Synthetic data in machine learning for medicine and healthcare

Computational pathology: A survey review and the way forward

Computational Pathology (CPath) is an interdisciplinary science that augments developments of computational approaches to analyze and model medical histopathology images. The main objective for CPath is to develop infrastructure and workflows of digital diagnostics as an assistive CAD system for clinical pathology, facilitating transformational changes in the diagnosis and treatment of cancer that are mainly address by CPath tools. With evergrowing developments in deep learning and computer vision algorithms, and the ease of the data flow from digital pathology, currently CPath is witnessing a paradigm shift. Despite the sheer volume of engineering and scientific works being introduced for cancer image analysis, there is still a considerable gap of adopting and integrating these algorithms in clinical practice. This raises a significant question regarding the direction and trends that are undertaken in CPath. In this article we provide a comprehensive review of more than 800 papers to address the challenges faced in problem design all-the-way to the application and implementation viewpoints. We have catalogued each paper into a model-card by examining the key works and challenges faced to layout the current landscape in CPath. We hope this helps the community to locate relevant works and facilitate understanding of the field's future directions. In a nutshell, we oversee the CPath developments in cycle of stages which are required to be cohesively linked together to address the challenges associated with such multidisciplinary science. We overview this cycle from different perspectives of data-centric, model-centric, and application-centric problems. We finally sketch remaining challenges and provide directions for future technical developments and clinical integration of CPath. For updated information on this survey review paper and accessing to the original model cards repository, please refer to GitHub. Updated version of this draft can also be found from arXiv.

Integrating machine learning and artificial intelligence in life-course epidemiology: pathways to innovative public health solutions

The integration of machine learning (ML) and artificial intelligence (AI) techniques in life-course epidemiology offers remarkable opportunities to advance our understanding of the complex interplay between biological, social, and environmental factors that shape health trajectories across the lifespan. This perspective summarizes the current applications, discusses future potential and challenges, and provides recommendations for harnessing ML and AI technologies to develop innovative public health solutions. ML and AI have been increasingly applied in epidemiological studies, demonstrating their ability to handle large, complex datasets, identify intricate patterns and associations, integrate multiple and multimodal data types, improve predictive accuracy, and enhance causal inference methods. In life-course epidemiology, these techniques can help identify sensitive periods and critical windows for intervention, model complex interactions between risk factors, predict individual and population-level disease risk trajectories, and strengthen causal inference in observational studies. By leveraging the five principles of life-course research proposed by Elder and Shanahan-lifespan development, agency, time and place, timing, and linked lives-we discuss a framework for applying ML and AI to uncover novel insights and inform targeted interventions. However, the successful integration of these technologies faces challenges related to data quality, model interpretability, bias, privacy, and equity. To fully realize the potential of ML and AI in life-course epidemiology, fostering interdisciplinary collaborations, developing standardized guidelines, advocating for their integration in public health decision-making, prioritizing fairness, and investing in training and capacity building are essential. By responsibly harnessing the power of ML and AI, we can take significant steps towards creating healthier and more equitable futures across the life course.

Data augmentation for generating synthetic electrogastrogram time series

To address an emerging need for large number of diverse datasets for rigor evaluation of signal processing techniques, we developed and evaluated a new method for generating synthetic electrogastrogram time series. We used electrogastrography (EGG) data from an open database to set model parameters and statistical tests to evaluate synthesized data. Additionally, we illustrated method customization for generating artificial EGG time series alterations caused by the simulator sickness. Proposed data augmentation method generates synthetic EGG data with specified duration, sampling frequency, recording state (postprandial or fasting state), overall noise and breathing artifact injection, and pauses in the gastric rhythm (arrhythmia occurrence) with statistically significant difference between postprandial and fasting states in > 70% cases while not accounting for individual differences. Features obtained from the synthetic EGG signal resembling simulator sickness occurrence displayed expected trends. The code for generation of synthetic EGG time series is not only freely available and can be further customized to assess signal processing algorithms but also may be used to increase data diversity for training artificial intelligence (AI) algorithms. The proposed approach is customized for EGG data synthesis but can be easily utilized for other biosignals with similar nature such as electroencephalogram.

A vision-language foundation model for the generation of realistic chest X-ray images

The paucity of high-quality medical imaging datasets could be mitigated by machine learning models that generate compositionally diverse images that faithfully represent medical concepts and pathologies. However, large vision-language models are trained on natural images, and the diversity distribution of the generated images substantially differs from that of medical images. Moreover, medical language involves specific and semantically rich vocabulary. Here we describe a domain-adaptation strategy for large vision-language models that overcomes distributional shifts. Specifically, by leveraging publicly available datasets of chest X-ray images and the corresponding radiology reports, we adapted a latent diffusion model pre-trained on pairs of natural images and text descriptors to generate diverse and visually plausible synthetic chest X-ray images (as confirmed by board-certified radiologists) whose appearance can be controlled with free-form medical text prompts. The domain-adaptation strategy for the text-conditioned synthesis of medical images can be used to augment training datasets and is a viable alternative to the sharing of real medical images for model training and fine-tuning.

Metadata-conditioned generative models to synthesize anatomically-plausible 3D brain MRIs

Recent advances in generative models have paved the way for enhanced generation of natural and medical images, including synthetic brain MRIs. However, the mainstay of current AI research focuses on optimizing synthetic MRIs with respect to visual quality (such as signal-to-noise ratio) while lacking insights into their relevance to neuroscience. To generate high-quality T1-weighted MRIs relevant for neuroscience discovery, we present a two-stage Diffusion Probabilistic Model (called BrainSynth) to synthesize high-resolution MRIs conditionally-dependent on metadata (such as age and sex). We then propose a novel procedure to assess the quality of BrainSynth according to how well its synthetic MRIs capture macrostructural properties of brain regions and how accurately they encode the effects of age and sex. Results indicate that more than half of the brain regions in our synthetic MRIs are anatomically plausible, i.e., the effect size between real and synthetic MRIs is small relative to biological factors such as age and sex. Moreover, the anatomical plausibility varies across cortical regions according to their geometric complexity. As is, the MRIs generated by BrainSynth significantly improve the training of a predictive model to identify accelerated aging effects in an independent study. These results indicate that our model accurately capture the brain's anatomical information and thus could enrich the data of underrepresented samples in a study. The code of BrainSynth will be released as part of the MONAI project at https://github.com/Project-MONAI/GenerativeModels.

On the evaluation of synthetic longitudinal electronic health records

BACKGROUND

Synthetic Electronic Health Records (EHRs) are becoming increasingly popular as a privacy enhancing technology. However, for longitudinal EHRs specifically, little research has been done into how to properly evaluate synthetically generated samples. In this article, we provide a discussion on existing methods and recommendations when evaluating the quality of synthetic longitudinal EHRs.

METHODS

We recommend to assess synthetic EHR quality through similarity to real EHRs in low-dimensional projections, accuracy of a classifier discriminating synthetic from real samples, performance of synthetic versus real trained algorithms in clinical tasks, and privacy risk through risk of attribute inference. For each metric we discuss strengths and weaknesses, next to showing how it can be applied on a longitudinal dataset.

RESULTS

To support the discussion on evaluation metrics, we apply discussed metrics on a dataset of synthetic EHRs generated from the Medical Information Mart for Intensive Care-IV (MIMIC-IV) repository.

CONCLUSIONS

The discussion on evaluation metrics provide guidance for researchers on how to use and interpret different metrics when evaluating the quality of synthetic longitudinal EHRs.

Towards equitable AI in oncology

Artificial intelligence (AI) stands at the threshold of revolutionizing clinical oncology, with considerable potential to improve early cancer detection and risk assessment, and to enable more accurate personalized treatment recommendations. However, a notable imbalance exists in the distribution of the benefits of AI, which disproportionately favour those living in specific geographical locations and in specific populations. In this Perspective, we discuss the need to foster the development of equitable AI tools that are both accurate in and accessible to a diverse range of patient populations, including those in low-income to middle-income countries. We also discuss some of the challenges and potential solutions in attaining equitable AI, including addressing the historically limited representation of diverse populations in existing clinical datasets and the use of inadequate clinical validation methods. Additionally, we focus on extant sources of inequity including the type of model approach (such as deep learning, and feature engineering-based methods), the implications of dataset curation strategies, the need for rigorous validation across a variety of populations and settings, and the risk of introducing contextual bias that comes with developing tools predominantly in high-income countries.

Anti-HER2 therapy response assessment for guiding treatment (de-)escalation in early HER2-positive breast cancer using a novel deep learning radiomics model

OBJECTIVES

Anti-HER2 targeted therapy significantly reduces risk of relapse in HER2 + breast cancer. New measures are needed for a precise risk stratification to guide (de-)escalation of anti-HER2 strategy.

METHODS

A total of 726 HER2 + cases who received no/single/dual anti-HER2 targeted therapies were split into three respective cohorts. A deep learning model (DeepTEPP) based on preoperative breast magnetic resonance (MR) was developed. Patients were scored and categorized into low-, moderate-, and high-risk groups. Recurrence-free survival (RFS) was compared in patients with different risk groups according to the anti-HER2 treatment they received, to validate the value of DeepTEPP in predicting treatment efficacy and guiding anti-HER2 strategy.

RESULTS

DeepTEPP was capable of risk stratification and guiding anti-HER2 treatment strategy: DeepTEPP-Low patients (60.5%) did not derive significant RFS benefit from trastuzumab (p = 0.144), proposing an anti-HER2 de-escalation. DeepTEPP-Moderate patients (19.8%) significantly benefited from trastuzumab (p = 0.048), but did not obtain additional improvements from pertuzumab (p = 0.125). DeepTEPP-High patients (19.7%) significantly benefited from dual HER2 blockade (p = 0.045), suggesting an anti-HER2 escalation.

CONCLUSIONS

DeepTEPP represents a pioneering MR-based deep learning model that enables the non-invasive prediction of adjuvant anti-HER2 effectiveness, thereby providing valuable guidance for anti-HER2 (de-)escalation strategies. DeepTEPP provides an important reference for choosing the appropriate individualized treatment in HER2 + breast cancer patients, warranting prospective validation.

CLINICAL RELEVANCE STATEMENT

We built an MR-based deep learning model DeepTEPP, which enables the non-invasive prediction of adjuvant anti-HER2 effectiveness, thus guiding anti-HER2 (de-)escalation strategies in early HER2-positive breast cancer patients.

KEY POINTS

• DeepTEPP is able to predict anti-HER2 effectiveness and to guide treatment (de-)escalation. • DeepTEPP demonstrated an impressive prognostic efficacy for recurrence-free survival and overall survival. • To our knowledge, this is one of the very few, also the largest study to test the efficacy of a deep learning model extracted from breast MR images on HER2-positive breast cancer survival and anti-HER2 therapy effectiveness prediction.

Novel Domain Knowledge-Encoding Algorithm Enables Label-Efficient Deep Learning for Cardiac CT Segmentation to Guide Atrial Fibrillation Treatment in a Pilot Dataset

Background: Segmenting computed tomography (CT) is crucial in various clinical applications, such as tailoring personalized cardiac ablation for managing cardiac arrhythmias. Automating segmentation through machine learning (ML) is hindered by the necessity for large, labeled training data, which can be challenging to obtain. This article proposes a novel approach for automated, robust labeling using domain knowledge to achieve high-performance segmentation by ML from a small training set. The approach, the domain knowledge-encoding (DOKEN) algorithm, reduces the reliance on large training datasets by encoding cardiac geometry while automatically labeling the training set. The method was validated in a hold-out dataset of CT results from an atrial fibrillation (AF) ablation study. Methods: The DOKEN algorithm parses left atrial (LA) structures, extracts "anatomical knowledge" by leveraging digital LA models (available publicly), and then applies this knowledge to achieve high ML segmentation performance with a small number of training samples. The DOKEN-labeled training set was used to train a nnU-Net deep neural network (DNN) model for segmenting cardiac CT in N = 20 patients. Subsequently, the method was tested in a hold-out set with N = 100 patients (five times larger than training set) who underwent AF ablation. Results: The DOKEN algorithm integrated with the nn-Unet model achieved high segmentation performance with few training samples, with a training to test ratio of 1:5. The Dice score of the DOKEN-enhanced model was 96.7% (IQR: 95.3% to 97.7%), with a median error in surface distance of boundaries of 1.51 mm (IQR: 0.72 to 3.12) and a mean centroid-boundary distance of 1.16 mm (95% CI: -4.57 to 6.89), similar to expert results (r = 0.99; p < 0.001). In digital hearts, the novel DOKEN approach segmented the LA structures with a mean difference for the centroid-boundary distances of -0.27 mm (95% CI: -3.87 to 3.33; r = 0.99; p < 0.0001). Conclusions: The proposed novel domain knowledge-encoding algorithm was able to perform the segmentation of six substructures of the LA, reducing the need for large training data sets. The combination of domain knowledge encoding and a machine learning approach could reduce the dependence of ML on large training datasets and could potentially be applied to AF ablation procedures and extended in the future to other imaging, 3D printing, and data science applications.

NeuroSynth: MRI-Derived Neuroanatomical Generative Models and Associated Dataset of 18,000 Samples

Availability of large and diverse medical datasets is often challenged by privacy and data sharing restrictions. For successful application of machine learning techniques for disease diagnosis, prognosis, and precision medicine, large amounts of data are necessary for model building and optimization. To help overcome such limitations in the context of brain MRI, we present NeuroSynth: a collection of generative models of normative regional volumetric features derived from structural brain imaging. NeuroSynth models are trained on real brain imaging regional volumetric measures from the iSTAGING consortium, which encompasses over 40,000 MRI scans across 13 studies, incorporating covariates such as age, sex, and race. Leveraging NeuroSynth, we produce and offer 18,000 synthetic samples spanning the adult lifespan (ages 22-90 years), alongside the model's capability to generate unlimited data. Experimental results indicate that samples generated from NeuroSynth agree with the distributions obtained from real data. Most importantly, the generated normative data significantly enhance the accuracy of downstream machine learning models on tasks such as disease classification. Data and models are available at: https://huggingface.co/spaces/rongguangw/neuro-synth.

Introducing SPINE: A Holistic Approach to Synthetic Pulmonary Imaging Evaluation Through End-to-End Data and Model Management

In the evolving field of medical imaging and machine learning (ML), this paper introduces a novel framework for evaluating synthetic pulmonary imaging aiming to assess synthetic data quality and applicability. Our study concentrates on synthetic X-ray chest images, crucial for diagnosing respiratory diseases. We employ SPINE (Synthetic Pulmonary Imaging Evaluation) framework, a threefold synthetic images evaluation method including expert domain assessment, statistical data analysis and adversarial evaluation. In order to replicate and validate our methodology, we followed an End-to-End data and model management process which begins with a dataset of Normal and Pneumonia chest X-rays, generating synthetic images using Generative Adversarial Networks (GANs) and training a baseline classifier, essential in the adversarial evaluation axis, testing synthetic images against real data assessing their predictive value. The critical outcome of our approach is the post-market analysis of synthetic images. This innovative method evaluates synthetic images using clinical, statistical, and scientific criteria independently from traditional generation performance metrics. This independent evaluation provides deep insights into the clinical and research effectiveness of the synthetic data. By ensuring these images mirror real data's statistical properties and maintain clinical accuracy, our framework establishes a new standard for the ethical and reliable use of synthetic data in medical imaging and research.

Artificial intelligence-enhanced patient evaluation: bridging art and science

The advent of digital health and artificial intelligence (AI) has promised to revolutionize clinical care, but real-world patient evaluation has yet to witness transformative changes. As history taking and physical examination continue to rely on long-established practices, a growing pipeline of AI-enhanced digital tools may soon augment the traditional clinical encounter into a data-driven process. This article presents an evidence-backed vision of how promising AI applications may enhance traditional practices, streamlining tedious tasks while elevating diverse data sources, including AI-enabled stethoscopes, cameras, and wearable sensors, to platforms for personalized medicine and efficient care delivery. Through the lens of traditional patient evaluation, we illustrate how digital technologies may soon be interwoven into routine clinical workflows, introducing a novel paradigm of longitudinal monitoring. Finally, we provide a skeptic's view on the practical, ethical, and regulatory challenges that limit the uptake of such technologies.

Latent Diffusion Models with Image-Derived Annotations for Enhanced AI-Assisted Cancer Diagnosis in Histopathology

Artificial Intelligence (AI)-based image analysis has immense potential to support diagnostic histopathology, including cancer diagnostics. However, developing supervised AI methods requires large-scale annotated datasets. A potentially powerful solution is to augment training data with synthetic data. Latent diffusion models, which can generate high-quality, diverse synthetic images, are promising. However, the most common implementations rely on detailed textual descriptions, which are not generally available in this domain. This work proposes a method that constructs structured textual prompts from automatically extracted image features. We experiment with the PCam dataset, composed of tissue patches only loosely annotated as healthy or cancerous. We show that including image-derived features in the prompt, as opposed to only healthy and cancerous labels, improves the Fréchet Inception Distance (FID) by 88.6. We also show that pathologists find it challenging to detect synthetic images, with a median sensitivity/specificity of 0.55/0.55. Finally, we show that synthetic data effectively train AI models.

Deep generative model of the distal tibial classic metaphyseal lesion in infants: assessment of synthetic images

Background

The classic metaphyseal lesion (CML) is a distinctive fracture highly specific to infant abuse. To increase the size and diversity of the training CML database for automated deep-learning detection of this fracture, we developed a mask conditional diffusion model (MaC-DM) to generate synthetic images with and without CMLs.

Purpose

To objectively and subjectively assess the synthetic radiographic images with and without CMLs generated by MaC-DM.

Materials and Methods

For retrospective testing, we randomly chose 100 real images (50 normals and 50 with CMLs; 39 infants, male = 22, female = 17; mean age = 4.1 months; SD = 3.1 months) from an existing distal tibia dataset (177 normal, 73 with CMLs), and generated 100 synthetic distal tibia images via MaC-DM (50 normals and 50 with CMLs). These test images were shown to 3 blinded radiologists. In the first session, radiologists determined if the images were normal or had CMLs. In the second session, they determined if the images were real or synthetic. We analyzed the radiologists' interpretations and employed t-distributed stochastic neighbor embedding technique to analyze the data distribution of the test images.

Results

When presented with the 200 images (100 synthetic, 100 with CMLs), radiologists reliably and accurately diagnosed CMLs (kappa = 0.90, 95% CI = [0.88-0.92]; accuracy = 92%, 95% CI = [89-97]). However, they were inaccurate in differentiating between real and synthetic images (kappa = 0.05, 95% CI = [0.03-0.07]; accuracy = 53%, 95% CI = [49-59]). The t-distributed stochastic neighbor embedding analysis showed substantial differences in the data distribution between normal images and those with CMLs (area under the curve = 0.996, 95% CI = [0.992-1.000], P < .01), but minor differences between real and synthetic images (area under the curve = 0.566, 95% CI = [0.486-0.647], P = .11).

Conclusion

Radiologists accurately diagnosed images with distal tibial CMLs but were unable to distinguish real from synthetically generated ones, indicating that our generative model could synthesize realistic images. Thus, MaC-DM holds promise as an effective strategy for data augmentation in training machine-learning models for diagnosis of distal tibial CMLs.

Improving the performance of machine learning penicillin adverse drug reaction classification with synthetic data and transfer learning

BACKGROUND

Machine learning may assist with the identification of potentially inappropriate penicillin allergy labels. Strategies to improve the performance of existing models for this task include the use of additional training data, synthetic data and transfer learning.

AIMS

The aims of this study were to investigate the use of additional training data and novel machine learning strategies, namely synthetic data and transfer learning, to improve the performance of penicillin adverse drug reaction (ADR) machine learning classification.

METHODS

Machine learning natural language processing was applied to free-text penicillin ADR data extracted from a public health system electronic health record (EHR). The models were developed by training on various labelled data sets. ADR entries were split into training and testing data sets and used to develop and test a variety of machine learning models. The effect of training on additional data and synthetic data versus the use of transfer learning was analysed.

RESULTS

Following the application of these techniques, the area under the receiver operator curve of best-performing models for the classification of penicillin allergy (vs intolerance) and high-risk allergy (vs low-risk allergy) improved to 0.984 (using the artificial neural network model) and 0.995 (with the transfer learning approach) respectively.

CONCLUSIONS

Machine learning models demonstrate high levels of accuracy in the classification and risk stratification of penicillin ADR labels using the reaction documented in the EHR. The model can be further optimised by incorporating additional training data and using transfer learning. Practical applications include automating case detection for penicillin allergy delabelling programmes.

Generative artificial intelligence in ophthalmology: current innovations, future applications and challenges

The rapid advancements in generative artificial intelligence are set to significantly influence the medical sector, particularly ophthalmology. Generative adversarial networks and diffusion models enable the creation of synthetic images, aiding the development of deep learning models tailored for specific imaging tasks. Additionally, the advent of multimodal foundational models, capable of generating images, text and videos, presents a broad spectrum of applications within ophthalmology. These range from enhancing diagnostic accuracy to improving patient education and training healthcare professionals. Despite the promising potential, this area of technology is still in its infancy, and there are several challenges to be addressed, including data bias, safety concerns and the practical implementation of these technologies in clinical settings.

From data to diagnosis: skin cancer image datasets for artificial intelligence

Artificial intelligence (AI) solutions for skin cancer diagnosis continue to gain momentum, edging closer towards broad clinical use. These AI models, particularly deep-learning architectures, require large digital image datasets for development. This review provides an overview of the datasets used to develop AI algorithms and highlights the importance of dataset transparency for the evaluation of algorithm generalizability across varying populations and settings. Current challenges for curation of clinically valuable datasets are detailed, which include dataset shifts arising from demographic variations and differences in data collection methodologies, along with inconsistencies in labelling. These shifts can lead to differential algorithm performance, compromise of clinical utility, and the propagation of discriminatory biases when developed algorithms are implemented in mismatched populations. Limited representation of rare skin cancers and minoritized groups in existing datasets are highlighted, which can further skew algorithm performance. Strategies to address these challenges are presented, which include improving transparency, representation and interoperability. Federated learning and generative methods, which may improve dataset size and diversity without compromising privacy, are also examined. Lastly, we discuss model-level techniques that may address biases entrained through the use of datasets derived from routine clinical care. As the role of AI in skin cancer diagnosis becomes more prominent, ensuring the robustness of underlying datasets is increasingly important.

A Clinician's Guide to Sharing Data for AI in Ophthalmology

Data is the cornerstone of using AI models, because their performance directly depends on the diversity, quantity, and quality of the data used for training. Using AI presents unique potential, particularly in medical applications that involve rich data such as ophthalmology, encompassing a variety of imaging methods, medical records, and eye-tracking data. However, sharing medical data comes with challenges because of regulatory issues and privacy concerns. This review explores traditional and nontraditional data sharing methods in medicine, focusing on previous works in ophthalmology. Traditional methods involve direct data transfer, whereas newer approaches prioritize security and privacy by sharing derived datasets, creating secure research environments, or using model-to-data strategies. We examine each method's mechanisms, variations, recent applications in ophthalmology, and their respective advantages and disadvantages. By empowering medical researchers with insights into data sharing methods and considerations, this review aims to assist informed decision-making while upholding ethical standards and patient privacy in medical AI development.

Synthesis of Hybrid Data Consisting of Chest Radiographs and Tabular Clinical Records Using Dual Generative Models for COVID-19 Positive Cases

To generate synthetic medical data incorporating image-tabular hybrid data by merging an image encoding/decoding model with a table-compatible generative model and assess their utility. We used 1342 cases from the Stony Brook University Covid-19-positive cases, comprising chest X-ray radiographs (CXRs) and tabular clinical data as a private dataset (pDS). We generated a synthetic dataset (sDS) through the following steps: (I) dimensionally reducing CXRs in the pDS using a pretrained encoder of the auto-encoding generative adversarial networks (αGAN) and integrating them with the correspondent tabular clinical data; (II) training the conditional tabular GAN (CTGAN) on this combined data to generate synthetic records, encompassing encoded image features and clinical data; and (III) reconstructing synthetic images from these encoded image features in the sDS using a pretrained decoder of the αGAN. The utility of sDS was assessed by the performance of the prediction models for patient outcomes (deceased or discharged). For the pDS test set, the area under the receiver operating characteristic (AUC) curve was calculated to compare the performance of prediction models trained separately with pDS, sDS, or a combination of both. We created an sDS comprising CXRs with a resolution of 256 × 256 pixels and tabular data containing 13 variables. The AUC for the outcome was 0.83 when the model was trained with the pDS, 0.74 with the sDS, and 0.87 when combining pDS and sDS for training. Our method is effective for generating synthetic records consisting of both images and tabular clinical data.

Using histopathology latent diffusion models as privacy-preserving dataset augmenters improves downstream classification performance

Latent diffusion models (LDMs) have emerged as a state-of-the-art image generation method, outperforming previous Generative Adversarial Networks (GANs) in terms of training stability and image quality. In computational pathology, generative models are valuable for data sharing and data augmentation. However, the impact of LDM-generated images on histopathology tasks compared to traditional GANs has not been systematically studied. We trained three LDMs and a styleGAN2 model on histology tiles from nine colorectal cancer (CRC) tissue classes. The LDMs include 1) a fine-tuned version of stable diffusion v1.4, 2) a Kullback-Leibler (KL)-autoencoder (KLF8-DM), and 3) a vector quantized (VQ)-autoencoder deploying LDM (VQF8-DM). We assessed image quality through expert ratings, dimensional reduction methods, distribution similarity measures, and their impact on training a multiclass tissue classifier. Additionally, we investigated image memorization in the KLF8-DM and styleGAN2 models. All models provided a high image quality, with the KLF8-DM achieving the best Frechet Inception Distance (FID) and expert rating scores for complex tissue classes. For simpler classes, the VQF8-DM and styleGAN2 models performed better. Image memorization was negligible for both styleGAN2 and KLF8-DM models. Classifiers trained on a mix of KLF8-DM generated and real images achieved a 4% improvement in overall classification accuracy, highlighting the usefulness of these images for dataset augmentation. Our systematic study of generative methods showed that KLF8-DM produces the highest quality images with negligible image memorization. The higher classifier performance in the generatively augmented dataset suggests that this augmentation technique can be employed to enhance histopathology classifiers for various tasks.

Evaluating synthetic neuroimaging data augmentation for automatic brain tumour segmentation with a deep fully-convolutional network

Gliomas observed in medical images require expert neuro-radiologist evaluation for treatment planning and monitoring, motivating development of intelligent systems capable of automating aspects of tumour evaluation. Deep learning models for automatic image segmentation rely on the amount and quality of training data. In this study we developed a neuroimaging synthesis technique to augment data for training fully-convolutional networks (U-nets) to perform automatic glioma segmentation. We used StyleGAN2-ada to simultaneously generate fluid-attenuated inversion recovery (FLAIR) magnetic resonance images and corresponding glioma segmentation masks. Synthetic data were successively added to real training data (n = 2751) in fourteen rounds of 1000 and used to train U-nets that were evaluated on held-out validation (n = 590) and test sets (n = 588). U-nets were trained with and without geometric augmentation (translation, zoom and shear), and Dice coefficients were computed to evaluate segmentation performance. We also monitored the number of training iterations before stopping, total training time, and time per iteration to evaluate computational costs associated with training each U-net. Synthetic data augmentation yielded marginal improvements in Dice coefficients (validation set +0.0409, test set +0.0355), whereas geometric augmentation improved generalization (standard deviation between training, validation and test set performances of 0.01 with, and 0.04 without geometric augmentation). Based on the modest performance gains for automatic glioma segmentation we find it hard to justify the computational expense of developing a synthetic image generation pipeline. Future work may seek to optimize the efficiency of synthetic data generation for augmentation of neuroimaging data.

Machine Learning-Based Predictive Models for Patients with Venous Thromboembolism: A Systematic Review

BACKGROUND

 Venous thromboembolism (VTE) is a chronic disorder with a significant health and economic burden. Several VTE-specific clinical prediction models (CPMs) have been used to assist physicians in decision-making but have several limitations. This systematic review explores if machine learning (ML) can enhance CPMs by analyzing extensive patient data derived from electronic health records. We aimed to explore ML-CPMs' applications in VTE for risk stratification, outcome prediction, diagnosis, and treatment.

METHODS

 Three databases were searched: PubMed, Google Scholar, and IEEE electronic library. Inclusion criteria focused on studies using structured data, excluding non-English publications, studies on non-humans, and certain data types such as natural language processing and image processing. Studies involving pregnant women, cancer patients, and children were also excluded. After excluding irrelevant studies, a total of 77 studies were included.

RESULTS

 Most studies report that ML-CPMs outperformed traditional CPMs in terms of receiver operating area under the curve in the four clinical domains that were explored. However, the majority of the studies were retrospective, monocentric, and lacked detailed model architecture description and external validation, which are essential for quality audit. This review identified research gaps and highlighted challenges related to standardized reporting, reproducibility, and model comparison.

CONCLUSION

 ML-CPMs show promise in improving risk assessment and individualized treatment recommendations in VTE. Apparently, there is an urgent need for standardized reporting and methodology for ML models, external validation, prospective and real-world data studies, as well as interventional studies to evaluate the impact of artificial intelligence in VTE.

An Update on the Use of Artificial Intelligence in Digital Pathology for Oral Epithelial Dysplasia Research

INTRODUCTION

Oral epithelial dysplasia (OED) is a precancerous histopathological finding which is considered the most important prognostic indicator for determining the risk of malignant transformation into oral squamous cell carcinoma (OSCC). The gold standard for diagnosis and grading of OED is through histopathological examination, which is subject to inter- and intra-observer variability, impacting accurate diagnosis and prognosis. The aim of this review article is to examine the current advances in digital pathology for artificial intelligence (AI) applications used for OED diagnosis.

MATERIALS AND METHODS

We included studies that used AI for diagnosis, grading, or prognosis of OED on histopathology images or intraoral clinical images. Studies utilizing imaging modalities other than routine light microscopy (e.g., scanning electron microscopy), or immunohistochemistry-stained histology slides, or immunofluorescence were excluded from the study. Studies not focusing on oral dysplasia grading and diagnosis, e.g., to discriminate OSCC from normal epithelial tissue were also excluded.

RESULTS

A total of 24 studies were included in this review. Nineteen studies utilized deep learning (DL) convolutional neural networks for histopathological OED analysis, and 4 used machine learning (ML) models. Studies were summarized by AI method, main study outcomes, predictive value for malignant transformation, strengths, and limitations.

CONCLUSION

ML/DL studies for OED grading and prediction of malignant transformation are emerging as promising adjunctive tools in the field of digital pathology. These adjunctive objective tools can ultimately aid the pathologist in more accurate diagnosis and prognosis prediction. However, further supportive studies that focus on generalization, explainable decisions, and prognosis prediction are needed.

Machine learning-based amide proton transfer imaging using partially synthetic training data

PURPOSE

Machine learning (ML) has been increasingly used to quantify CEST effect. ML models are typically trained using either measured data or fully simulated data. However, training with measured data often lacks sufficient training data, whereas training with fully simulated data may introduce bias because of limited simulations pools. This study introduces a new platform that combines simulated and measured components to generate partially synthetic CEST data, and to evaluate its feasibility for training ML models to predict amide proton transfer (APT) effect.

METHODS

Partially synthetic CEST signals were created using an inverse summation of APT effects from simulations and the other components from measurements. Training data were generated by varying APT simulation parameters and applying scaling factors to adjust the measured components, achieving a balance between simulation flexibility and fidelity. First, tissue-mimicking CEST signals along with ground truth information were created using multiple-pool model simulations to validate this method. Second, an ML model was trained individually on partially synthetic data, in vivo data, and fully simulated data, to predict APT effect in rat brains bearing 9 L tumors.

RESULTS

Experiments on tissue-mimicking data suggest that the ML method using the partially synthetic data is accurate in predicting APT. In vivo experiments suggest that our method provides more accurate and robust prediction than the training using in vivo data and fully synthetic data.

CONCLUSION

Partially synthetic CEST data can address the challenges in conventional ML methods.

A Systematic Review of ChatGPT and Other Conversational Large Language Models in Healthcare

Background

The launch of the Chat Generative Pre-trained Transformer (ChatGPT) in November 2022 has attracted public attention and academic interest to large language models (LLMs), facilitating the emergence of many other innovative LLMs. These LLMs have been applied in various fields, including healthcare. Numerous studies have since been conducted regarding how to employ state-of-the-art LLMs in health-related scenarios to assist patients, doctors, and public health administrators.

Objective

This review aims to summarize the applications and concerns of applying conversational LLMs in healthcare and provide an agenda for future research on LLMs in healthcare.

Methods

We utilized PubMed, ACM, and IEEE digital libraries as primary sources for this review. We followed the guidance of Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRIMSA) to screen and select peer-reviewed research articles that (1) were related to both healthcare applications and conversational LLMs and (2) were published before September 1st, 2023, the date when we started paper collection and screening. We investigated these papers and classified them according to their applications and concerns.

Results

Our search initially identified 820 papers according to targeted keywords, out of which 65 papers met our criteria and were included in the review. The most popular conversational LLM was ChatGPT from OpenAI (60), followed by Bard from Google (1), Large Language Model Meta AI (LLaMA) from Meta (1), and other LLMs (5). These papers were classified into four categories in terms of their applications: 1) summarization, 2) medical knowledge inquiry, 3) prediction, and 4) administration, and four categories of concerns: 1) reliability, 2) bias, 3) privacy, and 4) public acceptability. There are 49 (75%) research papers using LLMs for summarization and/or medical knowledge inquiry, and 58 (89%) research papers expressing concerns about reliability and/or bias. We found that conversational LLMs exhibit promising results in summarization and providing medical knowledge to patients with a relatively high accuracy. However, conversational LLMs like ChatGPT are not able to provide reliable answers to complex health-related tasks that require specialized domain expertise. Additionally, no experiments in our reviewed papers have been conducted to thoughtfully examine how conversational LLMs lead to bias or privacy issues in healthcare research.

Conclusions

Future studies should focus on improving the reliability of LLM applications in complex health-related tasks, as well as investigating the mechanisms of how LLM applications brought bias and privacy issues. Considering the vast accessibility of LLMs, legal, social, and technical efforts are all needed to address concerns about LLMs to promote, improve, and regularize the application of LLMs in healthcare.

CE, Gevaert O, Rao A. Multimodal Machine Learning in Image-Based and Clinical Biomedicine: Survey and Prospects

Machine learning (ML) applications in medical artificial intelligence (AI) systems have shifted from traditional and statistical methods to increasing application of deep learning models. This survey navigates the current landscape of multimodal ML, focusing on its profound impact on medical image analysis and clinical decision support systems. Emphasizing challenges and innovations in addressing multimodal representation, fusion, translation, alignment, and co-learning, the paper explores the transformative potential of multimodal models for clinical predictions. It also highlights the need for principled assessments and practical implementation of such models, bringing attention to the dynamics between decision support systems and healthcare providers and personnel. Despite advancements, challenges such as data biases and the scarcity of "big data" in many biomedical domains persist. We conclude with a discussion on principled innovation and collaborative efforts to further the mission of seamless integration of multimodal ML models into biomedical practice.

Generating Synthetic Electronic Health Record Data Using Generative Adversarial Networks: Tutorial

Synthetic electronic health record (EHR) data generation has been increasingly recognized as an important solution to expand the accessibility and maximize the value of private health data on a large scale. Recent advances in machine learning have facilitated more accurate modeling for complex and high-dimensional data, thereby greatly enhancing the data quality of synthetic EHR data. Among various approaches, generative adversarial networks (GANs) have become the main technical path in the literature due to their ability to capture the statistical characteristics of real data. However, there is a scarcity of detailed guidance within the domain regarding the development procedures of synthetic EHR data. The objective of this tutorial is to present a transparent and reproducible process for generating structured synthetic EHR data using a publicly accessible EHR data set as an example. We cover the topics of GAN architecture, EHR data types and representation, data preprocessing, GAN training, synthetic data generation and postprocessing, and data quality evaluation. We conclude this tutorial by discussing multiple important issues and future opportunities in this domain. The source code of the entire process has been made publicly available.

Inherent Bias in Electronic Health Records: A Scoping Review of Sources of Bias

Objectives

1.1Biases inherent in electronic health records (EHRs), and therefore in medical artificial intelligence (AI) models may significantly exacerbate health inequities and challenge the adoption of ethical and responsible AI in healthcare. Biases arise from multiple sources, some of which are not as documented in the literature. Biases are encoded in how the data has been collected and labeled, by implicit and unconscious biases of clinicians, or by the tools used for data processing. These biases and their encoding in healthcare records undermine the reliability of such data and bias clinical judgments and medical outcomes. Moreover, when healthcare records are used to build data-driven solutions, the biases are further exacerbated, resulting in systems that perpetuate biases and induce healthcare disparities. This literature scoping review aims to categorize the main sources of biases inherent in EHRs.

Methods

1.2We queried PubMed and Web of Science on January 19th, 2023, for peer-reviewed sources in English, published between 2016 and 2023, using the PRISMA approach to stepwise scoping of the literature. To select the papers that empirically analyze bias in EHR, from the initial yield of 430 papers, 27 duplicates were removed, and 403 studies were screened for eligibility. 196 articles were removed after the title and abstract screening, and 96 articles were excluded after the full-text review resulting in a final selection of 116 articles.

Results

1.3Systematic categorizations of diverse sources of bias are scarce in the literature, while the effects of separate studies are often convoluted and methodologically contestable. Our categorization of published empirical evidence identified the six main sources of bias: a) bias arising from past clinical trials; b) data-related biases arising from missing, incomplete information or poor labeling of data; human-related bias induced by c) implicit clinician bias, d) referral and admission bias; e) diagnosis or risk disparities bias and finally, (f) biases in machinery and algorithms.

Conclusions

1.4Machine learning and data-driven solutions can potentially transform healthcare delivery, but not without limitations. The core inputs in the systems (data and human factors) currently contain several sources of bias that are poorly documented and analyzed for remedies. The current evidence heavily focuses on data-related biases, while other sources are less often analyzed or anecdotal. However, these different sources of biases add to one another exponentially. Therefore, to understand the issues holistically we need to explore these diverse sources of bias. While racial biases in EHR have been often documented, other sources of biases have been less frequently investigated and documented (e.g. gender-related biases, sexual orientation discrimination, socially induced biases, and implicit, often unconscious, human-related cognitive biases). Moreover, some existing studies lack causal evidence, illustrating the different prevalences of disease across groups, which does not per se prove the causality. Our review shows that data-, human- and machine biases are prevalent in healthcare and they significantly impact healthcare outcomes and judgments and exacerbate disparities and differential treatment. Understanding how diverse biases affect AI systems and recommendations is critical. We suggest that researchers and medical personnel should develop safeguards and adopt data-driven solutions with a "bias-in-mind" approach. More empirical evidence is needed to tease out the effects of different sources of bias on health outcomes.

Finding new analgesics: Computational pharmacology faces drug discovery challenges

Despite the worldwide prevalence and huge burden of pain, pain is an undertreated phenomenon. Currently used analgesics have several limitations regarding their efficacy and safety. The discovery of analgesics possessing a novel mechanism of action has faced multiple challenges, including a limited understanding of biological processes underpinning pain and analgesia and poor animal-to-human translation. Computational pharmacology is currently employed to face these challenges. In this review, we discuss the theory, methods, and applications of computational pharmacology in pain research. Computational pharmacology encompasses a wide variety of theoretical concepts and practical methodological approaches, with the overall aim of gaining biological insight through data acquisition and analysis. Data are acquired from patients or animal models with pain or analgesic treatment, at different levels of biological organization (molecular, cellular, physiological, and behavioral). Distinct methodological algorithms can then be used to analyze and integrate data. This helps to facilitate the identification of biological molecules and processes associated with pain phenotype, build quantitative models of pain signaling, and extract translatable features between humans and animals. However, computational pharmacology has several limitations, and its predictions can provide false positive and negative findings. Therefore, computational predictions are required to be validated experimentally before drawing solid conclusions. In this review, we discuss several case study examples of combining and integrating computational tools with experimental pain research tools to meet drug discovery challenges.

GAN-based generation of realistic 3D volumetric data: A systematic review and taxonomy

With the massive proliferation of data-driven algorithms, such as deep learning-based approaches, the availability of high-quality data is of great interest. Volumetric data is very important in medicine, as it ranges from disease diagnoses to therapy monitoring. When the dataset is sufficient, models can be trained to help doctors with these tasks. Unfortunately, there are scenarios where large amounts of data is unavailable. For example, rare diseases and privacy issues can lead to restricted data availability. In non-medical fields, the high cost of obtaining enough high-quality data can also be a concern. A solution to these problems can be the generation of realistic synthetic data using Generative Adversarial Networks (GANs). The existence of these mechanisms is a good asset, especially in healthcare, as the data must be of good quality, realistic, and without privacy issues. Therefore, most of the publications on volumetric GANs are within the medical domain. In this review, we provide a summary of works that generate realistic volumetric synthetic data using GANs. We therefore outline GAN-based methods in these areas with common architectures, loss functions and evaluation metrics, including their advantages and disadvantages. We present a novel taxonomy, evaluations, challenges, and research opportunities to provide a holistic overview of the current state of volumetric GANs.

Generative models improve fairness of medical classifiers under distribution shifts

Domain generalization is a ubiquitous challenge for machine learning in healthcare. Model performance in real-world conditions might be lower than expected because of discrepancies between the data encountered during deployment and development. Underrepresentation of some groups or conditions during model development is a common cause of this phenomenon. This challenge is often not readily addressed by targeted data acquisition and 'labeling' by expert clinicians, which can be prohibitively expensive or practically impossible because of the rarity of conditions or the available clinical expertise. We hypothesize that advances in generative artificial intelligence can help mitigate this unmet need in a steerable fashion, enriching our training dataset with synthetic examples that address shortfalls of underrepresented conditions or subgroups. We show that diffusion models can automatically learn realistic augmentations from data in a label-efficient manner. We demonstrate that learned augmentations make models more robust and statistically fair in-distribution and out of distribution. To evaluate the generality of our approach, we studied three distinct medical imaging contexts of varying difficulty: (1) histopathology, (2) chest X-ray and (3) dermatology images. Complementing real samples with synthetic ones improved the robustness of models in all three medical tasks and increased fairness by improving the accuracy of clinical diagnosis within underrepresented groups, especially out of distribution.

Cross2SynNet: cross-device-cross-modal synthesis of routine brain MRI sequences from CT with brain lesion

OBJECTIVES

CT and MR are often needed to determine the location and extent of brain lesions collectively to improve diagnosis. However, patients with acute brain diseases cannot complete the MRI examination within a short time. The aim of the study is to devise a cross-device and cross-modal medical image synthesis (MIS) method Cross2SynNet for synthesizing routine brain MRI sequences of T1WI, T2WI, FLAIR, and DWI from CT with stroke and brain tumors.

MATERIALS AND METHODS

For the retrospective study, the participants covered four different diseases of cerebral ischemic stroke (CIS-cohort), cerebral hemorrhage (CH-cohort), meningioma (M-cohort), glioma (G-cohort). The MIS model Cross2SynNet was established on the basic architecture of conditional generative adversarial network (CGAN), of which, the fully convolutional Transformer (FCT) module was adopted into generator to capture the short- and long-range dependencies between healthy and pathological tissues, and the edge loss function was to minimize the difference in gradient magnitude between synthetic image and ground truth. Three metrics of mean square error (MSE), peak signal-to-noise ratio (PSNR), and structure similarity index measure (SSIM) were used for evaluation.

RESULTS

A total of 230 participants (mean patient age, 59.77 years ± 13.63 [standard deviation]; 163 men [71%] and 67 women [29%]) were included, including CIS-cohort (95 participants between Dec 2019 and Feb 2022), CH-cohort (69 participants between Jan 2020 and Dec 2021), M-cohort (40 participants between Sep 2018 and Dec 2021), and G-cohort (26 participants between Sep 2019 and Dec 2021). The Cross2SynNet achieved averaged values of MSE = 0.008, PSNR = 21.728, and SSIM = 0.758 when synthesizing MRIs from CT, outperforming the CycleGAN, pix2pix, RegGAN, Pix2PixHD, and ResViT. The Cross2SynNet could synthesize the brain lesion on pseudo DWI even if the CT image did not exhibit clear signal in the acute ischemic stroke patients.

CONCLUSIONS

Cross2SynNet could achieve routine brain MRI synthesis of T1WI, T2WI, FLAIR, and DWI from CT with promising performance given the brain lesion of stroke and brain tumor.

Generation of synthetic whole-slide image tiles of tumours from RNA-sequencing data via cascaded diffusion models

Training machine-learning models with synthetically generated data can alleviate the problem of data scarcity when acquiring diverse and sufficiently large datasets is costly and challenging. Here we show that cascaded diffusion models can be used to synthesize realistic whole-slide image tiles from latent representations of RNA-sequencing data from human tumours. Alterations in gene expression affected the composition of cell types in the generated synthetic image tiles, which accurately preserved the distribution of cell types and maintained the cell fraction observed in bulk RNA-sequencing data, as we show for lung adenocarcinoma, kidney renal papillary cell carcinoma, cervical squamous cell carcinoma, colon adenocarcinoma and glioblastoma. Machine-learning models pretrained with the generated synthetic data performed better than models trained from scratch. Synthetic data may accelerate the development of machine-learning models in scarce-data settings and allow for the imputation of missing data modalities.

Potential of artificial intelligence in injury prevention research and practice

There is increasing interest and use of artificial Intelligence algorithms and methods in biomedical research and practice, particularly as the technology has made significant advances in the past decade and become more accessible to more disciplines. This editorial briefly reviews this technology and its potential for injury prevention research and practice, proposing ways that it can be used to advance the discipline, as well as the potential pitfalls, concerns and biases that accompany it.

Mimicking clinical trials with synthetic acute myeloid leukemia patients using generative artificial intelligence

Clinical research relies on high-quality patient data, however, obtaining big data sets is costly and access to existing data is often hindered by privacy and regulatory concerns. Synthetic data generation holds the promise of effectively bypassing these boundaries allowing for simplified data accessibility and the prospect of synthetic control cohorts. We employed two different methodologies of generative artificial intelligence - CTAB-GAN+ and normalizing flows (NFlow) - to synthesize patient data derived from 1606 patients with acute myeloid leukemia, a heterogeneous hematological malignancy, that were treated within four multicenter clinical trials. Both generative models accurately captured distributions of demographic, laboratory, molecular and cytogenetic variables, as well as patient outcomes yielding high performance scores regarding fidelity and usability of both synthetic cohorts (n = 1606 each). Survival analysis demonstrated close resemblance of survival curves between original and synthetic cohorts. Inter-variable relationships were preserved in univariable outcome analysis enabling explorative analysis in our synthetic data. Additionally, training sample privacy is safeguarded mitigating possible patient re-identification, which we quantified using Hamming distances. We provide not only a proof-of-concept for synthetic data generation in multimodal clinical data for rare diseases, but also full public access to synthetic data sets to foster further research.

Generative AI in Medical Practice: In-Depth Exploration of Privacy and Security Challenges

As advances in artificial intelligence (AI) continue to transform and revolutionize the field of medicine, understanding the potential uses of generative AI in health care becomes increasingly important. Generative AI, including models such as generative adversarial networks and large language models, shows promise in transforming medical diagnostics, research, treatment planning, and patient care. However, these data-intensive systems pose new threats to protected health information. This Viewpoint paper aims to explore various categories of generative AI in health care, including medical diagnostics, drug discovery, virtual health assistants, medical research, and clinical decision support, while identifying security and privacy threats within each phase of the life cycle of such systems (ie, data collection, model development, and implementation phases). The objectives of this study were to analyze the current state of generative AI in health care, identify opportunities and privacy and security challenges posed by integrating these technologies into existing health care infrastructure, and propose strategies for mitigating security and privacy risks. This study highlights the importance of addressing the security and privacy threats associated with generative AI in health care to ensure the safe and effective use of these systems. The findings of this study can inform the development of future generative AI systems in health care and help health care organizations better understand the potential benefits and risks associated with these systems. By examining the use cases and benefits of generative AI across diverse domains within health care, this paper contributes to theoretical discussions surrounding AI ethics, security vulnerabilities, and data privacy regulations. In addition, this study provides practical insights for stakeholders looking to adopt generative AI solutions within their organizations.

Generative artificial intelligence: synthetic datasets in dentistry

INTRODUCTION

Artificial Intelligence (AI) algorithms, particularly Deep Learning (DL) models are known to be data intensive. This has increased the demand for digital data in all domains of healthcare, including dentistry. The main hindrance in the progress of AI is access to diverse datasets which train DL models ensuring optimal performance, comparable to subject experts. However, administration of these traditionally acquired datasets is challenging due to privacy regulations and the extensive manual annotation required by subject experts. Biases such as ethical, socioeconomic and class imbalances are also incorporated during the curation of these datasets, limiting their overall generalizability. These challenges prevent their accrual at a larger scale for training DL models.

METHODS

Generative AI techniques can be useful in the production of Synthetic Datasets (SDs) that can overcome issues affecting traditionally acquired datasets. Variational autoencoders, generative adversarial networks and diffusion models have been used to generate SDs. The following text is a review of these generative AI techniques and their operations. It discusses the chances of SDs and challenges with potential solutions which will improve the understanding of healthcare professionals working in AI research.

CONCLUSION

Synthetic data customized to the need of researchers can be produced to train robust AI models. These models, having been trained on such a diverse dataset will be applicable for dissemination across countries. However, there is a need for the limitations associated with SDs to be better understood, and attempts made to overcome those concerns prior to their widespread use.

CHeart: A Conditional Spatio-Temporal Generative Model for Cardiac Anatomy

Two key questions in cardiac image analysis are to assess the anatomy and motion of the heart from images; and to understand how they are associated with non-imaging clinical factors such as gender, age and diseases. While the first question can often be addressed by image segmentation and motion tracking algorithms, our capability to model and answer the second question is still limited. In this work, we propose a novel conditional generative model to describe the 4D spatio-temporal anatomy of the heart and its interaction with non-imaging clinical factors. The clinical factors are integrated as the conditions of the generative modelling, which allows us to investigate how these factors influence the cardiac anatomy. We evaluate the model performance in mainly two tasks, anatomical sequence completion and sequence generation. The model achieves high performance in anatomical sequence completion, comparable to or outperforming other state-of-the-art generative models. In terms of sequence generation, given clinical conditions, the model can generate realistic synthetic 4D sequential anatomies that share similar distributions with the real data. The code and the trained generative model are available at https://github.com/MengyunQ/CHeart.

A Generative Method for a Laryngeal Biosignal

The Glottal Area Waveform (GAW) is an important component in quantitative clinical voice assessment, providing valuable insights into vocal fold function. In this study, we introduce a novel method employing Variational Autoencoders (VAEs) to generate synthetic GAWs. Our approach enables the creation of synthetic GAWs that closely replicate real-world data, offering a versatile tool for researchers and clinicians. We elucidate the process of manipulating the VAE latent space using the Glottal Opening Vector (GlOVe). The GlOVe allows precise control over the synthetic closure and opening of the vocal folds. By utilizing the GlOVe, we generate synthetic laryngeal biosignals. These biosignals accurately reflect vocal fold behavior, allowing for the emulation of realistic glottal opening changes. This manipulation extends to the introduction of arbitrary oscillations in the vocal folds, closely resembling real vocal fold oscillations. The range of factor coefficient values enables the generation of diverse biosignals with varying frequencies and amplitudes. Our results demonstrate that this approach yields highly accurate laryngeal biosignals, with the Normalized Mean Absolute Error values for various frequencies ranging from 9.6 ⋅ 10-3 to 1.20 ⋅ 10-2 for different experimented frequencies, alongside a remarkable training effectiveness, reflected in reductions of up to approximately 89.52% in key loss components. This proposed method may have implications for downstream speech synthesis and phonetics research, offering the potential for advanced and natural-sounding speech technologies.

Data encoding for healthcare data democratization and information leakage prevention

The lack of data democratization and information leakage from trained models hinder the development and acceptance of robust deep learning-based healthcare solutions. This paper argues that irreversible data encoding can provide an effective solution to achieve data democratization without violating the privacy constraints imposed on healthcare data and clinical models. An ideal encoding framework transforms the data into a new space where it is imperceptible to a manual or computational inspection. However, encoded data should preserve the semantics of the original data such that deep learning models can be trained effectively. This paper hypothesizes the characteristics of the desired encoding framework and then exploits random projections and random quantum encoding to realize this framework for dense and longitudinal or time-series data. Experimental evaluation highlights that models trained on encoded time-series data effectively uphold the information bottleneck principle and hence, exhibit lesser information leakage from trained models.

A scalable federated learning solution for secondary care using low-cost microcomputing: privacy-preserving development and evaluation of a COVID-19 screening test in UK hospitals

BACKGROUND

Multicentre training could reduce biases in medical artificial intelligence (AI); however, ethical, legal, and technical considerations can constrain the ability of hospitals to share data. Federated learning enables institutions to participate in algorithm development while retaining custody of their data but uptake in hospitals has been limited, possibly as deployment requires specialist software and technical expertise at each site. We previously developed an artificial intelligence-driven screening test for COVID-19 in emergency departments, known as CURIAL-Lab, which uses vital signs and blood tests that are routinely available within 1 h of a patient's arrival. Here we aimed to federate our COVID-19 screening test by developing an easy-to-use embedded system-which we introduce as full-stack federated learning-to train and evaluate machine learning models across four UK hospital groups without centralising patient data.

METHODS

We supplied a Raspberry Pi 4 Model B preloaded with our federated learning software pipeline to four National Health Service (NHS) hospital groups in the UK: Oxford University Hospitals NHS Foundation Trust (OUH; through the locally linked research University, University of Oxford), University Hospitals Birmingham NHS Foundation Trust (UHB), Bedfordshire Hospitals NHS Foundation Trust (BH), and Portsmouth Hospitals University NHS Trust (PUH). OUH, PUH, and UHB participated in federated training, training a deep neural network and logistic regressor over 150 rounds to form and calibrate a global model to predict COVID-19 status, using clinical data from patients admitted before the pandemic (COVID-19-negative) and testing positive for COVID-19 during the first wave of the pandemic. We conducted a federated evaluation of the global model for admissions during the second wave of the pandemic at OUH, PUH, and externally at BH. For OUH and PUH, we additionally performed local fine-tuning of the global model using the sites' individual training data, forming a site-tuned model, and evaluated the resultant model for admissions during the second wave of the pandemic. This study included data collected between Dec 1, 2018, and March 1, 2021; the exact date ranges used varied by site. The primary outcome was overall model performance, measured as the area under the receiver operating characteristic curve (AUROC). Removable micro secure digital (microSD) storage was destroyed on study completion.

FINDINGS

Clinical data from 130 941 patients (1772 COVID-19-positive), routinely collected across three hospital groups (OUH, PUH, and UHB), were included in federated training. The evaluation step included data from 32 986 patients (3549 COVID-19-positive) attending OUH, PUH, or BH during the second wave of the pandemic. Federated training of a global deep neural network classifier improved upon performance of models trained locally in terms of AUROC by a mean of 27·6% (SD 2·2): AUROC increased from 0·574 (95% CI 0·560-0·589) at OUH and 0·622 (0·608-0·637) at PUH using the locally trained models to 0·872 (0·862-0·882) at OUH and 0·876 (0·865-0·886) at PUH using the federated global model. Performance improvement was smaller for a logistic regression model, with a mean increase in AUROC of 13·9% (0·5%). During federated external evaluation at BH, AUROC for the global deep neural network model was 0·917 (0·893-0·942), with 89·7% sensitivity (83·6-93·6) and 76·6% specificity (73·9-79·1). Site-specific tuning of the global model did not significantly improve performance (change in AUROC <0·01).

INTERPRETATION

We developed an embedded system for federated learning, using microcomputing to optimise for ease of deployment. We deployed full-stack federated learning across four UK hospital groups to develop a COVID-19 screening test without centralising patient data. Federation improved model performance, and the resultant global models were generalisable. Full-stack federated learning could enable hospitals to contribute to AI development at low cost and without specialist technical expertise at each site.

FUNDING

The Wellcome Trust, University of Oxford Medical and Life Sciences Translational Fund.

[Artificial intelligence and secure use of health data in the KI-FDZ project: anonymization, synthetization, and secure processing of real-world data]

The increasing digitization of the healthcare system is leading to a growing volume of health data. Leveraging this data beyond its initial collection purpose for secondary use can provide valuable insights into diagnostics, treatment processes, and the quality of care. The Health Data Lab (HDL) will provide infrastructure for this purpose. Both the protection of patient privacy and optimal analytical capabilities are of central importance in this context, and artificial intelligence (AI) provides two opportunities. First, it enables the analysis of large volumes of data with flexible models, which means that hidden correlations and patterns can be discovered. Second, synthetic - that is, artificial - data generated by AI can protect privacy.This paper describes the KI-FDZ project, which aims to investigate innovative technologies that can support the secure provision of health data for secondary research purposes. A multi-layered approach is investigated in which data-level measures can be combined in different ways with processing in secure environments. To this end, anonymization and synthetization methods, among others, are evaluated based on two concrete application examples. Moreover, it is examined how the creation of machine learning pipelines and the execution of AI algorithms can be supported in secure processing environments. Preliminary results indicate that this approach can achieve a high level of protection while maintaining data validity. The approach investigated in the project can be an important building block in the secure secondary use of health data.

SpasticSim: a synthetic data generation method for upper limb spasticity modelling in neurorehabilitation

In neurorehabilitation, assessment of functional problems is essential to define optimal rehabilitation treatments. Usually, this assessment process requires distinguishing between impaired and non-impaired behavior of limbs. One of the common muscle motor disorders affecting limbs is spasticity, which is complicated to quantify objectively due to the complex nature of motor control. Thus, the lack of heterogeneous samples of patients constituting an acceptable amount of data is an obstacle which is relevant to understanding the behavior of spasticity and, consequently, quantifying it. In this article, we use the 3D creation suite Blender combined with the MBLab add-on to generate synthetic samples of human body models, aiming to be as sufficiently representative as possible to real human samples. Exporting these samples to OpenSim and performing four specific upper limb movements, we analyze the muscle behavior by simulating the six degrees of spasticity contemplated by the Modified Ashworth Scale (MAS). The complete dataset of patients and movements is open-source and available for future research. This approach advocates the potential to generate synthetic data for testing and validating musculoskeletal models.

Rapid single-particle chemical imaging of nanoplastics by SRS microscopy

Plastics are now omnipresent in our daily lives. The existence of microplastics (1 µm to 5 mm in length) and possibly even nanoplastics (<1 μm) has recently raised health concerns. In particular, nanoplastics are believed to be more toxic since their smaller size renders them much more amenable, compared to microplastics, to enter the human body. However, detecting nanoplastics imposes tremendous analytical challenges on both the nano-level sensitivity and the plastic-identifying specificity, leading to a knowledge gap in this mysterious nanoworld surrounding us. To address these challenges, we developed a hyperspectral stimulated Raman scattering (SRS) imaging platform with an automated plastic identification algorithm that allows micro-nano plastic analysis at the single-particle level with high chemical specificity and throughput. We first validated the sensitivity enhancement of the narrow band of SRS to enable high-speed single nanoplastic detection below 100 nm. We then devised a data-driven spectral matching algorithm to address spectral identification challenges imposed by sensitive narrow-band hyperspectral imaging and achieve robust determination of common plastic polymers. With the established technique, we studied the micro-nano plastics from bottled water as a model system. We successfully detected and identified nanoplastics from major plastic types. Micro-nano plastics concentrations were estimated to be about 2.4 ± 1.3 × 105 particles per liter of bottled water, about 90% of which are nanoplastics. This is orders of magnitude more than the microplastic abundance reported previously in bottled water. High-throughput single-particle counting revealed extraordinary particle heterogeneity and nonorthogonality between plastic composition and morphologies; the resulting multidimensional profiling sheds light on the science of nanoplastics.

Large language models to identify social determinants of health in electronic health records

Social determinants of health (SDoH) play a critical role in patient outcomes, yet their documentation is often missing or incomplete in the structured data of electronic health records (EHRs). Large language models (LLMs) could enable high-throughput extraction of SDoH from the EHR to support research and clinical care. However, class imbalance and data limitations present challenges for this sparsely documented yet critical information. Here, we investigated the optimal methods for using LLMs to extract six SDoH categories from narrative text in the EHR: employment, housing, transportation, parental status, relationship, and social support. The best-performing models were fine-tuned Flan-T5 XL for any SDoH mentions (macro-F1 0.71), and Flan-T5 XXL for adverse SDoH mentions (macro-F1 0.70). Adding LLM-generated synthetic data to training varied across models and architecture, but improved the performance of smaller Flan-T5 models (delta F1 + 0.12 to +0.23). Our best-fine-tuned models outperformed zero- and few-shot performance of ChatGPT-family models in the zero- and few-shot setting, except GPT4 with 10-shot prompting for adverse SDoH. Fine-tuned models were less likely than ChatGPT to change their prediction when race/ethnicity and gender descriptors were added to the text, suggesting less algorithmic bias (p < 0.05). Our models identified 93.8% of patients with adverse SDoH, while ICD-10 codes captured 2.0%. These results demonstrate the potential of LLMs in improving real-world evidence on SDoH and assisting in identifying patients who could benefit from resource support.

Synthetic Data Improve Survival Status Prediction Models in Early-Onset Colorectal Cancer

PURPOSE

In artificial intelligence-based modeling, working with a limited number of patient groups is challenging. This retrospective study aimed to evaluate whether applying synthetic data generation methods to the clinical data of small patient groups can enhance the performance of prediction models.

MATERIALS AND METHODS

A data set collected by the Cancer Registry Library Project from the Yonsei Cancer Center (YCC), Severance Hospital, between January 2008 and October 2020 was reviewed. Patients with colorectal cancer younger than 50 years who started their initial treatment at YCC were included. A Bayesian network-based synthesizing model was used to generate a synthetic data set, combined with the differential privacy (DP) method.

RESULTS

A synthetic population of 5,005 was generated from a data set of 1,253 patients with 93 clinical features. The Hellinger distance and correlation difference metric were below 0.3 and 0.5, respectively, indicating no statistical difference. The overall survival by disease stage did not differ between the synthetic and original populations. Training with the synthetic data and validating with the original data showed the highest performances of 0.850, 0.836, and 0.790 for the Decision Tree, Random Forest, and XGBoost models, respectively. Comparison of synthetic data sets with different epsilon parameters from the original data sets showed improved performance >0.1%. For extremely small data sets, models using synthetic data outperformed those using only original data sets. The reidentification risk measures demonstrated that the epsilons between 0.1 and 100 fell below the baseline, indicating a preserved privacy state.

CONCLUSION

The synthetic data generation approach enhances predictive modeling performance by maintaining statistical and clinical integrity, and simultaneously reduces privacy risks through the application of DP techniques.

Survey of Medical Applications of Federated Learning

OBJECTIVES

Medical artificial intelligence (AI) has recently attracted considerable attention. However, training medical AI models is challenging due to privacy-protection regulations. Among the proposed solutions, federated learning (FL) stands out. FL involves transmitting only model parameters without sharing the original data, making it particularly suitable for the medical field, where data privacy is paramount. This study reviews the application of FL in the medical domain.

METHODS

We conducted a literature search using the keywords "federated learning" in combination with "medical," "healthcare," or "clinical" on Google Scholar and PubMed. After reviewing titles and abstracts, 58 papers were selected for analysis. These FL studies were categorized based on the types of data used, the target disease, the use of open datasets, the local model of FL, and the neural network model. We also examined issues related to heterogeneity and security.

RESULTS

In the investigated FL studies, the most commonly used data type was image data, and the most studied target diseases were cancer and COVID-19. The majority of studies utilized open datasets. Furthermore, 72% of the FL articles addressed heterogeneity issues, while 50% discussed security concerns.

CONCLUSIONS

FL in the medical domain appears to be in its early stages, with most research using open data and focusing on specific data types and diseases for performance verification purposes. Nonetheless, medical FL research is anticipated to be increasingly applied and to become a vital component of multi-institutional research.

Improving mixed-integer temporal modeling by generating synthetic data using conditional generative adversarial networks: A case study of fluid overload prediction in the intensive care unit

OBJECTIVE

The challenge of mixed-integer temporal data, which is particularly prominent for medication use in the critically ill, limits the performance of predictive models. The purpose of this evaluation was to pilot test integrating synthetic data within an existing dataset of complex medication data to improve machine learning model prediction of fluid overload.

MATERIALS AND METHODS

This retrospective cohort study evaluated patients admitted to an ICU ≥ 72 h. Four machine learning algorithms to predict fluid overload after 48-72 h of ICU admission were developed using the original dataset. Then, two distinct synthetic data generation methodologies (synthetic minority over-sampling technique (SMOTE) and conditional tabular generative adversarial network (CTGAN)) were used to create synthetic data. Finally, a stacking ensemble technique designed to train a meta-learner was established. Models underwent training in three scenarios of varying qualities and quantities of datasets.

RESULTS

Training machine learning algorithms on the combined synthetic and original dataset overall increased the performance of the predictive models compared to training on the original dataset. The highest performing model was the meta-model trained on the combined dataset with 0.83 AUROC while it managed to significantly enhance the sensitivity across different training scenarios.

DISCUSSION

The integration of synthetically generated data is the first time such methods have been applied to ICU medication data and offers a promising solution to enhance the performance of machine learning models for fluid overload, which may be translated to other ICU outcomes. A meta-learner was able to make a trade-off between different performance metrics and improve the ability to identify the minority class.

The perils and promises of generative artificial intelligence in neurointerventional surgery

Generative artificial intelligence (AI) holds great promise in neurointerventional surgery by providing clinicians with powerful tools for improving surgical precision, accuracy of diagnoses, and treatment planning. However, potential perils include biases or inaccuracies in the data used to train the algorithms, over-reliance on generative AI without human oversight, patient privacy concerns, and ethical implications of using AI in medical decision-making. Careful regulation and oversight are needed to ensure that the promises of generative AI in neurointerventional surgery are realized while minimizing its potential perils.[ChatGPT authored summary using the prompt "In one paragraph summarize the promises and perils of generative AI in neurointerventional surgery".].