ABSTRACT
BACKGROUND: The temporal relationships across cardiometabolic diseases (CMDs) were recently conceptualized as the cardiometabolic continuum (CMC), sequence of cardiovascular events that stem from gene-environmental interactions, unhealthy lifestyle influences, and metabolic diseases such as diabetes, and hypertension. While the physiological pathways linking metabolic and cardiovascular diseases have been investigated, the study of the sex and population differences in the CMC have still not been described. METHODS: We present a machine learning approach to model the CMC and investigate sex and population differences in two distinct cohorts: the UK Biobank (17,700 participants) and the Brazilian Longitudinal Study of Adult Health (ELSA-Brasil) (7162 participants). We consider the following CMDs: hypertension (Hyp), diabetes (DM), heart diseases (HD: angina, myocardial infarction, or heart failure), and stroke (STK). For the identification of the CMC patterns, individual trajectories with the time of disease occurrence were clustered using k-means. Based on clinical, sociodemographic, and lifestyle characteristics, we built multiclass random forest classifiers and used the SHAP methodology to evaluate feature importance. RESULTS: Five CMC patterns were identified across both sexes and cohorts: EarlyHyp, FirstDM, FirstHD, Healthy, and LateHyp, named according to prevalence and disease occurrence time that depicted around 95%, 78%, 75%, 88% and 99% of individuals, respectively. Within the UK Biobank, more women were classified in the Healthy cluster and more men in all others. In the EarlyHyp and LateHyp clusters, isolated hypertension occurred earlier among women. Smoking habits and education had high importance and clear directionality for both sexes. For ELSA-Brasil, more men were classified in the Healthy cluster and more women in the FirstDM. The diabetes occurrence time when followed by hypertension was lower among women. Education and ethnicity had high importance and clear directionality for women, while for men these features were smoking, alcohol, and coffee consumption. CONCLUSIONS: There are clear sex differences in the CMC that varied across the UK and Brazilian cohorts. In particular, disadvantages regarding incidence and the time to onset of diseases were more pronounced in Brazil, against woman. The results show the need to strengthen public health policies to prevent and control the time course of CMD, with an emphasis on women.
Subject(s)
Cardiovascular Diseases , Machine Learning , Adult , Aged , Female , Humans , Male , Middle Aged , Brazil/epidemiology , Cardiometabolic Risk Factors , Cardiovascular Diseases/epidemiology , Cohort Studies , Longitudinal Studies , Sex Factors , UK Biobank , United Kingdom/epidemiologyABSTRACT
Explainable artificial intelligence (XAI) elucidates the decision-making process of complex AI models and is important in building trust in model predictions. XAI explanations themselves require evaluation as to accuracy and reasonableness and in the context of use of the underlying AI model. This review details the evaluation of XAI in cardiac AI applications and has found that, of the studies examined, 37% evaluated XAI quality using literature results, 11% used clinicians as domain-experts, 11% used proxies or statistical analysis, with the remaining 43% not assessing the XAI used at all. We aim to inspire additional studies within healthcare, urging researchers not only to apply XAI methods but to systematically assess the resulting explanations, as a step towards developing trustworthy and safe models. Supplementary Information: The online version contains supplementary material available at 10.1007/s10462-024-10852-w.
ABSTRACT
The rapid evolution of highly adaptable and reusable artificial intelligence models facilitates the implementation of digital twinning and has the potential to redefine cardiovascular risk prevention. Digital twinning combines vast amounts of data from diverse sources to construct virtual models of an individual. Emerging artificial intelligence models, called generalist AI, enable the processing of different types of data, including data from electronic health records, laboratory results, medical texts, imaging, genomics, or graphs. Among their unprecedented capabilities are an easy adaptation of a model to previously unseen medical tasks and the ability to reason and explain output using precise medical language derived from scientific literature, medical guidelines, or knowledge graphs. The proposed combination of a digital twinning approach with generalist AI is a path to accelerate the implementation of precision medicine and enhance early recognition and prevention of cardiovascular disease. This proposed strategy may extend to other domains to advance predictive, preventive, and precision medicine and also boost health research discoveries.
ABSTRACT
Accurate segmentation of the fetal head and pubic symphysis in intrapartum ultrasound images and measurement of fetal angle of progression (AoP) are critical to both outcome prediction and complication prevention in delivery. However, due to poor quality of perinatal ultrasound imaging with blurred target boundaries and the relatively small target of the public symphysis, fully automated and accurate segmentation remains challenging. In this paper, we propse a dual-path boundary-guided residual network (DBRN), which is a novel approach to tackle these challenges. The model contains a multi-scale weighted module (MWM) to gather global context information, and enhance the feature response within the target region by weighting the feature map. The model also incorporates an enhanced boundary module (EBM) to obtain more precise boundary information. Furthermore, the model introduces a boundary-guided dual-attention residual module (BDRM) for residual learning. BDRM leverages boundary information as prior knowledge and employs spatial attention to simultaneously focus on background and foreground information, in order to capture concealed details and improve segmentation accuracy. Extensive comparative experiments have been conducted on three datasets. The proposed method achieves average Dice score of 0.908 ±0.05 and average Hausdorff distance of 3.396 ±0.66 mm. Compared with state-of-the-art competitors, the proposed DBRN achieves better results. In addition, the average difference between the automatic measurement of AoPs based on this model and the manual measurement results is 6.157 °, which has good consistency and has broad application prospects in clinical practice.
Subject(s)
Head , Pubic Symphysis , Ultrasonography, Prenatal , Humans , Pregnancy , Female , Ultrasonography, Prenatal/methods , Head/diagnostic imaging , Pubic Symphysis/diagnostic imaging , AlgorithmsABSTRACT
Artificial intelligence (AI) is revolutionizing the field of medical imaging, holding the potential to shift medicine from a reactive "sick-care" approach to a proactive focus on healthcare and prevention. The successful development of AI in this domain relies on access to large, comprehensive, and standardized real-world datasets that accurately represent diverse populations and diseases. However, images and data are sensitive, and as such, before using them in any way the data needs to be modified to protect the privacy of the patients. This paper explores the approaches in the domain of five EU projects working on the creation of ethically compliant and GDPR-regulated European medical imaging platforms, focused on cancer-related data. It presents the individual approaches to the de-identification of imaging data, and describes the problems and the solutions adopted in each case. Further, lessons learned are provided, enabling future projects to optimally handle the problem of data de-identification. CRITICAL RELEVANCE STATEMENT: This paper presents key approaches from five flagship EU projects for the de-identification of imaging and clinical data offering valuable insights and guidelines in the domain. KEY POINTS: ΑΙ models for health imaging require access to large amounts of data. Access to large imaging datasets requires an appropriate de-identification process. This paper provides de-identification guidelines from the AI for health imaging (AI4HI) projects.
ABSTRACT
Image quality assessment of magnetic resonance imaging (MRI) data is an important factor not only for conventional diagnosis and protocol optimization but also for fairness, trustworthiness, and robustness of artificial intelligence (AI) applications, especially on large heterogeneous datasets. Information on image quality in multi-centric studies is important to complement the contribution profile from each data node along with quantity information, especially when large variability is expected, and certain acceptance criteria apply. The main goal of this work was to present a tool enabling users to assess image quality based on both subjective criteria as well as objective image quality metrics used to support the decision on image quality based on evidence. The evaluation can be performed on both conventional and dynamic MRI acquisition protocols, while the latter is also checked longitudinally across dynamic series. The assessment provides an overall image quality score and information on the types of artifacts and degrading factors as well as a number of objective metrics for automated evaluation across series (BRISQUE score, Total Variation, PSNR, SSIM, FSIM, MS-SSIM). Moreover, the user can define specific regions of interest (ROIs) to calculate the regional signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR), thus individualizing the quality output to specific use cases, such as tissue-specific contrast or regional noise quantification.
ABSTRACT
Understanding the diversity in cancer research priorities and the correlations among different treatment modalities is essential to address the evolving landscape of oncology. This study, conducted in collaboration with the European Cancer Patient Coalition (ECPC) and Childhood Cancer International-Europe (CCI-E) as part of the "UNCAN.eu" initiative, analyzed data from a comprehensive survey to explore the complex interplay of demographics, time since cancer diagnosis, and types of treatments received. Demographic analysis revealed intriguing trends, highlighting the importance of tailoring cancer research efforts to specific age groups and genders. Individuals aged 45-69 exhibited highly aligned research priorities, emphasizing the need to address the unique concerns of middle-aged and older populations. In contrast, patients over 70 years demonstrated a divergence in research priorities, underscoring the importance of recognising the distinct needs of older individuals in cancer research. The analysis of correlations among different types of cancer treatments underscored the multidisciplinary approach to cancer care, with surgery, radiotherapy, chemotherapy, precision therapy, and biological therapies playing integral roles. These findings support the need for personalized and combined treatment strategies to achieve optimal outcomes. In conclusion, this study provides valuable insights into the complexity of cancer research priorities and treatment correlations in a European context. It emphasizes the importance of a multifaceted, patient-centred approach to cancer research and treatment, highlighting the need for ongoing support, adaptation, and collaboration to address the ever-changing landscape of oncology.
Subject(s)
Neoplasms , Humans , Neoplasms/therapy , Male , Aged , Middle Aged , Female , Biomedical Research , Adult , Demography , Research , EuropeABSTRACT
A significant level of stigma and inequality exists in mental healthcare, especially in under-served populations. Inequalities are reflected in the data collected for scientific purposes. When not properly accounted for, machine learning (ML) models learned from data can reinforce these structural inequalities or biases. Here, we present a systematic study of bias in ML models designed to predict depression in four different case studies covering different countries and populations. We find that standard ML approaches regularly present biased behaviors. We also show that mitigation techniques, both standard and our own post-hoc method, can be effective in reducing the level of unfair bias. There is no one best ML model for depression prediction that provides equality of outcomes. This emphasizes the importance of analyzing fairness during model selection and transparent reporting about the impact of debiasing interventions. Finally, we also identify positive habits and open challenges that practitioners could follow to enhance fairness in their models.
Subject(s)
Depression , Habits , Humans , Depression/diagnosis , Bias , Health Facilities , Machine LearningABSTRACT
BACKGROUND: Developing trustworthy artificial intelligence (AI) models for clinical applications requires access to clinical and imaging data cohorts. Reusing of publicly available datasets has the potential to fill this gap. Specifically in the domain of breast cancer, a large archive of publicly accessible medical images along with the corresponding clinical data is available at The Cancer Imaging Archive (TCIA). However, existing datasets cannot be directly used as they are heterogeneous and cannot be effectively filtered for selecting specific image types required to develop AI models. This work focuses on the development of a homogenized dataset in the domain of breast cancer including clinical and imaging data. METHODS: Five datasets were acquired from the TCIA and were harmonized. For the clinical data harmonization, a common data model was developed and a repeatable, documented "extract-transform-load" process was defined and executed for their homogenization. Further, Digital Imaging and COmmunications in Medicine (DICOM) information was extracted from magnetic resonance imaging (MRI) data and made accessible and searchable. RESULTS: The resulting harmonized dataset includes information about 2,035 subjects with breast cancer. Further, a platform named RV-Cherry-Picker enables search over both the clinical and diagnostic imaging datasets, providing unified access, facilitating the downloading of all study imaging that correspond to specific series' characteristics (e.g., dynamic contrast-enhanced series), and reducing the burden of acquiring the appropriate set of images for the respective AI model scenario. CONCLUSIONS: RV-Cherry-Picker provides access to the largest, publicly available, homogenized, imaging/clinical dataset for breast cancer to develop AI models on top. RELEVANCE STATEMENT: We present a solution for creating merged public datasets supporting AI model development, using as an example the breast cancer domain and magnetic resonance imaging images. KEY POINTS: ⢠The proposed platform allows unified access to the largest, homogenized public imaging dataset for breast cancer. ⢠A methodology for the semantically enriched homogenization of public clinical data is presented. ⢠The platform is able to make a detailed selection of breast MRI data for the development of AI models.
Subject(s)
Breast Neoplasms , Humans , Female , Breast Neoplasms/diagnostic imaging , Artificial Intelligence , BreastSubject(s)
Deep Learning , Humans , Heart Ventricles/diagnostic imaging , Radiography, Thoracic , CardiologyABSTRACT
Next-generation sequencing (NGS) and liquid biopsy (LB) showed positive results in the fight against different cancer types. This paper aims to assess the uptake of advanced molecular diagnostics/NGS for quick and efficient genetic profiles of tumour cells. For that purpose, the European Alliance for Personalised Medicine conducted a series of expert interviews to ascertain the current status across member states. One stakeholder meeting was additionally conducted to prioritize relevant factors by stakeholders. Seven common pillars were identified, and twenty-five measures were defined based on these pillars. Results showed that a multi-faceted approach is necessary for successful NGS implementation and that regional differences may be influenced by healthcare policies, resources, and infrastructure. It is important to consider different correlations when interpreting the results and to use them as a starting point for further discussion.
ABSTRACT
Improvements in cancer care require a new degree of collaboration beyond the purely medical sphere, extending deeply into the world of other stakeholders-preeminently patients but also the other stakeholders in the hardware and software of care. Cancer remains a global health challenge, necessitating collaborative efforts to understand, prevent, and treat this complex disease. To achieve this goal, a comprehensive analysis was conducted, aligning the prioritization of cancer research measures in 13 European countries with 13 key recommendations for conquering cancer in the region. The study utilized a survey involving both patients and citizens, alongside data from IQVIA, a global healthcare data provider, to assess the availability and access to single-biomarker tests in multiple European countries. The results revealed a focused approach toward understanding, preventing, and treating cancer, with each country emphasizing specific research measures tailored to its strengths and healthcare objectives. This analysis highlights the intricate relationship between research priorities, access to biomarker tests, and financial support. Timely access to tests and increased availability positively influence research areas such as cancer prevention, early detection, ageing, and data utilization. The alignment of these country-specific measures with 13 recommendations for conquering cancer in Europe underscores the importance of tailored strategies for understanding, preventing, and treating cancer.
ABSTRACT
AIMS: The regional and temporal differences in the associations between cardiovascular disease (CVD) and its classic risk factors are unknown. The current study examined these associations in different European regions over a 30-year period. METHODS AND RESULTS: The study sample comprised 553 818 individuals from 49 cohorts in 11 European countries (baseline: 1982-2012) who were followed up for a maximum of 10 years. Risk factors [sex, smoking, diabetes, non-HDL cholesterol, systolic blood pressure (BP), and body mass index (BMI)] and CVD events (coronary heart disease or stroke) were harmonized across cohorts. Risk factor-outcome associations were analysed using multivariable-adjusted Cox regression models, and differences in associations were assessed using meta-regression. The differences in the risk factor-CVD associations between central Europe, northern Europe, southern Europe, and the UK were generally small. Men had a slightly higher hazard ratio (HR) in southern Europe (P = 0.043 for overall difference), and those with diabetes had a slightly lower HR in central Europe (P = 0.022 for overall difference) compared with the other regions. Of the six CVD risk factors, minor HR decreases per decade were observed for non-HDL cholesterol [7% per mmol/L; 95% confidence interval (CI), 3-10%] and systolic BP (4% per 20â mmHg; 95% CI, 1-8%), while a minor HR increase per decade was observed for BMI (7% per 10â kg/m2; 95% CI, 1-13%). CONCLUSION: The results demonstrate that all classic CVD risk factors are still relevant in Europe, irrespective of regional area. Preventive strategies should focus on risk factors with the greatest population attributable risk.
All classic cardiovascular disease (CVD) risk factors are still relevant in Europe, irrespective of regional area. The differences in the associations of CVD risk factors with overt CVD between regions of Europe are generally small. Minor temporal hazard decreases were observed for non-HDL cholesterol and systolic blood pressure, while a minor hazard increase was observed for body mass index.
Subject(s)
Cardiovascular Diseases , Diabetes Mellitus , Male , Humans , Cardiovascular Diseases/diagnosis , Cardiovascular Diseases/epidemiology , Cardiovascular Diseases/etiology , Risk Factors , Cholesterol , Europe/epidemiology , Diabetes Mellitus/diagnosis , Diabetes Mellitus/epidemiologyABSTRACT
OBJECTIVE: Depression and obesity are 2 highly prevalent and often comorbid conditions. Exposure to early-life stress (ELS) has been associated with both depression and obesity in adulthood, as well as their preclinical manifestations during development. However, it remains unclear whether (1) associations differ depending on the timing of stress exposure (prenatal vs postnatal), and whether (2) ELS is a shared risk factor underlying the comorbidity between the 2 conditions. METHOD: Leveraging data from 2 large population-based birth cohorts (ALSPAC: n = 8,428 [52% male participants]; Generation R: n = 4,268 [48% male participants]), we constructed comprehensive cumulative measures of prenatal (in utero) and postnatal (from birth to 10 years) ELS. At age 13.5 years, we assessed the following: internalizing symptoms (using maternal reports); fat mass percentage (using dual-energy X-ray absorptiometry); and their comorbidity, defined as the co-occurrence of high internalizing and high adiposity. RESULTS: Both prenatal (total effect [95% CI] = 0.20 [0.16; 0.22]) and postnatal stress (ß [95%CI] = 0.22 [0.17; 0.25]) were associated with higher internalizing symptoms, with evidence of a more prominent role of postnatal stress. A weaker association (driven primarily by prenatal stress) was observed between stress and adiposity (prenatal: 0.07 [0.05; 0.09]; postnatal: 0.04 [0.01; 0.07]). Both prenatal (odds ratio [95%CI] = 1.70 [1.47; 1.97]) and postnatal (1.87 [1.61; 2.17]) stress were associated with an increased risk of developing comorbidity. CONCLUSION: We found evidence of timing and shared causal effects of ELS on psycho-cardiometabolic health in adolescence; however, future research is warranted to clarify how these associations may unfold over time. DIVERSITY & INCLUSION STATEMENT: We worked to ensure sex and gender balance in the recruitment of human participants. We worked to ensure race, ethnic, and/or other types of diversity in the recruitment of human participants. We worked to ensure that the study questionnaires were prepared in an inclusive way. One or more of the authors of this paper self-identifies as a member of one or more historically underrepresented racial and/or ethnic groups in science. One or more of the authors of this paper self-identifies as a member of one or more historically underrepresented sexual and/or gender groups in science. We actively worked to promote sex and gender balance in our author group.
Subject(s)
Adiposity , Adverse Childhood Experiences , Female , Pregnancy , Adolescent , Humans , Male , Obesity , Risk Factors , ComorbidityABSTRACT
ABSTRACT Artificial intelligence (AI) generative models driven by the integration of AI and natural language processing technologies, such as OpenAI's chatbot generative pre-trained transformer large language model (LLM), are receiving much public attention and have the potential to transform personalized medicine. Dialysis patients are highly dependent on technology and their treatment generates a challenging large volume of data that has to be analyzed for knowledge extraction. We argue that, by integrating the data acquired from hemodialysis treatments with the powerful conversational capabilities of LLMs, nephrologists could personalize treatments adapted to patients' lifestyles and preferences. We also argue that this new conversational AI integrated with a personalized patient-computer interface will enhance patients' engagement and self-care by providing them with a more personalized experience. However, generative AI models require continuous and accurate updates of data, and expert supervision and must address potential biases and limitations. Dialysis patients can also benefit from other new emerging technologies such as Digital Twins with which patients' care can also be addressed from a personalized medicine perspective. In this paper, we will revise LLMs potential strengths in terms of their contribution to personalized medicine, and, in particular, their potential impact, and limitations in nephrology. Nephrologists' collaboration with AI academia and companies, to develop algorithms and models that are more transparent, understandable, and trustworthy, will be crucial for the next generation of dialysis patients. The combination of technology, patient-specific data, and AI should contribute to create a more personalized and interactive dialysis process, improving patients' quality of life.
ABSTRACT
OBJECTIVES: To use pericardial adipose tissue (PAT) radiomics phenotyping to differentiate existing and predict future heart failure (HF) cases in the UK Biobank. METHODS: PAT segmentations were derived from cardiovascular magnetic resonance (CMR) studies using an automated quality-controlled model to define the region-of-interest for radiomics analysis. Prevalent (present at time of imaging) and incident (first occurrence after imaging) HF were ascertained using health record linkage. We created balanced cohorts of non-HF individuals for comparison. PyRadiomics was utilised to extract 104 radiomics features, of which 28 were chosen after excluding highly correlated ones (0.8). These features, plus sex and age, served as predictors in binary classification models trained separately to detect (1) prevalent and (2) incident HF. We tested seven modeling methods using tenfold nested cross-validation and examined feature importance with explainability methods. RESULTS: We studied 1204 participants in total, 297 participants with prevalent (60 ± 7 years, 21% female) and 305 with incident (61 ± 6 years, 32% female) HF, and an equal number of non-HF comparators. We achieved good discriminative performance for both prevalent (voting classifier; AUC: 0.76; F1 score: 0.70) and incident (light gradient boosting machine: AUC: 0.74; F1 score: 0.68) HF. Our radiomics models showed marginally better performance compared to PAT area alone. Increased PAT size (maximum 2D diameter in a given column or slice) and texture heterogeneity (sum entropy) were important features for prevalent and incident HF classification models. CONCLUSIONS: The amount and character of PAT discriminate individuals with prevalent HF and predict incidence of future HF. CLINICAL RELEVANCE STATEMENT: This study presents an innovative application of pericardial adipose tissue (PAT) radiomics phenotyping as a predictive tool for heart failure (HF), a major public health concern. By leveraging advanced machine learning methods, the research uncovers that the quantity and characteristics of PAT can be used to identify existing cases of HF and predict future occurrences. The enhanced performance of these radiomics models over PAT area alone supports the potential for better personalised care through earlier detection and prevention of HF. KEY POINTS: â¢PAT radiomics applied to CMR was used for the first time to derive binary machine learning classifiers to develop models for discrimination of prevalence and prediction of incident heart failure. â¢Models using PAT area provided acceptable discrimination between cases of prevalent or incident heart failure and comparator groups. â¢An increased PAT volume (increased diameter using shape features) and greater texture heterogeneity captured by radiomics texture features (increased sum entropy) can be used as an additional classifier marker for heart failure.
ABSTRACT
Objectives: To assess the feasibility of extracting radiomics signal intensity based features from the myocardium using cardiovascular magnetic resonance (CMR) imaging stress perfusion sequences. Furthermore, to compare the diagnostic performance of radiomics models against standard-of-care qualitative visual assessment of stress perfusion images, with the ground truth stenosis label being defined by invasive Fractional Flow Reserve (FFR) and quantitative coronary angiography. Methods: We used the Dan-NICAD 1 dataset, a multi-centre study with coronary computed tomography angiography, 1,5â T CMR stress perfusion, and invasive FFR available for a subset of 148 patients with suspected coronary artery disease. Image segmentation was performed by two independent readers. We used the Pyradiomics platform to extract radiomics first-order (n = 14) and texture (n = 75) features from the LV myocardium (basal, mid, apical) in rest and stress perfusion images. Results: Overall, 92 patients (mean age 62 years, 56 men) were included in the study, 39 with positive FFR. We double-cross validated the model and, in each inner fold, we trained and validated a per territory model. The conventional analysis results reported sensitivity of 41% and specificity of 84%. Our final radiomics model demonstrated an improvement on these results with an average sensitivity of 53% and specificity of 86%. Conclusion: In this proof-of-concept study from the Dan-NICAD dataset, we demonstrate the feasibility of radiomics analysis applied to CMR perfusion images with a suggestion of superior diagnostic performance of radiomics models over conventional visual analysis of perfusion images in picking up perfusion defects defined by invasive coronary angiography.
ABSTRACT
BACKGROUND: The human exposome encompasses all exposures that individuals encounter throughout their lifetime. It is now widely acknowledged that health outcomes are influenced not only by genetic factors but also by the interactions between these factors and various exposures. Consequently, the exposome has emerged as a significant contributor to the overall risk of developing major diseases, such as cardiovascular disease (CVD) and diabetes. Therefore, personalized early risk assessment based on exposome attributes might be a promising tool for identifying high-risk individuals and improving disease prevention. OBJECTIVE: Develop and evaluate a novel and fair machine learning (ML) model for CVD and type 2 diabetes (T2D) risk prediction based on a set of readily available exposome factors. We evaluated our model using internal and external validation groups from a multi-center cohort. To be considered fair, the model was required to demonstrate consistent performance across different sub-groups of the cohort. METHODS: From the UK Biobank, we identified 5,348 and 1,534 participants who within 13 years from the baseline visit were diagnosed with CVD and T2D, respectively. An equal number of participants who did not develop these pathologies were randomly selected as the control group. 109 readily available exposure variables from six different categories (physical measures, environmental, lifestyle, mental health events, sociodemographics, and early-life factors) from the participant's baseline visit were considered. We adopted the XGBoost ensemble model to predict individuals at risk of developing the diseases. The model's performance was compared to that of an integrative ML model which is based on a set of biological, clinical, physical, and sociodemographic variables, and, additionally for CVD, to the Framingham risk score. Moreover, we assessed the proposed model for potential bias related to sex, ethnicity, and age. Lastly, we interpreted the model's results using SHAP, a state-of-the-art explainability method. RESULTS: The proposed ML model presents a comparable performance to the integrative ML model despite using solely exposome information, achieving a ROC-AUC of 0.78±0.01 and 0.77±0.01 for CVD and T2D, respectively. Additionally, for CVD risk prediction, the exposome-based model presents an improved performance over the traditional Framingham risk score. No bias in terms of key sensitive variables was identified. CONCLUSIONS: We identified exposome factors that play an important role in identifying patients at risk of CVD and T2D, such as naps during the day, age completed full-time education, past tobacco smoking, frequency of tiredness/unenthusiasm, and current work status. Overall, this work demonstrates the potential of exposome-based machine learning as a fair CVD and T2D risk assessment tool.
Subject(s)
Cardiovascular Diseases , Diabetes Mellitus, Type 2 , Exposome , Humans , Diabetes Mellitus, Type 2/epidemiology , Risk Factors , Cardiovascular Diseases/epidemiology , Cardiovascular Diseases/etiology , Machine LearningABSTRACT
Artificial intelligence (AI) generative models driven by the integration of AI and natural language processing technologies, such as OpenAI's chatbot generative pre-trained transformer large language model (LLM), are receiving much public attention and have the potential to transform personalized medicine. Dialysis patients are highly dependent on technology and their treatment generates a challenging large volume of data that has to be analyzed for knowledge extraction. We argue that, by integrating the data acquired from hemodialysis treatments with the powerful conversational capabilities of LLMs, nephrologists could personalize treatments adapted to patients' lifestyles and preferences. We also argue that this new conversational AI integrated with a personalized patient-computer interface will enhance patients' engagement and self-care by providing them with a more personalized experience. However, generative AI models require continuous and accurate updates of data, and expert supervision and must address potential biases and limitations. Dialysis patients can also benefit from other new emerging technologies such as Digital Twins with which patients' care can also be addressed from a personalized medicine perspective. In this paper, we will revise LLMs potential strengths in terms of their contribution to personalized medicine, and, in particular, their potential impact, and limitations in nephrology. Nephrologists' collaboration with AI academia and companies, to develop algorithms and models that are more transparent, understandable, and trustworthy, will be crucial for the next generation of dialysis patients. The combination of technology, patient-specific data, and AI should contribute to create a more personalized and interactive dialysis process, improving patients' quality of life.