Patients' reactions to the implementation of Artificial Intelligence (AI) in healthcare range from adverse to favourable. While AI holds the promise of revolutionising healthcare by enhancing, accelerating, and improving the precision of care services, our understanding of patients' reactions to these paradigm shifts remains limited. In particular, little is known about the extent to which patients are receptive to independently use AI-enabled applications for diagnosis. This research seeks to develop a holistic, context-specific model capturing both the negative and positive cognitive responses of patients utilising AI-powered diagnostic services. Employing a sequential mixed-methods approach, the study draws on Behavioural Reasoning Theory to decode patients' cognitive reactions, including their reasons for and reasons giants using such applications. The research begins with a qualitative exploration, analysing user reviews to identify context-specific barriers and motivators. Building on these qualitative insights, the model's empirical validity is tested through a quantitative phase involving survey data analysis. Our findings provide a nuanced understanding of the context-dependent factors shaping patients' cognitive responses to AI-enabled diagnostic services, offering valuable insights for the design and implementation of patient-centred AI solutions in healthcare.

Population health programs rely on healthcare predictive models to allocate resources, yet models can perpetuate biases that exacerbate health disparities among marginalized communities.

We developed the Bias-reduction and Equity Framework for Assessing, Implementing, and Redesigning (BE-FAIR) healthcare predictive models, an applied framework tested within a large health system using a population health predictive model, aiming to minimize bias and enhance equity.

Retrospective cohort study conducted at an academic medical center. Data collected from September 30, 2020, to October 1, 2022, were analyzed to assess bias resulting from model use.

Primary care or payer-attributed patients at the medical center identified through electronic health records and claims data. Participants were stratified by race-ethnicity, gender, and social vulnerability defined by the Healthy Places Index (HPI).

BE-FAIR implementation involved steps such as an anti-racism lens application, de-siloed team structure, historical intervention review, disaggregated data analysis, and calibration evaluation.

The primary outcome was the calibration and discrimination of the model across different demographic groups, measured by logistic regression and area under the receiver operating characteristic curve (AUROC).

The study population consisted of 114,311 individuals with a mean age of 43.4 years (SD 24.0 years), 55.4% female, and 59.5% white/Caucasian. Calibration differed by race-ethnicity and HPI with significantly lower predicted probabilities of hospitalization for African Americans (0.129±0.051, p=0.016), Hispanics (0.133±0.047, p=0.004), AAPI (0.120±0.051, p=0.018), and multi-race (0.245±0.087, p=0.005) relative to white/Caucasians and for individuals in low HPI areas (0 - 25%, 0.178±0.042, p<0.001; 25 - 50%, 0.129±0.044, p=0.003). AUROC values varied among demographic groups.

The BE-FAIR framework offers a practical approach to address bias in healthcare predictive models, guiding model development, and implementation. By identifying and mitigating biases, BE-FAIR enhances the fairness and equity of healthcare delivery, particularly for minoritized groups, paving the way for more inclusive and effective population health strategies.

Periprosthetic joint infection leads to significant morbidity and mortality after total knee arthroplasty. Preoperative and perioperative risk prediction and assessment tools are lacking in Asia. This study developed the first machine learning model for individualized prediction of periprosthetic joint infection following primary total knee arthroplasty in this demographic.

A retrospective analysis was conducted on 3,483 primary total knee arthroplasty (81 with periprosthetic joint infection) from 1998 to 2021 in a Chinese tertiary and quaternary referral academic center. We gathered 60 features, encompassing patient demographics, operation-related variables, laboratory findings, and comorbidities. Six of them were selected after univariate and multivariate analysis. Five machine learning models were trained with stratified 10-fold cross-validation and assessed by discrimination and calibration analysis to determine the optimal predictive model.

The balanced random forest model demonstrated the best predictive capability with average metrics of 0.963 for the area under the receiver operating characteristic curve, 0.920 for balanced accuracy, 0.938 for sensitivity, and 0.902 for specificity. The significant risk factors identified were long operative time (OR, 9.07; p = 0.018), male gender (OR, 3.11; p < 0.001), ASA > 2 (OR, 1.68; p = 0.028), history of anemia (OR, 2.17; p = 0.023), and history of septic arthritis (OR, 4.35; p = 0.030). Spinal anesthesia emerged as a protective factor (OR, 0.55; p = 0.022).

Our study presented the first machine learning model in Asia to predict periprosthetic joint infection following primary total knee arthroplasty. We enhanced the model's usability by providing global and local interpretations. This tool provides preoperative and perioperative risk assessment for periprosthetic joint infection and opens the potential for better individualized optimization before total knee arthroplasty.

In this Matters Arising, we critically examine the data processing and computational challenges highlighted under the high-resolution, three-dimensional reconstruction of human cortical tissue by Shapson-Coe et al. While the study represents a technical milestone in connectomics, involving a 1.4-petabyte dataset derived from mapping a cubic millimeter of temporal cortex, the findings also reveal the substantial obstacles inherent in scaling such approaches to the entire human brain. Beyond the application of artificial intelligence (AI) for segmentation and synapse detection, the study underscores the immense complexity of data acquisition, cleaning, alignment, and visualization at this scale. This article contextualizes these challenges by comparing the computational and infrastructural requirements of the Shapson-Coe work to other large-scale neuroscience initiatives, such as the fruit fly brain atlas, and explores emerging technologies like quantum computing and neuromorphic hardware as potential solutions. Additionally, we discuss the ethical and logistical implications of managing zettabyte-scale datasets and emphasize the necessity of international collaboration to achieve the ambitious goal of mapping the human connectome. By critically addressing these challenges and potential solutions, this article aims to guide future advancements in the field of connectomics and their transformative applications in neuroscience, artificial intelligence, and medicine.

The European CORE-MD consortium (Coordinating Research and Evidence for Medical Devices) proposes a score for medical devices incorporating artificial intelligence or machine learning algorithms. Its domains are summarised as valid clinical association, technical performance, and clinical performance. High scores indicate that extensive clinical investigations should be undertaken before regulatory approval, whereas lower scores indicate devices for which less pre-market clinical evaluation may be balanced by more post-market evidence.

Community acquired pneumonia (CAP) is a common cause of hospital admission. CAP carries significant risk of adverse outcomes including organ dysfunction, intensive care unit (ICU) admission and death. Earlier admission to ICU for those with severe CAP is associated with better outcomes. Traditional prediction models are used in clinical practice to predict the severity of CAP. However, accuracy of predicting severity may be improved by using machine learning (ML) based models with added advantages of automation and speed. This systematic review evaluates the evidence base of ML-prediction tools in predicting CAP severity.

MEDLINE, EMBASE and PubMed were systematically searched for studies that used ML-based models to predict mortality and/or ICU admission in CAP patients, where a performance metric was reported.

11 papers including a total of 351,365 CAP patients were included. All papers predicted severity and four predicted ICU admission. Most papers applied multiple ML algorithms to datasets and derived area under the receiver operator characteristic curve (AUROC) of 0.98 at best performance and 0.57 at worst, with a mixed performance against traditional prediction tools.

Although ML models showed good performance at predicting CAP severity, the variables selected for inclusion in each model varied significantly which limited comparisons between models and there was a lack of reproducible data, limiting validity. Future research should focus on validating ML predication models in multiple cohorts to derive robust, reproducible performance measures, and to demonstrate a benefit in terms of patient outcomes and resource use.

Biliary abnormalities in autoimmune hepatitis (AIH) and interface hepatitis in primary biliary cholangitis (PBC) occur frequently, and misinterpretation may lead to therapeutic mistakes with a negative impact on patients. This study investigates the use of a deep learning (DL)-based pipeline for the diagnosis of AIH and PBC to aid differential diagnosis.

We conducted a multicenter study across six European referral centers, and built a library of digitized liver biopsy slides dating from 1997 to 2023. A training set of 354 cases (266 AIH and 102 PBC) and an external validation set of 92 cases (62 AIH and 30 PBC) were available for analysis. A novel DL model, the autoimmune liver neural estimator (ALNE), was trained on whole-slide images (WSIs) with H&E staining, without human annotations. The ALNE model was evaluated against clinico-pathological diagnoses and tested for interobserver variability among general pathologists.

The ALNE model demonstrated high accuracy in differentiating AIH from PBC, achieving an area under the receiver operating characteristic curve of 0.81 in external validation. Attention heatmaps showed that ALNE tends to focus more on areas with increased inflammation, associating such patterns predominantly with AIH. A multivariate explainable ML model revealed that PBC cases misclassified as AIH more often had ALP values between 1 × upper limit of normal (ULN) and 2 × ULN, coupled with AST values above 1 × ULN. Inconsistency among general pathologists was noticed when evaluating a random sample of the same cases (Fleiss's kappa value 0.09).

The ALNE model is the first system generating a quantitative and accurate differential diagnosis between cases with AIH or PBC.

This study demonstrates the significant potential of the autoimmune liver neural estimator model, a transformer-based deep learning system, in accurately distinguishing between autoimmune hepatitis and primary biliary cholangitis using digitized liver biopsy slides without human annotation. The scientific justification for this work lies in addressing the challenge of differentiating these conditions, which often present with overlapping features and can lead to therapeutic mistakes. In addition, there is need for quantitative assessment of information embedded in liver biopsies, which are currently evaluated on qualitative or semi-quantitative methods. The results of this study are crucial for pathologists, researchers, and clinicians, providing a reliable diagnostic tool that reduces interobserver variability and improves diagnostic accuracy of these conditions. Potential methodological limitations, such as the diversity in scanning techniques and slide colorations, were considered, ensuring the robustness and generalizability of the findings.

The timing of a woman's final menstrual period (FMP) in relation to her age is considered a valuable indicator of overall health, being associated with cardiovascular, bone health, reproductive, and general mortality outcomes.

This work aimed to evaluate the relationship between hormones and the "time to FMP" when daily hormone trajectories are characterized by their 1) entropy, and 2) deviation from premenopausal/stable cycle patterns (representing a textbook "gold standard"; GS).

As part of the Study of Women's Health Across the Nation, urinary luteinizing hormone (LH), follicle-stimulating hormone (FSH), estrogen conjugates (E1C), and pregnanediol glucuronide (PDG) were measured daily from a multiracial sample of 549 mid-life women for the duration of one menstrual cycle. Hormone trajectories were mapped onto a plane with axes representing Fuzzy entropy (FuzzEn) and the normalized dynamic time warping distance (DTW) from the GS.

Viewing FSH, E1C, PDG, and LH through this lens reveals that, contrary to existing wisdom, PDG stands out as a powerful predictor/descriptor of "time to FMP." Using cluster analyses to discretize PDG on the DTW/FuzzEn plane yields statistically different survival curves, and Cox proportional hazards analyses confirm that this separation persists in the presence of known covariates of FSH, antimüllerian hormone, age, body mass index, financial hardship, smoking status, and cycle length.

Since PDG is generally not considered a predictor/descriptor of ovarian aging, this work validates the DTW/FuzzEn analytical framework and introduces another metric/hormone to be used in FMP-related preventive care.

Accurate prediction of clinical outcomes following percutaneous coronary intervention (PCI) is essential for mitigating risk and peri-procedural planning. Traditional risk models have demonstrated a modest predictive value. Machine learning (ML) models offer an alternative risk stratification that may provide improved predictive accuracy.

This study was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses, Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies and Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis guidelines. PubMed, EMBASE, Web of Science, and Cochrane databases were searched until 1 November 2023 for studies comparing ML models with traditional statistical methods for event prediction after PCI. The primary outcome was comparative discrimination measured by C-statistics with 95% confidence intervals (CIs) between ML models and traditional methods in estimating the risk of all-cause mortality, major bleeding, and the composite outcome major adverse cardiovascular events (MACE). Thirty-four models were included across 13 observational studies (4 105 916 patients). For all-cause mortality, the pooled C-statistic for top-performing ML models was 0.89 (95%CI, 0.84-0.91), compared with 0.86 (95% CI, 0.80-0.93) for traditional methods (P = 0.54). For major bleeding, the pooled C-statistic for ML models was 0.80 (95% CI, 0.77-0.84), compared with 0.78 (95% CI, 0.77-0.79) for traditional methods (P = 0.02). For MACE, the C-statistic for ML models was 0.83 (95% CI, 0.75-0.91), compared with 0.71 (95% CI, 0.69-0.74) for traditional methods (P = 0.007). Out of all included models, only one model was externally validated. Calibration was inconsistently reported across all models. Prediction Model Risk of Bias Assessment Tool demonstrated a high risk of bias across all studies.

Machine learning models marginally outperformed traditional risk scores in the discrimination of MACE and major bleeding following PCI. While integration of ML algorithms into electronic healthcare systems has been hypothesized to improve peri-procedural risk stratification, immediate implementation in the clinical setting remains uncertain. Further research is required to overcome methodological and validation limitations.

With an increasing focus on precision medicine in medical research, numerous studies have been conducted in recent years to clarify the relationship between treatment effects and patient characteristics. The treatment effects for patients with different characteristics are always heterogeneous, and therefore, various heterogeneous treatment effect machine learning estimation methods have been proposed owing to their flexibility and high estimation accuracy. However, most machine learning methods rely on black-box models, preventing direct interpretation of the relationship between patient characteristics and treatment effects. Moreover, most of these studies have focused on continuous or binary outcomes, although survival outcomes are also important in medical research. To address these challenges, we propose a heterogeneous treatment effect estimation method for survival data based on RuleFit, an interpretable machine learning method. Numerical simulation results confirmed that the prediction performance of the proposed method was comparable to that of existing methods. We also applied a dataset from an HIV study, the AIDS Clinical Trials Group Protocol 175 dataset, to illustrate the interpretability of the proposed method using real data. Consequently, the proposed survival causal rule ensemble method provides an interpretable model with sufficient estimation accuracy.

Artificial Intelligence (AI) is the capability of computational systems to perform tasks that require human-like cognitive functions, such as reasoning, learning, and decision-making. Unlike human intelligence, AI does not involve sentience or consciousness but focuses on data processing, pattern recognition, and prediction through algorithms and learned experiences. In healthcare including neuroscience, AI is valuable for improving prevention, diagnosis, prognosis, and surveillance.

This qualitative study aimed to investigate the acceptability of AI in Medicine (AIIM) and to elucidate any technical and scientific, as well as social and ethical issues involved. Twenty-five doctors from various specialties were carefully interviewed regarding their views, experience, knowledge, and attitude toward AI in healthcare.

Content analysis confirmed the key ethical principles involved: confidentiality, beneficence, and non-maleficence. Honesty was the least invoked principle. A thematic analysis established four salient topic areas, i.e., advantages, risks, restrictions, and precautions. Alongside the advantages, there were many limitations and risks. The study revealed a perceived need for precautions to be embedded in healthcare policies to counter the risks discussed. These precautions need to be multi-dimensional.

The authors conclude that AI should be rationally guided, function transparently, and produce impartial results. It should assist human healthcare professionals collaboratively. This kind of AI will permit fairer, more innovative healthcare which benefits patients and society whilst preserving human dignity. It can foster accuracy and precision in medical practice and reduce the workload by assisting physicians during clinical tasks. AIIM that functions transparently and respects the public interest can be an inspiring scientific innovation for humanity.

This study assessed whether deep learning applied to routine outpatient chest X-rays (CXRs) can identify individuals at high risk for incident chronic obstructive pulmonary disease (COPD).

Using cancer screening trial data, we previously developed a convolutional neural network (CXR-Lung-Risk) to predict lung-related mortality from a CXR image. In this study, we externally validated CXR-Lung-Risk to predict incident COPD from routine CXRs. We identified outpatients without lung cancer, COPD, or emphysema who had a CXR taken from 2013-2014 at a Mass General Brigham site in Boston, Massachusetts. The primary outcome was 6-year incident COPD. Discrimination was assessed using AUC compared to the TargetCOPD clinical risk score. All analyses were stratified by smoking status. A secondary analysis was conducted in the Project Baseline Health Study (PBHS) to test associations between CXR-Lung-Risk with pulmonary function and protein abundance.

The primary analysis consisted of 12,550 ever-smokers (mean age 62·4±6·8 years, 48.9% male, 12.4% rate of 6-year COPD) and 15,298 never-smokers (mean age 63·0±8·1 years, 42.8% male, 3.8% rate of 6-year COPD). CXR-Lung-Risk had additive predictive value beyond the TargetCOPD score for 6-year incident COPD in both ever-smokers (CXR-Lung-Risk + TargetCOPD AUC: 0·73 [95% CI: 0·72-0·74] vs. TargetCOPD alone AUC: 0·66 [0·65-0·68], p<0·01) and never-smokers (CXR-Lung-Risk + TargetCOPD AUC: 0·70 [0·67-0·72] vs. TargetCOPD AUC: 0·60 [0·57-0·62], p<0·01). In secondary analyses of 2,097 individuals in the PBHS, CXR-Lung-Risk was associated with worse pulmonary function and with abundance of SCGB3A2 (secretoglobin family 3A member 2) and LYZ (lysozyme), proteins involved in pulmonary physiology.

In external validation, a deep learning model applied to a routine CXR image identified individuals at high risk for incident COPD, beyond known risk factors.

The Project Baseline Health Study and this analysis were funded by Verily Life Sciences, San Francisco, California.

NCT03154346.

The development and implementation of artificial intelligence (AI), particularly deep learning (DL) models, has generated significant interest across various fields of gastroenterology. While research in luminal endoscopy has seen rapid translation to clinical practice with approved AI devices, its potential extends far beyond, offering promising benefits for biliopancreatic endoscopy like optical characterization of strictures during cholangioscopy or detection and classification of pancreatic lesions during diagnostic endoscopic ultrasound (EUS). This narrative review provides an up-to-date of the latest literature and available studies in this field. Serving as a comprehensive guide to the current landscape of AI in biliopancreatic endoscopy, emphasizing technological advancements, main applications, ethical considerations, and future directions for research and clinical implementation.

Syndromic heritable thoracic aortic diseases (sHTAD), such as Marfan (MFS) or Loeys-Dietz (LDS) syndromes, involve high risk of life threatening aortic events. Diagnosis of syndromic features alone is difficult, and negative genetic tests do not necessarily exclude a genetic or hereditary condition. Periodic 3D imaging of the aorta is recommended in patients with aortic disease. Thus, an imaging-based approach aimed at identifying unique features of aortic geometry can be highly effective for diagnosing sHTAD and assessing risk. In this study, we present a method that can help identify the manifestations of sHTAD by focusing on the entire geometry of the thoracic aorta, rather than only using measurements of dilation of the aortic root. We analyze the geometric phenotype of 97 patients with genetically confirmed sHTAD (79 MF and 18 LDS) and of 45 healthy volunteers, using 3D aorta meshes obtained from phase contrast-enhanced magnetic resonance angiograms computed from 4D flow cardiac magnetic resonance. We build a geometric encoding of the aorta, based on a vessel coordinate system, and use several mathematical models to discriminate between controls and patients with sHTAD: a baseline scenario, based on aortic root dimensions only, a descriptor typically used in sHTAD patients; a low dimensional scenario, with a reduce encoding using principal component analysis; and a high-dimensional scenario, which included the full coefficient representation for geometry encoding, aiming to capture finer geometric details. The results indicate that considering the anatomy of the whole thoracic aorta can improve predictive ability. We achieve precision and sensitivity values over 0.8, with a specificity of over 70% in all the models used, while a single value classifiers (based only on aortic root diameter) demonstrated a trade-off between sensitivity and specificity. Using the mathematical properties of the vessel coordinate system representation, feature importance is mapped onto a set of anatomical traits that are used by the models to do the classification, thus providing interpretability of the results. This analysis indicates that in addition to the diameter of the aortic root, aortic elongation and a narrowing of the descending thoracic aorta may be markers of positive sHTAD.

Identification of mature sperm at microdissection testicular sperm extraction (mTESE) is a crucial step of sperm retrieval to help patients with non-obstructive azoospermia (NOA) proceed to intracytoplasmic sperm injection. Touch print smear (TPS) cytology allows immediate interpretation and prompt sperm identification intraoperatively. In this study, we leverage machine learning (ML) to facilitate TPS reading and conquer the learning curve for new operators.

One hundred seventy-six microscopic TPS images from the testicular specimen of patients with azoospermia at Taipei Veterans General Hospital were retrospectively collected, including categories of Sertoli cell, primary spermatocytes, round spermatids, elongated spermatids, immature sperm, and mature sperm. Among them, 118 images were assigned as the training set and 29 images as the validation set. RetinaNet (Lin et al. in IEEE Trans Pattern Anal Mach Intell. 42:318-327, 2020), a one-stage detection framework, was adopted for cell detection. The performance was evaluated at the cell level with average precision (AP) and recall, and the precision-recall (PR) curve was displayed among an independent testing set that contains 29 images that aim to assess the model.

The training set consisted of 4772 annotated cells, including 1782 Sertoli cells, 314 primary spermatocytes, 443 round spermatids, 279 elongated spermatids, 504 immature sperm, and 1450 mature sperm. This study demonstrated the performance of each category and the overall AP and recall on the validation set, which were 80.47% and 96.69%. The overall AP and recall were 79.48% and 93.63% on the testing set, while increased to 85.29% and 93.80% once the post-meiotic cells were merged into one category.

This study proposed an innovative approach that leveraged ML methods to facilitate the diagnosis of spermatogenesis at mTESE for patients with NOA. With the assistance of ML techniques, surgeons could determine the stages of spermatogenesis and provide timely histopathological diagnosis for infertile males.

Pathological complete response (pCR) serves as a critical measure of the success of neoadjuvant chemotherapy (NAC) in breast cancer, directly influencing subsequent therapeutic decisions. With the continuous advancement of artificial intelligence, methods for early and accurate prediction of pCR are being extensively explored. In this study, we propose a cross-modal multi-pathway automated prediction model that integrates temporal and spatial information. This model fuses digital pathology images from biopsy specimens and multi-temporal ultrasound (US) images to predict pCR status early in NAC. The model demonstrates exceptional predictive efficacy. Our findings lay the foundation for developing personalized treatment paradigms based on individual responses. This approach has the potential to become a critical auxiliary tool for the early prediction of NAC response in breast cancer patients.

This study aimed to (1) develop and validate a natural language processing model to identify the presence of pulmonary embolism (PE) based on real-time radiology reports and (2) identify low-risk PE patients based on previously validated risk stratification scores using variables extracted from the electronic health record at the time of diagnosis. The combination of these approaches yielded an natural language processing-based clinical decision support tool that can identify patients presenting to the emergency department (ED) with low-risk PE as candidates for outpatient management.

Data were curated from all patients who received a PE-protocol computed tomography pulmonary angiogram (PE-CTPA) imaging study in the ED of a 3-hospital academic health system between June 1, 2018 and December 31, 2020 (n=12,183). The "preliminary" radiology reports from these imaging studies made available to ED clinicians at the time of diagnosis were adjudicated as positive or negative for PE by the clinical team. The reports were then divided into development, internal validation, and temporal validation cohorts in order to train, test, and validate an natural language processing model that could identify the presence of PE based on unstructured text. For risk stratification, patient- and encounter-level data elements were curated from the electronic health record and used to compute a real-time simplified pulmonary embolism severity (sPESI) score at the time of diagnosis. Chart abstraction was performed on all low-risk PE patients admitted for inpatient management.

When applied to the internal validation and temporal validation cohorts, the natural language processing model identified the presence of PE from radiology reports with an area under the receiver operating characteristic curve of 0.99, sensitivity of 0.86 to 0.87, and specificity of 0.99. Across cohorts, 10.5% of PE-CTPA studies were positive for PE, of which 22.2% were classified as low-risk by the sPESI score. Of all low-risk PE patients, 74.3% were admitted for inpatient management.

This study demonstrates that a natural language processing-based model utilizing real-time radiology reports can accurately identify patients with PE. Further, this model, used in combination with a validated risk stratification score (sPESI), provides a clinical decision support tool that accurately identifies patients in the ED with low-risk PE as candidates for outpatient management.

In the coming years, artificial intelligence (AI) and machine learning will likely give rise to profound changes in the field of emergency medicine, and medicine more broadly. This article discusses these anticipated changes in terms of 3 overlapping yet distinct stages of AI development. It reviews some fundamental concepts in AI and explores their relation to clinical practice, with a focus on emergency medicine. In addition, it describes some of the applications of AI in disease diagnosis, prognosis, and treatment, as well as some of the practical issues that they raise, the barriers to their implementation, and some of the legal and regulatory challenges they create.

Artificial intelligence (AI) has emerged as a potentially transformative force, particularly in the realm of emergency medicine (EM). The implementation of AI in emergency departments (ED) has the potential to improve patient care through various modalities. However, the implementation of AI in the ED presents unique challenges that influence its clinical adoption. This scoping review summarizes the current literature exploring the barriers and facilitators of the clinical implementation of AI in the ED.

We systematically searched Embase (Ovid), MEDLINE (Ovid), Web of Science, and Engineering Village. All articles were published in English through November 20th, 2023. Two reviewers screened the search results, with disagreements resolved through third-party adjudication.

A total of 8172 studies were included in the preliminary search, with 22 selected for the final data extraction. 10 studies were reviews and the remaining 12 were primary quantitative, qualitative, and mixed-methods studies. Out of the 22, 13 studies investigated a specific AI tool or application. Common barriers to implementation included a lack of model interpretability and explainability, encroachment on physician autonomy, and medicolegal considerations. Common facilitators to implementation included educating staff on the model, efficient integration into existing workflows, and sound external validation.

There is increasing literature on AI implementation in the ED. Our research suggests that the most common barrier facing AI implementation in the ED is model interpretability and explainability. More primary research investigating the implementation of specific AI tools should be undertaken to help facilitate their successful clinical adoption in the ED.

Interpretability and intuitive visualization facilitate medical knowledge generation through big data. In addition, robustness to high-dimensional and missing data is a requirement for statistical approaches in the medical domain. A method tailored to the needs of physicians must meet all the abovementioned criteria.

This study aims to develop an accessible tool for visual data exploration without the need for programming knowledge, adjusting complex parameterizations, or handling missing data. We sought to use statistical analysis using the setting of disease and control cohorts familiar to clinical researchers. We aimed to guide the user by identifying and highlighting data patterns associated with disease and reveal relations between attributes within the data set.

We introduce the attribute association graph, a novel graph structure designed for visual data exploration using robust statistical metrics. The nodes capture frequencies of participant attributes in disease and control cohorts as well as deviations between groups. The edges represent conditional relations between attributes. The graph is visualized using the Neo4j (Neo4j, Inc) data platform and can be interactively explored without the need for technical knowledge. Nodes with high deviations between cohorts and edges of noticeable conditional relationship are highlighted to guide the user during the exploration. The graph is accompanied by a dashboard visualizing variable distributions. For evaluation, we applied the graph and dashboard to the Hamburg City Health Study data set, a large cohort study conducted in the city of Hamburg, Germany. All data structures can be accessed freely by researchers, physicians, and patients. In addition, we developed a user test conducted with physicians incorporating the System Usability Scale, individual questions, and user tasks.

We evaluated the attribute association graph and dashboard through an exemplary data analysis of participants with a general cardiovascular disease in the Hamburg City Health Study data set. All results extracted from the graph structure and dashboard are in accordance with findings from the literature, except for unusually low cholesterol levels in participants with cardiovascular disease, which could be induced by medication. In addition, 95% CIs of Pearson correlation coefficients were calculated for all associations identified during the data analysis, confirming the results. In addition, a user test with 10 physicians assessing the usability of the proposed methods was conducted. A System Usability Scale score of 70.5% and average successful task completion of 81.4% were reported.

The proposed attribute association graph and dashboard enable intuitive visual data exploration. They are robust to high-dimensional as well as missing data and require no parameterization. The usability for clinicians was confirmed via a user test, and the validity of the statistical results was confirmed by associations known from literature and standard statistical inference.

Musculoskeletal injuries (MSKIs) are endemic in military populations. Thus, it is essential to identify and mitigate MSKI risks. Time-to-event machine learning models utilizing self-reported questionnaires or existing data (e.g., electronic health records) may aid in creating efficient risk screening tools.

A total of 4,222 U.S. Army Service members completed a self-report MSKI risk screen as part of their unit's standard in-processing. Additionally, participants' MSKI and demographic data were abstracted from electronic health record data. Survival machine learning models (Cox proportional hazard regression (COX), COX with splines, conditional inference trees, and random forest) were deployed to develop a predictive model on the training data (75%; n = 2,963) for MSKI risk over varying time horizons (30, 90, 180, and 365 days) and were evaluated on the testing data (25%; n = 987). Probability of predicted risk (0.00-1.00) from the final model stratified Service members into quartiles based on MSKI risk.

The COX model demonstrated the best model performance over the time horizons. The time-dependent area under the curve ranged from 0.73 to 0.70 at 30 and 180 days. The index prediction accuracy (IPA) was 12% better at 180 days than the IPA of the null model (0 variables). Within the COX model, "other" race, more self-reported pain items during the movement screens, female gender, and prior MSKI demonstrated the largest hazard ratios. When predicted probability was binned into quartiles, at 180 days, the highest risk bin had an MSKI incidence rate of 2,130.82 ± 171.15 per 1,000 person-years and incidence rate ratio of 4.74 (95% confidence interval: 3.44, 6.54) compared to the lowest risk bin.

Self-reported questionnaires and existing data can be used to create a machine learning algorithm to identify Service members' MSKI risk profiles. Further research should develop more granular Service member-specific MSKI screening tools and create MSKI risk mitigation strategies based on these screenings.

Real world evidence is crucial to understanding the diffusion of new oncologic therapies, monitoring cancer outcomes, and detecting unexpected toxicities. In practice, real world evidence is challenging to collect rapidly and comprehensively, often requiring expensive and time-consuming manual case-finding and annotation of clinical text. In this Review, we summarise recent developments in the use of artificial intelligence to collect and analyze real world evidence in oncology.

We performed a narrative review of the major current trends and recent literature in artificial intelligence applications in oncology.

Artificial intelligence (AI) approaches are increasingly used to efficiently phenotype patients and tumors at large scale. These tools also may provide novel biological insights and improve risk prediction through multimodal integration of radiographic, pathological, and genomic datasets. Custom language processing pipelines and large language models hold great promise for clinical prediction and phenotyping.

Despite rapid advances, continued progress in computation, generalizability, interpretability, and reliability as well as prospective validation are needed to integrate AI approaches into routine clinical care and real-time monitoring of novel therapies.

Artificial intelligence (AI) methods, especially machine learning and natural language processing, are increasingly affecting health professions education (HPE), including the medical school application and selection processes, assessment, and scholarship production. The rise of large language models over the past 18 months, such as ChatGPT, has raised questions about how best to incorporate these methods into HPE. The lack of training in AI among most HPE faculty and scholars poses an important challenge in facilitating such discussions. In this commentary, the authors provide a primer on the AI methods most often used in the practice and scholarship of HPE, discuss the most pressing challenges and opportunities these tools afford, and underscore that these methods should be understood as part of the larger set of statistical tools available.Despite their ability to process huge amounts of data and their high performance completing some tasks, AI methods are only as good as the data on which they are trained. Of particular importance is that these models can perpetuate the biases that are present in those training datasets, and they can be applied in a biased manner by human users. A minimum set of expectations for the application of AI methods in HPE practice and scholarship is discussed in this commentary, including the interpretability of the models developed and the transparency needed into the use and characteristics of such methods.The rise of AI methods is affecting multiple aspects of HPE including raising questions about how best to incorporate these models into HPE practice and scholarship. In this commentary, we provide a primer on the AI methods most often used in HPE and discuss the most pressing challenges and opportunities these tools afford.

This study explored dental students' and dentists' perceptions and attitudes toward artificial intelligence (AI) and analyzed differences according to professional seniority.

In September to November 2022, online surveys using Google Forms were conducted at 2 dental colleges and on 2 dental websites. The questionnaire consisted of general information (8 or 10 items) and participants' perceptions, confidence, predictions, and perceived future prospects regarding AI (17 items). A multivariate logistic regression analysis was performed on 4 questions representing perceptions and attitudes toward AI to identify highly influential factors according to position, age, sex, residence, and self-reported knowledge level about AI of respondents. Participants were reclassified into 2 subgroups based on students' years in school and 4 subgroups based on dentists' years of experience. The chi-square test or Fisher's exact test was used to determine differences between dental students and dentists and between subgroups for all 17 questions.

The study included 120 dental students and 96 dentists. Participants with high level of AI knowledge were more likely to be interested in AI compared to those with moderate or low level (adjusted OR 24.345, p < 0.001). Most dental students (60.8%) and dentists (67.7%) predicted that dental AI would complement human limitations. Dental students responded that they would actively use AI in almost all cases (40.8%), while dentists responded that they would use AI only when necessary (44.8%). Dentists with 11-20 years of experience were the most likely to disagree that AI could outperform skilled dentists (50.0%), and respondents with longer careers had higher response rates regarding the need for AI education in schools.

Knowledge level about AI emerged as the factor influencing perceptions and attitudes toward AI, with both dental students and dentists showing similar views on recognizing the potential of AI as an auxiliary tool. However, students' and dentists' willingness to use AI differed. Although dentists differed in their confidence in the abilities of AI, all dentists recognized the need for education on AI. AI adoption is becoming a reality in dentistry, which requires proper awareness, proper use, and comprehensive AI education.

Mismatch repair deficiency (d-MMR)/microsatellite instability (MSI), KRAS, and BRAF mutational status are crucial for treating advanced colorectal cancer patients. Traditional methods like immunohistochemistry or polymerase chain reaction (PCR) can be challenged by artificial intelligence (AI) based on whole slide images (WSI) to predict tumor status. In this systematic review, we evaluated the role of AI in predicting MSI status, KRAS, and BRAF mutations in colorectal cancer. Studies published in PubMed up to June 2023 were included (n = 17), and we reported the risk of bias and the performance for each study. Some studies were impacted by the reduced number of slides included in the data set and the lack of external validation cohorts. Deep learning models for the d-MMR/MSI status showed a good performance in training cohorts (mean AUC = 0.89, [0.74-0.97]) but slightly less than expected in the validation cohort when available (mean AUC = 0.82, [0.63-0.98]). Contrary to the MSI status, the prediction of KRAS and BRAF mutations was less explored with a less robust methodology. The performance was lower, with a maximum of 0.77 in the training cohort, 0.58 in the validation cohort for KRAS, and 0.82 AUC in the training cohort for BRAF.

Tuberculosis is a chronic infectious disease caused by mycobacterium tuberculosis (MTB) and is the ninth leading cause of death worldwide. It is still difficult to distinguish active TB from latent TB,but it is very important for individualized management and treatment to distinguish whether patients are active or latent tuberculosis infection.

A total of 220 subjects, including active TB patients (ATB, n = 97) and latent TB patients (LTB, n = 113), were recruited in this study .46 features about blood routine indicators and the VCS parameters (volume, conductivity, light scatter) of neutrophils(NE), monocytes(MO), and lymphocytes(LY) were collected and was constructed classification model by four machine learning algorithms(logistic regression(LR), random forest(RF), support vector machine(SVM) and k-nearest neighbor(KNN)). And the area under the precision-recall curve (AUPRC) and the area under the receiver operating characteristic curve (AUROC) to estimate of the model's predictive performance for dentifying active and latent tuberculosis infection.

After verification,among the four classifications, LR and RF had the best performance (AUROC = 1, AUPRC = 1), followed by SVM (AUROC = 0.967, AUPRC = 0.971), KNN (AUROC = 0.943, AUPRC = 0.959) in the training set. And LR had the best performance (AUROC = 0.977, AUPRC = 0.957), followed by SVM (AUROC = 0.962, AUPRC = 0.949), RF (AUROC = 0.903, AUPRC = 0.922),KNN(AUROC = 0.883, AUPRC = 0.901) in the testing set.

The machine learning algorithm classifier based on leukocyte VCS parameters is of great value in identifying active and latent tuberculosis infection.

Applying machine-based learning and synthetic cognition, commonly referred to as artificial intelligence (AI), to medicine intimates prescient knowledge. The ability of these algorithms to potentially unlock secrets held within vast data sets makes them invaluable to healthcare. Complex computer algorithms are routinely used to enhance diagnoses in fields like oncology, cardiology, and neurology. These algorithms have found utility in making healthcare decisions that are often complicated by seemingly endless relationships between exogenous and endogenous variables. They have also found utility in the allocation of limited healthcare resources and the management of end-of-life issues. With the increase in computing power and the ability to test a virtually unlimited number of relationships, scientists and engineers have the unprecedented ability to increase the prognostic confidence that comes from complex data analysis. While these systems present exciting opportunities for the democratization and precision of healthcare, their use raises important moral and ethical considerations around Christian concepts of autonomy and hope. The purpose of this essay is to explore some of the practical limitations associated with AI in medicine and discuss some of the potential theological implications that machine-generated diagnoses may present. Specifically, this article examines how these systems may disrupt the patient and healthcare provider relationship emblematic of Christ's healing mission. Finally, this article seeks to offer insights that might help in the development of a more robust ethical framework for the application of these systems in the future.

Extranodal extension (ENE) in head and neck squamous cell carcinoma (HNSCC) correlates to poor prognoses and influences treatment strategies. Deep learning may yield promising performance of predicting ENE in HNSCC but lack of transparency and interpretability. This work proposes an evolutionary learning method, called EL-ENE, to establish a more interpretable ENE prediction model for aiding clinical diagnosis.

There were 364 HNSCC patients who underwent neck lymph node (LN) dissection with pre-operative contrast-enhanced computerized tomography images. All the 778 LNs were divided into training and test sets with the ratio 8:2. EL-ENE uses an inheritable bi-objective combinatorial genetic algorithm for optimal feature selection and parameter setting of support vector machine. The diagnostic performances of the ENE prediction model and radiologists were compared using independent test datasets.

The EL-ENE model achieved the test accuracy of 80.00%, sensitivity of 81.13%, and specificity of 79.44% for ENE detection. The three radiologists achieved the mean diagnostic accuracy of 70.4%, sensitivity of 75.6%, and specificity of 67.9%. The features of gray-level texture and 3D morphology of LNs played essential roles in predicting ENE.

The EL-ENE method provided an accurate, comprehensible, and robust model to predict ENE in HNSCC with interpretable radiomic features for expanding clinical knowledge. The proposed transparent prediction models are more trustworthy and may increase their acceptance in daily clinical practice.

The promise of artificial intelligence (AI) in health care has propelled a significant uptrend in the number of clinical trials in AI and global market spending in this novel technology. In vascular surgery, this technology has the ability to diagnose disease, predict disease outcomes, and assist with image-guided surgery. As we enter an era of rapid change, it is critical to evaluate the ethical concerns of AI, particularly as it may impact patient safety and privacy. This is particularly important to discuss in the early stages of AI, as technology frequently outpaces the policies and ethical guidelines regulating it. Issues at the forefront include patient privacy and confidentiality, protection of patient autonomy and informed consent, accuracy and applicability of this technology, and propagation of health care disparities. Vascular surgeons should be equipped to work with AI, as well as discuss its novel risks to patient safety and privacy.

To identify the risk factors for 5-year cancer-specific (CSS) and overall survival (OS) and to compare the accuracy of logistic regression (LR) and artificial neural network (ANN) in the prediction of survival outcomes in T1 non-muscle-invasive bladder cancer.

This is a population-based analysis using the Surveillance, Epidemiology, and End Results database. Patients with T1 bladder cancer (BC) who underwent transurethral resection of the tumour (TURBT) between 2004 and 2015 were included in the analysis. The predictive abilities of LR and ANN were compared.

Overall 32,060 patients with T1 BC were randomly assigned to training and validation cohorts in the proportion of 70:30. There were 5691 (17.75%) cancer-specific deaths and 18,485 (57.7%) all-cause deaths within a median of 116 months of follow-up (IQR 80-153). Multivariable analysis with LR revealed that age, race, tumour grade, histology variant, the primary character, location and size of the tumour, marital status, and annual income constitute independent risk factors for CSS. In the validation cohort, LR and ANN yielded 79.5% and 79.4% accuracy in 5-year CSS prediction respectively. The area under the ROC curve for CSS predictions reached 73.4% and 72.5% for LR and ANN respectively.

Available risk factors might be useful to estimate the risk of CSS and OS and thus facilitate optimal treatment choice. The accuracy of survival prediction is still moderate. T1 BC with adverse features requires more aggressive treatment after initial TURBT.

The rapid adoption of artificial intelligence (AI) into everyday use has presented multiple issues for surgical educators to consider. In this article, the authors discuss some of the ethical aspects of academic integrity and the use of AI. These issues include the importance of understanding the current limits of AI and the inherent biases of the technology. The authors further discuss the ethical considerations of the use of AI in surgical training and in clinical use, with an emphasis on vascular surgery.

Endoscopic ultrasound (EUS) has emerged as a widely utilized tool in the diagnosis of digestive diseases. In recent years, the potential of artificial intelligence (AI) in healthcare has been gradually recognized, and its superiority in the field of EUS is becoming apparent. Machine learning (ML) and deep learning (DL) are the two main AI algorithms. This paper aims to outline the applications and prospects of artificial intelligence-assisted endoscopic ultrasound (EUS-AI) in digestive diseases over the past decade. The results demonstrated that EUS-AI has shown superiority or at least equivalence to traditional methods in the diagnosis, prognosis, and quality control of subepithelial lesions, early esophageal cancer, early gastric cancer, and pancreatic diseases including pancreatic cystic lesions, autoimmune pancreatitis, and pancreatic cancer. The implementation of EUS-AI has opened up new avenues for individualized precision medicine and has introduced novel diagnostic and treatment approaches for digestive diseases.

The COVID-19 patients in the convalescent stage noticeably have pulmonary diffusing capacity impairment (PDCI). The pulmonary diffusing capacity is a frequently-used indicator of the COVID-19 survivors' prognosis of pulmonary function, but the current studies focusing on prediction of the pulmonary diffusing capacity of these people are limited. The aim of this study was to develop and validate a machine learning (ML) model for predicting PDCI in the COVID-19 patients using routinely available clinical data, thus assisting the clinical diagnosis.

Collected from a follow-up study from August to September 2021 of 221 hospitalized survivors of COVID-19 18 months after discharge from Wuhan, including the demographic characteristics and clinical examination, the data in this study were randomly separated into a training (80%) data set and a validation (20%) data set. Six popular machine learning models were developed to predict the pulmonary diffusing capacity of patients infected with COVID-19 in the recovery stage. The performance indicators of the model included area under the curve (AUC), Accuracy, Recall, Precision, Positive Predictive Value(PPV), Negative Predictive Value (NPV) and F1. The model with the optimum performance was defined as the optimal model, which was further employed in the interpretability analysis. The MAHAKIL method was utilized to balance the data and optimize the balance of sample distribution, while the RFECV method for feature selection was utilized to select combined features more favorable to machine learning.

A total of 221 COVID-19 survivors were recruited in this study after discharge from hospitals in Wuhan. Of these participants, 117 (52.94%) were female, with a median age of 58.2 years (standard deviation (SD) = 12). After feature selection, 31 of the 37 clinical factors were finally selected for use in constructing the model. Among the six tested ML models, the best performance was accomplished in the XGBoost model, with an AUC of 0.755 and an accuracy of 78.01% after experimental verification. The SHAPELY Additive explanations (SHAP) summary analysis exhibited that hemoglobin (Hb), maximal voluntary ventilation (MVV), severity of illness, platelet (PLT), Uric Acid (UA) and blood urea nitrogen (BUN) were the top six most important factors affecting the XGBoost model decision-making.

The XGBoost model reported here showed a good prognostic prediction ability for PDCI of COVID-19 survivors during the recovery period. Among the interpretation methods based on the importance of SHAP values, Hb and MVV contributed the most to the prediction of PDCI outcomes of COVID-19 survivors in the recovery period.

Global surgery broadly refers to a rapidly expanding multidisciplinary field concerned with providing better and equitable surgical care across international health systems. Global surgery initiatives primarily focus on capacity building, advocacy, education, research, and policy development in low- and middle-income countries (LMICs). The inadequate surgical, anesthetic, and obstetric care currently contributes to 18 million preventable deaths each year. Hence, there is a growing interest in the rapid growth of artificial intelligence (AI) that provides a distinctive opportunity to enhance surgical services in LMICs. AI modalities have been used for personalizing surgical education, automating administrative tasks, and developing realistic and cost-effective simulation-training programs with provisions for people with special needs. Furthermore, AI may assist with providing insights for governance, infrastructure development, and monitoring/predicting stock take or logistics failure that can help in strengthening global surgery pillars. Numerous AI-assisted telemedicine-based platforms have allowed healthcare professionals to virtually assist in complex surgeries that may help to improve surgical accessibility across LMICs. Challenges in implementing AI technology include the misrepresentation of minority populations in the datasets leading to discriminatory bias. Human hesitancy, employment uncertainty, automation bias, and role of confounding factors need to be further studied for equitable utilization of AI. With a focused and evidence-based approach, AI could help several LMICs overcome bureaucratic inefficiency and develop more efficient surgical systems.

Machine learning is a promising and powerful technology with increasing use in orthopedics. Periprosthetic joint infection following total knee arthroplasty results in increased morbidity and mortality. This systematic review investigated the use of machine learning in preventing periprosthetic joint infection.

A systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. PubMed was searched in November 2022. All studies that investigated the clinical applications of machine learning in the prevention of periprosthetic joint infection following total knee arthroplasty were included. Non-English studies, studies with no full text available, studies focusing on non-clinical applications of machine learning, reviews and meta-analyses were excluded. For each included study, its characteristics, machine learning applications, algorithms, statistical performances, strengths and limitations were summarized. Limitations of the current machine learning applications and the studies, including their 'black box' nature, overfitting, the requirement of a large dataset, the lack of external validation, and their retrospective nature were identified.

Eleven studies were included in the final analysis. Machine learning applications in the prevention of periprosthetic joint infection were divided into four categories: prediction, diagnosis, antibiotic application and prognosis.

Machine learning may be a favorable alternative to manual methods in the prevention of periprosthetic joint infection following total knee arthroplasty. It aids in preoperative health optimization, preoperative surgical planning, the early diagnosis of infection, the early application of suitable antibiotics, and the prediction of clinical outcomes. Future research is warranted to resolve the current limitations and bring machine learning into clinical settings.

To develop and assess the performance of machine learning (ML) models based on magnetic resonance imaging (MRI) radiomics analysis for knee osteoarthritis (KOA) diagnosis.

This retrospective study analysed 148 consecutive patients (72 with KOA and 76 without) with available MRI image data, where radiomics features in cartilage portions were extracted and then filtered. Intraclass correlation coefficient (ICC) was calculated to quantify the reproducibility of features, and a threshold of 0.8 was set. The training and validation cohorts consisted of 117 and 31 cases, respectively. Least absolute shrinkage and selection operator (LASSO) regression method was employed for feature selection. The ML classifiers were logistic regression (LR), K-nearest neighbour (KNN) and support vector machine (SVM). In each algorithm, ten models derived from all available planes of three joint compartments and their various combinations were, respectively, constructed for comparative analysis. The performance of classifiers was mainly evaluated and compared by receiver operating characteristic (ROC) analysis.

All models achieved satisfying performances, especially the Final model, where accuracy and area under ROC curve (AUC) of LR classifier were 0.968, 0.983 (0.957-1.000, 95% CI) in the validation cohort, and 0.940, 0.984 (0.969-0.995, 95% CI) in the training cohort, respectively.

The MRI radiomics analysis represented promising performance in noninvasive and preoperative KOA diagnosis, especially when considering all available planes of all three compartments of knee joints.

Despite being in sustained and stable remission, patients with Crohn's disease (CD) stopping anti-tumour necrosis factor α (TNFα) show a high rate of relapse (~50% within 2 years). Characterising non-invasively the biological profiles of those patients is needed to better guide the decision of anti-TNFα withdrawal.

Ninety-two immune-related proteins were measured by proximity extension assay in serum of patients with CD (n=102) in sustained steroid-free remission and stopping anti-TNFα (infliximab). As previously shown, a stratification based on time to clinical relapse was used to characterise the distinct biological profiles of relapsers (short-term relapsers: <6 months vs mid/long-term relapsers: >6 months). Associations between protein levels and time to clinical relapse were determined by univariable Cox model.

The risk (HR) of mid/long-term clinical relapse was specifically associated with a high serum level of proteins mainly expressed in lymphocytes (LAG3, SH2B3, SIT1; HR: 2.2-4.5; p<0.05), a low serum level of anti-inflammatory effectors (IL-10, HSD11B1; HR: 0.2-0.3; p<0.05) and cellular junction proteins (CDSN, CNTNAP2, CXADR, ITGA11; HR: 0.4; p<0.05). The risk of short-term clinical relapse was specifically associated with a high serum level of pro-inflammatory effectors (IL-6, IL12RB1; HR: 3.5-3.6; p<0.05) and a low or high serum level of proteins mainly expressed in antigen presenting cells (CLEC4A, CLEC4C, CLEC7A, LAMP3; HR: 0.4-4.1; p<0.05).

We identified distinct blood protein profiles associated with the risk of short-term and mid/long-term clinical relapse in patients with CD stopping infliximab. These findings constitute an advance for the development of non-invasive biomarkers guiding the decision of anti-TNFα withdrawal.

Little is known regarding the prediction of the risks of asthma exacerbation after stopping asthma biologics.

To develop and validate a predictive model for the risk of asthma exacerbations after stopping asthma biologics using machine learning models.

We identified 3057 people with asthma who stopped asthma biologics in the OptumLabs Database Warehouse and considered a wide range of demographic and clinical risk factors to predict subsequent outcomes. The primary outcome used to assess success after stopping was having no exacerbations in the 6 months after stopping the biologic. Elastic-net logistic regression (GLMnet), random forest, and gradient boosting machine models were used with 10-fold cross-validation within a development (80%) cohort and validation cohort (20%).

The mean age of the total cohort was 47.1 (SD, 17.1) years, 1859 (60.8%) were women, 2261 (74.0%) were White, and 1475 (48.3%) were in the Southern region of the United States. The elastic-net logistic regression model yielded an area under the curve (AUC) of 0.75 (95% confidence interval [CI], 0.71-0.78) in the development and an AUC of 0.72 in the validation cohort. The random forest model yielded an AUC of 0.75 (95% CI, 0.68-0.79) in the development cohort and an AUC of 0.72 in the validation cohort. The gradient boosting machine model yielded an AUC of 0.76 (95% CI, 0.72-0.80) in the development cohort and an AUC of 0.74 in the validation cohort.

Outcomes after stopping asthma biologics can be predicted with moderate accuracy using machine learning methods.