Background Multimodal generative artificial intelligence (AI) technologies can produce preliminary radiology reports, and validation with reader studies is crucial for understanding the clinical value of these technologies. Purpose To assess the clinical value of the use of a domain-specific multimodal generative AI tool for chest radiograph interpretation by means of a reader study. Materials and Methods A retrospective, sequential, multireader, multicase reader study was conducted using 758 chest radiographs from a publicly available dataset from 2009 to 2017. Five radiologists interpreted the chest radiographs in two sessions: without AI-generated reports and with AI-generated reports as preliminary reports. Reading times, reporting agreement (RADPEER), and quality scores (five-point scale) were evaluated by two experienced thoracic radiologists and compared between the first and second sessions from October to December 2023. Reading times, report agreement, and quality scores were analyzed using a generalized linear mixed model. Additionally, a subset of 258 chest radiographs was used to assess the factual correctness of the reports, and sensitivities and specificities were compared between the reports from the first and second sessions with use of the McNemar test. Results The introduction of AI-generated reports significantly reduced average reading times from 34.2 seconds ± 20.4 to 19.8 seconds ± 12.5 ( < .001). Report agreement scores shifted from a median of 5.0 (IQR, 4.0-5.0) without AI reports to 5.0 (IQR, 4.5-5.0) with AI reports ( < .001). Report quality scores changed from 4.5 (IQR, 4.0-5.0) without AI reports to 4.5 (IQR, 4.5-5.0) with AI reports ( < .001). From the subset analysis of factual correctness, the sensitivity for detecting various abnormalities increased significantly, including widened mediastinal silhouettes (84.3% to 90.8%; < .001) and pleural lesions (77.7% to 87.4%; < .001). While the overall diagnostic performance improved, variability among individual radiologists was noted. Conclusion The use of a domain-specific multimodal generative AI model increased the efficiency and quality of radiology report generation. © RSNA, 2025 See also the editorial by Babyn and Adams in this issue.
Download full-text PDF |
Source |
|---|---|
| http://dx.doi.org/10.1148/radiol.241646 | DOI Listing |
Publication Analysis
Top Keywords
Similar Publications
Comparative Assessment of Real-Time and Offline Short-Lag Spatial Coherence Imaging of Ultrasound Breast Masses.
Ultrasound Med Biol
March 2025
Department of Computer Science, Johns Hopkins University, Baltimore, MD, USA; Department of Electrical & Computer Engineering, Johns Hopkins University, Baltimore, MD, USA; Department of Biomedical Engineering, Johns Hopkins University, Baltimore, MD, USA. Electronic address:
Objective: To perform the first known investigation of differences between real-time and offline B-mode and short-lag spatial coherence (SLSC) images when evaluating fluid or solid content in 60 hypoechoic breast masses.
Methods: Real-time and retrospective (i.e.
Morphological advances and innovations in conjugated polymer films for high-performance gas sensors.
Talanta
March 2025
NTT Hi-Tech Institute, Nguyen Tat Thanh University, Ho Chi Minh City, Viet Nam. Electronic address:
Conjugated polymers (CPs) are considered one of the most important gas-sensing materials due to their unique features, combining the benefits of both metals and semiconductors, along with their outstanding mechanical properties and excellent processability. However, CPs with conventional morphological structures, such as largely amorphous and bulky matrices, face limitations in practical applications because of their inferior charge transport characteristics, low surface area, and insufficient sensitivity. Therefore, the design and development of novel morphological nanostructures in CPs have attracted significant attention as a promising strategy for improving morphological and electrical characteristics, thereby enabling a considerable increase in the sensing performance of corresponding gas sensors.
View Article and Find Full Text PDF6-methyladnosine of vRNA facilitates influenza A virus replication by promoting the interaction of vRNA with polymerase proteins.
Proc Natl Acad Sci U S A
March 2025
State Key Laboratory for Animal Disease Control and Prevention, College of Veterinary Medicine, Lanzhou University, Lanzhou Veterinary Research Institute, Chinese Academy of Agricultural Sciences, Lanzhou 730000, People's Republic of China.
6-methyladnosine (mA) modification is present in both positive- and negative-strand RNA of influenza A virus (IAV) and affects the replication and pathogenicity of IAV. However, little is known about the regulatory mechanism of mA in IAV RNA. In the present study, we identified the mA methylation of the viral RNA of different IAV subtypes and confirmed that mA modification promotes the polymerase activity and replication of IAV.
View Article and Find Full Text PDFModification of RNF183 via m6A Methylation Mediates Podocyte Dysfunction in Diabetic Nephropathy by Regulating PKM2 Ubiquitination and Degradation.
Cells
March 2025
Department of Pathology, Hebei Medical University, Shijiazhuang 050017, China.
Diabetic kidney disease (DKD) is a prevalent complication associated with diabetes in which podocyte dysfunction significantly contributes to the development and progression of the condition. Ring finger protein 183 (RNF183) is an ER-localized, transmembrane ring finger protein with classical E3 ligase activity. However, whether RNF183 is involved in glomerular podocyte dysfunction, which is the mechanism of action of DKD, is still poorly understood.
View Article and Find Full Text PDFDeep Learning-Based Contrast Boosting in Low-Contrast Media Pre-TAVR CT Imaging.
Can Assoc Radiol J
March 2025
Department of Radiology, Seoul St. Mary's Hospital, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea.
This study investigates the impact of deep learning-based contrast boosting (DL-CB) on image quality and measurement reliability in low-contrast media (low-CM) CT for pre-transcatheter aortic valve replacement (TAVR) assessment. This retrospective study included TAVR candidates with renal dysfunction who underwent low-CM (30-mL: 15-mL bolus of contrast followed by 50-mL of 30% iomeprol solution) pre-TAVR CT between April and December 2023, along with matched standard-CM controls (n = 68). Low-CM images were reconstructed as conventional, 50-keV, and DL-CB images.
View Article and Find Full Text PDFWant AI Summaries of new PubMed Abstracts delivered to your In-box?
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!