Publications by authors named "Geoffrey D Rubin"

Article Synopsis
  • Many machine learning and AI applications are being developed for cardiovascular imaging, but their implementation processes differ based on factors like imaging type, patient characteristics, and clinical uses.
  • The article proposes a framework to evaluate AI's value from an organizational standpoint, using value chain analysis to pinpoint areas where AI can enhance value creation in healthcare.
  • It emphasizes the need to consider various stakeholders' perspectives—such as clinicians, hospitals, and patients—to effectively integrate AI tools in real-world applications, showcasing the potential benefits of AI throughout the entire patient experience in cardiovascular care.
View Article and Find Full Text PDF
Article Synopsis
  • - The Coronary Artery Disease Reporting and Data System (CAD-RADS) was updated in 2022 to standardize reporting for coronary CT angiography (CCTA) and improve patient management based on new developments and clinical guidelines.
  • - The updated CAD-RADS 2.0 uses a classification system that assesses stenosis severity, plaque burden, and includes lesion-specific ischemia evaluations, with a scale ranging from CAD-RADS 0 (no disease) to CAD-RADS 5 (total blockage).
  • - The primary aim of CAD-RADS is to enhance communication between healthcare providers regarding test results and management suggestions, while also supporting education, research, and quality assurance in patient care.
View Article and Find Full Text PDF
Article Synopsis
  • The Coronary Artery Disease Reporting and Data System (CAD-RADS) was updated to version 2.0 in 2022 to improve the standardization of reporting for coronary CT angiography (CCTA) and guide patient management based on new technologies and clinical evidence.
  • * The updated system classifies coronary disease severity based on criteria such as stenosis, plaque burden, and additional factors like CT fractional-flow-reserve (CT-FFR) or myocardial CT perfusion (CTP), with a scale from CAD-RADS 0 (no disease) to CAD-RADS 5 (total occlusion).
  • * The CAD-RADS framework enhances communication between healthcare providers regarding test results, supports education and research, and aims to
View Article and Find Full Text PDF

Coronary Artery Disease Reporting and Data System (CAD-RADS) was created to standardize reporting system for patients undergoing coronary CT angiography (CCTA) and to guide possible next steps in patient management. The goal of this updated 2022 CAD-RADS 2.0 is to improve the initial reporting system for CCTA by considering new technical developments in cardiac CT, including data from recent clinical trials and new clinical guidelines.

View Article and Find Full Text PDF
Article Synopsis
  • The Coronary Artery Disease Reporting and Data System (CAD-RADS) was updated to version 2.0 in 2022 to standardize the reporting of coronary CT angiography (CCTA) results and improve patient management.
  • The updated system incorporates new technical advances, clinical trials, and guidelines, including assessments of plaque burden and ischemia using advanced measures like CT fractional-flow-reserve (CT-FFR).
  • CAD-RADS aims to enhance communication between healthcare providers by offering a clear classification of stenosis severity and plaque burden, ultimately focusing on better patient care and supporting education, research, and clinical practices.
View Article and Find Full Text PDF
Article Synopsis
  • The study aims to enhance AI systems for detecting various abnormalities in CT scans by creating efficient, automated multi-label annotators to reduce reliance on manual annotation.
  • The researchers developed rule-based algorithms to extract disease information from radiology reports for three organ systems and used attention-guided RNNs to improve classification accuracy.
  • Results showed high accuracy in the manual validation of the algorithms, and automated models successfully analyzed over 261,000 reports, demonstrating the potential for improved disease detection with AI.
View Article and Find Full Text PDF
Article Synopsis
  • The study aimed to create classifiers for identifying multiple diseases in body CT scans, using labels automatically extracted from radiology reports across three organ systems: lungs, liver, and kidneys.
  • It involved analyzing over 12,000 patient CT scans from 2012 to 2017 and utilized a 3D DenseVNet model to segment organs and classify disease presence or absence.
  • Results showed high accuracy in the label extraction and AUC values for the classifiers, indicating effectiveness in diagnosing various conditions like emphysema and kidney stones across the different organ systems.
View Article and Find Full Text PDF
Article Synopsis
  • Coronary CT angiography (CCTA) and contrast-enhanced thoracic CT (CECT) are different cancer screening procedures that receive the same reimbursement rate from Medicare, but this study aims to determine if their actual costs differ significantly.
  • Using a time-driven activity-based costing model, researchers analyzed the direct costs associated with both exams across various medical facilities, finding that CCTA takes significantly longer and costs more than CECT.
  • The study concluded that the reimbursement rates for CCTA do not accurately reflect the higher resources and costs required for the procedure, potentially impacting patient access.
View Article and Find Full Text PDF
Article Synopsis
  • - SARS-CoV-2 infection can range from no symptoms to severe disease, with serious cases often resulting from acute lung injury; imaging techniques like CT became more crucial early in the pandemic for diagnosing COVID-19 when RT-PCR testing was less reliable.
  • - Various classification systems for chest imaging were created to help identify potential COVID-19 cases during times of limited testing availability, and studies indicate a high agreement among observers in categorizing these images.
  • - Beyond respiratory issues, COVID-19 is linked to cardiovascular problems like thromboembolism, and research suggests that artificial intelligence could be useful in diagnosing and predicting outcomes for COVID-19 pneumonia using imaging techniques like radiography and CT.
View Article and Find Full Text PDF
Article Synopsis
  • The use of molecular targeting agents and immune checkpoint inhibitors in cancer treatment has led to an increase in lung toxicity, particularly drug-related pneumonitis (DRP).
  • Diagnosing DRP involves ruling out other causes, with symptoms varying from mild to potentially fatal, and the assessment of lung imaging can reveal different radiologic patterns indicative of DRP severity.
  • Treatment typically includes stopping the offending drug, administering immunosuppressive therapy, and providing supportive care, with recommendations for diagnosis and management aimed at various healthcare professionals.
View Article and Find Full Text PDF
Article Synopsis
  • * The framework utilized various image processing techniques and was tested on 132 clinical cases, divided into high and low quality based on expert evaluations, achieving a remarkable accuracy of 97.7%.
  • * The results indicate that the framework can effectively predict image quality and support the goals of the Quantitative Imaging Biomarker Alliance (QIBA) initiative, validating its application across diverse clinical settings.
View Article and Find Full Text PDF
Article Synopsis
  • The use of molecular targeting agents and immune checkpoint inhibitors (ICIs) in cancer treatment has led to increased lung toxicity, specifically drug-related pneumonitis (DRP), highlighting the importance of recognizing its incidence and risk factors.
  • Diagnosis of DRP typically involves ruling out other causes, and symptoms can vary significantly, from mild to potentially life-threatening.
  • Imaging features of DRP can include various radiological patterns, and treatment usually involves stopping the offending drug, administering immunosuppressive therapy, and providing supportive care, including oxygen if needed.
View Article and Find Full Text PDF
Article Synopsis
  • Pulmonary hypertension (PH) is characterized by elevated pressure in the pulmonary artery (over 20 mmHg) and is categorized into five groups based on similar causes and treatment approaches.
  • Radiologists are crucial in the assessment and management of PH, working alongside other specialists to provide comprehensive care.
  • A working group from the Fleischner Society focused on imaging techniques like CT, MRI, and nuclear medicine to determine their effectiveness in diagnosing PH, understanding its causes, assessing severity, and planning treatment.
View Article and Find Full Text PDF
Article Synopsis
  • Pulmonary hypertension (PH) is characterized by a mean pulmonary artery pressure exceeding 20 mm Hg and is divided into five groups based on similar mechanisms and treatment approaches.
  • A specialized working group within the Fleischner Society is investigating the role of imaging techniques like CT, MRI, and nuclear medicine in diagnosing and managing PH, focusing on questions about noninvasive imaging, identifying causes, assessing severity, and planning treatment.
  • This systematic review emphasizes the critical function of imaging in identifying, evaluating, and following up on patients with PH, with the same content published jointly in two different journals.
View Article and Find Full Text PDF
Article Synopsis
  • A comprehensive chest CT data set of 36,316 volumes from nearly 20,000 patients was created, making it the largest multiply-annotated medical imaging data set to date.
  • A rule-based method with a high accuracy (F-score of 0.976) was developed to automatically label abnormalities from free-text radiology reports.
  • A deep convolutional neural network (CNN) model achieved strong classification performance, with an AUROC above 0.90 for 18 abnormalities, and demonstrated that more training labels significantly improved overall model performance.
View Article and Find Full Text PDF

Background: A number of organizations, including the US Preventive Services Task Force (USPSTF), recommend lung cancer screening (LCS) with low-dose CT (LDCT) imaging for high-risk current and former smokers. In 2015, Medicare issued a decision to cover LCS as a preventive health benefit; however, utilization by the Medicare population has not been thoroughly examined.

Research Question: Our objective was to evaluate the early use of LCS in the Medicare fee-for-service (FFS) population and determine the relationship(s) among beneficiary sociodemographic characteristics, geographic location, and use.

View Article and Find Full Text PDF

With more than 900,000 confirmed cases worldwide and nearly 50,000 deaths during the first 3 months of 2020, the coronavirus disease 2019 (COVID-19) pandemic has emerged as an unprecedented health care crisis. The spread of COVID-19 has been heterogeneous, resulting in some regions having sporadic transmission and relatively few hospitalized patients with COVID-19 and others having community transmission that has led to overwhelming numbers of severe cases. For these regions, health care delivery has been disrupted and compromised by critical resource constraints in diagnostic testing, hospital beds, ventilators, and health care workers who have fallen ill to the virus exacerbated by shortages of personal protective equipment.

View Article and Find Full Text PDF

With more than 900 000 confirmed cases worldwide and nearly 50 000 deaths during the first 3 months of 2020, the coronavirus disease 2019 (COVID-19) pandemic has emerged as an unprecedented health care crisis. The spread of COVID-19 has been heterogeneous, resulting in some regions having sporadic transmission and relatively few hospitalized patients with COVID-19 and others having community transmission that has led to overwhelming numbers of severe cases. For these regions, health care delivery has been disrupted and compromised by critical resource constraints in diagnostic testing, hospital beds, ventilators, and health care workers who have fallen ill to the virus exacerbated by shortages of personal protective equipment.

View Article and Find Full Text PDF

We sought to characterize local lung complexity in chest computed tomography (CT) and to characterize its impact on the detectability of pulmonary nodules. Forty volumetric chest CT scans were created by embedding between three and five simulated 5-mm lung nodules into one of three volumetric chest CT datasets. Thirteen radiologists evaluated 157 nodules, resulting in 2041 detection opportunities.

View Article and Find Full Text PDF

Objective: The association between access to CT facilities for lung cancer screening and population characteristics is understudied. We aimed to determine the relationship between census tract-level socioeconomic characteristics (SEC) and driving distance to an ACR-accredited CT facility.

Methods: Census tract-level SEC were determined from the US Census Bureau.

View Article and Find Full Text PDF

The purpose of this study is to (1) develop metrics to characterize the regional anatomical complexity of the lungs, and (2) relate these metrics with lung nodule detection in chest CT. A free-scrolling reader-study with virtually inserted nodules (13 radiologists × 157 total nodules = 2041 responses) is used to characterize human detection performance. Metrics of complexity based on the local density and orientation of distracting vasculature are developed for two-dimensional (2-D) and three-dimensional (3-D) considerations of the image volume.

View Article and Find Full Text PDF

Purpose: Spatial access to health care resources is a requisite for utilization. Our purpose was to determine, at a census tract level, the geographic distribution of US smokers and their driving distance to an ACR-accredited CT facility.

Methods: The number of smokers per US census tract was determined from US Census Bureau data (American Community Survey, 2011-2015) and census tract smoking prevalence estimates.

View Article and Find Full Text PDF