Search

BackgroundRadiological imaging guidelines are crucial for accurate diagnosis and optimal patient care as they result in standardized procedures and thus reduce inappropriate imaging studies. In the present study, we investigated the potential to support clinical decision-making using an interactive chatbot designed to provide personalized imaging recommendations based on indexed and vectorized American College of Radiology (ACR) appropriateness criteria documents.MethodsWe utilized 209 ACR appropriateness criteria documents as specialized knowledge base and employed LlamaIndex and the ChatGPT 3.5-Turbo to create an appropriateness criteria contexted chatbot (accGPT). Fifty clinical case files were used to compare the accGPT’s performance against radiologists at varying experience levels and to generic ChatGPT 3.5 and 4.0.ResultsAll chatbots reached at least human performance level. For the 50 case files, the accGPT provided a median of 83% (95% CI 82-84) ‘usually appropriate’ recommendations, while radiologists provided a median of 66% (95% CI 62-70). GPT 3.5-Turbo 70% (95% CI 67-73) and GPT 4 79% (95% CI 76-81) correct answers. Consistency was highest for the accGPT with almost perfect Fleiss’ Kappa of 0.82. Further, the chatbots provided substantial time and cost savings, with an average decision time of 5 minutes and a cost of 0.19 Euro for all cases, compared to 50 minutes and 29.99 Euro for radiologists (both p < 0.01).ConclusionChatGPT-based algorithms have the potential to substantially improve the decision-making for clinical imaging studies in accordance with ACR guidelines. Specifically, a context-based algorithm performed superior to its generic counterpart, demonstrating the value of tailoring AI solutions to specific healthcare applications.

Background Whole-body magnetic resonance imaging (MRI) permits non-invasive, non-ionizing phenotyping of the human body and ideally complements the epidemiological assessment of the NAKO participants. As such, it allows for the detection of morphologic or functional predisposition, early disease stages prior to overt clinical events as well as evident pathological changes. The assessment of progression and regression of imaging phenotypes over time will provide the basis to identify and understand the relevance of imaging-based risk factor profiles for disease development. Methods Integrated in the general NAKO study program and managed by a central Imaging Core, study participants underwent whole-body imaging at five dedicated MR imaging sites. Imaging was performed on five identical 3 Tesla scanners (Magnetom Skyra, Siemens Healthineers, Erlangen, Germany) applying a one hour protocol, including sequences for the brain, the cardiovascular and musculoskeletal system as well as for the thorax and abdomen. Comprehensive measures to assure high image quality and management of incidental findings were established. Results As part of the baseline examination program, a total of 30,861 participants successfully underwent the MR imaging program. All measures of quality assurance and incidental findings management were successfully employed throughout the study period and obtained image quality and completeness of all MR sequences was excellent (>94.2% completeness). While MR imaging as part of the first re-examination is ongoing, baseline MRI data is currently accessible for scientific analyses. Conclusions The MRI-Study of the NAKO will provide a comprehensive imaging phenotypic biobank covering different organ systems with highest morphological and functional detail. MRI data analysis will gain novel insights into the natural history of disease development, the role of subclinical disease burden, and revolutionize our understanding of imaging biomarkers of risk.

Contrast-enhanced computed tomography (CT) is a convenient method to visualize left atrial appendage (LAA) thrombi. We determined whether diagnostic accuracy improves by including dual-energy as compared to transesophageal echocardiography (TEE). Furthermore, the influence of protocol parameters on radiation dose were quantified.Patients were assigned to the different CT protocols. All CTs were assessed qualitatively for presence of LAA thrombi and dual-energy CT scans quantitatively for iodine concentration. TEE was assessed qualitatively for the presence of thrombi.Of 32 enrolled patients, 6 had a thrombus in TEE. Qualitative CT assessment yielded 83% sensitivity and 88% specificity. In the 26 patients who underwent dual-energy CT, median iodine concentration was 8.6 mg/cmCT provides good diagnostic accuracy for detecting LAA thrombi, which can further be improved if iodine density measurements by dual-energy are incorporated. With an optimized protocol, reasonably low radiation dose can be achieved.

This paper introduces the application of a coextrusion printing as a contact and mask-free fine line metallization technology for direct printing of narrow grid lines with a strongly enhanced aspect ratio. Coextrusion printed grid lines enhance short circuit current and lower series resistance through a higher number of filigree fingers. The silver and sacrificial paste that merge inside the nozzle are simultaneously extruded onto the wafer and form a silver paste structure smaller than every design feature inside the nozzle. After contact firing the grid lines become as narrow as 35 µm with an aspect ratio up to 0.7. A further shrinkage potential down to 23.7 µm width with an aspect ratio up to 0.62 has been demonstrated. Coextrusion printed grid lines outperform screen printed grid lines in a whole slew of geometrical and optical properties. As for example grid line width and height variation is as low as ±0.5 µm and the optical effective finger width is far smaller than the screen printed effective finger width. Due to these low effective finger widths the absolute effective shading of 96 coextrusion printed grid lines is less than half of the 80 screen printed grid lines. The calculated short circuit current gain, due to the lower effective shading was experimentally confirmed accordingly. Furthermore a benchmark test in a production like environment was performed. Hereby an absolute cell efficiency gain of +0.25% over screen printed metallization was confirmed and a cell efficiency gain of over +0.35% to +0.4% abs. can be assumed to be achieved by using a direct contacting coextrusion silver paste, which is comparable to state of the art screen printing pastes.

This paper presents a theoretical analysis of an optical nanocircuit connected with broadband nanoantenna composed of electric and magnetic nanodipoles with feeding of the electric dipole only. The numerical analysis of the nanoantenna is performed by method of moments (MoM). The calculated results are input impedance, current distribution, reflection coefficient, radiation efficiency, bandwidth, voltage reflection coefficient and electric near field. Comsol Multiphysics software calculates some of these results. The results show that the loop inserted into the circuit increases the bandwidth of nanoantenna and decreases the overall voltage reflection coefficient of the optical nanocircuit.