Radiology. Imaging cancer最新文献

This retrospective, single-center study aimed to assess the diagnostic performance and agreement of two large language models (LLMs), ChatGPT-4 Turbo (OpenAI) and Llama 3 (Meta AI), in assigning Prostate Imaging Reporting and Data System (PI-RADS) scores to prostate MRI reports and to compare their performance with two abdominal radiologists. Structured prostate MRI reports (n = 500) obtained between January and December 2022, with original PI-RADS scores removed, were processed with LLMs using a standardized prompt to extract PI-RADS version 2.1 scores. Two abdominal radiologists independently assigned scores, with a third adjudicating discrepancies. Prostate biopsy results served as the reference standard for diagnostic performance assessment. There was high agreement between both models and radiologists: ChatGPT-4 Turbo and Llama 3 achieved 97.7% agreement (κ = 0.95), agreement between the LLM and radiologists ranged from 94.7% to 95.7% (κ = 0.89-0.91), and interradiologist agreement was 94.4% (κ = 0.88). ChatGPT-4 Turbo assigned significantly higher scores than radiologists (P < .005), while differences with Llama 3 were not statistically significant (P = .08). ChatGPT-4 Turbo and the original MRI reports achieved an area under the receiver operating characteristic curve (AUC) of 0.79 for predicting prostate cancer, and radiologists and Llama 3 achieved AUCs of 0.78 each. These results suggest that LLMs could improve prostate MRI reporting through accurate and consistent PI-RADS scoring. Keywords: Prostate, Oncology, Large Language Model Supplemental material is available for this article. © RSNA, 2025.

Purpose To evaluate predictors of mammography engagement at the county level to better understand the challenges associated with population-based screening in the United States. Materials and Methods This retrospective, geospatial, cross-sectional study conducted from March 2024 to September 2024 used data from the 2023 National Cancer Institute's Small Area Estimates, County Health Rankings, and the Atlas of Rural and Small-Town America. Univariable linear regression analyses were conducted to evaluate the relationship between sociodemographic variables and county-level mammography screening rates. Multivariable linear regression was performed to create an average model of mammography screening rates as a function of county-level sociodemographic variables, which was then used to assess county screening performance after adjusting for these factors. Results Information obtained from 3121 counties was included in this study. Higher percentages of White residents, residents older than 18 years, women, rural residents, and high school graduates were positively associated with county-level screening rates, while higher percentages of non-Black minority residents, residents with limited English proficiency, and uninsured residents were negatively associated with county-level mammography screening rates (P < .001). After accounting for sociodemographic composition, counties with a higher proportion of Black residents (P < .001), residents with limited English proficiency (P = .009), insured residents (P < .001), high school graduates (P < .001), and residents with higher median income (P < .001) exceeded expected screening rates based on national county-level data. Conclusion In this national county-level model, population-based characteristics associated with mammography engagement at the county level could be used to estimate expected screening rates based on each county's sociodemographic profile. Keywords: Mammography, Breast, Screening, Socioeconomic Issues, Oncology, Epidemiology, Statistics, Health Policy and Practice ©RSNA, 2025.

Purpose To examine the use of propensity score matching (PSM) to compare the safety and efficacy of US-guided thermal ablation (TA) as a first-line treatment for initial thyroid nodules versus as a second-line treatment for recurrent nodules following thyroid surgery. Materials and Methods This retrospective study included data from 693 patients with benign thyroid nodules (BTNs) who underwent US-guided TA between January 2015 and July 2023. These included 619 patients in the initial nodule group and 74 patients in the recurrent nodule group. PSM was employed to balance preablation data and minimize confounding bias between the two groups. The perioperative and post-TA complication rates of the two matched patient groups were compared. Nodule volume and volume reduction rate (VRR) were compared at 1, 3, 6, and 12 months after TA. Thyroid function, symptoms, and cosmetic scores were also compared. Results Successful matching of 222 patients (mean age ± SD, 55.92 years ± 12.08; 172 female) (2:1 ratio) was achieved, with 148 patients in the initial nodule group and 74 patients in the recurrent nodule group. There was no evidence of a difference between the two groups regarding incidence of major complications (P > .99), minor complications (P = .82), or adverse effects (P = .51). The difference in VRR between the two groups at each follow-up point was not statistically significant (P = .30, .28, .33, .20, .33, respectively). At the last follow-up, symptom and cosmetic scores were significantly reduced in both groups (both P < .01), whereas serum thyroid hormone levels were not significantly different from those before TA treatment (all P > .05). Conclusion US-guided TA treatment exhibited a favorable safety and efficacy profile for patients with BTNs, regardless of prior thyroid surgery. Supplemental material is available for this article. © RSNA, 2025 Keywords: Head/Neck, Thyroid, Treatment Effects, Efficacy Studies, Ablation Techniques (Radiofrequency, Thermal, Chemical).