Ingenuity Pathway Analysis revealed the top relevant network linked to Post-Translational Modification, Protein Degradation, and Protein Synthesis, where MMP3 was at the core of the network. Another important network was related to cell cycle regulation, TESC being a component of this. We should also not underestimate the complex regulatory mechanisms mediated by the interplay of the multiple other regulatory molecules, emphasizing the interconnection with molecules related to invasion and migration involved in COAD, that might serve as the basis for the development of new biomarkers and therapeutic targets. Full article

Background: Pulmonary hypertension (PH) is a complex disease caused by a wide range of underlying conditions, Tanshinone IIA (Tan IIA) has been widely used in PH patients. The study aimed to explore the possible molecular mechanism of Tan IIA against PH by network pharmacology and molecular docking. Methods: Tan IIA and PH-related targets were retrieved from public databases. Gene ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis, and protein–protein interaction (PPI) network were used to investigate the protein targets and mechanism. The binding activity of core targets and Tan IIA were verified by molecular docking. Results: A total of 26 overlapping target proteins between Tan IIA and PH were screened. PPI network identified HSP90AA1, PTPN11, ATM, CA2, TERT, PRKDC, and APEX1 as key pharmacological targets. The results of GO function enrichment analysis included regulation of smooth muscle cell proliferation and migration, regulation of mitotic cell cycle, and regulation of G1/S transition of mitotic cell cycle. KEGG pathway analysis showed that nitrogen metabolism, NF-kappa B signaling pathway, cell cycle, necroptosis, apoptosis, and JAK-STAT signaling pathway were associated with Tan IIA in PH. The molecular docking results showed that Tan IIA can closely bind three core targets (HSP90AA1, PTPN11, and CA2). Conclusions: The present work initially clarified the effective therapeutic targets, biological processes, and signaling pathways of Tan IIA treatment of PH, which lay a foundation for further research on the pharmacological effects of Tan IIA. Full article

The lack of consistent presentation of results in published studies on the association between a quantitative explanatory variable and a quantitative dependent variable has been a long-term issue in evaluating the reported findings. Studies are analyzed and reported in a variety of ways. The main purpose of this review is to illustrate the procedures in summarizing and synthesizing research results from multivariate models with a quantitative outcome variable. The review summarizes the application of the standardized regression coefficient as an effect size index in the context of meta-analysis and describe how it can be estimated and converted from data presented in original research articles. An example of synthesis is provided using research articles on the association between childhood body mass index and carotid intima-media thickness in adult life. Finally, the paper shares practical recommendations for meta-analysts wanting to use the standardized regression coefficient in pooling findings. Full article

Background: Endoplasmic reticulum stress (ERS) is involved in the etiology of non-alcoholic fatty liver disease (NAFLD). Thus, the current study was designed to identify key ERS-associated genes in NAFLD. Methods: RNA-Seq data of NAFLD and controls were sourced from the Gene Expression Omnibus (GEO) database. Differentially expressed genes (DEGs) in NAFLD and controls were identified by limma. By overlapping DEGs and ERS-related genes, ERS-related DEGs were identified. The function of ERS-related DEGs was characterized by clusterProfiler. Next, the protein–protein interaction (PPI) network was created using the Cytoscape software and the STRING database to identify key ERS-related genes in NAFLD. Furthermore, the correlations among key ERS-related genes were calculated. Results: A total of 8965 DEGs were identified between NAFLD and controls in the GSE126848 dataset. After overlapping these DEGs and ERS-related genes, 20 genes were identified as ERS-related DEGs in NAFLD. Functional analysis revealed that the genes mainly participated in ER-related functions, such as the ER–nucleus signaling pathway, regulation of ERS response, and protein processing in ER. The PPI network revealed the interactions among 17 ERS-related DEGs, including ERN1, ATF6, and EIF2S1 as the key genes. The expressions of ERN1, ATF6, and EIF2S1 were significantly down-regulated in NAFLD and were strongly positively correlated with each other. Further, the expression of ERN1 and ATFA6 was also similar in the GSE89632 datasets. Conclusion: The present study identified ERN1, ATF6, and EIF2S1 as key ERS-related genes in NAFLD. These findings may provide a molecular basis for the role of ERS in NAFLD. Full article

Dengue endemicity has become regular in recent times across the world. The numbers of cases and deaths have been alarmingly increasing over the years. In addition to this, there are no direct medications or vaccines to treat this viral infection. Thus, monitoring and controlling the carriers of this virus which are the Aedes mosquitoes become specially demanding to combat the endemicity, as killing all the mosquitoes regardless of their species would destroy ecosystems. The current approach requires collecting a larva sample from the hatching sites and, then, an expert entomologist manually examining it using a microscope in the laboratory to identify the Aedes vector. This is time-consuming, labor-intensive, subjective, and impractical. Several automated Aedes larvae detection systems have been proposed previously, but failed to achieve sufficient accuracy and reliability. We propose an automated system utilizing ensemble learning, which detects Aedes larvae effectively from a low-magnification image with an accuracy of over 99%. The proposed system outperformed all the previous methods with respect to accuracy. The practical usability of the system is also demonstrated. Full article

Since the beginning of the COVID-19 pandemic, a large number of epidemiological models have been developed. The principal objective of the present study is to provide a new six-compartment model for the COVID-19 pandemic, which takes into account both the possibility of re-infection and the differentiation between asymptomatic and symptomatic infected subjects. The model, denoted as θ-SI(R)D, is a six-compartment model, described by as many ordinary differential equations. The six compartments are denoted as Susceptible ($S$), Symptomatic Infected ($I_s$), Asymptomatic Infected ($I_a$), Recovered from Asymptomatic fraction ($R_a$), Recovered from Symptomatic fraction ($R_s$), and Deceased ($D$). Such a model has no analytical solutions, so we performed both a simulation and a model validation ($R^2=0.829)$. Based on the results of our simulations (and, on the other hand, on the results of most of the models in the scientific literature), it is possible to draw the reasonable conclusion that the epidemic tends, even without vaccination, to a steady state. Full article

Acute myeloid leukemia (AML), a hematologic malignancy that results in bone marrow failure, is the most common acute leukemia in adults. The presence of FMS-related tyrosine kinase 3 (FLT3) mutations is associated with a poor prognosis, making the evaluation of FLT3-inhibitors an imperative goal in clinical trials. Midostaurin was the first FLT3-inhibitor approved by the FDA and EMA for the treatment of FLT3-mutated AML, and it showed a significant improvement in overall survival for newly diagnosed patients treated with midostaurin, in combination with standard chemotherapy (RATIFY study). The main interest of midostaurin has been the FLT3-specific inhibition, but little is known about its role as a multikinase inhibitor and whether it may be used in relapse and maintenance therapy. Here, we used systems biology- and machine learning-based approaches to deepen the potential benefits of the multitarget activity of midostaurin and to better understand its anti-leukemic effect on FLT3-mutated AML. The resulting in silico study revealed that the multikinase activity of midostaurin may play a role in the treatment’s efficacy. Additionally, we propose a series of molecular mechanisms that support a potential benefit of midostaurin as a maintenance therapy in FLT3-mutated AML, by regulating the microenvironment. The obtained results are backed up using independent gene expression data. Full article

In the context of improving clinical treatments and certifying clinics, guideline-compliant care has become more important. However, verifying the compliance of treatment procedures with Clinical Guidelines remains difficult, as guidelines are mostly available in non-computer interpretable form and previous computer-interpretable approaches neglect the process perspective with its potential to gain medical insight. In this paper, we present our transformation framework CGK4PM, which addresses the procedural nature of treatment processes and which guides the transformation of clinical explicit and implicit guideline knowledge into process models. The procedural representation enables the use of process mining techniques such as conformance checking to verify guideline compliance and the opportunity to gain insights from complex clinical treatment processes. In collaboration with physicians from Münster University Hospital, the practical applicability of the framework is demonstrated in a case study by transforming the guideline for the treatment of malignant melanoma. The case study findings demonstrate the need for structured and guided transformation and highlight the difficulties in developing a guideline-based process model. Full article

Breast cancer is a prevalent disease that affects mostly women, and early diagnosis will expedite the treatment of this ailment. Recently, machine learning (ML) techniques have been employed in biomedical and informatics to help fight breast cancer. Extracting information from data to support the clinical diagnosis of breast cancer is a tedious and time-consuming task. The use of machine learning and feature extraction techniques has significantly changed the whole process of a breast cancer diagnosis. This research work proposed a machine learning model for the classification of breast cancer. To achieve this, a support vector machine (SVM) was employed for the classification, and linear discriminant analysis (LDA) was employed for feature extraction. We measured our model’s feature extraction performance in principal component analysis (PCA) and random forest for classification. A comparative analysis of the proposed model was performed to show the effectiveness of the feature extraction, and we computed missing values based on the classifier’s accuracy, precision, and recall. The original Wisconsin Breast Cancer dataset (WBCD) and Wisconsin Prognostic Breast Cancer dataset (WPBC) were used. We evaluated performance in two phases: In phase 1, rows containing missing values were computed using the mean, and in phase 2, rows containing missing values were computed using the median. LDA–SVM when median was used to compute missing values has better results, with accuracy of 99.2%, recall of 98.0% and precision of 98.0% on the WBCD dataset and an accuracy of 79.5%, recall of 76.0% and precision of 59.0% on the WPBC dataset. The SVM classifier had a better performance in handling classification problems when LDA was applied and the median was used as a method for computing missing values. Full article