Diagnosis Of Disease Research Articles

The Role Of Machine Learning In Transforming Healthcare: A Systematic Review

This systematic review explores the transformative role of machine learning (ML) in healthcare, focusing on its applications in disease diagnosis, personalized medicine, healthcare operations, and patient monitoring. By analyzing 167 peer-reviewed studies published between 2015 and 2024, the review highlights how ML algorithms have significantly improved early disease detection, treatment personalization, and operational efficiencies in healthcare settings. Findings reveal that ML-powered diagnostic tools, especially deep learning models, demonstrate accuracy levels comparable to human experts in areas like cancer detection. Additionally, ML models tailored for personalized medicine have shown promise in optimizing treatment protocols based on genetic profiles, reducing adverse reactions and improving patient outcomes. In healthcare operations, ML applications have streamlined hospital resource management, enhanced workflow efficiency, and reduced administrative burdens through predictive analytics and natural language processing (NLP). Moreover, ML-driven remote patient monitoring systems have enabled proactive interventions for chronic disease management. Despite these advancements, challenges related to data privacy, algorithmic transparency, and regulatory compliance persist. Addressing these barriers is critical for realizing the full potential of ML in healthcare, ensuring ethical deployment and equitable access. The review concludes by emphasizing the need for robust regulatory frameworks and further research to enhance the integration of ML technologies into clinical practice.

Explainable early detection of Alzheimer's disease using ROIs and an ensemble of 138 3D vision transformers.

Early detection and accurate diagnosis of brain morphological abnormalities are essential for the effective management and treatment of Alzheimer's disease (AD) and mild cognitive impairment (MCI). Structural magnetic resonance imaging (MRI) is a powerful support tool to aid in disease diagnosis and prediction. In this research study, we present an innovative approach to predict Alzheimer's disease (AD) and mild cognitive impairment (MCI) using MRI data, which integrates regional interest (ROI)-based methodology and deep learning within a comprehensible framework. The proposed method involves dividing the brain into 138 predetermined sections based on anatomical information. Next, we apply three-dimensional vision transformers (3D-ViTs) to each ROI individually, harnessing the power of deep learning. To improve prediction accuracy, we employ a deep belief network (DBN) as an ensemble learning model. Evaluating our approach on the baseline structural MRI dataset obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort, and comparing it against five other competing models, we demonstrate its performance across four binary classification tasks and a three-class classification test (AD vs MCI vs CN (Cognitively Normal)). The proposed system outperforms existing models and provides interpretable insights into the brain regions that significantly contribute to solving each classification problem. Our findings align with the existing body of literature and hold promise for guiding future research directions in this domain.

Bimodal Array-Based Fluorescence Sensor and Microfluidic Technology for Protein Fingerprinting and Clinical Diagnosis.

Proteins play a crucial role in determining disease states in humans, making them prime targets for the development of diagnostic sensors. The developed sensor array is used to investigate global proteomic changes by fingerprinting multifactorial disease states in model urine simulating phenylketonuria and in serum from preeclamptic pregnant women. Here, we report a fluorescence-based chemical sensing array that exploits the host-guest interaction between cucurbit[7]uril (CB[7]) and fluorescent triphenylamine derivatives (TPA) to detect a range of proteins. Using linear discriminant analysis, we identify fluorescence fingerprints of 14 proteins with over 98% accuracy in buffer and human serum. The array is optimized on an automated droplet microfluidic-based platform, for high-throughput sensing with controlled composition and lower sample volumes. This sensor enables the discrimination of proteins in physiological buffer and human serum, with promising applications in disease diagnosis.

Hotspots and future trends of dermatofibrosarcoma protuberans

ObjectiveDermatofibrosarcoma protuberans (DFSP) is a moderately malignant soft tissue sarcoma with localized infiltrative growth. The extensive surgical scope and high recurrence rate of DFSP brings dysfunctional, aesthetic, psychological and economic problems to patients. The aim of this study is to explore the global publication characteristics, research hotspots and future trends of DFSP over the past 32 years via an intuitive visualized way.MethodsThe Web of Science Core Collection (WoSCC) database was searched for papers related to DFSP published from 1990 to 2022. Then bibliometric analysis of these publications was performed, including collaborative networks, co-citation analysis of journals and references, and cluster analysis of keywords.ResultsA total of 1588 papers were retrieved between 1990-2022. The United States was the most prolific country, followed by China. The article Imatinib Mesylate in Advanced Dermatofibrosarcoma Protuberans: Pooled Analysis of Two Phase II Clinical Trials, received most citations. Research hotspots and future trends are mainly focused on disease diagnosis, COL1A1-PDGFB fusion gene, drug and surgical treatment, fibrosarcomatous transformation, and immunotherapy.ConclusionThe research on DFSP faces several clinical challenges. This study provides novel insights into future research directions and scientific decisions for DFSP.

Diagnosis of heart disease using an advanced triple hybrid algorithm combining machine learning techniques

Purpose This study aims to improve the detection and quantification of cardiac issues, which are a leading cause of mortality globally. By leveraging past data and using knowledge mining strategies, this study seeks to develop a technique that could assess and predict the onset of cardiac sickness in real time. The use of a triple algorithm, combining particle swarm optimization (PSO), artificial bee colony (ABC) and support vector machine (SVM), is proposed to enhance the accuracy of predictions. The purpose is to contribute to the existing body of knowledge on cardiac disease prognosis and improve overall performance in health care. Design/methodology/approach This research uses a knowledge-mining strategy to enhance the detection and quantification of cardiac issues. Decision trees are used to form predictions of cardiovascular disorders, and these predictions are evaluated using training data and test results. The study has also introduced a novel triple algorithm that combines three different combination processes: PSO, ABC and SVM to process and merge the data. A neural network is then used to classify the data based on these three approaches. Real data on various aspects of cardiac disease are incorporated into the simulation. Findings The results of this study suggest that the proposed triple algorithm, using the combination of PSO, ABC and SVM, significantly improves the accuracy of predictions for cardiac disease. By processing and merging data using the triple algorithm, the neural network was able to effectively classify the data. The incorporation of real data on various aspects of cardiac disease in the simulation further enhanced the findings. This research contributes to the existing knowledge on cardiac disease prognosis and highlights the potential of leveraging past data for strategic forecasting in the health-care sector. Originality/value The originality of this research lies in the development of the triple algorithm, which combines multiple data mining strategies to improve prognosis accuracy for cardiac diseases. This approach differs from existing methods by using a combination of PSO, ABC, SVM, information gain, genetic algorithms and bacterial foraging optimization with the Gray Wolf Optimizer. The proposed technique offers a novel and valuable contribution to the field, enhancing the competitive position and overall performance of businesses in the health-care sector.

Can flow cytometry be a key in the difficult diagnosis of coeliac disease?

We present the clinical case of a 39-year-old woman with a medical history of juvenile idiopathic arthritis and uveitis, evaluated for abdominal pain and distension, difficulty gaining weight, and diarrhea. Blood tests were requested, including antibodies which were negative and common genetic markers, showing positivity for HLA DQ 7.5. A gastroscopy with duodenal biopsies revealed findings consistent with stage 3b of the modified Marsh classification. Due to diagnostic uncertainty, an IEL study was performed, showing a characteristic CD pattern, leading to the initiation of a strict gluten-free diet (GFD). During follow-up, the patient showed partial improvement, with persistent digestive symptoms. A repeat endoscopic study revealed persistent atrophy, but refractory disease type 2 was ruled out. During a flare of rheumatologic disease, Tocilizumab was initiated by Rheumatology, resulting in improvement of duodenal atrophy.

Multi-omics insights implicate the remodeling of the intestinal structure and microbiome in aging

BackgroundAging can impair the ability of elderly individuals to fight infections and trigger persistent systemic inflammation, a condition known as inflammaging. However, the mechanisms underlying the development of inflammaging remain unknown.MethodsWe conducted 16S rRNA sequencing of intestinal contents from young and old C57BL/6J mice to elucidate changes in gut microbiota diversity and microbial community composition after aging. Aging-related differential bacterial taxa were then identified, and their abundance trends were validated in human samples. The variances in intestinal barrier function and circulating endotoxin between groups were also assessed. Furthermore, widely targeted metabolomics was conducted to characterize metabolic profiles after aging and to investigate the key metabolic pathways enriched by the differential metabolites.ResultsOur findings demonstrated an increase in relative proportion of pathogenic bacteria with age, a trend also revealed in healthy populations of different age groups. Additionally, aging individuals exhibited reduced intestinal barrier function and increased circulating endotoxin levels. Widely targeted metabolomics revealed a significant increase in various secondary bile acid metabolites after aging, positively correlated with the relative abundance of several aging-related bacterial taxa. Furthermore, old group had lower levels of various anti-inflammatory or beneficial metabolites. Enrichment analysis identified the starch and sucrose metabolism pathway as potentially the most significantly impacted signaling pathway during aging.ConclusionThis study aimed to provide insights into the complex interactions involved in organismal inflammaging through microbial multi-omics. These findings lay a solid foundation for future research aimed at identifying novel biomarkers for the clinical diagnosis of aging-related diseases or potential therapeutic targets.

Predictive Modeling and Early Detection of Parkinson's Disease Using Machine Learning

Parkinson's disease is a progressive neurodegenerative disorder that impacts millions of citizens in the USA, mainly targeting the motor system and causing debilitating symptoms such as rigidity, tremors, and bradykinesia. The diagnosis of Parkinson's disease is presently heavily dependent on clinical evaluations and neurological examinations, targeting the detection of motor dysfunction. The principal aim of this study was to use machine learning as a means for early detection and prediction of Parkinson's disease. The dataset utilized for this study was the Parkinson’s Disease Dataset, retrieved from the UCI Machine Learning Repository, which included comprehensive biomedical voice measurements from a cohort of individuals, 23 of whom are diagnosed with Parkinson’s disease and 8 who are healthy controls. This dataset included a set of features extracted from voice recordings. These include parameters like fundamental frequency (pitch), amplitude variation, jitter, shimmer, and several phonation-related measures known to reflect early vocal impairments associated with the disease in question. This research project deployed three credible and proven algorithms, namely, logistic regression, random forest, and the Support Vector Machines. Besides, this study employed a combination of metrics, including accuracy, precision, recall, F1-score, and ROC-AUC, which are more holistic toward the model performance. According to the metric performance results of the three models, several key insights were drawn into the early detection of Parkinson's Disease. Particularly, it was clear that the Random Forest model had superior accuracy and was the most reliable in classifying positive cases and healthy patients, which could be that this model turned out to be most reliable in an early detection setting. In that respect, predictive modeling in Parkinson's Disease is a capability frontier that has the potential to make much difference in clinical decision-making. Supported by Machine Learning algorithms, clinicians can trace the minute patterns in patient data, which are not easily visible through any other diagnostic means. In such cases, early detection of Parkinson's Disorder with vocal biomarkers or motor assessment may afford healthcare professionals opportunities for early interventions that might retard the disease process.

Abstract 4138718: Teen Pregnancy and Risk of Premature Major Adverse Cardiovascular Events

Introduction: Teen pregnancy (TP), defined as pregnancy between 10 and 19 years, is associated with adverse pregnancy outcomes and perpetuates socioeconomic disadvantage, but the impact in later life is not well studied. We investigated the association of TP with premature major adverse cardiovascular events (PMACE) Methods: We used data from the National Health and Nutrition Examination Survey 1999–2018. Surveys were linked with the National Death Index to ascertain mortality records. Only adult females aged 20 years and older were included in the analysis. Participants were divided into 3 groups based on age at their first life birth: 19 or younger (TP), 20 and above (non-TP), and nulliparous women (NP). TP was ascertained by age at first life birth ≤ 19 years, and PMACE was defined as any self-reported diagnosis of coronary artery disease, stroke, heart failure, or all-cause mortality occurring before age 65. We assessed the association between these groups and PMACE using univariate and multivariate logistic regression models. Result: Among 19,737 (weighted frequency of 83,450,668) adult females, 24.2% had a history of TP, 54.3% had a history of non-TP and 21.5% were NP. Compared with NP females, females with TP had a higher odd of PMACE with an unadjusted odd ratio of 3.75 (95%CI, 2.80 - 5.01) versus 2.07 (1.55 – 2.75) among non-TP. Adjusting for age, race, alcohol use, smoking, and social determinants of health factors such as income, educational status, insurance status, and food insecurity, TP remained independently associated with higher odds of PMACE with an adjusted odd ratio of 1.66 (95%CI, 1.19 – 2.30); p=0.003, but this association was not statistically significant among non-TP females with aOR 1.32 (0.95 – 1.84); p=0.10. Conclusion: Teen pregnancy was associated with higher likelihood of PMACE. This association may be indicative of the enduring impact of social determinants of health factors associated with teen pregnancy into adulthood.

Abstract 4137594: Intrinsic Capacity Trajectories and Cardiovascular Disease Among Chinese Older Adults: A Population-Based Longitudinal Study

Background: WHO defined intrinsic capacity (IC), a composite of physical and mental capacities, is a marker of healthy aging. Monitoring IC trajectories in older adults may guide preventative strategies and delay negative health outcomes. With the rapidly increasing proportion of older adults and the high prevalence of cardiovascular disease (CVD) in China, this study aimed to identify IC trajectories over time and determine whether IC trajectories can predict CVD outcomes. Hypothesis: The IC trajectory among Chinese older adults is heterogeneous; the decline of IC will be associated with increased CVD. Methods: The study used data from five waves (2011 to 2020) of the China Health and Retirement Longitudinal Study (CHARLS). A total of 3,336 adults aged 60 years or older without CVD during the baseline survey in 2011 (T1), who completed the follow-ups in 2013 (T2) and 2015 (T3) were included to determine IC trajectory. IC was assessed through five domains (locomotion, sensory, vitality, cognition, and psychology) using the WHO framework. The CVD outcomes were determined by self-reported diagnoses of heart disease or stroke in the 2020 (T5) wave. A group-based trajectory model (GBTM) was used to identify IC trajectory, and logistic regression was employed to explore the association between IC trajectory groups and CVD outcomes adjusting for covariates. Results: Three IC trajectory groups were identified (Fig 1): high IC-stable group (13.40%), medium IC-stable group (56.12%), and low IC-decline group (30.48%). By the 2020 follow-up, 1204 participants (36.1%) reported CVD diagnosis. Adjusted for sociodemographic, lifestyle factors (smoking, drinking), and cardiovascular metabolic factors (hypertension, diabetes, dyslipidemia), compared to the high IC-stable group, the low IC-decline group had 63.7% higher odds of CVD (OR=1.637, 95%CI=1.268-2.114), and medium IC-stable group had 7.1% higher odds of CVD (OR=1.071, 95%CI=0.851-1.349). Conclusion: IC trajectory among Chinese older adults is heterogeneous, and IC declines predict higher odds of CVD onset. Regular monitoring of IC and tailored intervention to improve IC may facilitate CVD prevention.

Clinical-histological correlation of various symptoms in the diagnosis of pediatric Crohn's disease.

Crohn's disease (CD) is a relapsing-remitting inflammatory bowel disease (IBD) affecting the gastrointestinal tract. Its symptoms are nonspecific, with weight loss, diarrhea, and abdominal pain being classic, though their diagnostic utility is controversial (2). We evaluated the clinical-pathological association of different symptoms in CD through an analytical study of patients under 15 years of age who underwent esophagogastroduodenoscopy and ileocolonoscopy in a pediatric hospital between 2015-2022 due to suspected IBD.

Abstract 4123370: Distance to a Cardiac Imaging Center and Likelihood of Imaging Receipt: Defining Imaging Deserts

Background: Cardiovascular imaging is critical for the diagnosis and management of cardiovascular disease. Whether geographic distance to a cardiovascular imaging center (CVIC) impacts access to care is unknown. Methods: We evaluated all Medicare beneficiaries (Fee-for-service and Medicare Advantage) from 2018-2022, excluding those < 65, without continuous Medicare enrollment in the previous year, individuals missing zip code of residence, or CVICs missing zip codes. We estimated distance from the centroid of an individual’s zip code of residence to the centroid of the zip code with the nearest CVIC (defined as > 10 imaging procedures/year). Multivariable regression models were used to estimate likelihood of imaging receipt by distance. Results/Data: A total of 64,260,530 individuals (age 73.0 ± 8 years, 54.6% female, 80.1% White) received care at 3,886 CVICS. The vast majority (95%) lived within 16 miles of a CVIC, but varied considerably by modality with substantially greater distances for cardiac CT, cardiac MRI, and PET (Median [IQR] distance in miles to CVIC, Echo: 3.4 [0.4-7.0]; SPECT: 3.8 [1.3-7.9]; cardiac CT: 8.1 [3.7-21.3]; cardiac MRI: 17.4 [7.3-43.3]; PET: 88.9 [26.2-194.6]). CVICs performing cardiac CT, cardiac MRI, or PET were nearly all in metropolitan locations. Likelihood of imaging receipt was triphasic: lowest within 0-10 miles of a CVIC (mainly metropolitan locations), greater within 10-15 miles, and then increasingly lower beyond 15 miles ( Figure 1 ). Sites beyond the 16-mile threshold were predominantly in the Midwest and West ( Figure 2 ). Conclusions: In this study of 64 million Medicare beneficiaries, 2018-2022, 95% of the US population was found to live within 16 miles of a CVIC, though varied by modality. While those living > 16 miles from a CVIC received less imaging, the lowest use was identified among individuals living in the same zip code as a CVIC, suggesting a possible disconnect between access to imaging care and proximity.

Challenges in a Biological Definition of Alzheimer Disease.

It has been suggested that the diagnostic landscape of Alzheimer disease (AD) is undergoing a profound transformation, marked by a shift toward a biomarker-based approach, as proposed by the Revised Criteria for Diagnosis and Staging of Alzheimer's Disease. These criteria advocate for diagnosing AD solely on biomarkers, without requiring clinical symptoms. This article explores the drivers behind this transition, primarily influenced by the Food and Drug Administration's approval of amyloid-lowering treatments. We evaluate the proposed criteria, which allow for an AD diagnosis based on amyloid "A" or phosphorylated tau "T1" positivity through surrogate amyloid PET imaging, CSF, or plasma biomarkers, and consider the arguments for and against their use. The merits of the new criteria include a clearer definition of AD, which is currently used interchangeably to refer to both the presence of neuropathology and the clinical syndrome. We argue that a purely biological definition risks a category error and emphasize the need for longitudinal data to establish the lifetime risk of dementia in amyloid-positive and tau-positive individuals. We also caution against limiting the scope of biomarker-based AD diagnosis to amyloid and tau alone. In conclusion, we recommend that the criteria remain within the research domain for the present while advocating for the considered adoption of plasma biomarkers in clinical practice.

Lyme serology in non-endemic countries: Why do we request it and what do we find?

The primary objectives of our research were to analyze Lyme serology results from pediatric patients, identify the clinical reasons for ordering these tests, and assess the clinical relevance of the serology results in the context of Lyme disease. Our study, conducted at a reference pediatric hospital in a non-endemic region for Lyme disease, included all pediatric patients for whom Lyme serology was requested. ELISA and Western blot results were documented. Presenting complaints and findings at the time of admission were recorded. Diagnostic tests were categorized as first-step tests if conducted during the initial visit. Subsequent tests, sent due to the inability to establish a diagnosis based on the results of the first-line tests, were defined as second and third-step tests, respectively. 219 patients, for whom Lyme serology had been requested for varying reasons, were included in our study. The most prevalent complaints and indications observed in the presentations of patients with serology requests were as follows: headache(24.7%), paresis/paresthesia(14.6%), and painful or blurred vision(13.2%). Serology was primarily requested in the presence of neurological(59.4%) and ocular symptoms(13.2%). 68% of the tests were requisitioned during the initial consultation. Nevertheless, it was discerned that no patient received a diagnosis of Lyme disease. According to the guidelines, Lyme serology should only be performed when there is a realistic possibility of exposure to infected ticks, particularly in patients who have had untreated erythema migrans or those with a history of a tick bite presenting with unexplained joint or neurological symptoms. Our data, in line with these guidelines, suggests that unnecessary Lyme serology testing in non-endemic areas, where exposure is highly unlikely, may lead to false-positive results and unnecessary follow-up testing, as illustrated by the high rate of false-positive ELISAs in our cohort.

Abstract 4136403: Impact of Coronary Artery Calcium Scores on Cardiovascular Risk and Preventive Therapies: A Systematic Review and Meta-Analysis

Introduction: Coronary artery calcium (CAC) serves as a key marker of atherosclerosis and is widely utilized in the noninvasive evaluation of coronary artery disease. We aim to compare cardiovascular risk outcomes and the likelihood of initiating or continuing pharmacological and lifestyle preventive therapies in asymptomatic and symptomatic patients with a CAC score greater than zero and those with a CAC score of zero with no established diagnosis of coronary artery disease. Methods: We performed systematic searches of PubMed, Scopus, Google Scholar, Cochrane Central, and ClinicalTrials.gov from inception to May 2024. Our search focused on randomized controlled trials and observational studies comparing outcomes between patients with a CAC score greater than zero and those with a CAC score of zero. Results: We included 53 observational studies, incorporating data from 210,672 asymptomatic and 32,477 symptomatic patients. Over a mean follow-up duration of 8.6 years, we observed that a CAC score greater than zero significantly correlated with an elevated risk of major adverse cardiovascular events (MACE) (OR: 5.58; 95% CI: 4.05–7.69; P < 0.00001). This association extended to all-cause mortality (OR: 3.90; 95% CI: 2.78–5.48; P < 0.00001), myocardial infarction (OR: 4.01; 95% CI: 3.45–4.65; P < 0.00001), and the need for revascularization (OR: 11.90; 95% CI: 6.82–20.78; P < 0.00001). Patients with a CAC score greater than zero were more likely to commence and continue aspirin therapy (OR: 2.45; 95% CI: 1.56–3.83; P < 0.0001 for initiation and OR: 1.71; 95% CI: 1.31–2.23; P < 0.0001 for continuation). Similar trends were noted for lipid-lowering medications (OR: 2.44; 95% CI: 1.57–3.79; P < 0.0001 for initiation and OR: 2.48; 95% CI: 1.53–4.01; P = 0.0002 for continuation), as well as the initiation of blood pressure medications (OR: 1.94; 95% CI: 1.61–2.33; P < 0.00001). Furthermore, a CAC score greater than zero was linked to increased adoption of lifestyle modifications, including enhanced physical activity (OR: 1.84; 95% CI: 1.41–2.41; P < 0.00001) and dietary changes (OR: 1.94; 95% CI: 1.52–2.49; P < 0.00001). Conclusion: Patients with elevated CAC scores are more likely to initiate and continue preventive pharmacological therapies and adopt beneficial lifestyle modifications. These findings underscore the importance of CAC scoring in guiding the management and prevention of cardiovascular disease in both asymptomatic and symptomatic populations.

Anion and Substituents Effect on Spectral-Kinetic and Biological Characteristics of Spiropyran Salts.

Spiropyran salts containing a cationic vinyl-3H-indolium moiety are characterized by NIR absorption and fluorescence of their merocyanine forms. This feature makes them promising fluorescent probes and markers for bioimaging. The article focuses on the synthesis and study of the spectral, kinetic and toxic characteristics of such compounds with respect to various substituents in different moieties and the type of anion. A detailed analysis of the acquired data made it possible to draw some important conclusions regarding the influence of structural factors, which will be very useful for the further rational design of such derivatives. In particular, it was shown that the counterion has minimal effect on the spectral and kinetic characteristics of the dyes but dramatically affects the toxicity of the compounds. Following selection of the most appropriate compounds, bioimaging experiments were carried out to visualize planktonic bacteria and bacterial biofilms of E. coli and A. calcoaceticus. The ability to visualize biofilms is critical for the diagnosis of chronic diseases. By the results of molecular docking a theoretical interaction pattern between spiropyran molecules and DNA was proposed.

Diagnosis of Alzheimer's disease using plasma biomarkers adjusted to clinical probability.

Recently approved anti-amyloid immunotherapies for Alzheimer's disease (AD) require evidence of amyloid-β pathology from positron emission tomography (PET) or cerebrospinal fluid (CSF) before initiating treatment. Blood-based biomarkers promise to reduce the need for PET or CSF testing; however, their interpretation at the individual level and the circumstances requiring confirmatory testing are poorly understood. Individual-level interpretation of diagnostic test results requires knowledge of disease prevalence in relation to clinical presentation (clinical pretest probability). Here, in a study of 6,896 individuals evaluated from 11 cohort studies from six countries, we determined the positive and negative predictive value of five plasma biomarkers for amyloid-β pathology in cognitively impaired individuals in relation to clinical pretest probability. We observed that p-tau217 could rule in amyloid-β pathology in individuals with probable AD dementia (positive predictive value above 95%). In mild cognitive impairment, p-tau217 interpretation depended on patient age. Negative p-tau217 results could rule out amyloid-β pathology in individuals with non-AD dementia syndromes (negative predictive value between 90% and 99%). Our findings provide a framework for the individual-level interpretation of plasma biomarkers, suggesting that p-tau217 combined with clinical phenotyping can identify patients where amyloid-β pathology can be ruled in or out without the need for PET or CSF confirmatory testing.

Abstract 4123530: Disparities in Hypertension Treatment Intensification among Adults in a Large United States Healthcare System

Background: There are racial/ethnic and socioeconomic status (SES) disparities in hypertension control, even in populations with regular care. It is unclear if clinical inertia by clinicians contributes to these disparities. Objectives: To quantify racial/ethnic and SES disparities in treatment intensification (TI) among adults with hypertension. Methods: We performed a retrospective cohort study using electronic health records from Johns Hopkins primary care practices from January 2018 to December 2022. We included non-Hispanic Black (Black) and non-Hispanic White (White) adults with at least 2 elevated blood pressure readings (>=140/90 mm Hg) in the 12 months prior to the index primary care visit and a diagnosis of essential hypertension. We excluded patients with diagnoses of end-stage renal disease, liver disease, dementia, and metastatic solid tumor. Patients could contribute multiple index visits. We measured SES using the national Area Deprivation Index, which comprises 17 neighborhood-level SES indicators, and dichotomized the population into high and low median SES groups. We defined TI as a change in antihypertensive class or dose increase within 14 days of the index visit. We balanced covariates reflecting clinical need at the index visit (age, sex, blood pressure readings, cardiometabolic comorbidities, body mass index) and the index visit’s calendar time across racial/ethnic and SES groups using inverse probability weighting. We estimated annual race/ethnic and SES specific TI rates and risk difference measures of disparity with 95% confidence intervals (CI) using a clustered bootstrap. Results: The cohort included 25,287 patients (11,800 Black, 13,487 White) with 79,898 eligible visits. Black patients were more often younger, female, and residing in low SES areas compared to White patients. In 2018, TI rates for Black patients (37%, 95% CI: 36-38%) were slightly lower than rates for White patients (39%, 95% CI: 38-40%). In 2022, the disparity persisted, with the Black patients’ TI rate lower by 3% (95% CI: -5%, -1%). Similarly, low SES patients had lower TI rates than high SES patients by 2% to 3% over the study period. Conclusion: Black patients and those from low SES consistently had modestly lower rates of TI compared to White patients and those from high SES. It will be important to identify the drivers and clinical implications of these treatment disparities.

Abstract 4135025: Inflammation-related 5-hydroxymethylation signatures as markers for clinical presentations of coronary artery disease

Background: Coronary angiography, an invasive method, remains the gold standard technique for coronary artery disease (CAD) diagnosis in clinical practice. However, due to its invasive nature, it cannot be used for mass population screening. There is an urgent need in the clinical to identify new liquid biopsy molecular markers for screening patients with CAD. Methods: We utilized 5hmC-Seal to generate genome-wide 5hmC profiles in plasma cell-free DNA (cfDNA) from 724 CAD patients. To investigate the correlation between 5hmC modifications and CAD, we separated patients into training and validation cohorts and developed a 5hmC-based logistic regression model from the training cohort to predict CAD patients in the validation cohort. Furthermore, we incorporated an external cohort comprising 167 samples to validate the reliability of our model. Results: We found that the differentially hydroxymethylated genes (DhMGs) in people with the normal coronary artery (NCA), nonfatal stable CAD (sCAD), non-ST-elevation myocardial infarction (NSTEMI), as well as ST-elevation myocardial infarction (STEMI) enriched in neutrophil-mediated inflammatory response pathways, and inflammation-related DhMGs can reflect the progression of atherosclerotic plaque. Furthermore, we used machine learning algorithms to develop a predictive model for predicting the occurrence of myocardial infarction (MI) based on 19 inflammatory markers, and the predictive model could effectively predict MI (NSTEMI and STEMI) (AUC = 0.913). Simultaneously, our model has demonstrated strong performance in external validation. Conclusions: We have identified a novel biomarker, the inflammation-associated 5hmC markers, which can reflect the clinical manifestations of CAD and can effectively predict the occurrence of MI.

An attention-based deep network for plant disease classification

Plant disease classification using machine learning in a real agricultural field environment is a difficult task. Often, an automated plant disease diagnosis method might fail to capture and interpret discriminatory information due to small variations among leaf sub-categories. Yet, modern Convolutional Neural Networks (CNNs) have achieved decent success in discriminating various plant diseases using leave images. A few existing methods have applied additional pre-processing modules or sub-networks to tackle this challenge. Sometimes, the feature maps ignore partial information for holistic description by part-mining. A deep CNN that emphasizes integration of partial descriptiveness of leaf regions is proposed in this work. The efficacious attention mechanism is integrated with high-level feature map of a base CNN for enhancing feature representation. The proposed method focuses on important diseased areas in leaves, and employs an attention weighting scheme for utilizing useful neighborhood information. The proposed Attention-based network for Plant Disease Classification (APDC) method has achieved state-of-the-art performances on four public plant datasets containing visual/thermal images. The best top-1 accuracies attained by the proposed APDC are: PlantPathology 97.74%, PaddyCrop 99.62%, PaddyDoctor 99.65%, and PlantVillage 99.97%. These results justify the suitability of proposed method.