Lessons Learned in Building Expertly Annotated Multi-Institution Datasets and Hosting the RSNA AI Challenges

The Radiological Society of North America (RSNA) has held artificial intelligence competitions to tackle real-world medical imaging problems at least annually since 2017. This article examines the challenges and processes involved in organizing these competitions, with a specific emphasis on the creation and curation of high-quality datasets. The collection of diverse and representative medical imaging data involves dealing with issues of patient privacy and data security. Furthermore, ensuring quality and consistency in data, which includes expert labeling and accounting for various patient and imaging characteristics, necessitates substantial planning and resources. Overcoming these obstacles requires meticulous project management and adherence to strict timelines. The article also highlights the potential of crowdsourced annotation to progress medical imaging research. Through the RSNA competitions, an effective global engagement has been realized, resulting in innovative solutions to complex medical imaging problems, thus potentially transforming health care by enhancing diagnostic accuracy and patient outcomes. Keywords: Use of AI in Education, Artificial Intelligence © RSNA, 2024.

GenomeMUSter mouse genetic variation service enables multitrait, multipopulation data integration and analysis

Hundreds of inbred mouse strains and intercross populations have been used to characterize the function of genetic variants that contribute to disease. Thousands of disease-relevant traits have been characterized in mice and made publicly available. New strains and populations including consomics, the collaborative cross, expanded BXD, and inbred wild-derived strains add to existing complex disease mouse models, mapping populations, and sensitized backgrounds for engineered mutations. The genome sequences of inbred strains, along with dense genotypes from others, enable integrated analysis of trait-variant associations across populations, but these analyses are hampered by the sparsity of genotypes available. Moreover, the data are not readily interoperable with other resources. To address these limitations, we created a uniformly dense variant resource by harmonizing multiple data sets. Missing genotypes were imputed using the Viterbi algorithm with a data-driven technique that incorporates local phylogenetic information, an approach that is extendable to other model organisms. The result is a web- and programmatically accessible data service called GenomeMUSter, comprising single-nucleotide variants covering 657 strains at 106.8 million segregating sites. Interoperation with phenotype databases, analytic tools, and other resources enable a wealth of applications, including multitrait, multipopulation meta-analysis. We show this in cross-species comparisons of type 2 diabetes and substance use disorder meta-analyses, leveraging mouse data to characterize the likely role of human variant effects in disease. Other applications include refinement of mapped loci and prioritization of strain backgrounds for disease modeling to further unlock extant mouse diversity for genetic and genomic studies in health and disease.

Performance of the Winning Algorithms of the RSNA 2022 Cervical Spine Fracture Detection Challenge

Purpose To evaluate and report the performance of the winning algorithms of the Radiological Society of North America Cervical Spine Fracture AI Challenge. Materials and Methods The competition was open to the public on Kaggle from July 28 to October 27, 2022. A sample of 3112 CT scans with and without cervical spine fractures (CSFx) were assembled from multiple sites (12 institutions across six continents) and prepared for the competition. The test set had 1093 scans (private test set: n = 789; mean age, 53.40 years ± 22.86 [SD]; 509 males; public test set: n = 304; mean age, 52.51 years ± 20.73; 189 males) and 847 fractures. The eight top-performing artificial intelligence (AI) algorithms were retrospectively evaluated, and the area under the receiver operating characteristic curve (AUC) value, F1 score, sensitivity, and specificity were calculated. Results A total of 1108 contestants composing 883 teams worldwide participated in the competition. The top eight AI models showed high performance, with a mean AUC value of 0.96 (95% CI: 0.95, 0.96), mean F1 score of 90% (95% CI: 90%, 91%), mean sensitivity of 88% (95% Cl: 86%, 90%), and mean specificity of 94% (95% CI: 93%, 96%). The highest values reported for previous models were an AUC of 0.85, F1 score of 81%, sensitivity of 76%, and specificity of 97%. Conclusion The competition successfully facilitated the development of AI models that could detect and localize CSFx on CT scans with high performance outcomes, which appear to exceed known values of previously reported models. Further study is needed to evaluate the generalizability of these models in a clinical environment. Keywords: Cervical Spine, Fracture Detection, Machine Learning, Artificial Intelligence Algorithms, CT, Head/Neck Supplemental material is available for this article. © RSNA, 2024.

Mouse phenome database: curated data repository with interactive multi-population and multi-trait analyses

The Mouse Phenome Database continues to serve as a curated repository and analysis suite for measured attributes of members of diverse mouse populations. The repository includes annotation to community standard ontologies and guidelines, a database of allelic states for 657 mouse strains, a collection of protocols, and analysis tools for flexible, interactive, user directed analyses that increasingly integrates data across traits and populations. The database has grown from its initial focus on a standard set of inbred strains to include heterogeneous mouse populations such as the Diversity Outbred and mapping crosses and well as Collaborative Cross, Hybrid Mouse Diversity Panel, and recombinant inbred strains. Most recently the system has expanded to include data from the International Mouse Phenotyping Consortium. Collectively these data are accessible by API and provided with an interactive tool suite that enables users' persistent selection, storage, and operation on collections of measures. The tool suite allows basic analyses, advanced functions with dynamic visualization including multi-population meta-analysis, multivariate outlier detection, trait pattern matching, correlation analyses and other functions. The data resources and analysis suite provide users a flexible environment in which to explore the basis of phenotypic variation in health and disease across the lifespan.

The Ontology of Biological Attributes (OBA)-computational traits for the life sciences

Existing phenotype ontologies were originally developed to represent phenotypes that manifest as a character state in relation to a wild-type or other reference. However, these do not include the phenotypic trait or attribute categories required for the annotation of genome-wide association studies (GWAS), Quantitative Trait Loci (QTL) mappings or any population-focussed measurable trait data. The integration of trait and biological attribute information with an ever increasing body of chemical, environmental and biological data greatly facilitates computational analyses and it is also highly relevant to biomedical and clinical applications. The Ontology of Biological Attributes (OBA) is a formalised, species-independent collection of interoperable phenotypic trait categories that is intended to fulfil a data integration role. OBA is a standardised representational framework for observable attributes that are characteristics of biological entities, organisms, or parts of organisms. OBA has a modular design which provides several benefits for users and data integrators, including an automated and meaningful classification of trait terms computed on the basis of logical inferences drawn from domain-specific ontologies for cells, anatomical and other relevant entities. The logical axioms in OBA also provide a previously missing bridge that can computationally link Mendelian phenotypes with GWAS and quantitative traits. The term components in OBA provide semantic links and enable knowledge and data integration across specialised research community boundaries, thereby breaking silos.

The RSNA Cervical Spine Fracture CT Dataset

This dataset is composed of cervical spine CT images with annotations related to fractures; it is available at https://www.kaggle.com/competitions/rsna-2022-cervical-spine-fracture-detection/.

GenomeMUSter mouse genetic variation service enables multi-trait, multi-population data integration and analyses

Hundreds of inbred laboratory mouse strains and intercross populations have been used to functionalize genetic variants that contribute to disease. Thousands of disease relevant traits have been characterized in mice and made publicly available. New strains and populations including the Collaborative Cross, expanded BXD and inbred wild-derived strains add to set of complex disease mouse models, genetic mapping resources and sensitized backgrounds against which to evaluate engineered mutations. The genome sequences of many inbred strains, along with dense genotypes from others could allow integrated analysis of trait - variant associations across populations, but these analyses are not feasible due to the sparsity of genotypes available. Moreover, the data are not readily interoperable with other resources. To address these limitations, we created a uniformly dense data resource by harmonizing multiple variant datasets. Missing genotypes were imputed using the Viterbi algorithm with a data-driven technique that incorporates local phylogenetic information, an approach that is extensible to other model organism species. The result is a web- and programmatically-accessible data service called GenomeMUSter ( https://muster.jax.org ), comprising allelic data covering 657 strains at 106.8M segregating sites. Interoperation with phenotype databases, analytic tools and other resources enable a wealth of applications including multi-trait, multi-population meta-analysis. We demonstrate this in a cross-species comparison of the meta-analysis of Type 2 Diabetes and of substance use disorders, resulting in the more specific characterization of the role of human variant effects in light of mouse phenotype data. Other applications include refinement of mapped loci and prioritization of strain backgrounds for disease modeling to further unlock extant mouse diversity for genetic and genomic studies in health and disease.

Variation in histone configurations correlates with gene expression across nine inbred strains of mice

The Diversity Outbred (DO) mice and their inbred founders are widely used models of human disease. However, although the genetic diversity of these mice has been well documented, their epigenetic diversity has not. Epigenetic modifications, such as histone modifications and DNA methylation, are important regulators of gene expression and, as such, are a critical mechanistic link between genotype and phenotype. Therefore, creating a map of epigenetic modifications in the DO mice and their founders is an important step toward understanding mechanisms of gene regulation and the link to disease in this widely used resource. To this end, we performed a strain survey of epigenetic modifications in hepatocytes of the DO founders. We surveyed four histone modifications (H3K4me1, H3K4me3, H3K27me3, and H3K27ac), as well as DNA methylation. We used ChromHMM to identify 14 chromatin states, each of which represents a distinct combination of the four histone modifications. We found that the epigenetic landscape is highly variable across the DO founders and is associated with variation in gene expression across strains. We found that epigenetic state imputed into a population of DO mice recapitulated the association with gene expression seen in the founders, suggesting that both histone modifications and DNA methylation are highly heritable mechanisms of gene expression regulation. We illustrate how DO gene expression can be aligned with inbred epigenetic states to identify putative cis-regulatory regions. Finally, we provide a data resource that documents strain-specific variation in the chromatin state and DNA methylation in hepatocytes across nine widely used strains of laboratory mice.

A genetic locus complements resistance to Bordetella pertussis-induced histamine sensitization

Histamine plays pivotal role in normal physiology and dysregulated production of histamine or signaling through histamine receptors (HRH) can promote pathology. Previously, we showed that Bordetella pertussis or pertussis toxin can induce histamine sensitization in laboratory inbred mice and is genetically controlled by Hrh1/HRH1. HRH1 allotypes differ at three amino acid residues with P263-V313-L331 and L263-M313-S331, imparting sensitization and resistance respectively. Unexpectedly, we found several wild-derived inbred strains that carry the resistant HRH1 allotype (L263-M313-S331) but exhibit histamine sensitization. This suggests the existence of a locus modifying pertussis-dependent histamine sensitization. Congenic mapping identified the location of this modifier locus on mouse chromosome 6 within a functional linkage disequilibrium domain encoding multiple loci controlling sensitization to histamine. We utilized interval-specific single-nucleotide polymorphism (SNP) based association testing across laboratory and wild-derived inbred mouse strains and functional prioritization analyses to identify candidate genes for this modifier locus. Atg7, Plxnd1, Tmcc1, Mkrn2, Il17re, Pparg, Lhfpl4, Vgll4, Rho and Syn2 are candidate genes within this modifier locus, which we named Bphse, enhancer of Bordetella pertussis induced histamine sensitization. Taken together, these results identify, using the evolutionarily significant diversity of wild-derived inbred mice, additional genetic mechanisms controlling histamine sensitization.

Mouse Phenome Database: towards a more FAIR-compliant and TRUST-worthy data repository and tool suite for phenotypes and genotypes

The Mouse Phenome Database (MPD; https://phenome.jax.org; RRID:SCR_003212), supported by the US National Institutes of Health, is a Biomedical Data Repository listed in the Trans-NIH Biomedical Informatics Coordinating Committee registry. As an increasingly FAIR-compliant and TRUST-worthy data repository, MPD accepts phenotype and genotype data from mouse experiments and curates, organizes, integrates, archives, and distributes those data using community standards. Data are accompanied by rich metadata, including widely used ontologies and detailed protocols. Data are from all over the world and represent genetic, behavioral, morphological, and physiological disease-related characteristics in mice at baseline or those exposed to drugs or other treatments. MPD houses data from over 6000 strains and populations, representing many reproducible strain types and heterogenous populations such as the Diversity Outbred where each mouse is unique but can be genotyped throughout the genome. A suite of analysis tools is available to aggregate, visualize, and analyze these data within and across studies and populations in an increasingly traceable and reproducible manner. We have refined existing resources and developed new tools to continue to provide users with access to consistent, high-quality data that has translational relevance in a modernized infrastructure that enables interaction with a suite of bioinformatics analytic and data services.

Electronic decision support and diarrhoeal disease guideline adherence (mHDM): a cluster randomised controlled trial

BACKGROUND

Acute diarrhoeal disease management often requires rehydration alone without antibiotics. However, non-indicated antibiotics are frequently ordered and this is an important driver of antimicrobial resistance. The mHealth Diarrhoea Management (mHDM) trial aimed to establish whether electronic decision support improves rehydration and antibiotic guideline adherence in resource-limited settings.

METHODS

A cluster randomised controlled trial was done at ten district hospitals in Bangladesh. Inclusion criteria were patients aged 2 months or older with uncomplicated acute diarrhoea. Admission orders were observed without intervention in the pre-intervention period, followed by randomisation to electronic (rehydration calculator) or paper formatted WHO guidelines for the intervention period. The primary outcome was rate of intravenous fluid ordered as a binary variable. Generalised linear mixed-effect models, accounting for hospital clustering, served as the analytical framework; the analysis was intention to treat. The trial is registered with ClinicalTrials.gov (NCT03154229) and is completed.

FINDINGS

From March 11 to Sept 10, 2018, 4975 patients (75·6%) of 6577 screened patients were enrolled. The intervention effect for the primary outcome showed no significant differences in rates of intravenous fluids ordered as a function of decision-support type. Intravenous fluid orders decreased by 0·9 percentage points for paper electronic decision support and 4·2 percentage points for electronic decision support, with a 4·2-point difference between decision-support types in the intervention period (paper 98·7% [95% CI 91·8-99·8] vs electronic 94·5% [72·2-99·1]; p_interaction=0·31). Adverse events such as complications and mortality events were uncommon and could not be statistically estimated.

INTERPRETATION

Although intravenous fluid orders did not change, electronic decision support was associated with increases in the volume of intravenous fluid ordered and decreases in antibiotics ordered, which are consistent with WHO guidelines.

FUNDING

US National Institutes of Health.

The RSNA Pulmonary Embolism CT Dataset

Supplemental material is available for this article.

CheXaid: deep learning assistance for physician diagnosis of tuberculosis using chest x-rays in patients with HIV

Tuberculosis (TB) is the leading cause of preventable death in HIV-positive patients, and yet often remains undiagnosed and untreated. Chest x-ray is often used to assist in diagnosis, yet this presents additional challenges due to atypical radiographic presentation and radiologist shortages in regions where co-infection is most common. We developed a deep learning algorithm to diagnose TB using clinical information and chest x-ray images from 677 HIV-positive patients with suspected TB from two hospitals in South Africa. We then sought to determine whether the algorithm could assist clinicians in the diagnosis of TB in HIV-positive patients as a web-based diagnostic assistant. Use of the algorithm resulted in a modest but statistically significant improvement in clinician accuracy (p = 0.002), increasing the mean clinician accuracy from 0.60 (95% CI 0.57, 0.63) without assistance to 0.65 (95% CI 0.60, 0.70) with assistance. However, the accuracy of assisted clinicians was significantly lower (p < 0.001) than that of the stand-alone algorithm, which had an accuracy of 0.79 (95% CI 0.77, 0.82) on the same unseen test cases. These results suggest that deep learning assistance may improve clinician accuracy in TB diagnosis using chest x-rays, which would be valuable in settings with a high burden of HIV/TB co-infection. Moreover, the high accuracy of the stand-alone algorithm suggests a potential value particularly in settings with a scarcity of radiological expertise.

Great variability in donor heart acceptance practices across the United States

Disparities in organ acceptance practices exacerbate donor heart nonuse and lead to increased waiting times and mortality for heart transplant candidates. We studied disparities in donor heart acceptance among US transplant centers and their relations to posttransplant outcomes. Candidate, potential transplant recipient match run, and deceased donor data were obtained from the United Network for Organ Sharing. We analyzed donor, candidate, and transplant center characteristics with respect to organ acceptance, offer acceptance, number of offers before acceptance (organ sequence number), and association with posttransplant mortality. A total of 693 420 donor heart offers made between April 2007 and December 2015 were included. We identified great variability in donor heart acceptance practices among US heart transplant centers. We identified donor and recipient characteristics that were strongly associated with heart organ and offer acceptance, and organ sequence number, and identified inconsistencies among centers with respect to how these characteristics influenced acceptance decisions. Finally, we identified characteristics that were highly predictive of donor heart nonuse and were not associated with increased recipient mortality, which may guide future efforts aimed at increasing use of available hearts for transplantation.

PENet-a scalable deep-learning model for automated diagnosis of pulmonary embolism using volumetric CT imaging

Pulmonary embolism (PE) is a life-threatening clinical problem and computed tomography pulmonary angiography (CTPA) is the gold standard for diagnosis. Prompt diagnosis and immediate treatment are critical to avoid high morbidity and mortality rates, yet PE remains among the diagnoses most frequently missed or delayed. In this study, we developed a deep learning model-PENet, to automatically detect PE on volumetric CTPA scans as an end-to-end solution for this purpose. The PENet is a 77-layer 3D convolutional neural network (CNN) pretrained on the Kinetics-600 dataset and fine-tuned on a retrospective CTPA dataset collected from a single academic institution. The PENet model performance was evaluated in detecting PE on data from two different institutions: one as a hold-out dataset from the same institution as the training data and a second collected from an external institution to evaluate model generalizability to an unrelated population dataset. PENet achieved an AUROC of 0.84 [0.82-0.87] on detecting PE on the hold out internal test set and 0.85 [0.81-0.88] on external dataset. PENet also outperformed current state-of-the-art 3D CNN models. The results represent successful application of an end-to-end 3D CNN model for the complex task of PE diagnosis without requiring computationally intensive and time consuming preprocessing and demonstrates sustained performance on data from an external institution. Our model could be applied as a triage tool to automatically identify clinically important PEs allowing for prioritization for diagnostic radiology interpretation and improved care pathways via more efficient diagnosis.

Dose-Response Relationship Between Intracoronary Acetylcholine and Minimal Lumen Diameter in Coronary Endothelial Function Testing of Women and Men With Angina and No Obstructive Coronary Artery Disease

BACKGROUND

Intracoronary acetylcholine (Ach) provocation testing is the gold standard for assessing coronary endothelial function. However, dosing regimens of Ach are quite varied in the literature, and there are limited data evaluating the optimal dose. We evaluated the dose-response relationship between Ach and minimal lumen diameter (MLD) by sex and studied whether incremental intracoronary Ach doses given during endothelial function testing improve its diagnostic utility.

METHODS

We evaluated 65 men and 212 women with angina and no obstructive coronary artery disease who underwent endothelial function testing using the highest tolerable dose of intracoronary Ach, up to 200 μg. Epicardial endothelial dysfunction was defined as a decrease in MLD >20% after intracoronary Ach by quantitative coronary angiography. We used a linear mixed effects model to evaluate the dose-response relationship. Deming regression analysis was done to compare the %MLD constriction after incremental doses of intracoronary Ach.

RESULTS

The mean age was 53.5 years. Endothelial dysfunction was present in 186 (68.1%). Among men with endothelial dysfunction, there was a significant decrease in MLD/10 µg of Ach at doses above 50 μg and 100 µg, while this decrease in MLD was not observed in women (P<0.001). The %MLD constriction at 20 μg versus 50 μg and 50 μg versus 100 μg were not equivalent while the %MLD constriction at 100 μg versus 200 μg were equivalent.

CONCLUSIONS

Women and men appear to have different responses to Ach during endothelial function testing. In addition to having a greater response to intracoronary Ach at all doses, men also demonstrate an Ach-MLD dose-response relationship with doses up to 200 μg, while women have minimal change in MLD with doses above 50 µg. An incremental dosing regimen during endothelial function testing appears to improve the diagnostic utility of the test and should be adjusted based on the sex of the patient.

Impact of a deep learning assistant on the histopathologic classification of liver cancer

Artificial intelligence (AI) algorithms continue to rival human performance on a variety of clinical tasks, while their actual impact on human diagnosticians, when incorporated into clinical workflows, remains relatively unexplored. In this study, we developed a deep learning-based assistant to help pathologists differentiate between two subtypes of primary liver cancer, hepatocellular carcinoma and cholangiocarcinoma, on hematoxylin and eosin-stained whole-slide images (WSI), and evaluated its effect on the diagnostic performance of 11 pathologists with varying levels of expertise. Our model achieved accuracies of 0.885 on a validation set of 26 WSI, and 0.842 on an independent test set of 80 WSI. Although use of the assistant did not change the mean accuracy of the 11 pathologists (p = 0.184, OR = 1.281), it significantly improved the accuracy (p = 0.045, OR = 1.499) of a subset of nine pathologists who fell within well-defined experience levels (GI subspecialists, non-GI subspecialists, and trainees). In the assisted state, model accuracy significantly impacted the diagnostic decisions of all 11 pathologists. As expected, when the model's prediction was correct, assistance significantly improved accuracy (p = 0.000, OR = 4.289), whereas when the model's prediction was incorrect, assistance significantly decreased accuracy (p = 0.000, OR = 0.253), with both effects holding across all pathologist experience levels and case difficulty levels. Our results highlight the challenges of translating AI models into the clinical setting, and emphasize the importance of taking into account potential unintended negative consequences of model assistance when designing and testing medical AI-assistance tools.

Asymmetric dimethylarginine predicts impaired epicardial coronary vasomotion in patients with angina in the absence of obstructive coronary artery disease

BACKGROUND

Impaired epicardial coronary vasomotion is a potential mechanism of angina and a predictor of adverse cardiovascular outcomes in patients without angiographic evidence of obstructive coronary artery disease (CAD). We sought to evaluate the association of asymmetric dimethylarginine (ADMA)-a marker of nitric oxide-mediated vascular dysfunction-with epicardial coronary vasomotor dysfunction in this select population.

METHODS

Invasive testing for epicardial vasomotor dysfunction was performed using intracoronary acetylcholine in the left anterior descending coronary artery. Impaired vasomotor response was defined as a luminal constriction of >20% on quantitative coronary angiography. Plasma ADMA levels were measured using high performance liquid chromatography. A robust multivariate linear mixed-effect model approach and Akaike information criterion were used to determine predictors of vasomotor dysfunction.

RESULTS

In 191 patients with angina in the absence of obstructive CAD, abnormal epicardial vasomotion was observed in 137 (71.7%) patients. Median ADMA rose as the extent of impairment progressed: none (0.48 [0.44-0.59] μM), any (0.51 [0.46-0.60] μM, p = 0.12), focal (0.54 [0.49,0.61] μM, p = 0.17), and diffuse (0.55 [0.49,0.63] μM, p = 0.02). In unadjusted analysis, ADMA was highly predictive of vasomotor dysfunction (χ²=15.1, p = 0.002). Notably, ADMA remained a significant predictor even after adjusting for other factors in the best fit model (χ²=10.0, p = 0.02).

CONCLUSIONS

ADMA is an independent predictor of epicardial coronary vasomotor dysfunction in patients with angina in the absence of obstructive CAD. These data support a very early mechanistic role of ADMA in the continuum of atherosclerotic heart disease.

Deep Learning-Assisted Diagnosis of Cerebral Aneurysms Using the HeadXNet Model

IMPORTANCE

Deep learning has the potential to augment clinician performance in medical imaging interpretation and reduce time to diagnosis through automated segmentation. Few studies to date have explored this topic.

OBJECTIVE

To develop and apply a neural network segmentation model (the HeadXNet model) capable of generating precise voxel-by-voxel predictions of intracranial aneurysms on head computed tomographic angiography (CTA) imaging to augment clinicians' intracranial aneurysm diagnostic performance.

DESIGN, SETTING, AND PARTICIPANTS

In this diagnostic study, a 3-dimensional convolutional neural network architecture was developed using a training set of 611 head CTA examinations to generate aneurysm segmentations. Segmentation outputs from this support model on a test set of 115 examinations were provided to clinicians. Between August 13, 2018, and October 4, 2018, 8 clinicians diagnosed the presence of aneurysm on the test set, both with and without model augmentation, in a crossover design using randomized order and a 14-day washout period. Head and neck examinations performed between January 3, 2003, and May 31, 2017, at a single academic medical center were used to train, validate, and test the model. Examinations positive for aneurysm had at least 1 clinically significant, nonruptured intracranial aneurysm. Examinations with hemorrhage, ruptured aneurysm, posttraumatic or infectious pseudoaneurysm, arteriovenous malformation, surgical clips, coils, catheters, or other surgical hardware were excluded. All other CTA examinations were considered controls.

MAIN OUTCOMES AND MEASURES

Sensitivity, specificity, accuracy, time, and interrater agreement were measured. Metrics for clinician performance with and without model augmentation were compared.

RESULTS

The data set contained 818 examinations from 662 unique patients with 328 CTA examinations (40.1%) containing at least 1 intracranial aneurysm and 490 examinations (59.9%) without intracranial aneurysms. The 8 clinicians reading the test set ranged in experience from 2 to 12 years. Augmenting clinicians with artificial intelligence-produced segmentation predictions resulted in clinicians achieving statistically significant improvements in sensitivity, accuracy, and interrater agreement when compared with no augmentation. The clinicians' mean sensitivity increased by 0.059 (95% CI, 0.028-0.091; adjusted P = .01), mean accuracy increased by 0.038 (95% CI, 0.014-0.062; adjusted P = .02), and mean interrater agreement (Fleiss κ) increased by 0.060, from 0.799 to 0.859 (adjusted P = .05). There was no statistically significant change in mean specificity (0.016; 95% CI, -0.010 to 0.041; adjusted P = .16) and time to diagnosis (5.71 seconds; 95% CI, 7.22-18.63 seconds; adjusted P = .19).

CONCLUSIONS AND RELEVANCE

The deep learning model developed successfully detected clinically significant intracranial aneurysms on CTA. This suggests that integration of an artificial intelligence-assisted diagnostic model may augment clinician performance with dependable and accurate predictions and thereby optimize patient care.

Factors influencing infarct growth including collateral status assessed using computed tomography in acute stroke patients with large artery occlusion

In major ischemic stroke caused by a large artery occlusion, neuronal loss varies considerably across individuals without revascularization. This study aims to identify which patient characteristics are most highly associated with this variability. Demographic and clinical information were retrospectively collected on a registry of 878 patients. Imaging biomarkers including Alberta Stroke Program Early CT score, noncontrast head computed tomography infarct volume, perfusion computed tomography infarct core and penumbra, occlusion site, collateral score, and recanalization status were evaluated on the baseline and early follow-up computed tomography images. Infarct growth rates were calculated by dividing infarct volumes by the time elapsed between the computed tomography scan and the symptom onset. Collateral score was graded into four levels (0, 1, 2, and 3) in comparison with the normal side. Correlation of perfusion computed tomography and noncontrast head computed tomography infarct volumes and infarct growth rates were estimated with the nonparametric Spearman's rank correlation. Conditional inference trees were used to identify the clinical and imaging biomarkers that were most highly associated with the infarct growth rate and modified Rankin Scale at 90 days. Two hundred and thirty-two patients met the inclusion criteria for this study. The median infarct growth rates for perfusion computed tomography and noncontrast head computed tomography were 11.2 and 6.2 ml/log(min) in logarithmic model, and 18.9 and 10.4 ml/h in linear model, respectively. Noncontrast head computed tomography and perfusion computed tomography infarct volumes and infarct growth rates were significantly correlated (rho=0.53; P < 0.001). Collateral status was the strongest predictor for infarct growth rates. For collateral=0, the perfusion computed tomography and noncontrast head computed tomography infarct growth rate were 31.56 and 16.86 ml/log(min), respectively. Patients who had collateral >0 and penumbra volumes>92 ml had the lowest predicted perfusion computed tomography infarct growth rates (6.61 ml/log(min)). Collateral status was closely related to the diversity of infarct growth rates, poor collaterals were associated with a faster infarct growth rates and vice versa.

A statistical approach to identify optimal inclusion criteria: An application to acute stroke clinical trials

Purpose

To develop a statistical approach that compares patient selection strategies across clinical trials and apply this approach to acute ischemic stroke clinical trials to identify the optimal inclusion criteria.

Methods

We developed a statistical approach that compares the number needed to treat to achieve one success (NNT) along with the number needed to screen to achieve one success (NNS) and assesses if there are significant differences in inclusion criteria, treatment course, and clinical outcome among patients that may have been included/excluded in the trials. We applied this approach to the study population from four recent positive acute stroke clinical trials: MR CLEAN, EXTEND-IA, ESCAPE, and SWIFT PRIME, applying published trial criteria to an independent registry of 612 acute stroke patients, since we did not have access to the complete trial data.

Results

Although reported NNT were similar for EXTEND-IA, SWIFT PRIME and ESCAPE, and somewhat higher for MR CLEAN, NNS varied across the trials from 21 for EXTEND-IA, 27 for MR CLEAN, to 46 for ESCAPE and 64 for SWIFT PRIME, reflecting less and more stringent inclusion criteria, respectively. Although there were significant differences in imaging biomarkers and other clinical characteristics among patients that may have been included/excluded in the trials, these differences did not translate to significant differences in treatment course or clinical outcomes.

Conclusions

Our study proposes a robust statistical approach that can be applied to a larger pooled trial dataset, if made available, to objectively compare across clinical trials and inform inclusion criteria of future trials. Pooled analysis of the acute stroke trial data is needed to determine which imaging biomarker inclusion criteria are critical and which may be relaxed. If this procedure were applied across the pooled trial data, it could decrease costs and refine the design of future trials to be the most efficacious for the greatest number of patients.

Uncoupling of transcriptomic and cytological differentiation in mouse spermatocytes with impaired meiosis

Cell differentiation is driven by changes in gene expression that manifest as changes in cellular phenotype or function. Altered cellular phenotypes, stemming from genetic mutations or other perturbations, are widely assumed to directly correspond to changes in the transcriptome and vice versa. Here, we exploited the cytologically well-defined Prdm9 mutant mouse as a model of developmental arrest to test whether parallel programs of cellular differentiation and gene expression are tightly coordinated, or can be disassociated. By comparing cytological phenotype markers and transcriptomes in wild-type and mutant spermatocytes, we identified multiple instances of cellular and molecular uncoupling in Prdm9^–/– mutants. Most notably, although Prdm9^–/– germ cells undergo cytological arrest in a late-leptotene/zygotene stage, they nevertheless develop gene expression signatures characteristic of later developmental substages. These findings suggest that transcriptomic changes may not reliably map to cellular phenotypes in developmentally perturbed systems.

Assessing the Relationship Between American Heart Association Atherosclerotic Cardiovascular Disease Risk Score and Coronary Artery Imaging Findings

OBJECTIVE

The aim of this study was to characterize the relationship between computed tomography angiography imaging characteristics of coronary artery and atherosclerotic cardiovascular disease (ASCVD) score.

METHODS

We retrospectively identified all patients who underwent a coronary computed tomography angiography at our institution from December 2013 to July 2016, then we calculated the 10-year ASCVD score. We characterized the relationship between coronary artery imaging findings and ASCVD risk score.

RESULTS

One hundred fifty-one patients met our inclusion criteria. Patients with a 10-year ASCVD score of 7.5% or greater had significantly more arterial segments showing stenosis (46.4%, P = 0.008) and significantly higher maximal plaque thickness (1.25 vs 0.53, P = 0.001). However, among 56 patients with a 10-year ASCVD score of 7.5% or greater, 30 (53.6%) had no arterial stenosis. Furthermore, among the patients with a 10-year ASCVD score of less than 7.5%, 24 (25.3%) had some arterial stenosis.

CONCLUSIONS

There is some concordance but not a perfect overlap between 10-year ASCVD risk scores and coronary artery imaging findings.

Deep learning for chest radiograph diagnosis: A retrospective comparison of the CheXNeXt algorithm to practicing radiologists

BACKGROUND

Chest radiograph interpretation is critical for the detection of thoracic diseases, including tuberculosis and lung cancer, which affect millions of people worldwide each year. This time-consuming task typically requires expert radiologists to read the images, leading to fatigue-based diagnostic error and lack of diagnostic expertise in areas of the world where radiologists are not available. Recently, deep learning approaches have been able to achieve expert-level performance in medical image interpretation tasks, powered by large network architectures and fueled by the emergence of large labeled datasets. The purpose of this study is to investigate the performance of a deep learning algorithm on the detection of pathologies in chest radiographs compared with practicing radiologists.

METHODS AND FINDINGS

We developed CheXNeXt, a convolutional neural network to concurrently detect the presence of 14 different pathologies, including pneumonia, pleural effusion, pulmonary masses, and nodules in frontal-view chest radiographs. CheXNeXt was trained and internally validated on the ChestX-ray8 dataset, with a held-out validation set consisting of 420 images, sampled to contain at least 50 cases of each of the original pathology labels. On this validation set, the majority vote of a panel of 3 board-certified cardiothoracic specialist radiologists served as reference standard. We compared CheXNeXt's discriminative performance on the validation set to the performance of 9 radiologists using the area under the receiver operating characteristic curve (AUC). The radiologists included 6 board-certified radiologists (average experience 12 years, range 4-28 years) and 3 senior radiology residents, from 3 academic institutions. We found that CheXNeXt achieved radiologist-level performance on 11 pathologies and did not achieve radiologist-level performance on 3 pathologies. The radiologists achieved statistically significantly higher AUC performance on cardiomegaly, emphysema, and hiatal hernia, with AUCs of 0.888 (95% confidence interval [CI] 0.863-0.910), 0.911 (95% CI 0.866-0.947), and 0.985 (95% CI 0.974-0.991), respectively, whereas CheXNeXt's AUCs were 0.831 (95% CI 0.790-0.870), 0.704 (95% CI 0.567-0.833), and 0.851 (95% CI 0.785-0.909), respectively. CheXNeXt performed better than radiologists in detecting atelectasis, with an AUC of 0.862 (95% CI 0.825-0.895), statistically significantly higher than radiologists' AUC of 0.808 (95% CI 0.777-0.838); there were no statistically significant differences in AUCs for the other 10 pathologies. The average time to interpret the 420 images in the validation set was substantially longer for the radiologists (240 minutes) than for CheXNeXt (1.5 minutes). The main limitations of our study are that neither CheXNeXt nor the radiologists were permitted to use patient history or review prior examinations and that evaluation was limited to a dataset from a single institution.

CONCLUSIONS

In this study, we developed and validated a deep learning algorithm that classified clinically important abnormalities in chest radiographs at a performance level comparable to practicing radiologists. Once tested prospectively in clinical settings, the algorithm could have the potential to expand patient access to chest radiograph diagnostics.

Deep-learning-assisted diagnosis for knee magnetic resonance imaging: Development and retrospective validation of MRNet

BACKGROUND

Magnetic resonance imaging (MRI) of the knee is the preferred method for diagnosing knee injuries. However, interpretation of knee MRI is time-intensive and subject to diagnostic error and variability. An automated system for interpreting knee MRI could prioritize high-risk patients and assist clinicians in making diagnoses. Deep learning methods, in being able to automatically learn layers of features, are well suited for modeling the complex relationships between medical images and their interpretations. In this study we developed a deep learning model for detecting general abnormalities and specific diagnoses (anterior cruciate ligament [ACL] tears and meniscal tears) on knee MRI exams. We then measured the effect of providing the model's predictions to clinical experts during interpretation.

METHODS AND FINDINGS

Our dataset consisted of 1,370 knee MRI exams performed at Stanford University Medical Center between January 1, 2001, and December 31, 2012 (mean age 38.0 years; 569 [41.5%] female patients). The majority vote of 3 musculoskeletal radiologists established reference standard labels on an internal validation set of 120 exams. We developed MRNet, a convolutional neural network for classifying MRI series and combined predictions from 3 series per exam using logistic regression. In detecting abnormalities, ACL tears, and meniscal tears, this model achieved area under the receiver operating characteristic curve (AUC) values of 0.937 (95% CI 0.895, 0.980), 0.965 (95% CI 0.938, 0.993), and 0.847 (95% CI 0.780, 0.914), respectively, on the internal validation set. We also obtained a public dataset of 917 exams with sagittal T1-weighted series and labels for ACL injury from Clinical Hospital Centre Rijeka, Croatia. On the external validation set of 183 exams, the MRNet trained on Stanford sagittal T2-weighted series achieved an AUC of 0.824 (95% CI 0.757, 0.892) in the detection of ACL injuries with no additional training, while an MRNet trained on the rest of the external data achieved an AUC of 0.911 (95% CI 0.864, 0.958). We additionally measured the specificity, sensitivity, and accuracy of 9 clinical experts (7 board-certified general radiologists and 2 orthopedic surgeons) on the internal validation set both with and without model assistance. Using a 2-sided Pearson's chi-squared test with adjustment for multiple comparisons, we found no significant differences between the performance of the model and that of unassisted general radiologists in detecting abnormalities. General radiologists achieved significantly higher sensitivity in detecting ACL tears (p-value = 0.002; q-value = 0.019) and significantly higher specificity in detecting meniscal tears (p-value = 0.003; q-value = 0.019). Using a 1-tailed t test on the change in performance metrics, we found that providing model predictions significantly increased clinical experts' specificity in identifying ACL tears (p-value < 0.001; q-value = 0.006). The primary limitations of our study include lack of surgical ground truth and the small size of the panel of clinical experts.

CONCLUSIONS

Our deep learning model can rapidly generate accurate clinical pathology classifications of knee MRI exams from both internal and external datasets. Moreover, our results support the assertion that deep learning models can improve the performance of clinical experts during medical imaging interpretation. Further research is needed to validate the model prospectively and to determine its utility in the clinical setting.

Assessing the Relationship between Atherosclerotic Cardiovascular Disease Risk Score and Carotid Artery Imaging Findings

BACKGROUND AND PURPOSE

To characterize the relationship between computed tomography angiography (CTA) imaging characteristics of carotid artery and the 10-year risk of atherosclerotic cardiovascular disease (ASCVD) score.

METHODS

We retrospectively identified all patients who underwent a cervical CTA at our institution from January 2013 to July 2016, extracted clinical information, and calculated the 10-year ASCVD score using the Pooled Cohort Equations from the 2013 ACC/AHA guidelines. We compared the imaging features of artery atherosclerosis derived from the CTAs between low and high risk.

RESULTS

One hundred forty-six patients met our inclusion criteria. Patients with an ASCVD score ≥7.5% (64.4%) had significantly more arterial stenosis than patients with an ASCVD score <7.5% (35.6%, P < .001). Maximal plaque thickness was significantly higher (mean 2.33 vs. .42 mm, P < .001) and soft plaques (55.3% vs. 13.5%, P < .001) were significantly more frequent in patients with an ASCVD score ≥7.5%. However, among patients with a 10-year ASCVD score ≥7.5%, 33 (35.1%) had no arterial stenosis, 35 (37.2%) had a maximal plaque thickness less than. 9 mm, and 42 (44.7%) had no soft plaque. Furthermore, among the patients with a 10-year ASCVD score <7.5%, 8 (15.4%) had some arterial stenosis, 8 (15.4%) had a maximal plaque thickness more than. 9 mm, and 7 (13.5%) had soft plaque.

CONCLUSION

There is some concordance but not a perfect overlap between the 10-year ASCVD risk scores calculated from clinical and blood assessment and carotid artery imaging findings.

Deep Learning to Classify Radiology Free-Text Reports

Purpose To evaluate the performance of a deep learning convolutional neural network (CNN) model compared with a traditional natural language processing (NLP) model in extracting pulmonary embolism (PE) findings from thoracic computed tomography (CT) reports from two institutions. Materials and Methods Contrast material-enhanced CT examinations of the chest performed between January 1, 1998, and January 1, 2016, were selected. Annotations by two human radiologists were made for three categories: the presence, chronicity, and location of PE. Classification of performance of a CNN model with an unsupervised learning algorithm for obtaining vector representations of words was compared with the open-source application PeFinder. Sensitivity, specificity, accuracy, and F1 scores for both the CNN model and PeFinder in the internal and external validation sets were determined. Results The CNN model demonstrated an accuracy of 99% and an area under the curve value of 0.97. For internal validation report data, the CNN model had a statistically significant larger F1 score (0.938) than did PeFinder (0.867) when classifying findings as either PE positive or PE negative, but no significant difference in sensitivity, specificity, or accuracy was found. For external validation report data, no statistical difference between the performance of the CNN model and PeFinder was found. Conclusion A deep learning CNN model can classify radiology free-text reports with accuracy equivalent to or beyond that of an existing traditional NLP model. ^© RSNA, 2017 Online supplemental material is available for this article.

Patient Outcomes and Cerebral Infarction after Ruptured Anterior Communicating Artery Aneurysm Treatment

BACKGROUND AND PURPOSE

Anterior communicating artery aneurysm rupture and treatment is associated with high rates of dependency, which are more severe after clipping compared with coiling. To determine whether ischemic injury might account for these differences, we characterized cerebral infarction burden, infarction patterns, and patient outcomes after surgical or endovascular treatment of ruptured anterior communicating artery aneurysms.

MATERIALS AND METHODS

We performed a retrospective cohort study of consecutive patients with ruptured anterior communicating artery aneurysms. Patient data and neuroimaging studies were reviewed. A propensity score for outcome measures was calculated to account for the nonrandom assignment to treatment. Primary outcome was the frequency of frontal lobe and striatum ischemic injury. Secondary outcomes were patient mortality and clinical outcome at discharge and at 3 months.

RESULTS

Coiled patients were older (median, 55 versus 50 years; P = .03), presented with a worse clinical status (60% with Hunt and Hess Score >2 versus 34% in clipped patients; P = .02), had a higher modified Fisher grade (P = .01), and were more likely to present with intraventricular hemorrhage (78% versus 56%; P = .03). Ischemic frontal lobe infarction (OR, 2.9; 95% CI, 1.1-8.4; P = .03) and recurrent artery of Heubner infarction (OR, 20.9; 95% CI, 3.5-403.7; P < .001) were more common in clipped patients. Clipped patients were more likely to be functionally dependent at discharge (OR, 3.2; P = .05) compared with coiled patients. Mortality and clinical outcome at 3 months were similar between coiled and clipped patients.

CONCLUSIONS

Frontal lobe and recurrent artery of Heubner infarctions are more common after surgical clipping of ruptured anterior communicating artery aneurysms, and are associated with poorer clinical outcomes at discharge.

Prevalence of Imaging Biomarkers to Guide the Planning of Acute Stroke Reperfusion Trials

BACKGROUND AND PURPOSE

Imaging biomarkers are increasingly used as selection criteria for stroke clinical trials. The goal of our study was to determine the prevalence of commonly studied imaging biomarkers in different time windows after acute ischemic stroke onset to better facilitate the design of stroke clinical trials using such biomarkers for patient selection.

METHODS

This retrospective study included 612 patients admitted with a clinical suspicion of acute ischemic stroke with symptom onset no more than 24 hours before completing baseline imaging. Patients with subacute/chronic/remote infarcts and hemorrhage were excluded from this study. Imaging biomarkers were extracted from baseline imaging, which included a noncontrast head computed tomography (CT), perfusion CT, and CT angiography. The prevalence of dichotomized versions of each of the imaging biomarkers in several time windows (time since symptom onset) was assessed and statistically modeled to assess time dependence (not lack thereof).

RESULTS

We created tables showing the prevalence of the imaging biomarkers pertaining to the core, the penumbra and the arterial occlusion for different time windows. All continuous imaging features vary over time. The dichotomized imaging features that vary significantly over time include: noncontrast head computed tomography Alberta Stroke Program Early CT (ASPECT) score and dense artery sign, perfusion CT infarct volume, and CT angiography collateral score and visible clot. The dichotomized imaging features that did not vary significantly over time include the thresholded perfusion CT penumbra volumes.

CONCLUSIONS

As part of the feasibility analysis in stroke clinical trials, this analysis and the resulting tables can help investigators determine sample size and the number needed to screen.

Evaluation of a Smartphone Decision-Support Tool for Diarrheal Disease Management in a Resource-Limited Setting

The emergence of mobile technology offers new opportunities to improve clinical guideline adherence in resource-limited settings. We conducted a clinical pilot study in rural Bangladesh to evaluate the impact of a smartphone adaptation of the World Health Organization (WHO) diarrheal disease management guidelines, including a modality for age-based weight estimation. Software development was guided by end-user input and evaluated in a resource-limited district and sub-district hospital during the fall 2015 cholera season; both hospitals lacked scales which necessitated weight estimation. The study consisted of a 6 week pre-intervention and 6 week intervention period with a 10-day post-discharge follow-up. Standard of care was maintained throughout the study with the exception that admitting clinicians used the tool during the intervention. Inclusion criteria were patients two months of age and older with uncomplicated diarrheal disease. The primary outcome was adherence to guidelines for prescriptions of intravenous (IV) fluids, antibiotics and zinc. A total of 841 patients were enrolled (325 pre-intervention; 516 intervention). During the intervention, the proportion of prescriptions for IV fluids decreased at the district and sub-district hospitals (both p < 0.001) with risk ratios (RRs) of 0.5 and 0.2, respectively. However, when IV fluids were prescribed, the volume better adhered to recommendations. The proportion of prescriptions for the recommended antibiotic azithromycin increased (p < 0.001 district; p = 0.035 sub-district) with RRs of 6.9 (district) and 1.6 (sub-district) while prescriptions for other antibiotics decreased; zinc adherence increased. Limitations included an absence of a concurrent control group and no independent dehydration assessment during the pre-intervention. Despite limitations, opportunities were identified to improve clinical care, including better assessment, weight estimation, and fluid/ antibiotic selection. These findings demonstrate that a smartphone-based tool can improve guideline adherence. This study should serve as a catalyst for a randomized controlled trial to expand on the findings and address limitations.

Predicting "heart age" using electrocardiography

Knowledge of a patient's cardiac age, or "heart age", could prove useful to both patients and physicians for better encouraging lifestyle changes potentially beneficial for cardiovascular health. This may be particularly true for patients who exhibit symptoms but who test negative for cardiac pathology. We developed a statistical model, using a Bayesian approach, that predicts an individual's heart age based on his/her electrocardiogram (ECG). The model is tailored to healthy individuals, with no known risk factors, who are at least 20 years old and for whom a resting ~5 min 12-lead ECG has been obtained. We evaluated the model using a database of ECGs from 776 such individuals. Secondarily, we also applied the model to other groups of individuals who had received 5-min ECGs, including 221 with risk factors for cardiac disease, 441 with overt cardiac disease diagnosed by clinical imaging tests, and a smaller group of highly endurance-trained athletes. Model-related heart age predictions in healthy non-athletes tended to center around body age, whereas about three-fourths of the subjects with risk factors and nearly all patients with proven heart diseases had higher predicted heart ages than true body ages. The model also predicted somewhat higher heart ages than body ages in a majority of highly endurance-trained athletes, potentially consistent with possible fibrotic or other anomalies recently noted in such individuals.

Comparative analysis of somatic copy-number alterations across different human cancer types reveals two distinct classes of breakpoint hotspots

Somatic copy-number alterations (SCNAs) play a crucial role in the development of human cancer. However, it is not well understood what evolutionary mechanisms contribute to the global patterns of SCNAs in cancer genomes. Taking advantage of data recently available through The Cancer Genome Atlas, we performed a systematic analysis on genome-wide SCNA breakpoint data for eight cancer types. First, we observed a high degree of overall similarity among the SCNA breakpoint landscapes of different cancer types. Then, we compiled 19 genomic features and evaluated their effects on the observed SCNA patterns. We found that evolutionary indel and substitution rates between species (i.e. humans and chimpanzees) consistently show the strongest correlations with breakpoint frequency among all the surveyed features; whereas the effects of some features are quite cancer-type dependent. Focusing on SCNA breakpoint hotspots, we found that cancer-type-specific breakpoint hotspots and common hotspots show distinct patterns. Cancer-type-specific hotspots are enriched with known cancer genes but are poorly predicted from genomic features; whereas common hotspots show the opposite patterns. This contrast suggests that explaining high-frequency SCNAs in cancer may require different evolutionary models: positive selection driven by cancer genes, and non-adaptive evolution related to an intrinsically unstable genomic context. Our results not only present a systematic view of the effects of genetic factors on genome-wide SCNA patterns, but also provide deep insights into the evolutionary process of SCNAs in cancer.

Predicting the lethal phenotype of the knockout mouse by integrating comprehensive genomic data

MOTIVATION

The phenotypes of knockout mice provide crucial information for understanding the biological functions of mammalian genes. Among various knockout phenotypes, lethality is of great interest because those involved genes play essential roles. With the availability of large-scale genomic data, we aimed to assess how well the integration of various genomic features can predict the lethal phenotype of single-gene knockout mice.

RESULTS

We first assembled a comprehensive list of 491 candidate genomic features derived from diverse data sources. Using mouse genes with a known phenotype as the training set, we integrated the informative genomic features to predict the knockout lethality through three machine learning methods. Based on cross-validation, our models could achieve a good performance (accuracy = 73% and recall = 63%). Our results serve as a valuable practical resource in the mouse genetics research community, and also accelerate the translation of the knowledge of mouse genes into better strategies for studying human disease.

Alkaline phosphatase as an indicator of true ejaculation in the rhinoceros

The objective was to determine if seminal alkaline phosphatase (ALP) can serve as an indicator of true ejaculation in the rhinoceros. Concentrations of ALP activity were determined in seminal fractions collected from African black rhinos (Diceros bicornis), an African white rhino (Ceratotherium simum), and an Indian rhino (Rhinoceros unicornis) during electroejaculation. In addition, seminal fractions collected during penile massage of a Sumatran rhino (Dicerorhinus sumatrensis) were assessed. Correlations between ALP activity and sperm concentration, fraction pH, and fraction osmolality were evaluated in the Indian rhino and black rhino. Concentrations of ALP activity in rhino ejaculate fractions ranged from < 5 to 11,780 U/L and were positively correlated (P < 0.05) with sperm concentration for both Indian rhino (r = 0.995) and black rhino (r = 0.697), but did not exhibit a strong correlation with either pH or osmolality (P > 0.05). Data were insufficient for establishing meaningful correlation coefficients in the Sumatran rhino and white rhino, but preliminary results were in accordance with findings in the Indian rhino and black rhino. We concluded that ALP was present in rhinoceros semen, likely originated from the epididymides and/or testes, and could serve as a useful tool for assessing the production of ejaculatory versus pre-ejaculatory fluid in the rhinoceros.

Origin, splicing, and expression of rodent amelogenin exon 8

Amelogenin RNA transcripts undergo extensive alternative splicing, and MMP-20 processes the isoforms following their secretion. Since amelogenins have been ascribed cell-signaling activities, we asked if a lack of proteolytic processing by MMP-20 affects amelogenin signaling and consequently alters amelogenin splice site selection. RT-PCR analyses of amelogenin mRNA between control and Mmp20(-/-)mice revealed no differences in the splicing pattern. We characterized 3 previously unidentified amelogenin alternatively spliced transcripts and demonstrated that exon-8-encoded amelogenin isoforms are processed by MMP-20. Transcripts with exon 8 were expressed approximately five-fold less than those with exon 7. Analyses of the mouse and rat amelogenin gene structures confirmed that exon 8 arose in a duplication of exons 4 through 5, with translocation of the copy downstream of exon 7. No downstream genomic sequences homologous to exons 4-5 were present in the bovine or human amelogenin genes, suggesting that this translocation occurred only in rodents.

Construction of a database of benzene biological monitoring

Biological monitoring of occupational exposure to benzene has been conducted in the petroleum, steel and chemical industries. The urinary benzene-specific biomarker, S-phenylmercapturic acid (PMA), was quantified in post-shift samples using a sensitive enzyme-linked immunosorbent assay (ELISA) and expressed as a function of urinary creatinine concentration. The assay, based on a PMA-specific antiserum, is sufficiently sensitive to measure PMA levels in non-occupationally exposed control subjects. The assay delivers batch results in a timely manner which may be as short as 3 h. Samples were analysed from groups of workers engaged in coke oven combustion processes, petroleum refining and decontamination of a benzene land spill. The construction of a database of results provides an index of benzene uptake as a consequence of the respective work processes and tasks and readily enables benchmarking exercises aimed at comparing degrees of exposure across segments of industry.

Development and validation of a competitive immunoassay for urinary S-phenylmercapturic acid and its application in benzene biological monitoring

An immunoassay that quantifies urinary S-phenylmercapturic acid (PMA), a benzenespecific biomarker, has been developed and its potential usefulness as a screening tool for monitoring occupational exposure to benzene has been demonstrated. Analytical reliability has been confirmed by correlation of results with gas chromatography-mass spectrometry (GC/MS) data (R = 0.92). The assay has been configured as a competitive enzyme-linked immunosorbent assay (ELISA) to facilitate rapid throughput of samples. The ELISA has a working range of 40-1200 nmoll-1 urinary PMA and appears to be unaffected by the presence of structurally related urinary metabolites. Background levels of 0-1.9 mumol PMA/mol creatinine (mean 0.9 mumol mol-1, n = 32) were measured in nonsmoking control subjects. Recent exposures to benzene (8 h time-weighted averages-TWA), during diverse industrial processes, over the range 0-4.8 ppm were identified by application of the assay in biological monitoring programmes.

Mycobacterium avium subsp. paratuberculosis infection in an addax (Addax nasomaculatus)

Mycobacterium avium subsp. paratuberculosis was cultured from a single fecal sample collected from a 10-yr-old, captive-bred male addax (Addax nasomaculatus). Attempts to confirm infection with additional fecal cultures, serology, semen culture, and tissue biopsy were unsuccessful. There were no gross lesions on necropsy. On histopathology there were neither acid-fast organisms nor microscopic changes suggestive of active or clinical Johne's disease. Mycobacterium avium subsp. paratuberculosis was isolated from four organ tissues: ileum, jejunum, colon, and mesenteric lymph node.

Cerebrospinal nematodiasis in a white-handed gibbon (Hylobates lar) due to Baylisascaris sp

An adult white-handed gibbon (Hylobates lar) at a zoo in eastern Kansas was euthanized after developing a head tremor, generalized motor incoordination, and partial paresis of the right arm that persisted over 2 yr. Magnetic resonance imaging early in the course of the disease demonstrated a localized left frontal lobe cerebritis. Larvae morphologically consistent with a Baylisascaris species were seen in tissue sections of the cerebrum and cerebellum. Epizootiologic investigation, which included qualitative fecal flotations, evaluation of soil samples for nematode eggs, and necropsy examination of livetrapped raccoons (Procyon lotor), indicated that Baylisascaris procyonis was most likely to have caused the cerebrospinal nematodiasis in this gibbon.

Water immiscible solvent based immunoassay

A competitive monoclonal antibody-based immunoassay which quantifies a hydrophobic hapten (Rx) in water immiscible solvents, obviating the need of a pre-extraction step, has been developed. Approximately linear dose response profiles of analyte, over the range 1-20 ugml-1 in the hydrophobic solvents, hexane, toluene and xylene were obtained. UV spectrophotometric analyses of Rx dosed hexane confirm the phenomenon of antibody-mediated transfer of analyte from the organic to the aqueous milieu. Preliminary data on the effect of water immiscible solvents on the immunoreactivity of a monoclonal antibody in free solution are presented. The potential industrial applications of water immiscible solvent based immunoassays are discussed.

Thrombin, epidermal growth factor, and phorbol myristate acetate stimulate tubulin polymerization in quiescent cells: a potential link to mitogenesis

Previous studies suggest that alterations in the microtubule (MT)-tubulin equilibrium during G0/G1 affect mitogenesis. To determine the effect of growth factors on the MT-tubulin equilibrium, we developed a radioactive monoclonal antibody binding assay (Ball et al.: J. Cell. Biol. 103:1033-1041, 1986). With this assay, 3H-Ab 1-1.1 binding to cytoskeletons in confluent populations of cultured cells is proportional to the number of tubulin subunits polymerized into MTs. We now show that purified alpha-thrombin increases 3H-Ab 1-1.1 binding to cytoskeletons of serum-arrested mouse embryo (ME) fibroblasts from 1.5- to 3-fold. This stimulation is dose-dependent and correlates with concentrations of thrombin required for initiation of DNA synthesis. Other mitogenic factors, epidermal growth factor (EGF) and phorbol 12-myristate 13-acetate (PMA), also stimulate MT polymerization. Addition of colchicine (0.3 microM) eight hours after growth factor addition, blocks stimulation of 3H-thymidine incorporation by thrombin, EGF, or PMA, suggesting that tubulin polymerization or subsequent events triggered by MT polymerization are required for cells to enter a proliferative cycle. Consistent with models for autoregulation of tubulin synthesis, thrombin, EGF, and PMA all increase tubulin synthesis 9 to 15 hr after growth factor addition, raising the possibility that the decrease in free tubulin and subsequent stimulation of tubulin synthesis is linked to progression of cells into a proliferative cycle. Colchicine addition to these cells also stimulates DNA synthesis, but colchicine-stimulated cells enter S phase 6 to 8 hr later than those stimulated by growth factors. This delayed stimulation may be related to the time required for degradation of tubulin-colchicine complexes below a critical level. These data suggest that regulation of cell proliferation may be linked to increased MT polymerization and the resulting decrease in free tubulin pools.

Taxol inhibits stimulation of cell DNA synthesis by human cytomegalovirus

The microtubule (MT)-stabilizing drug, taxol, inhibited human cytomegalovirus (CMV)-initiated cell DNA synthesis by up to 100% in serum-arrested mouse embryo (ME) fibroblasts that were abortively infected by CMV. Taxol concentrations known to increase MT polymerization and to stabilize existing MTs (10 to 20 micrograms/ml) blocked CMV-stimulated cell DNA synthesis, while taxol concentrations of 2.5 micrograms/ml, or less, did not. Taxol maximally inhibited CMV initiation of cell DNA synthesis when added 3 h after virus infection and inhibited this initiation by greater than 50% when added up to 12 h after CMV infection. Control experiments suggest that taxol specifically inhibited CMV-stimulated cell DNA synthesis. Pretreatment of CMV stock with taxol did not reduce the stimulatory effect of CMV on cell DNA synthesis and taxol had no detectable effect on CMV-specific early protein synthesis. Moreover, taxol did not appear to alter thymidine pool sizes, affect cell viability, or compromise the DNA synthetic machinery in CMV-infected cells. Since taxol increases tubulin polymerization and inhibits MT disassembly, these results suggest that dynamic changes in MTs or in the pool of free tubulin subunits are necessary for CMV to stimulate cell entry into a proliferative cycle.

Epidermal growth factor potentiates cyclic AMP accumulation in A-431 cells

Epidermal growth factor (EGF) treatment of A-431 cells potentiates up to 5-fold the intracellular cyclic AMP (cAMP) accumulation induced by isoproterenol, cholera toxin, forskolin, or 3-isobutyl-1-methylxanthine (IBMX). EGF potentiates cAMP accumulation in several epithelial cell lines which overexpress the EGF receptor including A-431 cells, HSC-1 cells, and MDA-468 cells, and in the A-431-29S clone which expresses a normal complement of EGF receptors. Although EGF potentiates cAMP accumulation, EGF by itself does not measurably alter the basal level of cAMP. EGF rapidly enhances cAMP accumulation (within 1 to 3 min) in A-431 cells treated with these cAMP-elevating agents. EGF potentiation of cAMP accumulation does not reflect enhancement of beta-adrenergic receptor activation and is not a consequence of intracellular cAMP elevation or the concomitant activation of cAMP-dependent protein kinase. Since EGF potentiates accumulation of both intracellular and extracellular cAMP in isoproterenol-treated A-431 cells, EGF does not potentiate intracellular cAMP accumulation by inhibition of cAMP export. EGF potentiation of cAMP accumulation is pertussis toxin-insensitive and does not result from EGF inhibition of cAMP degradation in A-431 cells. These results demonstrate that EGF transmembrane signaling includes an interaction with a component of the adenylate cyclase system and that this interaction stimulates cAMP synthesis resulting in enhancement of cAMP accumulation.

A radiolabeled monoclonal antibody binding assay for cytoskeletal tubulin in cultured cells

To detect changes in the extent of tubulin polymerization in cultured cells, we have developed a radioactive antibody binding assay that can be used to quantitate total cytoskeletal tubulin or specific antigenic subsets of polymerized tubulin. Fibroblastic cells, grown to confluence in multiwell plates, were permeabilized and extracted with 0.5% Triton X-100 in a microtubule-stabilizing buffer. These extracted cytoskeletons were then fixed and incubated with translationally radiolabeled monoclonal antitubulin antibody (Ab 1-1.1), an IgM antibody specific for the beta subunit of tubulin. Specific binding of Ab 1-1.1 to the cytoskeletons was saturable and of a single apparent affinity. All specific binding was blocked by preincubation of the radiolabeled antibody with excess purified brain tubulin. Specific Ab 1-1.1 binding appeared to represent binding to cytoskeletal tubulin inasmuch as: pretreatment of cells with colchicine decreased Ab 1-1.1 binding in a dose-dependent manner which correlated with the amount of polymerized tubulin visualized in parallel cultures by indirect immunofluorescence, taxol pretreatment alone caused an increase in Ab 1-1.1 binding and prevented in a dose-dependent manner the colchicine-induced decrease in antibody binding, in cells pretreated with colcemid and returned to fresh medium, Ab 1-1.1 binding decreased and recovered in parallel with the depolymerization and regrowth of microtubules in these cells, and comparison of maximal antibody binding per cell between primary mouse embryo, 3T3, and human foreskin fibroblasts correlated with immunofluorescence visualization of microtubules in these cells. Thus, this assay can be used to measure relative changes in the level of polymerized cytoskeletal tubulin. Moreover, by Scatchard-type analysis of the binding data it is possible to estimate the total number of antibody binding sites per cell. Therefore, depending on the stoichiometry of antibody binding, this type of assay may be used for quantitating total cytoskeletal tubulin, specific antigenic subsets of cytoskeletal tubulin, or other cytoskeletal proteins.