Background Large language models (LLMs) hold substantial promise for medical imaging interpretation. However, there is a lack of studies on their feasibility in handling reasoning questions associated with medical diagnosis. Purpose To investigate the viability of leveraging three publicly available LLMs to enhance consistency and diagnostic accuracy in medical imaging based on standardized reporting, with pathology as the reference standard. Materials and Methods US images of thyroid nodules with pathologic results were retrospectively collected from a tertiary referral hospital between July 2022 and December 2022 and used to evaluate malignancy diagnoses generated by three LLMs-OpenAI's ChatGPT 3.5, ChatGPT 4.0, and Google's Bard. Inter- and intra-LLM agreement of diagnosis were evaluated. Then, diagnostic performance, including accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC), was evaluated and compared for the LLMs and three interactive approaches: human reader combined with LLMs, image-to-text model combined with LLMs, and an end-to-end convolutional neural network model. Results A total of 1161 US images of thyroid nodules (498 benign, 663 malignant) from 725 patients (mean age, 42.2 years ± 14.1 [SD]; 516 women) were evaluated. ChatGPT 4.0 and Bard displayed substantial to almost perfect intra-LLM agreement (κ range, 0.65-0.86 [95% CI: 0.64, 0.86]), while ChatGPT 3.5 showed fair to substantial agreement (κ range, 0.36-0.68 [95% CI: 0.36, 0.68]). ChatGPT 4.0 had an accuracy of 78%-86% (95% CI: 76%, 88%) and sensitivity of 86%-95% (95% CI: 83%, 96%), compared with 74%-86% (95% CI: 71%, 88%) and 74%-91% (95% CI: 71%, 93%), respectively, for Bard. Moreover, with ChatGPT 4.0, the image-to-text-LLM strategy exhibited an AUC (0.83 [95% CI: 0.80, 0.85]) and accuracy (84% [95% CI: 82%, 86%]) comparable to those of the human-LLM interaction strategy with two senior readers and one junior reader and exceeding those of the human-LLM interaction strategy with one junior reader. Conclusion LLMs, particularly integrated with image-to-text approaches, show potential in enhancing diagnostic medical imaging. ChatGPT 4.0 was optimal for consistency and diagnostic accuracy when compared with Bard and ChatGPT 3.5. © RSNA, 2024 Supplemental material is available for this article.

Background. Shear stress is an effective modulator of endothelial progenitor cells (EPCs) and has been suggested to play an important role in angiogenesis. The phosphatase and tensin homolog (PTEN)/Akt and guanosine triphosphate cyclohydrolase (GTPCH)/tetrahydrobiopterin (BH4) pathways regulate the function of early EPCs. However, the role of these pathways in the shear stress-induced angiogenesis of late EPCs remains poorly understood. Therefore, we aim to investigate whether shear stress could upregulate the angiogenesis capacity of late EPCs and to further explore the possible underlying mechanisms. Methods. Late EPCs were subjected to laminar shear stress (LSS), and their in vitro migration, proliferation, and tube formation capacity were determined. In addition, the in vivo angiogenesis capacity was explored, along with the expression of molecules involved in the PTEN/Akt and GTPCH/BH4 pathways. Results. LSS elevated the in vitro activities of late EPCs, which were accompanied by downregulated PTEN expression, accelerated Akt phosphorylation, and GTPCH/BH4 pathway activation (all P<0.05). Following Akt inhibition, LSS-induced upregulated GTPCH expression, BH4, and NO level of EPCs were suppressed. LSS significantly improved the migration, proliferation, and tube formation ability (15 dyn/cm² LSS vs. stationary: 72.2±5.5 vs. 47.3±7.3, 0.517±0.05 vs. 0.367±0.038, and 1.664±0.315 vs. 1±0, respectively; all P<0.05) along with the in vivo angiogenesis capacity of late EPCs, contributing to the recovery of limb ischemia. These effects were also blocked by Akt inhibition or GTPCH knockdown (P<0.05, respectively). Conclusions. This study provides the first evidence that shear stress triggers angiogenesis in late EPCs via the PTEN/Akt/GTPCH/BH4 pathway, providing a potential nonpharmacologic therapeutic strategy for promoting angiogenesis in ischemia-related diseases.

The adaptive dual-task deep learning model, ThyNet-S, demonstrated exceptional performance in enhancing thyroid cancer US screening efficiency and improved clinical decision-making.

Background. Studies show that aliskiren exerts favourable effects not only on endothelial progenitor cells (EPCs) but also on endothelial function. However, the mechanism of the favourable effect of aliskiren on EPCs from patients with hypertension is unclear and remains to be further studied. Methods. The object of this study was to investigate and assess the in vitro function of EPCs pretreated with aliskiren. After treated with aliskiren, the human EPCs were transplanted into a nude mouse model of carotid artery injury, and the in vivo reendothelialization of injured artery was estimated by staining denuded areas with Evans blue dye via tail vein injection. Results. We found that aliskiren increased the in vitro migration, proliferation, and adhesion of EPCs from patients with hypertension in a dose-dependent manner and improved the reendothelialization capability of these EPCs. Furthermore, aliskiren increased the phosphorylation of Tie2, Akt, and eNOS. After the blockade of the Tie2 signalling pathway, the favourable effects of aliskiren on the in vitro function and in vivo reendothelialization capability of EPCs were suppressed. Conclusions. This study demonstrates that aliskiren can improve the in vitro function and in vivo reendothelialization capability of EPCs from patients with hypertension via the activation of the Tie2/PI3k/Akt/eNOS signalling pathway. These findings further indicate that aliskiren is an effective pharmacological treatment for cell-based repair in hypertension-related vascular injury.

Background/Aims. Sexual differences exist in endothelial progenitor cells (EPCs), and various cardiovascular risk factors are associated with the preservation of endothelial function in premenopausal women. However, it is unclear whether differences in endothelial function and circulating EPCs exist between overweight premenopausal women and age-matched men. Methods. We compared EPC counting and functions in normal-weight and overweight premenopausal women and men, evaluated endothelial function in each group, and detected the expression of the guanosine triphosphate cyclohydrolase I (GTPCH I) pathway. Results. The number of EPCs was lower in the male group than in the female group, regardless of normal-weight or overweight status, and there was no significant difference between the different weight groups among females or males. Endothelial function and EPC migration and proliferation were preserved in overweight premenopausal women compared with overweight men as were nitric oxide (NO) levels in plasma and secreted by EPCs. Endothelial function, the circulating EPC population, and NO levels were not different between normal-weight and overweight premenopausal women. Flow-mediated dilatation was significantly correlated with EPC function, plasma NO levels, and EPC-secreted NO. Conclusions. This investigation provides the first evidence for sex-based differences in EPC activity and endothelial function in overweight middle-aged individuals; these differences are associated with alterations in NO production and may partly occur through downregulation of the GTPCH I pathway. The present results provide new insights into the mechanism underlying the preserved endothelial function in overweight premenopausal women and may uncover a potential therapeutic target for endothelial repair in overweight population.

American Association for the Study of Liver Diseases (AASLD) version 2023, which incorporates visualization score VIS-C, increasing α-fetoprotein level, and lesion growth, improved sensitivity and negative predictive value for hepatocellular carcinoma surveillance, outperforming US Liver Imaging Reporting System version 2017 and AASLD version 2018.