ABSTRACT
BACKGROUND: The majority of out-of-hospital cardiac arrests (OHCAs) occur among individuals in the general population, for whom there is no established strategy to identify risk. In this study, we assess the use of electronic health record (EHR) data to identify OHCA in the general population and define salient factors contributing to OHCA risk. METHODS: The analytical cohort included 2366 individuals with OHCA and 23 660 age- and sex-matched controls receiving health care at the University of Washington. Comorbidities, electrocardiographic measures, vital signs, and medication prescription were abstracted from the EHR. The primary outcome was OHCA. Secondary outcomes included shockable and nonshockable OHCA. Model performance including area under the receiver operating characteristic curve and positive predictive value were assessed and adjusted for observed rate of OHCA across the health system. RESULTS: There were significant differences in demographic characteristics, vital signs, electrocardiographic measures, comorbidities, and medication distribution between individuals with OHCA and controls. In external validation, discrimination in machine learning models (area under the receiver operating characteristic curve 0.80-0.85) was superior to a baseline model with conventional cardiovascular risk factors (area under the receiver operating characteristic curve 0.66). At a specificity threshold of 99%, correcting for baseline OHCA incidence across the health system, positive predictive value was 2.5% to 3.1% in machine learning models compared with 0.8% for the baseline model. Longer corrected QT interval, substance abuse disorder, fluid and electrolyte disorder, alcohol abuse, and higher heart rate were identified as salient predictors of OHCA risk across all machine learning models. Established cardiovascular risk factors retained predictive importance for shockable OHCA, but demographic characteristics (minority race, single marital status) and noncardiovascular comorbidities (substance abuse disorder) also contributed to risk prediction. For nonshockable OHCA, a range of salient predictors, including comorbidities, habits, vital signs, demographic characteristics, and electrocardiographic measures, were identified. CONCLUSIONS: In a population-based case-control study, machine learning models incorporating readily available EHR data showed reasonable discrimination and risk enrichment for OHCA in the general population. Salient factors associated with OCHA risk were myriad across the cardiovascular and noncardiovascular spectrum. Public health and tailored strategies for OHCA prediction and prevention will require incorporation of this complexity.
Subject(s)
Electronic Health Records , Out-of-Hospital Cardiac Arrest , Humans , Male , Out-of-Hospital Cardiac Arrest/epidemiology , Out-of-Hospital Cardiac Arrest/diagnosis , Female , Middle Aged , Aged , Risk Factors , Adult , Predictive Value of Tests , Risk Assessment , Comorbidity , Electrocardiography , Machine Learning , Case-Control StudiesABSTRACT
Background: Human identification and kinship testing in forensic science rely on Short Tandem Repeat (STR) multiplex kits, typically containing loci recommended by standard sets. However, complementary kits with additional STR loci can be valuable in complex cases. Allele frequency databases specific to the population are essential for accurate forensic analysis.Aim: This study aimed to generate allele frequencies and population genetic data for 44 autosomal STR loci from SureID® PanGlobal and 27comp kits in English and Irish populations for forensic casework, human identification, and kinship testing.Subjects and methods: Buccal swab samples from 645 White Caucasians (365 English, 280 Irish) were collected. DNA was extracted and amplified using the mentioned kits. Quality control, statistical analysis, and genetic distance calculations were performed.Results: Both kits demonstrated robustness with no significant deviations from Hardy-Weinberg Equilibrium (HWE). Variant alleles and minor discordances between kits were observed. Syntenic STR pairs were identified but showed no significant linkage. A close genetic relationship was found between English and Irish populations, allowing for combined databases.Conclusions: The SureID® PanGlobal and 27comp kits showed high discriminatory power and reliability in the English and Irish populations. Care is needed when handling variant alleles, discordances, and syntenic loci. Combining data from both populations is feasible for a comprehensive database. Further studies are required to explore their effectiveness in diverse populations.
Subject(s)
DNA , Genetics, Population , Humans , Reproducibility of Results , Gene Frequency , DNA/genetics , Microsatellite Repeats/genetics , Genetic VariationABSTRACT
RATIONALE AND OBJECTIVES: Spinal osteoporotic compression fractures (OCFs) can be an early biomarker for osteoporosis but are often subtle, incidental, and underreported. To ensure early diagnosis and treatment of osteoporosis, we aimed to build a deep learning vertebral body classifier for OCFs as a critical component of our future automated opportunistic screening tool. MATERIALS AND METHODS: We retrospectively assembled a local dataset, including 1790 subjects and 15,050 vertebral bodies (thoracic and lumbar). Each vertebral body was annotated using an adaption of the modified-2 algorithm-based qualitative criteria. The Osteoporotic Fractures in Men (MrOS) Study dataset provided thoracic and lumbar spine radiographs of 5994 men from six clinical centers. Using both datasets, five deep learning algorithms were trained to classify each individual vertebral body of the spine radiographs. Classification performance was compared for these models using multiple metrics, including the area under the receiver operating characteristic curve (AUC-ROC), sensitivity, specificity, and positive predictive value (PPV). RESULTS: Our best model, built with ensemble averaging, achieved an AUC-ROC of 0.948 and 0.936 on the local dataset's test set and the MrOS dataset's test set, respectively. After setting the cutoff threshold to prioritize PPV, this model achieved a sensitivity of 54.5% and 47.8%, a specificity of 99.7% and 99.6%, and a PPV of 89.8% and 94.8%. CONCLUSION: Our model achieved an AUC-ROC>0.90 on both datasets. This testing shows some generalizability to real-world clinical datasets and a suitable performance for a future opportunistic osteoporosis screening tool.
Subject(s)
Deep Learning , Fractures, Compression , Osteoporosis , Spinal Fractures , Male , Humans , Fractures, Compression/diagnostic imaging , Retrospective Studies , Bone Density , Spinal Fractures/diagnostic imaging , Osteoporosis/complications , Osteoporosis/diagnostic imaging , Lumbar Vertebrae/diagnostic imaging , AlgorithmsABSTRACT
The current approach to using machine learning (ML) algorithms in healthcare is to either require clinician oversight for every use case or use their predictions without any human oversight. We explore a middle ground that lets ML algorithms abstain from making a prediction to simultaneously improve their reliability and reduce the burden placed on human experts. To this end, we present a general penalized loss minimization framework for training selective prediction-set (SPS) models, which choose to either output a prediction set or abstain. The resulting models abstain when the outcome is difficult to predict accurately, such as on subjects who are too different from the training data, and achieve higher accuracy on those they do give predictions for. We then introduce a model-agnostic, statistical inference procedure for the coverage rate of an SPS model that ensembles individual models trained using K-fold cross-validation. We find that SPS ensembles attain prediction-set coverage rates closer to the nominal level and have narrower confidence intervals for its marginal coverage rate. We apply our method to train neural networks that abstain more for out-of-sample images on the MNIST digit prediction task and achieve higher predictive accuracy for ICU patients compared to existing approaches.
Subject(s)
Machine Learning , Neural Networks, Computer , Humans , Reproducibility of Results , Algorithms , Research DesignABSTRACT
The efficiency of reduced volume PCR amplification was studied using the VeriFiler™ Express PCR Amplification Kit. Full (25 µL) and reduced (5 µL) volumes were tested in parallel to identify any differences in template DNA sensitivity and other electropherogram parameters. Both volumes produced full DNA profiles down to 0.08 ng/µL DNA concentration at 26 PCR cycles; however, reduced volume produced higher peak heights due to increased signal intensities. Significant difference (p-value ≤ 0.05) in heterozygote peak height ratios was observed between both volumes, where the reduced volume threshold was lowered to 0.6 to accommodate all data points. However, no significant difference (p-value > 0.05) was identified in the stutter ratios between both volumes. The analytical threshold for reduced volume was also determined to be 150 RFU with the presence of template DNA in PCR amplification. When the optimized reduced volume parameters were tested on DNA extracted from buccal swab samples using Prep-n-Go™ Buffer, good quality DNA profiles were produced. Overall, the reduced volume not only showed better results compared to the full volume, but also enable more samples to be processed with a PCR amplification kit, thus reduced the cost.
Subject(s)
DNA Fingerprinting , Microsatellite Repeats , DNA/genetics , DNA Fingerprinting/methods , Heterozygote , Polymerase Chain Reaction/methodsABSTRACT
The objective of this study was to identify factors that predict for sperm granuloma formation and the impact of sperm granuloma presence and quantity on vasectomy reversal (VR) outcomes. A cross sectional retrospective review of prospectively collected data, on the impact of granuloma on VR outcomes from a single academic center was performed. The impact of age, obstructive interval, intraoperative vasal fluid findings, anastomosis type, body mass index, tobacco use and total motile count (TMC) was determined. A total of 1550 men underwent VR between January 2000 and August 2019. Granulomas were present unilaterally in 23.3% (n = 361) and bilaterally in 14.2% (n = 220). On univariate analysis, increasing patient age negatively correlated with a larger number of granulomas (p = .011). Granuloma presence was associated with finding intact and motile sperm from the vasal stump intraoperatively (p = .001), and vasoepididymostomy anastomosis (p < .001). However, granuloma presence (and quantity) did not correlate with obstructive interval or maximum TMC. Tobacco use and body mass index (BMI) were not associated with granuloma presence. On multivariate analysis, granuloma quantity was not associated with TMC. Obstructive interval and vasovasostomy anastomosis were associated with higher TMC, while BMI was negatively associated with TMC. In conclusion, increasing age was negatively correlated with granuloma formation. Granuloma presence was associated with more favourable intraoperative fluid findings and anastomosis type, but not post-VR TMC, suggesting men with and without granulomas undergoing skilled microsurgery will have similar patency rates. Heavier men should be encouraged for weight loss prior to vasectomy reversal as increasing BMI was associated with lower TMC.
Subject(s)
Vasectomy , Vasovasostomy , Cross-Sectional Studies , Granuloma/etiology , Humans , Male , Microsurgery , Semen , SpermatozoaABSTRACT
Currently, there is a lack of racial/ethnic heterogeneity in research databases, exposing a systematic issue in studies exploring inflammation-mediated diseases, such as hidradenitis suppurativa (HS). HS is a chronic inflammatory skin condition that disrupts normal structure and functioning of terminal hair follicles, resulting in the formation of recurrent abscesses, nodules, and sinus tracts within intertriginous regions. Studies have described higher serum levels of inflammation-mediated C-reactive protein (CRP) in patients with HS, a disease that predominantly affects skin of color (SOC) populations. Herein, we explore the role of CRP levels in the context of HS disease presentation, management, and psychosocial implications in SOC patients to determine existing disparities in research studies.
Subject(s)
Hidradenitis Suppurativa , C-Reactive Protein/metabolism , Ethnicity , Hair Follicle/metabolism , Humans , InflammationABSTRACT
RATIONALE AND OBJECTIVES: Osteoporosis affects 9% of individuals over 50 in the United States and 200 million women globally. Spinal osteoporotic compression fractures (OCFs), an osteoporosis biomarker, are often incidental and under-reported. Accurate automated opportunistic OCF screening can increase the diagnosis rate and ensure adequate treatment. We aimed to develop a deep learning classifier for OCFs, a critical component of our future automated opportunistic screening tool. MATERIALS AND METHODS: The dataset from the Osteoporotic Fractures in Men Study comprised 4461 subjects and 15,524 spine radiographs. This dataset was split by subject: 76.5% training, 8.5% validation, and 15% testing. From the radiographs, 100,409 vertebral bodies were extracted, each assigned one of two labels adapted from the Genant semiquantitative system: moderate to severe fracture vs. normal/trace/mild fracture. GoogLeNet, a deep learning model, was trained to classify the vertebral bodies. The classification threshold on the predicted probability of OCF outputted by GoogLeNet was set to prioritize the positive predictive value (PPV) while balancing it with the sensitivity. Vertebral bodies with the top 0.75% predicted probabilities were classified as moderate to severe fracture. RESULTS: Our model yielded a sensitivity of 59.8%, a PPV of 91.2%, and an F1 score of 0.72. The areas under the receiver operating characteristic curve (AUC-ROC) and the precision-recall curve were 0.99 and 0.82, respectively. CONCLUSION: Our model classified vertebral bodies with an AUC-ROC of 0.99, providing a critical component for our future automated opportunistic screening tool. This could lead to earlier detection and treatment of OCFs.
Subject(s)
Deep Learning , Fractures, Compression , Osteoporosis , Spinal Fractures , Male , Female , Humans , Fractures, Compression/diagnostic imaging , Spinal Fractures/diagnostic imaging , Osteoporosis/diagnostic imaging , RadiographySubject(s)
COVID-19 , Hidradenitis Suppurativa , Hidradenitis Suppurativa/diagnosis , Humans , PandemicsABSTRACT
Allogeneic hematopoietic cell transplantation (allo-HCT) is potentially curative for certain hematologic malignancies and nonmalignant diseases. The field of allo-HCT has witnessed significant advances, including broadening indications for transplantation, availability of alternative donor sources, less toxic preparative regimens, new cell manipulation techniques, and novel GVHD prevention methods, all of which have expanded the applicability of the procedure. These advances have led to clinical practice conundrums when applying traditional definitions of hematopoietic recovery, graft rejection, graft failure, poor graft function, and donor chimerism, because these may vary based on donor type, cell source, cell dose, primary disease, graft-versus-host disease (GVHD) prophylaxis, and conditioning intensity, among other variables. To address these contemporary challenges, we surveyed a panel of allo-HCT experts in an attempt to standardize these definitions. We analyzed survey responses from adult and pediatric transplantation physicians separately. Consensus was achieved for definitions of neutrophil and platelet recovery, graft rejection, graft failure, poor graft function, and donor chimerism, but not for delayed engraftment. Here we highlight the complexities associated with the management of mixed donor chimerism in malignant and nonmalignant hematologic diseases, which remains an area for future research. We recognize that there are multiple other specific, and at times complex, clinical scenarios for which clinical management must be individualized.
Subject(s)
Chimerism , Hematopoietic Stem Cell Transplantation , Adult , Child , Graft Rejection/prevention & control , Hematopoietic Stem Cell Transplantation/adverse effects , Humans , Transplantation Conditioning , Transplantation, Homologous , United StatesABSTRACT
Immune-mediated cytopenias (IMC)-isolated or combined hemolytic anemia, thrombocytopenia, or neutropenia-are increasingly recognized as serious complications after allogeneic hematopoietic cell transplantation (HCT) for nonmalignant disorders (NMD). However, IMC incidence, duration, response to therapy, and risk factors are not well defined. This retrospective chart review identified cases of IMC with serologic confirmation among patients who underwent HCT for NMD at a single institution between 2010 and 2017. IMC after HCT for NMD in a large pediatric cohort (n = 271) was common with a cumulative incidence of 18%, identified at a median of 136 days after HCT. Treatment included prolonged immune suppression (>3 months) in 58% of all IMC cases, 91% when multiple cell lines were affected. Multiple therapeutic agents were used for the majority affected, and median time to resolution of IMC was 118 days from diagnosis. Fine-Gray competing risk multivariate regression analysis identified a combined risk factor of younger age (<3 years) and inherited metabolic disorder, as well as hemoglobinopathy (at any age) associated with 1-year incidence of IMC (P < .01). We expand these findings with the observation of declining donor T-lymphoid chimerism from day 60 to 100 and lower absolute CD4+ counts at day 100 (P < .01), before median onset of IMC, for patients with IMC compared to those without. In this cohort, 4 deaths (8%) were associated with IMC, including 2 requiring second transplantation for secondary graft failure. Although the pathogenesis of IMC post-HCT for NMD remains elusive, further research may identify approaches to prevent and better treat this HCT complication.
Subject(s)
Hematopoietic Stem Cell Transplantation , Thrombocytopenia , Child , Child, Preschool , Chimerism , Hematopoietic Stem Cell Transplantation/adverse effects , Humans , Retrospective Studies , Transplantation ConditioningABSTRACT
Background: Operative management of lateral epicondylitis can be managed with percutaneous, arthroscopic, or open surgical release. Extraarticular arthroscopic release is a new technique, and no study has compared its outcomes and risk profile. Methods: A 26-patient cohort was reviewed before and after extraarticular arthroscopic release, which was performed by the senior author. The Mayo Elbow Performance Scores were used as a functional outcome score and obtained via a phone interview. Results were analyzed using a paired t-test with a statistical significance set at P < .05. Results: Of the 26 patients, 10 were being treated under workers compensation. Preoperative Mayo Elbow Performance Score was 47.5, and the postoperative score was 90.2 with a significant difference of 42.7 (P value = .05). The workers compensation group scored 13.3 points lower postoperatively than the remainder of patients, which was shown to also be significant with a P value of .002. Discussion and Conclusion: The advantage of extraarticular arthroscopic release was better visualization of affected structures, which improved accuracy of debridement, and a small capsulotomy, which decreased the risk of a transient radial nerve palsy. Overall, extraarticular arthroscopic results were found to be good and comparable to the results of other operative techniques with the added advantage of a lower risk profile.
ABSTRACT
Cardiac angiosarcoma (AS) is an extremely rare, malignant vascular tumor with <10 cases reported in the pediatric literature. Prognosis is dismal with overall survival often <1 year from initial diagnosis. In this report, we present the case of a 10-year-old boy with metastatic cardiac AS who is currently alive and is the longest pediatric survivor of metastatic cardiac AS reported in the literature. This is the only published pediatric case to successfully use a combination of surgical resection, conventional chemotherapy, radiation and targeted therapies including bevacizumab and pazopanib for metastatic cardiac AS.
Subject(s)
Antineoplastic Combined Chemotherapy Protocols/therapeutic use , Chemoradiotherapy/methods , Heart Neoplasms/therapy , Hemangiosarcoma/therapy , Neoplasms, Second Primary/therapy , Surgical Procedures, Operative/methods , Bevacizumab/administration & dosage , Child , Combined Modality Therapy , Heart Neoplasms/pathology , Hemangiosarcoma/secondary , Humans , Indazoles , Male , Neoplasms, Second Primary/pathology , Prognosis , Pyrimidines/administration & dosage , Sulfonamides/administration & dosageABSTRACT
BACKGROUND: Autoimmune cytopenias (AICs) are rare, but serious complications of allogeneic hematopoietic cell transplantation (allo-HSCT). PROCEDURE: We performed a case-control study using 20 pediatric AIC cases and 40 controls, matched by stem cell source and primary indication comparing clinical and transplant characteristics, treatment, outcomes, and late effects. RESULTS: Cases were more likely to be human leukocyte antigen mismatched (P = 0.04). There was no difference in conditioning regimen, serotherapy use, graft-versus-host disease (GVHD) prophylaxis, incidence of acute or chronic GVHD, ABO compatibility, infections, and donor engraftment. The median time to AIC onset was 219 days (range, 97-1205 days) and AIC resolution was 365 days (range, 10 days to 2737.5 days). First-line therapies for AIC patients most commonly included corticosteroids (75%) and rituximab (55%). Only 25% of patients responded to first-line treatment. At a median of 611.5 days from last rituximab dose, 82.5% patients were still receiving intravenous immune globulin for hypogammaglobulinemia compared with 2.5% of controls (P < 0.0001). Iron overload was higher in AIC patients (P = 0.0004), as was avascular necrosis (P = 0.04). There was no difference in overall survival at one year after HSCT (85% vs 82.5%). Two patients with refractory autoimmune hemolytic anemia responded to daratumumab and had resolution of B-cell aplasia. CONCLUSIONS: In this study, we find poor initial responses to AIC-directed therapies and significant late effects.
Subject(s)
Anemia, Hemolytic, Autoimmune/mortality , Graft vs Host Disease/mortality , Hematologic Neoplasms/mortality , Hematopoietic Stem Cell Transplantation/mortality , Adolescent , Adult , Anemia, Hemolytic, Autoimmune/etiology , Anemia, Hemolytic, Autoimmune/pathology , Case-Control Studies , Child , Child, Preschool , Female , Follow-Up Studies , Graft vs Host Disease/etiology , Graft vs Host Disease/pathology , Hematologic Neoplasms/pathology , Hematologic Neoplasms/therapy , Hematopoietic Stem Cell Transplantation/adverse effects , Humans , Infant , Male , Prognosis , Retrospective Studies , Survival Rate , Young AdultSubject(s)
Coronavirus Infections , Hidradenitis Suppurativa , Pandemics , Pneumonia, Viral , Betacoronavirus , COVID-19 , Humans , SARS-CoV-2ABSTRACT
INTRODUCTION: Organophosphates are a class of insecticides used globally by the agricultural industry for insect control. Acute consequences of organophosphate exposures are well known, while there has been limited research on their long-term effects. The objective of this review was to discuss the health effects of chronic organophosphate exposure in farmers. METHODS: Medline, Scopus and Web of Science were searched to find the relevant articles. Articles published only in English and until December 2018 were reviewed. The selected articles were then categorised as neurological (neurobehaviour, neurodevelopmental, neurological signs and symptoms) or non-neurological subheadings. RESULTS: A total of 53 articles for neurological effects and 17 articles for non-neurological effects were identified. Chronic organophosphates exposure was associated with deficits in the neurobehaviour subsets of attention and short-term memory, increased incidence of neurodegenerative diseases and effects on peripheral nerves and neurodevelopment. However, research to support non-neurological effects such as respiratory symptoms, increased cancer risk, endocrine disruption, cardiac issues, chronic fatigue and infertility was limited. CONCLUSION: Chronic organophosphate exposure was found to affect four of the five areas of described neurological effects in the literature. A large proportion of the research in this area was not methodologically strong, therefore few recommendations can be conclusively made. Future research is warranted to investigate the non-neurological effects of chronic exposure to ensure the occupational risks of low-level chronic exposure are clearly communicated to farmers and farm workers.
Subject(s)
Agricultural Workers' Diseases/chemically induced , Farmers , Occupational Exposure/adverse effects , Organophosphates/adverse effects , Humans , Insecticides/adverse effectsABSTRACT
To optimize voriconazole dosing in pediatric hematopoietic cell transplantation (HCT), we conducted a phase I study with a modified 3 + 3 dose-escalation followed by an expansion cohort at the maximum tolerated, minimum efficacious dose (MTD/MED). Patients ≤21 years who required voriconazole for prevention or treatment of an invasive fungal infection were assigned to three age groups. Of the 59 evaluable patients, 13 were <2 years, 23 were 2-11, and 23 were 12-21. Therapeutic serum voriconazole troughs (1.5-5 µg/mL) drawn at 7 days after initiation determined efficacy. The MTD/MED was 12 mg/kg/dose q12 h × 2 loading doses, then 10 mg/kg/dose q12 h in patients <2, and was 10 mg/kg/dose q12 h in patients 2-11. The 12-21 age group had no dose-limiting toxicity at 8 mg/kg/dose q12 h; however, the MED was not reached. Drug-related AEs ≥grade 3 included increased bilirubin, transaminases, and creatinine, all occurring in <10%. There was no significant association between supra-therapeutic troughs and AEs. Five of 17 patients who had supra-therapeutic troughs (29%) had an AE, compared to 8 of 42 who did not (19%, p = 0.38). Observational population pharmacokinetic analysis demonstrated that inter-individual variability on voriconazole clearance was >100% CV, and clearance increased with age.