Adapted large language models can outperform medical experts in clinical text summarization.

Analyzing vast textual data and summarizing key information from electronic health records imposes a substantial burden on how clinicians allocate their time. Although large language models (LLMs) have shown promise in natural language processing (NLP) tasks, their effectiveness on a diverse range of clinical summarization tasks remains unproven. Here we applied adaptation methods to eight LLMs, spanning four distinct clinical summarization tasks: radiology reports, patient questions, progress notes and doctor-patient dialogue. Quantitative assessments with syntactic, semantic and conceptual NLP metrics reveal trade-offs between models and adaptation methods. A clinical reader study with 10 physicians evaluated summary completeness, correctness and conciseness; in most cases, summaries from our best-adapted LLMs were deemed either equivalent (45%) or superior (36%) compared with summaries from medical experts. The ensuing safety analysis highlights challenges faced by both LLMs and medical experts, as we connect errors to potential medical harm and categorize types of fabricated information. Our research provides evidence of LLMs outperforming medical experts in clinical text summarization across multiple tasks. This suggests that integrating LLMs into clinical workflows could alleviate documentation burden, allowing clinicians to focus more on patient care.

Prediction of pregnancy-related complications in women undergoing assisted reproduction, using machine learning methods.

OBJECTIVE: To use machine learning methods to develop prediction models of pregnancy complications in women who conceived with assisted reproductive techniques (ART). DESIGN: A nation-wide register-based cohort study with prospectively collected data. SETTING: Swedish national registers and nationwide quality IVF register. PATIENT(S): all nulliparous women who achieved birth within the first 3 ART treatment cycles between 2008 and 2016 in Sweden. INTERVENTION(S): Characteristics before the use of ART, such as demographics and medical history, were considered potential predictors in the development of before treatment prediction models. ART treatment details were further included in after treatment prediction models. MAIN OUTCOME MEASURE(S): Potential diagnoses of preeclampsia, placental complications (previa, accreta, and abruption), and postpartum hemorrhage were identified using the International Classification of Diseases recorded in the Swedish Medical Birth and Patient registers, respectively. Multiple prediction model algorithms were performed and compared for each outcome and treatment cycle, including logistic regression, decision tree model, naïve Bayes classification, support vector machine, random forest, and gradient boosting. The performance of each model was assessed with C statistic, and nested cross-validation was used to aid model selection and hyperparameter tuning. RESULT(S): A total of 14,732 women gave birth after the first (N = 7,302), second (N = 4,688), or third (N = 2,742) ART cycle, representing birth rates of 24.1%, 23.8%, and 22.0%. Overall prediction performance did not vary much across the different methods used. In the first cycle, the before treatment prediction performance was at best 66%, 66%, and 60% for preeclampsia, placental complications, and postpartum hemorrhage, respectively. Inclusion of after treatment characteristics conferred slight improvement (approximately 1%-5%), as did prediction in later cycles (approximately 1%-5%). The top influential and consistent predictors included age, region of residence, infertility diagnosis, and type of embryo transfer (fresh or frozen) in the later (2nd and 3rd) cycles. Body mass index was a top predictor of preeclampsia and was also influential for placental complications but not for postpartum hemorrhage. CONCLUSION(S): The combined use of demographics, medical history, and ART treatment information was not enough to confidently predict serious pregnancy complications in women who conceived with ART. Future studies are needed to assess if additional longitudinal follow-up during pregnancy can improve the prediction to allow clinical protocol development.

Clinical Text Summarization: Adapting Large Language Models Can Outperform Human Experts.

Sifting through vast textual data and summarizing key information from electronic health records (EHR) imposes a substantial burden on how clinicians allocate their time. Although large language models (LLMs) have shown immense promise in natural language processing (NLP) tasks, their efficacy on a diverse range of clinical summarization tasks has not yet been rigorously demonstrated. In this work, we apply domain adaptation methods to eight LLMs, spanning six datasets and four distinct clinical summarization tasks: radiology reports, patient questions, progress notes, and doctor-patient dialogue. Our thorough quantitative assessment reveals trade-offs between models and adaptation methods in addition to instances where recent advances in LLMs may not improve results. Further, in a clinical reader study with ten physicians, we show that summaries from our best-adapted LLMs are preferable to human summaries in terms of completeness and correctness. Our ensuing qualitative analysis highlights challenges faced by both LLMs and human experts. Lastly, we correlate traditional quantitative NLP metrics with reader study scores to enhance our understanding of how these metrics align with physician preferences. Our research marks the first evidence of LLMs outperforming human experts in clinical text summarization across multiple tasks. This implies that integrating LLMs into clinical workflows could alleviate documentation burden, empowering clinicians to focus more on personalized patient care and the inherently human aspects of medicine.

Self-supervised learning for medical image classification: a systematic review and implementation guidelines.

Advancements in deep learning and computer vision provide promising solutions for medical image analysis, potentially improving healthcare and patient outcomes. However, the prevailing paradigm of training deep learning models requires large quantities of labeled training data, which is both time-consuming and cost-prohibitive to curate for medical images. Self-supervised learning has the potential to make significant contributions to the development of robust medical imaging models through its ability to learn useful insights from copious medical datasets without labels. In this review, we provide consistent descriptions of different self-supervised learning strategies and compose a systematic review of papers published between 2012 and 2022 on PubMed, Scopus, and ArXiv that applied self-supervised learning to medical imaging classification. We screened a total of 412 relevant studies and included 79 papers for data extraction and analysis. With this comprehensive effort, we synthesize the collective knowledge of prior work and provide implementation guidelines for future researchers interested in applying self-supervised learning to their development of medical imaging classification models.

The requirements for performing artificial-intelligence-related research and model development.

Artificial intelligence research in health care has undergone tremendous growth in the last several years thanks to the explosion of digital health care data and systems that can leverage large amounts of data to learn patterns that can be applied to clinical tasks. In addition, given broad acceleration in machine learning across industries like transportation, media and commerce, there has been a significant growth in demand for machine-learning practitioners such as engineers and data scientists, who have skill sets that can be applied to health care use cases but who simultaneously lack important health care domain expertise. The purpose of this paper is to discuss the requirements of building an artificial-intelligence research enterprise including the research team, technical software/hardware, and procurement and curation of health care data.

Diagnostic accuracy of radiography, digital breast tomosynthesis, micro-CT and ultrasound for margin assessment during breast surgery: A systematic review and meta-analysis.

RATIONALE AND OBJECTIVES: Achieving adequate resection margins in breast conserving surgery is challenging and often demands more than one surgical procedure. We evaluated pooled diagnostic sensitivity, and specificity of radiological methods for intraoperative margin assessment and their impact on repeat surgery rate. MATERIALS AND METHODS: We included studies using radiography, digital breast tomosynthesis (DBT), micro-CT, and ultrasound for intraoperative margin assessment with the histological assessment as the reference method. A systematic search was performed in PubMed, Embase, Cochrane Library, Scopus, and Web of Science. Two investigators screened the studies for eligibility criteria and extracted data of the included studies independently. The quality assessment on diagnostic accuracy studies (QUADAS)-2 tool was used. A bivariate random effect model was used to obtained pooled sensitivity and specificity of the index tests in the meta-analysis. RESULTS: The systematic search resulted in screening of 798 unique records. Twenty-two articles with 29 radiological imaging methods were selected for meta-analysis. Pooled sensitivity and specificity and area under the curve were calculated for each of the 4 subgroups in the meta-analysis respectively: Radiography; 52%, 77%, 60%, DBT; 67%, 76%, 76%, micro-CT; 68%, 69%, 72%, and ultrasound; 72%, 78%, 80%. The repeat surgery rate was poorly reported in the included studies. CONCLUSION: Ultrasound showed the highest and radiography the lowest diagnostic performance for intraoperative margin assessment. However, the heterogeneity between studies was high and the subgroups small. The radiological methods for margin assessment need further improvement to provide reliable guidance in the clinical workflow and to prevent repeat surgeries.

CheXED: Comparison of a Deep Learning Model to a Clinical Decision Support System for Pneumonia in the Emergency Department.

PURPOSE: Patients with pneumonia often present to the emergency department (ED) and require prompt diagnosis and treatment. Clinical decision support systems for the diagnosis and management of pneumonia are commonly utilized in EDs to improve patient care. The purpose of this study is to investigate whether a deep learning model for detecting radiographic pneumonia and pleural effusions can improve functionality of a clinical decision support system (CDSS) for pneumonia management (ePNa) operating in 20 EDs. MATERIALS AND METHODS: In this retrospective cohort study, a dataset of 7434 prior chest radiographic studies from 6551 ED patients was used to develop and validate a deep learning model to identify radiographic pneumonia, pleural effusions, and evidence of multilobar pneumonia. Model performance was evaluated against 3 radiologists' adjudicated interpretation and compared with performance of the natural language processing of radiology reports used by ePNa. RESULTS: The deep learning model achieved an area under the receiver operating characteristic curve of 0.833 (95% confidence interval [CI]: 0.795, 0.868) for detecting radiographic pneumonia, 0.939 (95% CI: 0.911, 0.962) for detecting pleural effusions and 0.847 (95% CI: 0.800, 0.890) for identifying multilobar pneumonia. On all 3 tasks, the model achieved higher agreement with the adjudicated radiologist interpretation compared with ePNa. CONCLUSIONS: A deep learning model demonstrated higher agreement with radiologists than the ePNa CDSS in detecting radiographic pneumonia and related findings. Incorporating deep learning models into pneumonia CDSS could enhance diagnostic performance and improve pneumonia management.

Current and emerging artificial intelligence applications for pediatric interventional radiology.

Artificial intelligence in medicine can help improve the accuracy and efficiency of diagnostics, selection of therapies and prediction of outcomes. Machine learning describes a subset of artificial intelligence that utilizes algorithms that can learn modeling functions from datasets. More complex algorithms, or deep learning, can similarly learn modeling functions for a variety of tasks leveraging massive complex datasets. The aggregation of artificial intelligence tools has the potential to improve many facets of health care delivery, from mundane tasks such as scheduling appointments to more complex functions such as enterprise management modeling and in-suite procedural assistance. Within radiology, the roles and use cases for artificial intelligence (inclusive of machine learning and deep learning) continue to evolve. Significant resources have been devoted to diagnostic radiology tasks via national radiology societies, academic medical centers and hundreds of commercial entities. Despite the widespread interest in artificial intelligence radiology solutions, there remains a lack of applications and discussion for use cases in interventional radiology (IR). Even more relevant to this audience, specific technologies tailored to the pediatric IR space are lacking. In this review, we describe artificial intelligence technologies that have been developed within the IR suite, as well as some future work, with a focus on artificial intelligence's potential impact in pediatric interventional medicine.

Outcomes of acute respiratory distress syndrome in COVID-19 patients compared to the general population: a systematic review and meta-analysis.

INTRODUCTION: Acute respiratory distress syndrome (ARDS) due to coronavirus disease 2019 (COVID-19) often leads to mortality. Outcomes of patients with COVID-19-related ARDS compared to ARDS unrelated to COVID-19 is not well characterized. AREAS COVERED: We performed a systematic review of PubMed, Scopus, and MedRxiv 11/1/2019 to 3/1/2021, including studies comparing outcomes in COVID-19-related ARDS (COVID-19 group) and ARDS unrelated to COVID-19 (ARDS group). Outcomes investigated were duration of mechanical ventilation-free days, intensive care unit (ICU) length-of-stay (LOS), hospital LOS, and mortality. Random effects models were fit for each outcome measure. Effect sizes were reported as pooled median differences of medians (MDMs), mean differences (MDs), or odds ratios (ORs). EXPERT OPINION: Ten studies with 2,281 patients met inclusion criteria (COVID-19: 861 [37.7%], ARDS: 1420 [62.3%]). There were no significant differences between the COVID-19 and ARDS groups for median number of mechanical ventilator-free days (MDM: -7.0 [95% CI: -14.8; 0.7], p = 0.075), ICU LOS (MD: 3.1 [95% CI: -5.9; 12.1], p = 0.501), hospital LOS (MD: 2.5 [95% CI: -5.6; 10.7], p = 0.542), or all-cause mortality (OR: 1.25 [95% CI: 0.78; 1.99], p = 0.361). Compared to the general ARDS population, results did not suggest worse outcomes in COVID-19-related ARDS.

Multi-task weak supervision enables anatomically-resolved abnormality detection in whole-body FDG-PET/CT.

Computational decision support systems could provide clinical value in whole-body FDG-PET/CT workflows. However, limited availability of labeled data combined with the large size of PET/CT imaging exams make it challenging to apply existing supervised machine learning systems. Leveraging recent advancements in natural language processing, we describe a weak supervision framework that extracts imperfect, yet highly granular, regional abnormality labels from free-text radiology reports. Our framework automatically labels each region in a custom ontology of anatomical regions, providing a structured profile of the pathologies in each imaging exam. Using these generated labels, we then train an attention-based, multi-task CNN architecture to detect and estimate the location of abnormalities in whole-body scans. We demonstrate empirically that our multi-task representation is critical for strong performance on rare abnormalities with limited training data. The representation also contributes to more accurate mortality prediction from imaging data, suggesting the potential utility of our framework beyond abnormality detection and location estimation.

Multimodal fusion with deep neural networks for leveraging CT imaging and electronic health record: a case-study in pulmonary embolism detection.

Recent advancements in deep learning have led to a resurgence of medical imaging and Electronic Medical Record (EMR) models for a variety of applications, including clinical decision support, automated workflow triage, clinical prediction and more. However, very few models have been developed to integrate both clinical and imaging data, despite that in routine practice clinicians rely on EMR to provide context in medical imaging interpretation. In this study, we developed and compared different multimodal fusion model architectures that are capable of utilizing both pixel data from volumetric Computed Tomography Pulmonary Angiography scans and clinical patient data from the EMR to automatically classify Pulmonary Embolism (PE) cases. The best performing multimodality model is a late fusion model that achieves an AUROC of 0.947 [95% CI: 0.946-0.948] on the entire held-out test set, outperforming imaging-only and EMR-only single modality models.

Deep learning and its role in COVID-19 medical imaging.

COVID-19 is one of the greatest global public health challenges in history. COVID-19 is caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and is estimated to have an cumulative global case-fatality rate as high as 7.2% (Onder et al., 2020) [1]. As the SARS-CoV-2 spread across the globe it catalyzed new urgency in building systems to allow rapid sharing and dissemination of data between international healthcare infrastructures and governments in a worldwide effort focused on case tracking/tracing, identifying effective therapeutic protocols, securing healthcare resources, and in drug and vaccine research. In addition to the worldwide efforts to share clinical and routine population health data, there are many large-scale efforts to collect and disseminate medical imaging data, owing to the critical role that imaging has played in diagnosis and management around the world. Given reported false negative rates of the reverse transcriptase polymerase chain reaction (RT-PCR) of up to 61% (Centers for Disease Control and Prevention, Division of Viral Diseases, 2020; Kucirka et al., 2020) [2,3], imaging can be used as an important adjunct or alternative. Furthermore, there has been a shortage of test-kits worldwide and laboratories in many testing sites have struggled to process the available tests within a reasonable time frame. Given these issues surrounding COVID-19, many groups began to explore the benefits of 'big data' processing and algorithms to assist with the diagnosis and therapeutic development of COVID-19.

Fusion of medical imaging and electronic health records using deep learning: a systematic review and implementation guidelines.

Advancements in deep learning techniques carry the potential to make significant contributions to healthcare, particularly in fields that utilize medical imaging for diagnosis, prognosis, and treatment decisions. The current state-of-the-art deep learning models for radiology applications consider only pixel-value information without data informing clinical context. Yet in practice, pertinent and accurate non-imaging data based on the clinical history and laboratory data enable physicians to interpret imaging findings in the appropriate clinical context, leading to a higher diagnostic accuracy, informative clinical decision making, and improved patient outcomes. To achieve a similar goal using deep learning, medical imaging pixel-based models must also achieve the capability to process contextual data from electronic health records (EHR) in addition to pixel data. In this paper, we describe different data fusion techniques that can be applied to combine medical imaging with EHR, and systematically review medical data fusion literature published between 2012 and 2020. We conducted a systematic search on PubMed and Scopus for original research articles leveraging deep learning for fusion of multimodality data. In total, we screened 985 studies and extracted data from 17 papers. By means of this systematic review, we present current knowledge, summarize important results and provide implementation guidelines to serve as a reference for researchers interested in the application of multimodal fusion in medical imaging.

Magnetic Resonance Imaging of Tumor-Associated Macrophages: Clinical Translation.

Purpose: Tumor-associated macrophages (TAMs) in malignant tumors have been linked to tumor aggressiveness and represent a new target for cancer immunotherapy. As new TAM-targeted immunotherapies are entering clinical trials, it is important to detect and quantify TAM with noninvasive imaging techniques. The purpose of this study was to determine if ferumoxytol-enhanced MRI can detect TAM in lymphomas and bone sarcomas of pediatric patients and young adults.Experimental Design: In a first-in-patient, Institutional Review Board-approved prospective clinical trial, 25 pediatric and young adult patients with lymphoma or bone sarcoma underwent ferumoxytol-enhanced MRI. To confirm ferumoxytol enhancement, five pilot patients (two lymphoma and three bone sarcoma) underwent pre- and postcontrast MRI. Subsequently, 20 patients (10 lymphoma and 10 bone sarcoma) underwent ferumoxytol-enhanced MRI 24 to 48 hours after i.v. injection, followed by tumor biopsy/resection and macrophage staining. To determine if ferumoxytol-MRI can differentiate tumors with different TAM content, we compared T2* relaxation times of lymphomas and bone sarcomas. Tumor T2* values of 20 patients were correlated with CD68+ and CD163+ TAM quantities on histopathology.Results: Significant ferumoxytol tumor enhancement was noted on postcontrast scans compared with precontrast scans (P = 0.036). Bone sarcomas and lymphomas demonstrated significantly different MRI enhancement and TAM density (P < 0.05). Within each tumor group, T2* signal enhancement on MR images correlated significantly with the density of CD68+ and CD163+ TAM (P < 0.05).Conclusions: Ferumoxytol-enhanced MRI is immediately clinically applicable and could be used to stratify patients with TAM-rich tumors to immune-targeted therapies and to monitor tumor response to these therapies. Clin Cancer Res; 24(17); 4110-8. ©2018 AACR.

Whole-body PET/MRI of Pediatric Patients: The Details That Matter.

Integrated PET/MRI is a hybrid imaging technique enabling clinicians to acquire diagnostic images for tumor assessment and treatment monitoring with both high soft tissue contrast and added metabolic information. Integrated PET/MRI has shown to be valuable in the clinical setting and has many promising future applications. The protocol presented here will provide step-by-step instructions for the acquisition of whole-body 2-deoxy-2-(18F)fluoro-D-glucose (18F-FDG) PET/MRI data in children with cancer. It also provides instructions on how to combine a whole-body staging scan with a local tumor scan for evaluation of the primary tumor. The focus of this protocol is to be both comprehensive and time-efficient, which are two ubiquitous needs for clinical applications. This protocol was originally developed for children above 6 years, or old enough to comply with breath-hold instructions, but can also be applied to patients under general anesthesia. Similarly, this protocol can be modified to fit institutional preferences in terms of choice of MRI pulse sequences for both the whole-body scan and local tumor assessment.

Calyceal diverticulum: a benign imitator of serious pathology.

A 72-year-old man with lung cancer underwent positron emission tomography CT (PET-CT) as a part of cancer staging. As an incidental finding, the PET-CT revealed a renal mass with metabolic and morphological characteristics of a malignant tumour. A diagnostic CT scan revealed a Bosniak III renal cyst, and malignancy could not be excluded. For correct Bosniak classification, a multiphasic contrast-enhanced CT was performed and the renal mass was finally diagnosed as a calyceal diverticulum. This case report summarises how calyceal diverticula may mimic serious pathology, leading to diagnostic difficulties.