OBJECTIVE: Machine learning models have utilized semantic features, deep features, or both to assess lung nodule malignancy. However, their reliance on manual annotation during inference, limited interpretability, and sensitivity to imaging variations hinder their application in real-world clinical settings. Thus, this research aims to integrate semantic features derived from radiologists' assessments of nodules, guiding the model to learn clinically relevant, robust, and explainable imaging features for predicting lung cancer. METHODS: We obtained 938 low-dose CT scans from the National Lung Screening Trial (NLST) with 1,261 nodules and semantic features. Additionally, the Lung Image Database Consortium dataset contains 1,018 CT scans, with 2,625 lesions annotated for nodule characteristics. Three external datasets were obtained from UCLA Health, the LUNGx Challenge, and the Duke Lung Cancer Screening. For imaging input, we obtained 2D nodule slices in nine directions from 50×50×50mm nodule crop. We converted structured semantic features into sentences using Gemini. We fine-tuned a pretrained Contrastive Language-Image Pretraining (CLIP) model with a parameter-efficient fine-tuning approach to align imaging and semantic text features and predict the one-year lung cancer diagnosis. RESULTS: Our model outperformed the state-of-the-art (SOTA) models in the NLST test set with an AUROC of 0.901 and AUPRC of 0.776. It also showed robust results in external datasets. Using CLIP, we also obtained predictions on semantic features through zero-shot inference, such as nodule margin (AUROC: 0.807), nodule consistency (0.812), and pleural attachment (0.840). CONCLUSION: By incorporating semantic features into the vision-language model, our approach surpasses the SOTA models in predicting lung cancer from CT scans collected from diverse clinical settings. It provides explainable outputs, aiding clinicians in comprehending the underlying meaning of model predictions. The code is available at https://github.com/luotingzhuang/CLIP_nodule.

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.

The development and deployment of machine learning (ML) models for biomedical research and healthcare currently lacks standard methodologies. Although tools for model replication are numerous, without a unifying blueprint it remains difficult to scientifically reproduce predictive ML models for any number of reasons (e.g., assumptions regarding data distributions and preprocessing, unclear test metrics, etc.) and ultimately, questions around generalizability and transportability are not readily answered. To facilitate scientific reproducibility, we built upon the Predictive Model Markup Language (PMML) to capture essential information. As a key component of the PREdictive Model Index and Exchange REpository (PREMIERE) platform, we present the Automated Metadata Pipeline (AMP) for conversion of a given predictive ML model into an extended PMML file that autocompletes an ML-based checklist, assessing model elements for interoperability and reproducibility. We demonstrate this pipeline on multiple test cases with three different ML algorithms and health-related datasets, providing a foundation for future predictive model reproducibility, sharing, and comparison. Problem: The development and deployment of machine learning (ML) models for biomedical research and healthcare currently lack standard methodologies, leading to problems of scientific reproducibility and interoperability. What is Already Known: Although there are many tools for model replication, without a unifying blueprint, it remains difficult to scientifically reproduce predictive ML models for any number of reasons. Moreover, questions around generalizability and transportability are not readily answered. What this Paper Adds: This study builds upon the Predictive Model Markup Language (PMML) to capture essential information and presents the Automated Metadata Pipeline (AMP) for conversion of a given predictive ML model into an extended PMML file that auto-completes an ML-based checklist, assessing model elements for interoperability and reproducibility. We demonstrate this pipeline on multiple test cases with three different ML algorithms and health-related datasets. The proposed AMP provides a framework for automating the completion and evaluation of a comprehensive ML model checklist, increasing compliance and ultimately, predictive model reproducibility, sharing, and comparison of predictive models by ensuring all appropriate information is available.

X-ray computed tomography (CT) is commonly used to obtain vivo images to characterize diseases but results in radiation exposure to patients. Low-dose CT (LDCT) provides CT images of clinical quality with reduced cumulative radiation dose. Iterative image reconstruction methods with effective regularization are used for LDCT but generally require more computing resources and induce higher computational load than the conventional filtered backprojection (FBP) methods. The high computational demand of the iterative reconstruction (IR) with notably increased reconstruction time precludes its routine clinical application. In this work, we focus on the FPGA acceleration of a compute-intensive full IR (full-IR) algorithm based on the Mumford-Shah regularization. At the algorithmic level, we propose a beam-based asynchronous update algorithm to reduce the computational cost and alleviate the conflicts. At the hardware-level, we first present pipeline-friendly optimization for the original algorithm to increase the computation throughput. We then apply the LDCT-specific tiling strategy to improve the data reuse rate. The experimental results show that our implementation takes 8.5 min to reconstruct a typical physical phantom with the image quality comparable with the vendor's result. The FPGA implementation achieves 11.6× throughput against the state-of-the-art GPU version.

MOTIVATION: Cancer heterogeneity is observed at multiple biological levels. To improve our understanding of these differences and their relevance in medicine, approaches to link organ- and tissue-level information from diagnostic images and cellular-level information from genomics are needed. However, these 'radiogenomic' studies often use linear or shallow models, depend on feature selection, or consider one gene at a time to map images to genes. Moreover, no study has systematically attempted to understand the molecular basis of imaging traits based on the interpretation of what the neural network has learned. These studies are thus limited in their ability to understand the transcriptomic drivers of imaging traits, which could provide additional context for determining clinical outcomes. RESULTS: We present a neural network-based approach that takes high-dimensional gene expression data as input and performs non-linear mapping to an imaging trait. To interpret the models, we propose gene masking and gene saliency to extract learned relationships from radiogenomic neural networks. In glioblastoma patients, our models outperformed comparable classifiers (>0.10 AUC) and our interpretation methods were validated using a similar model to identify known relationships between genes and molecular subtypes. We found that tumor imaging traits had specific transcription patterns, e.g. edema and genes related to cellular invasion, and 10 radiogenomic traits were significantly predictive of survival. We demonstrate that neural networks can model transcriptomic heterogeneity to reflect differences in imaging and can be used to derive radiogenomic traits with clinical value. AVAILABILITY AND IMPLEMENTATION: https://github.com/novasmedley/deepRadiogenomics. CONTACT: [email protected]. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Computed tomography (CT) is a widely used screening and diagnostic tool that allows clinicians to obtain a high-resolution, volumetric image of internal structures in a non-invasive manner. Increasingly, efforts have been made to improve the image quality of low-dose CT (LDCT) to reduce the cumulative radiation exposure of patients undergoing routine screening exams. The resurgence of deep learning has yielded a new approach for noise reduction by training a deep multi-layer convolutional neural networks (CNN) to map the low-dose to normal-dose CT images. However, CNN-based methods heavily rely on convolutional kernels, which use fixed-size filters to process one local neighborhood within the receptive field at a time. As a result, they are not efficient at retrieving structural information across large regions. In this paper, we propose a novel 3D self-attention convolutional neural network for the LDCT denoising problem. Our 3D self-attention module leverages the 3D volume of CT images to capture a wide range of spatial information both within CT slices and between CT slices. With the help of the 3D self-attention module, CNNs are able to leverage pixels with stronger relationships regardless of their distance and achieve better denoising results. In addition, we propose a self-supervised learning scheme to train a domain-specific autoencoder as the perceptual loss function. We combine these two methods and demonstrate their effectiveness on both CNN-based neural networks and WGAN-based neural networks with comprehensive experiments. Tested on the AAPM-Mayo Clinic Low Dose CT Grand Challenge data set, our experiments demonstrate that self-attention (SA) module and autoencoder (AE) perceptual loss function can efficiently enhance traditional CNNs and can achieve comparable or better results than the state-of-the-art methods.

Biologists synthesize research articles into coherent models—ideally, causal models, which predict how systems will respond to interventions. But it is challenging to derive causal models from articles alone, without primary data. To enable causal discovery using only literature, we built software for annotating empirical results in free text and computing valid explanations, expressed as causal graphs. This paper presents our meta-analytic pipeline: with the “research map” schema, we annotate results in literature, which we convert into logical constraints on causal structure; with these constraints, we find consistent causal graphs using a state-of-the-art, causal discovery algorithm based on answer set programming. Because these causal graphs show which relations are underdetermined, biologists can use this pipeline to select their next experiment. To demonstrate this approach, we annotated neuroscience articles and applied a “degrees-of-freedom” analysis for concisely visualizing features of the causal graphs that remain consistent with the evidence—a model space that is often too large for a machine to compute quickly, or for a researcher to examine exhaustively.

Given the increasing emphasis on delivering high-quality, cost-efficient healthcare, improved methodologies are needed to measure the accuracy and utility of ordered diagnostic examinations in achieving the appropriate diagnosis. Here, we present a data-driven approach for performing automated quality assessment of radiologic interpretations using other clinical information (e.g., pathology) as a reference standard for individual radiologists, subspecialty sections, imaging modalities, and entire departments. Downstream diagnostic conclusions from the electronic medical record are utilized as "truth" to which upstream diagnoses generated by radiology are compared. The described system automatically extracts and compares patient medical data to characterize concordance between clinical sources. Initial results are presented in the context of breast imaging, matching 18 101 radiologic interpretations with 301 pathology diagnoses and achieving a precision and recall of 84% and 92%, respectively. The presented data-driven method highlights the challenges of integrating multiple data sources and the application of information extraction tools to facilitate healthcare quality improvement.

Clinicians need to routinely make management decisions about patients who are at risk for a disease such as breast cancer. This paper presents a novel clinical decision support tool that is capable of helping physicians make diagnostic decisions. We apply this support system to improve the specificity of breast cancer screening and diagnosis. The system utilizes clinical context (e.g., demographics, medical history) to minimize the false positive rates while avoiding false negatives. An online contextual learning algorithm is used to update the diagnostic strategy presented to the physicians over time. We analytically evaluate the diagnostic performance loss of the proposed algorithm, in which the true patient distribution is not known and needs to be learned, as compared with the optimal strategy where all information is assumed known, and prove that the false positive rate of the proposed learning algorithm asymptotically converges to the optimum. In addition, our algorithm also has the important merit that it can provide individualized confidence estimates about the accuracy of the diagnosis recommendation. Moreover, the relevancy of contextual features is assessed, enabling the approach to identify specific contextual features that provide the most value of information in reducing diagnostic errors. Experiments were conducted using patient data collected at a large academic medical center. Our proposed approach outperforms the current clinical practice by 36% in terms of false positive rate given a 2% false negative rate.

Breast cancer screening policies attempt to achieve timely diagnosis by regularly screening healthy women via various imaging tests. Various clinical decisions are needed to manage the screening process: selecting initial screening tests, interpreting test results, and deciding if further diagnostic tests are required. Current screening policies are guided by clinical practice guidelines (CPGs), which represent a “one-size-fits-all” approach, designed to work well (on average) for a population, and can only offer coarse expert-based patient stratification that is not rigorously validated through data. Since the risks and benefits of screening tests are functions of each patient's features,personalized screening policiestailored to the features of individuals are desirable. To address this issue, we developedConfidentCare: a computer-aided clinical decision support system that learns a personalized screening policy from electronic health record (EHR) data. By a “personalized screening policy,” we mean a clustering of women's features, and a set of customized screening guidelines for each cluster. ConfidentCare operates by computing clusters of patients with similar features, then learning the “best” screening procedure for each cluster using a supervised learning algorithm. The algorithm ensures that the learned screening policy satisfies a predefined accuracy requirement with a high level of confidence for every cluster. By applying ConfidentCare to real-world data, we show that it outperforms the current CPGs in terms of cost efficiency and false positive rates: a reduction of 31$\%$in the false positive rate can be achieved.

Hospitals are increasingly utilizing business intelligence and analytics tools to mine electronic health data to uncover inefficiencies in care delivery (e.g., slow turnaround times, high readmission rates). Given that the expertise and experience of healthcare providers may vary significantly, an area of potential improvement is optimizing the way patient cases are recommended to clinical experts (e.g., the pathologist who is most adept at diagnosing a rare cancer). In this paper, we propose an expert selection system that automatically matches a given patient case to the best available expert considering both the available contextual information about a patient (e.g., demographics, medical history, signs and symptoms, past interventions) and the congestion of the expert. We prove that as the number of patients grows, the proposed algorithm will discover the best expert to select for patients with a specific context. Moreover, the algorithm also provides confidence bounds on the diagnostic accuracy of the expert it selects. While the proposed system can be applied in many scenarios, we demonstrate its performance in the context of assigning mammography exams to individual radiologists for interpretation. We show that our proposed system can improve current clinical practice by improving overall sensitivity and specificity of screening exams compared to random assignment.Finally, since each expert can only take a certain number of diagnosis decisions on a daily basis, we show how our system can take the experts' workload into account as well as the expertise when deciding how to select experts.

Imaging has become a prevalent tool in the diagnosis and treatment of many diseases, providing a unique in vivo, multi-scale view of anatomic and physiologic processes. With the increased use of imaging and its progressive technical advances, the role of imaging informatics is now evolving--from one of managing images, to one of integrating the full scope of clinical information needed to contextualize and link observations across phenotypic and genotypic scales. Several challenges exist for imaging informatics, including the need for methods to transform clinical imaging studies and associated data into structured information that can be organized and analyzed. We examine some of these challenges in establishing imaging-based observational databases that can support the creation of comprehensive disease models. The development of these databases and ensuing models can aid in medical decision making and knowledge discovery and ultimately, transform the use of imaging to support individually-tailored patient care.

Biomedical informatics is the interdisciplinary field that studies and pursues the effective uses of biomedical data, information, and knowledge for scientific inquiry, problem solving, and decision making, motivated by efforts to improve human health.1 ,2 Not only do biomedical informaticians study and develop theories, methods, and processes for the generation, manipulation, and sharing of biomedical data, but they also investigate how to model and reason on these data in order to effect beneficial change in the healthcare enterprise. In addition, an important aspect associated with developments in this field is the consideration of social and behavioral sciences in the design and evaluation of technical solutions. As a subfield of biomedical informatics, biomedical imaging informatics (BMII) encompasses all of the aforementioned aspects from the perspective of imaging. BMII has emerged as one of the fastest growing research areas in recent years given the evolution of techniques in molecular imaging, anatomical imaging, and functional imaging and advancements in imaging biomarker generation. Developments have also been accelerated by efforts to realize precision medicine,3 which necessitates a multiscale understanding of diseases that integrate insights in areas such as radiology, pathology, and genetics. This focus issue highlights the growing impact of BMII, demonstrating the increasing breadth of imaging modalities (eg, optical, molecular, in addition to traditional diagnostic modalities) and the diversity of specialties that depend on imaging information (eg, dermatology, pathology, surgery). Early efforts in BMII can be traced to the 1980s when the rise in radiological imaging techniques such as CT and MRI necessitated a digital, filmless approach to acquiring and interpreting images. The ability to acquire and distribute images electronically using picture archiving and communication systems (PACS) spawned a variety of applications aimed at improving radiological practice, research, and education. Imaging informatics efforts resulted in the development of specialized standardized … Correspondence to Dr William Hsu, Medical Imaging Informatics (MII) Group, Department of Radiological Sciences, UCLA David Geffen School of Medicine, Los Angeles, CA 90024, USA; willhsu{at}mii.ucla.edu

Due to the increasingly data-intensive clinical environment, physicians now have unprecedented access to detailed clinical information from a multitude of sources. However, applying this information to guide medical decisions for a specific patient case remains challenging. One issue is related to presenting information to the practitioner: displaying a large (irrelevant) amount of information often leads to information overload. Next-generation interfaces for the electronic health record (EHR) should not only make patient data easily searchable and accessible, but also synthesize fragments of evidence documented in the entire record to understand the etiology of a disease and its clinical manifestation in individual patients. In this paper, we describe our efforts toward creating a context-based EHR, which employs biomedical ontologies and (graphical) disease models as sources of domain knowledge to identify relevant parts of the record to display. We hypothesize that knowledge (e.g., variables, relationships) from these sources can be used to standardize, annotate, and contextualize information from the patient record, improving access to relevant parts of the record and informing medical decision making. To achieve this goal, we describe a framework that aggregates and extracts findings and attributes from free-text clinical reports, maps findings to concepts in available knowledge sources, and generates a tailored presentation of the record based on the information needs of the user. We have implemented this framework in a system called Adaptive EHR, demonstrating its capabilities to present and synthesize information from neurooncology patients. This paper highlights the challenges and potential applications of leveraging disease models to improve the access, integration, and interpretation of clinical patient data.