Increasing acceptance of AI-generated digital twins through clinical trial applications.

Today's approach to medicine requires extensive trial and error to determine the proper treatment path for each patient. While many fields have benefited from technological breakthroughs in computer science, such as artificial intelligence (AI), the task of developing effective treatments is actually getting slower and more costly. With the increased availability of rich historical datasets from previous clinical trials and real-world data sources, one can leverage AI models to create holistic forecasts of future health outcomes for an individual patient in the form of an AI-generated digital twin. This could support the rapid evaluation of intervention strategies in silico and could eventually be implemented in clinical practice to make personalized medicine a reality. In this work, we focus on uses for AI-generated digital twins of clinical trial participants and contend that the regulatory outlook for this technology within drug development makes it an ideal setting for the safe application of AI-generated digital twins in healthcare. With continued research and growing regulatory acceptance, this path will serve to increase trust in this technology and provide momentum for the widespread adoption of AI-generated digital twins in clinical practice.

Standard machine learning approaches outperform deep representation learning on phenotype prediction from transcriptomics data.

BACKGROUND: The ability to confidently predict health outcomes from gene expression would catalyze a revolution in molecular diagnostics. Yet, the goal of developing actionable, robust, and reproducible predictive signatures of phenotypes such as clinical outcome has not been attained in almost any disease area. Here, we report a comprehensive analysis spanning prediction tasks from ulcerative colitis, atopic dermatitis, diabetes, to many cancer subtypes for a total of 24 binary and multiclass prediction problems and 26 survival analysis tasks. We systematically investigate the influence of gene subsets, normalization methods and prediction algorithms. Crucially, we also explore the novel use of deep representation learning methods on large transcriptomics compendia, such as GTEx and TCGA, to boost the performance of state-of-the-art methods. The resources and findings in this work should serve as both an up-to-date reference on attainable performance, and as a benchmarking resource for further research. RESULTS: Approaches that combine large numbers of genes outperformed single gene methods consistently and with a significant margin, but neither unsupervised nor semi-supervised representation learning techniques yielded consistent improvements in out-of-sample performance across datasets. Our findings suggest that using l2-regularized regression methods applied to centered log-ratio transformed transcript abundances provide the best predictive analyses overall. CONCLUSIONS: Transcriptomics-based phenotype prediction benefits from proper normalization techniques and state-of-the-art regularized regression approaches. In our view, breakthrough performance is likely contingent on factors which are independent of normalization and general modeling techniques; these factors might include reduction of systematic errors in sequencing data, incorporation of other data types such as single-cell sequencing and proteomics, and improved use of prior knowledge.

Machine learning for comprehensive forecasting of Alzheimer's Disease progression.

Most approaches to machine learning from electronic health data can only predict a single endpoint. The ability to simultaneously simulate dozens of patient characteristics is a crucial step towards personalized medicine for Alzheimer's Disease. Here, we use an unsupervised machine learning model called a Conditional Restricted Boltzmann Machine (CRBM) to simulate detailed patient trajectories. We use data comprising 18-month trajectories of 44 clinical variables from 1909 patients with Mild Cognitive Impairment or Alzheimer's Disease to train a model for personalized forecasting of disease progression. We simulate synthetic patient data including the evolution of each sub-component of cognitive exams, laboratory tests, and their associations with baseline clinical characteristics. Synthetic patient data generated by the CRBM accurately reflect the means, standard deviations, and correlations of each variable over time to the extent that synthetic data cannot be distinguished from actual data by a logistic regression. Moreover, our unsupervised model predicts changes in total ADAS-Cog scores with the same accuracy as specifically trained supervised models, additionally capturing the correlation structure in the components of ADAS-Cog, and identifies sub-components associated with word recall as predictive of progression.

ACR Appropriateness Criteria® Radiologic Management of Venous Thromboembolism-Inferior Vena Cava Filters.

Venous thromboembolism (VTE)-deep vein thrombosis and pulmonary embolism-is a common cause of morbidity and mortality. The mainstay of VTE prophylaxis and therapy is anticoagulation. In select patients with VTE, inferior vena cava (IVC) filters are used to prevent pulmonary embolism by trapping emboli as they pass from the lower extremity venous system through the IVC. These guidelines review the indications for placement of IVC filters in acute and chronic VTE, as well as the indications for retrieval of implanted IVC filters. The American College of Radiology Appropriateness Criteria are evidence-based guidelines for specific clinical conditions that are reviewed annually by a multidisciplinary expert panel. The guideline development and revision include an extensive analysis of current medical literature from peer reviewed journals and the application of well-established methodologies (RAND/UCLA Appropriateness Method and Grading of Recommendations Assessment, Development, and Evaluation or GRADE) to rate the appropriateness of imaging and treatment procedures for specific clinical scenarios. In those instances where evidence is lacking or equivocal, expert opinion may supplement the available evidence to recommend imaging or treatment.

Analysis of Cervical Angle in the Submental Muscular Medialization and Suspension Procedure.

IMPORTANCE: The submental muscular medialization and suspension (SMMS) procedure is an option for addressing an obtuse cervical angle in select patients vs the use of traditional rhytidectomy techniques. OBJECTIVE: To compare the change in position of the cervical point between groups undergoing SMMS vs the traditional rhytidectomy technique. DESIGN, SETTING, AND PARTICIPANTS: A retrospective review was performed of 141 patients undergoing rhytidectomy in an accredited private surgery center between January 1, 2013, and December 31, 2014, comparing cervical point depth between patients undergoing SMMS vs those undergoing traditional rhytidectomy with platysma plication. Statistical analysis was performed from November 11, 2017, to January 5, 2018. INTERVENTIONS: Patients underwent either SMMS or traditional rhytidectomy with platysma plication. MAIN OUTCOMES AND MEASURES: The primary end point was change in the cervical point distance between preoperative and postoperative standardized profile photos. RESULTS: A total of 141 patients were included in the analysis. A cohort of 46 patients (43 women and 3 men; mean age, 63.5 years [range, 49.0-79.0 years]) underwent neck contouring with the SMMS technique and a cohort of 95 patients (90 women and 5 men; mean age, 61.0 years [range, 48.0-73.0 years]) underwent traditional rhytidectomy with platysmaplasty. The cervical point distance of the SMMS cohort had a mean (SD) postoperative increase of 2.0 (1.05) cm (95% CI, 1.73-2.28; P < .001) compared with 0.78 (0.82) cm in the traditional rhytidectomy cohort (95% CI, 0.54-1.02; P < .001). CONCLUSIONS AND RELEVANCE: The results of this study suggest that submental muscular medialization and suspension appears to be an effective option to address the obtuse neck in select patients. LEVEL OF EVIDENCE: 3.

Characteristics of Dog Bites in Arkansas.

OBJECTIVE: Dog bite injuries are encountered frequently in emergency departments and can cause significant morbidity. The objective of this study was to explore the associations between the multiple variables at play during these occurrences (eg, the patient's age, the bite location, the bite severity, the dog's relationship with the patient, the breed of dog). METHODS: This two-institution study collected and analyzed dog bite data from Arkansas' only Level I trauma centers. The charts of 740 patients were included in our retrospective chart review. The chart review extracted data, including each individual patient's age, sex, dog bite location, and dog bite severity, as well as the patient's relationship to the dog and the dog's breed. To determine the relation between and among variables, contingency tables were created and analyzed to determine odds ratios (ORs) and confidence intervals (CIs). In addition, standard t tests were used in statistical comparisons of means and proportions. RESULTS: Of the 740 patient charts reviewed, 574 were for patients who presented to Arkansas Children's Hospital and 166 were for patients who presented to the University of Arkansas for Medical Sciences. Of the patients across both institutions, 267 (37.1%) required some form of repair, with 225 (30.4%) receiving closure within the emergency department and 42 (6.7%) requiring an operative intervention. Among children, those younger than age 5 years were >8 times as likely to require an operative repair (OR 8.1, 95% CI 2.77-23.58, P < 0.0001), >4 times as likely to be bitten on the head and neck (OR 4.30, 95% CI 3.00-6.16, P < 0.0001), and ≤3 times as likely to be bitten by a family dog (OR 2.97, 95% CI 2.10-4.20, P < 0.0001). Conversely, children older than age 12 years were >3 times as likely to be bitten on an extremity (OR 3.43, 95% CI 2.08-5.65, P < 0.0001). CONCLUSIONS: The results of this retrospective review are aligned mostly with the general trends found in previous national and global studies, supporting the notion that family dogs represent a more significant threat than often is realized and that, among the breeds identified, pit bulls are proportionally linked with more severe bite injuries. Our data further validate previous studies that note an increased risk of bites and bite severity in children younger than 5 years. In addition, our data show that bites to the head and neck occurred more frequently among children younger than 5 years than among older children, and that boys younger than 5 years were bitten more frequently than girls.

Big data in facial plastic and reconstructive surgery: from large databases to registries.

PURPOSE OF REVIEW: There are many limitations to performing clinical research with high levels of evidence in facial plastic and reconstructive surgery (FPRS), such as randomization into surgical groups and sample size recruitment. Therefore, additional avenues for exploring research should be explored using big data, from databases to registries. Other organizations have developed these tools in the evolving landscape of outcomes measurement and value in healthcare, which may serve as models for our specialty. RECENT FINDINGS: Over the last 5 years, FPRS literature of large-scale outcomes research, utilizing several administrative databases, has steadily grown. Our objectives are to describe key administrative databases, strengths and weaknesses of each, and identify recent FPRS publications utilizing big data. A registry with FPRS defined outcomes has the most potential. SUMMARY: Although FPRS research has trended to a more evidence-based approach in the modern healthcare era, gaps persist. Several large administrative databases or registries can address voids in outcomes research within FPRS.

Retropharyngeal and Parapharyngeal Abscesses Among Children and Adolescents in the United States: Epidemiology and Management Trends, 2003-2012.

INTRODUCTION: The epidemiology and hospital course of children with retropharyngeal abscess (RPA) or parapharyngeal abscess (PPA) have not been fully described at the national level in the United States. METHODS: Pediatric discharges for PPA and RPA were evaluated by using the Kids' Inpatient Database from 2003, 2006, 2009, and 2012. Cases were identified by using International Classification of Disease, Ninth Revision, Clinical Modification codes 478.22 and 478.24 for PPA and RPA, respectively. Nationally representative incidence data were calculated by using weighted case estimates and US census data. Demographic and cost analyses were conducted by using unweighted analyses. RESULTS: There were 2685 hospital discharges for PPA and 6233 hospital discharges for RPA during the 4 study years combined. The incidence of RPA increased from 2.98 per 100 000 population among children <20 years old in 2003 to 4.10 per 100 000 in 2012. The incidence of PPA peaked at 1.49 per 100 000 in 2006. Incidences were highest among children <5 years old and boys in all age groups for PPA and RPA. Winter-to-spring seasonality also was evident for both. PPA was managed surgically in 58.1% of the cases, and RPA was managed surgically in 46.7%. Surgery was performed most often on the day of admission or the following day, was more frequent at teaching hospitals, and was associated with higher hospital charges. The mean hospital length of stay was longer for children who had surgery versus those who did not (4.4 vs 3.1 days [for PPA] and 4.8 vs 3.2 days [for RPA], respectively; both P < .001). The median charges for RPA and PPA were similar. The proportions of children with RPA or PPA covered by Medicaid increased during the study period. CONCLUSION: PPA and RPA represent relatively common male-predominant childhood infections with similar epidemiologies. The incidence of hospital discharges with a diagnosis of RPA increased during the study period. Substantial proportions of children with PPA or RPA are now managed without surgery. Surgical drainage was associated with higher hospital charges and longer lengths of stay.

Cell-Based Therapy in TBI: Magnetic Retention of Neural Stem Cells In Vivo.

Stem cell therapy is under active investigation for traumatic brain injury (TBI). Noninvasive stem cell delivery is the preferred method, but retention of stem cells at the site of injury in TBI has proven challenging and impacts effectiveness. To investigate the effects of applying a magnetic field on cell homing and retention, we delivered human neuroprogenitor cells (hNPCs) labeled with a superparamagnetic nanoparticle into post-TBI animals in the presence of a static magnetic field. We have previously devised a method of loading hNPCs with ultrasmall superparamagnetic iron oxide (USPIO) nanoparticles Molday ION Rhodamine B (MIRB™). Labeling of hNPCs (MIRB-hNPCs) does not affect hNPC viability, proliferation, or differentiation. The 0.6 tesla (T) permanent magnet was placed ∼4 mm above the injured parietal cortex prior to intracarotid injection of 4 × 10(4) MIRB-hNPCs. Fluorescence imaging, Perls' Prussian blue histochemistry, immunocytochemistry with SC121, a human-specific antibody, and T2-weighted magnetic resonance imaging ex vivo revealed there was increased homing and retention of MIRB-hNPCs in the injured cortex as compared to the control group in which MIRB-hNPCs were injected in the absence of a static magnetic field. Fluoro-Jade C staining and immunolabeling with specific markers confirmed the viability status of MIRB-hNPCs posttransplantation. These results show that increased homing and retention of MIRB-hNPCs post-TBI by applying a static magnetic field is a promising technique to deliver cells into the CNS for treatment of neurological injuries and neurodegenerative diseases.

Prevalence of Clostridium difficile infection presenting to US EDs.

OBJECTIVE: The objective of the study is to determine the prevalence of Clostridium difficile infection (CDI) presenting to emergency departments (EDs) in the United States. Secondary objectives included defining the burden of CDI. METHODS: This is a retrospective, observational cohort study of 2006-2010 Nationwide Emergency Department Sample database of 980 US hospital EDs in 29 states. Prevalence, mortality rate, length of stay, hospital charges, and endemicity were measured. RESULTS: A total of 474513 patients with CDI-related ED visits were identified. From 2006 to 2010, the prevalence of CDI increased from 26.2 to 33.1 per 100,000 population (P<.001). The number of CDI-related ED cases increased 26.1% (P<.001) over the study period: 18.6% from 2006 to 2007 (P<.001), 4.3% from 2007 to 2008 (P=.46), 1.8% from 2008 to 2009 (P=.73), and 0.13% from 2009 to 2010 (P=.95). Emergency department visits occurred more frequently with individuals 85 years or older (relative risk [RR], 13.74; P<.001), females (RR, 1.77; P<.001) and in the northeast United States (RR, 1.42; P<.001). From 2009 to 2010, the mortality rate decreased 17.9% (P=.01). CONCLUSIONS: The prevalence of CDI presenting to EDs increased each year from 2006 to 2010; however, the rate of increase slowed from each year to the next. The mortality rate increased from 2006 to 2009 and decreased significantly from 2009 to 2010. C difficile infection visits presenting to EDs occurred more frequently with older individuals, females, and in the northeast.

Multi-cellular rosettes in the mouse visceral endoderm facilitate the ordered migration of anterior visceral endoderm cells.

The visceral endoderm (VE) is a simple epithelium that forms the outer layer of the egg-cylinder stage mouse embryo. The anterior visceral endoderm (AVE), a specialised subset of VE cells, is responsible for specifying anterior pattern. AVE cells show a stereotypic migratory behaviour within the VE, which is responsible for correctly orientating the anterior-posterior axis. The epithelial integrity of the VE is maintained during the course of AVE migration, which takes place by intercalation of AVE and other VE cells. Though a continuous epithelial sheet, the VE is characterised by two regions of dramatically different behaviour, one showing robust cell movement and intercalation (in which the AVE migrates) and one that is static, with relatively little cell movement and mixing. Little is known about the cellular rearrangements that accommodate and influence the sustained directional movement of subsets of cells (such as the AVE) within epithelia like the VE. This study uses an interdisciplinary approach to further our understanding of cell movement in epithelia. Using both wild-type embryos as well as mutants in which AVE migration is abnormal or arrested, we show that AVE migration is specifically linked to changes in cell packing in the VE and an increase in multi-cellular rosette arrangements (five or more cells meeting at a point). To probe the role of rosettes during AVE migration, we develop a mathematical model of cell movement in the VE. To do this, we use a vertex-based model, implemented on an ellipsoidal surface to represent a realistic geometry for the mouse egg-cylinder. The potential for rosette formation is included, along with various junctional rearrangements. Simulations suggest that while rosettes are not essential for AVE migration, they are crucial for the orderliness of this migration observed in embryos. Our simulations are similar to results from transgenic embryos in which Planar Cell Polarity (PCP) signalling is disrupted. Such embryos have significantly reduced rosette numbers, altered epithelial packing, and show abnormalities in AVE migration. Our results show that the formation of multi-cellular rosettes in the mouse VE is dependent on normal PCP signalling. Taken together, our model and experimental observations suggest that rosettes in the VE epithelium do not form passively in response to AVE migration. Instead, they are a PCP-dependent arrangement of cells that acts to buffer the disequilibrium in cell packing generated in the VE by AVE migration, enabling AVE cells to migrate in an orderly manner.

Incorporating chemical signalling factors into cell-based models of growing epithelial tissues.

In this paper we present a comprehensive computational framework within which the effects of chemical signalling factors on growing epithelial tissues can be studied. The method incorporates a vertex-based cell model, in conjunction with a solver for the governing chemical equations. The vertex model provides a natural mesh for the finite element method (FEM), with node movements determined by force laws. The arbitrary Lagrangian-Eulerian formulation is adopted to account for domain movement between iterations. The effects of cell proliferation and junctional rearrangements on the mesh are also examined. By implementing refinements of the mesh we show that the finite element (FE) approximation converges towards an accurate numerical solution. The potential utility of the system is demonstrated in the context of Decapentaplegic (Dpp), a morphogen which plays a crucial role in development of the Drosophila imaginal wing disc. Despite the presence of a Dpp gradient, growth is uniform across the wing disc. We make the growth rate of cells dependent on Dpp concentration and show that the number of proliferation events increases in regions of high concentration. This allows hypotheses regarding mechanisms of growth control to be rigorously tested. The method we describe may be adapted to a range of potential application areas, and to other cell-based models with designated node movements, to accurately probe the role of morphogens in epithelial tissues.