Employing heat maps to mine associations in structured routine care data

A unique color-coded visualization system with multimodal information fusion and deep learning in a longitudinal study of Alzheimer's disease

PURPOSE

Automated diagnosis and prognosis of Alzheimer's Disease remain a challenging problem that machine learning (ML) techniques have attempted to resolve in the last decade. This study introduces a first-of-its-kind color-coded visualization mechanism driven by an integrated ML model to predict disease trajectory in a 2-year longitudinal study. The main aim of this study is to help capture visually in 2D and 3D renderings the diagnosis and prognosis of AD, therefore augmenting our understanding of the processes of multiclass classification and regression analysis.

METHOD

The proposed method, Machine Learning for Visualizing AD (ML4VisAD), is designed to predict disease progression through a visual output. This newly developed model takes baseline measurements as input to generate a color-coded visual image that reflects disease progression at different time points. The architecture of the network relies on convolutional neural networks. With 1123 subjects selected from the ADNI QT-PAD dataset, we use a 10-fold cross-validation process to evaluate the method. Multimodal inputs* include neuroimaging data (MRI, PET), neuropsychological test scores (excluding MMSE, CDR-SB, and ADAS to avoid bias), cerebrospinal fluid (CSF) biomarkers with measures of amyloid beta (ABETA), phosphorylated tau protein (PTAU), total tau protein (TAU), and risk factors that include age, gender, years of education, and ApoE4 gene.

FINDINGS/RESULTS

Based on subjective scores reached by three raters, the results showed an accuracy of 0.82 ± 0.03 for a 3-way classification and 0.68 ± 0.05 for a 5-way classification. The visual renderings were generated in 0.08 msec for a 23 × 23 output image and in 0.17 ms for a 45 × 45 output image. Through visualization, this study (1) demonstrates that the ML visual output augments the prospects for a more accurate diagnosis and (2) highlights why multiclass classification and regression analysis are incredibly challenging. An online survey was conducted to gauge this visualization platform's merits and obtain valuable feedback from users. All implementation codes are shared online on GitHub.

CONCLUSION

This approach makes it possible to visualize the many nuances that lead to a specific classification or prediction in the disease trajectory, all in context to multimodal measurements taken at baseline. This ML model can serve as a multiclass classification and prediction model while reinforcing the diagnosis and prognosis capabilities by including a visualization platform.

Signatures of illness in children requiring unplanned intubation in the pediatric intensive care unit: A retrospective cohort machine-learning study

Acute respiratory failure requiring the initiation of invasive mechanical ventilation remains commonplace in the pediatric intensive care unit (PICU). Early recognition of patients at risk for respiratory failure may provide clinicians with the opportunity to intervene and potentially improve outcomes. Through the development of a random forest model to identify patients at risk for requiring unplanned intubation, we tested the hypothesis that subtle signatures of illness are present in physiological and biochemical time series of PICU patients in the early stages of respiratory decompensation. We included 116 unplanned intubation events as recorded in the National Emergency Airway Registry for Children in 92 PICU admissions over a 29-month period at our institution. We observed that children have a physiologic signature of illness preceding unplanned intubation in the PICU. Generally, it comprises younger age, and abnormalities in electrolyte, hematologic and vital sign parameters. Additionally, given the heterogeneity of the PICU patient population, we found differences in the presentation among the major patient groups - medical, cardiac surgical, and non-cardiac surgical. At four hours prior to the event, our random forest model demonstrated an area under the receiver operating characteristic curve of 0.766 (0.738 for medical, 0.755 for cardiac surgical, and 0.797 for non-cardiac surgical patients). The multivariable statistical models that captured the physiological and biochemical dynamics leading up to the event of urgent unplanned intubation in a PICU can be repurposed for bedside risk prediction.

Characterisation of the Phenolic Profile of Acacia retinodes and Acacia mearnsii Flowers' Extracts

Acacia spp. is an invasive species that is widespread throughout the Portuguese territory. Thus, it is pertinent to better understand this species in order to find different applications that will value its use. To evaluate the phenolic profile in Acacia flowers, ethanolic extracts obtained through an energized guided dispersive extraction were analysed, focusing on two species, Acacia retinodes and Acacia mearnsii, at two flowering stages. The phytochemical profile of each extract was determined by ultra-high performance liquid chromatography coupled with quadrupole time-of-flight mass spectrometry and high-performance liquid chromatography coupled with diode array detector. The FTIR-ATR technique was used to distinguish the different samples’ compositions. The results showed the presence of high concentrations of phenolic compounds (>300 mg GAE/g extract), among which are flavonoids (>136 mg QE/g extract), for all combinations of species/flowering stages. The phytochemical profile showed a complex composition with 21 compounds identified and quantified (the predominant ones being epicatechin, rutin, vanillin, and catechol). Both species and flowering stages presented significant variations regarding the presence and quantity of phenols and flavonoids, so much so that a principal component analysis performed with FTIR-ATR spectra data of the extracts was able to discriminate between species and flowering stages.

Domain Adaptation Using Convolutional Autoencoder and Gradient Boosting for Adverse Events Prediction in the Intensive Care Unit

More than 5 million patients have admitted annually to intensive care units (ICUs) in the United States. The leading causes of mortality are cardiovascular failures, multi-organ failures, and sepsis. Data-driven techniques have been used in the analysis of patient data to predict adverse events, such as ICU mortality and ICU readmission. These models often make use of temporal or static features from a single ICU database to make predictions on subsequent adverse events. To explore the potential of domain adaptation, we propose a method of data analysis using gradient boosting and convolutional autoencoder (CAE) to predict significant adverse events in the ICU, such as ICU mortality and ICU readmission. We demonstrate our results from a retrospective data analysis using patient records from a publicly available database called Multi-parameter Intelligent Monitoring in Intensive Care-II (MIMIC-II) and a local database from Children's Healthcare of Atlanta (CHOA). We demonstrate that after adopting novel data imputation on patient ICU data, gradient boosting is effective in both the mortality prediction task and the ICU readmission prediction task. In addition, we use gradient boosting to identify top-ranking temporal and non-temporal features in both prediction tasks. We discuss the relationship between these features and the specific prediction task. Lastly, we indicate that CAE might not be effective in feature extraction on one dataset, but domain adaptation with CAE feature extraction across two datasets shows promising results.

Visual Analytic Tools and Techniques in Population Health and Health Services Research: Scoping Review

Background

Visual analytics (VA) promotes the understanding of data with visual, interactive techniques, using analytic and visual engines. The analytic engine includes automated techniques, whereas common visual outputs include flow maps and spatiotemporal hot spots.

Objective

This scoping review aims to address a gap in the literature, with the specific objective to synthesize literature on the use of VA tools, techniques, and frameworks in interrelated health care areas of population health and health services research (HSR).

Methods

Using the 2018 PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) guidelines, the review focuses on peer-reviewed journal articles and full conference papers from 2005 to March 2019. Two researchers were involved at each step, and another researcher arbitrated disagreements. A comprehensive abstraction platform captured data from diverse bodies of the literature, primarily from the computer and health sciences.

Results

After screening 11,310 articles, findings from 55 articles were synthesized under the major headings of visual and analytic engines, visual presentation characteristics, tools used and their capabilities, application to health care areas, data types and sources, VA frameworks, frameworks used for VA applications, availability and innovation, and co-design initiatives. We found extensive application of VA methods used in areas of epidemiology, surveillance and modeling, health services access, use, and cost analyses. All articles included a distinct analytic and visualization engine, with varying levels of detail provided. Most tools were prototypes, with 5 in use at the time of publication. Seven articles presented methodological frameworks. Toward consistent reporting, we present a checklist, with an expanded definition for VA applications in health care, to assist researchers in sharing research for greater replicability. We summarized the results in a Tableau dashboard.

Conclusions

With the increasing availability and generation of big health care data, VA is a fast-growing method applied to complex health care data. What makes VA innovative is its capability to process multiple, varied data sources to demonstrate trends and patterns for exploratory analysis, leading to knowledge generation and decision support. This is the first review to bridge a critical gap in the literature on VA methods applied to the areas of population health and HSR, which further indicates possible avenues for the adoption of these methods in the future. This review is especially important in the wake of COVID-19 surveillance and response initiatives, where many VA products have taken center stage.

International Registered Report Identifier (IRRID)

RR2-10.2196/14019

Transcriptome Analysis of Oleoresin-Producing Tree Sindora Glabra and Characterization of Sesquiterpene Synthases

Terpenes serve important physiological and ecological functions in plants. Sindora glabra trees accumulate copious amounts of sesquiterpene-rich oleoresin in the stem. A transcriptome approach was used to determine the unique terpene biosynthesis pathway and to explore the different regulatory mechanisms responsible for the variation of terpene content among individuals. Analysis of de novo-assembled contigs revealed a complete set of genes for terpene biosynthesis. A total of 23,261 differentially expressed unigenes (DEGs) were discovered between high and low oil-yielding plants. DEG enrichment analysis suggested that the terpene biosynthesis process and the plant hormone signal transduction pathway may exert a major role in determining terpene variation in S. glabra. The expression patterns of candidate genes were further verified by quantitative RT-PCR experiments. Key genes involved in the terpene biosynthesis pathway were predominantly expressed in phloem and root tissues. Phylogenetic analysis and subcellular localization implied that S. glabra terpene synthases may evolve from a common ancestor. Furthermore, two sesquiterpene synthase genes, SgSTPS1 and SgSTPS2, were functionally characterized. SgSTPS1 mainly generated β-caryophyllene from farnesyl pyrophosphate. SgSTPS2 is a versatile enzyme that catalyzes the formation of 12 sequiterpenes from farnesyl pyrophosphate and synthesis of three monoterpenes using geranyl pyrophosphate. Together, these results provide large reservoir for elucidating the molecular mechanism of terpene biosynthesis and for exploring the ecological function of sesquiterpenes in S. glabra.

Comprehensive transcriptome profiling of Salvia miltiorrhiza for discovery of genes associated with the biosynthesis of tanshinones and phenolic acids

Tanshinones and phenolic acids are crucial bioactive compounds biosynthesized in Salvia miltiorrhiza. Methyl jasmonate (MeJA) is an effective elicitor to enhance the production of phenolic acids and tanshinones simultaneously, while yeast extract (YE) is used as a biotic elicitor that only induce tanshinones accumulation. However, little was known about the different molecular mechanism. To identify the downstream and regulatory genes involved in tanshinone and phenolic acid biosynthesis, we conducted comprehensive transcriptome profiling of S. miltiorrhiza hairy roots treated with either MeJA or YE. Total 55588 unigenes were assembled from about 1.72 billion clean reads, of which 42458 unigenes (76.4%) were successfully annotated. The expression patterns of 19 selected genes in the significantly upregulated unigenes were verified by quantitative real-time PCR. The candidate downstream genes and other cytochrome P450s involved in the late steps of tanshinone and phenolic acid biosynthesis pathways were screened from the RNA-seq dataset based on co-expression pattern analysis with specific biosynthetic genes. Additionally, 375 transcription factors were identified to exhibit a significant up-regulated expression pattern in response to induction. This study can provide us a valuable gene resource for elucidating the molecular mechanism of tanshinones and phenolic acids biosynthesis in hairy roots of S. miltiorrhiza.

Hierarchical cluster analysis in clinical research with heterogeneous study population: highlighting its visualization with R

Big data clinical research typically involves thousands of patients and there are numerous variables available. Conventionally, these variables can be handled by multivariable regression modeling. In this article, the hierarchical cluster analysis (HCA) is introduced. This method is used to explore similarity between observations and/or clusters. The result can be visualized using heat maps and dendrograms. Sometimes, it would be interesting to add scatter plot and smooth lines into the panels of the heat map. The inherent R heatmap package does not provide this function. A series of scatter plots can be created using lattice package, and then background color of each panel is mapped to the regression coefficient by using custom-made panel functions. This is the unique feature of the lattice package. Dendrograms and color keys can be added as the legend elements of the lattice system. The latticeExtra package provides some useful functions for the work.