A5037419111- Topic Modeling
- Radiomics and Machine Learning in Medical Imaging
- Natural Language Processing Techniques
- Machine Learning in Healthcare
- Multimodal Machine Learning Applications
- COVID-19 diagnosis using AI
- Medical Imaging Techniques and Applications
- Anesthesia and Sedative Agents
- Advanced X-ray and CT Imaging
- Medical Image Segmentation Techniques
- COVID-19 Clinical Research Studies
- Anesthesia and Pain Management
- Ophthalmology and Eye Disorders
- Intraoperative Neuromonitoring and Anesthetic Effects
- Geological Modeling and Analysis
- Maternal and Neonatal Healthcare
- Impact of AI and Big Data on Business and Society
- Case Reports on Hematomas
- Electronic Health Records Systems
- Blood donation and transfusion practices
- Advanced Neural Network Applications
- Cerebrospinal fluid and hydrocephalus
- Radiology practices and education
- Muscle and Compartmental Disorders
- Biomedical Text Mining and Ontologies
Stanford University
2023-2024
Hospital Israelita Albert Einstein
2020-2024
Association for the Advancement of Artificial Intelligence
2023-2024
Hospital São Paulo
2024
Universidade Estadual de Campinas (UNICAMP)
2003-2020
In-Q-Tel
2003
Artificial intelligence (AI) models for automatic generation of narrative radiology reports from images have the potential to enhance efficiency and reduce workload radiologists. However, evaluating correctness these requires metrics that can capture clinically pertinent differences. In this study, we investigate alignment between automated radiologists' scoring errors in report generation. We address limitations existing by proposing new metrics, RadGraph F1 RadCliQ, which demonstrate...
Abstract 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 processing (NLP) tasks, efficacy 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 four distinct tasks: radiology...
The ability of large language models (LLMs) to follow natural instructions with human-level fluency suggests many opportunities in healthcare reduce administrative burden and improve quality care. However, evaluating LLMs on realistic text generation tasks for remains challenging. Existing question answering datasets electronic health record (EHR) data fail capture the complexity information needs documentation burdens experienced by clinicians. To address these challenges, we introduce...
Over 85 million computed tomography (CT) scans are performed annually in the US, of which approximately one quarter focus on abdomen. Given current shortage both general and specialized radiologists, there is a large impetus to use artificial intelligence alleviate burden interpreting these complex imaging studies while simultaneously using images extract novel physiological insights. Prior state-of-the-art approaches for automated medical image interpretation leverage vision language models...
Objective This study describes epidemiological and clinical features of patients with confirmed infection by SARS-CoV-2 diagnosed treated at Hospital Israelita Albert Einstein , which admitted the first this condition in Brazil. Methods In retrospective, single-center study, we included all laboratory COVID-19 cases São Paulo, Brazil, from February until March 2020. Demographic, clinical, radiological data were analyzed. Results A total 510 a diagnosis [...]
Medicine, by its nature, is a multifaceted domain that requires the synthesis of information across various modalities. Medical generative vision-language models (VLMs) make first step in this direction and promise many exciting clinical applications. However, existing typically have to be fine-tuned on sizeable down-stream datasets, which poses significant limitation as medical applications data scarce, necessitating are capable learning from few examples real-time. Here we propose...
The Coronavirus disease 2019 (COVID-19) presents open questions in how we clinically diagnose and assess course. Recently, chest computed tomography (CT) has shown utility for COVID-19 diagnosis. In this study, developed Deep COVID DeteCT (DCD), a deep learning convolutional neural network (CNN) that uses the entire CT volume to automatically predict (COVID+) from non-COVID-19 (COVID-) pneumonia normal controls. We discuss training strategies differences performance across 13 international...
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 processing (NLP), effectiveness diverse range of clinical summarization tasks remains unproven. In this study, we apply adaptation methods to eight LLMs, spanning four distinct tasks: radiology reports, patient questions, progress notes, doctor-patient dialogue....
Chest X-rays (CXRs) are the most frequently performed imaging test in clinical practice. Recent advances development of vision-language foundation models (FMs) give rise to possibility performing automated CXR interpretation, which can assist physicians with decision-making and improve patient outcomes. However, developing FMs that accurately interpret CXRs is challenging due (1) limited availability large-scale datasets medical image domain, (2) lack vision language encoders capture...
Chest radiographs allow for the meticulous examination of a patient's chest but demands specialized training proper interpretation. Automated analysis medical imaging has become increasingly accessible with advent machine learning (ML) algorithms. Large labeled datasets are key elements and validation these ML solutions. In this paper we describe Brazilian x-ray dataset, BRAX: an automatically dataset designed to assist researchers in models. The contains 24,959 radiography studies from...
Background and objective: Subarachnoid block is a widely used technique for Caesarean section. Its quality can be improved by adding opioids to the local anaesthetics. We studied of its maternal-fetal repercussions when different doses sufentanil were combined with hyperbaric bupivacaine using intrathecal route in pregnant women undergoing Methods: This was prospective, randomized, double-blind, controlled trial 80 women, ASA I-II, who scheduled elective section under subarachnoid block....
SummaryBackground and objective: Subarachnoid block is a widely used technique for Caesarean section. Its quality can be improved by adding opioids to the local anaesthetics. We studied of its maternal–fetal repercussions when different doses sufentanil were combined with hyperbaric bupivacaine using intrathecal route in pregnant women undergoing section.Methods: This was prospective, randomized, double-blind, controlled trial 80 women, ASA I–II, who scheduled elective section under...
Diagnosing and managing a patient is complex, sequential decision making process that requires physicians to obtain information -- such as which tests perform act upon it. Recent advances in artificial intelligence (AI) large language models (LLMs) promise profoundly impact clinical care. However, current evaluation schemes overrely on static medical question-answering benchmarks, falling short interactive decision-making required real-life work. Here, we present AgentClinic: multimodal...
Abstract The application of AI to medical image interpretation tasks has largely been limited the identification a handful individual pathologies. In contrast, generation complete narrative radiology reports more closely matches how radiologists communicate diagnostic information in clinical workflows. Recent progress artificial intelligence (AI) on vision-language enabled possibility generating high-quality from images. Automated metrics evaluate quality generated attempt capture overlap...
Sarcopenia is commonly assessed on CT by use of the skeletal muscle index (SMI), which calculated as area (SMA) at L3 divided patient height squared (i.e., a scaling power 2).
Benjamin Yan, Ruochen Liu, David Kuo, Subathra Adithan, Eduardo Reis, Stephen Kwak, Vasantha Venugopal, Chloe O’Connell, Agustina Saenz, Pranav Rajpurkar, Michael Moor. Findings of the Association for Computational Linguistics: EMNLP 2023.
Brief hospital course (BHC) summaries are clinical documents that summarize a patient's stay. While large language models (LLMs) depict remarkable capabilities in automating real-world tasks, their for healthcare applications such as synthesizing BHCs from notes have not been shown. We introduce novel pre-processed dataset, the MIMIC-IV-BHC, encapsulating note and brief pairs to adapt LLMs BHC synthesis. Furthermore, we benchmark of summarization performance two general-purpose three...
The current gold standard for evaluating generated chest x-ray (CXR) reports is through radiologist annotations. However, this process can be extremely time-consuming and costly, especially when large numbers of reports. In work, we present FineRadScore, a Large Language Model (LLM)-based automated evaluation metric CXR Given candidate report ground-truth report, FineRadScore gives the minimum number line-by-line corrections required to go from report. Additionally, provides an error...
Recent research in deep learning (DL) has investigated the use of Fast Fourier Transform (FFT) to accelerate computations involved Convolutional Neural Networks (CNNs) by replacing spatial convolution with element-wise multiplications on spectral domain. These approaches mainly rely FFT reduce number operations, which can be further decreased adopting Real-Valued FFT. In this paper, we propose using phasor form, a polar representation complex numbers, as more efficient alternative...
Over 85 million computed tomography (CT) scans are performed annually in the US, of which approximately one quarter focus on abdomen. Given current radiologist shortage, there is a large impetus to use artificial intelligence alleviate burden interpreting these complex imaging studies. Prior state-of-the-art approaches for automated medical image interpretation leverage vision language models (VLMs). However, VLMs generally limited 2D images and short reports, do not electronic health record...
Radiologists play a crucial role by translating medical images into reports. However, the field faces staffing shortages and increasing workloads. While automated approaches using vision-language models (VLMs) show promise as assistants, they require exceptionally high accuracy. Most current VLMs in radiology rely solely on supervised fine-tuning (SFT). Meanwhile, general domain, additional preference has become standard practice. The challenge lies prohibitive cost of obtaining radiologist...