AI-integrated Screening to Replace Double Reading of Mammograms: A Population-wide Accuracy and Feasibility Study.

"Just Accepted" papers have undergone full peer review and have been accepted for publication in Radiology: Artificial Intelligence. This article will undergo copyediting, layout, and proof review before it is published in its final version. Please note that during production of the final copyedited article, errors may be discovered which could affect the content. Mammography screening supported by deep learning-based artificial intelligence (AI) solutions can potentially reduce workload without compromising breast cancer detection accuracy, but the site of deployment in the workflow might be crucial. This retrospective study compared three simulated AI-integrated screening scenarios with standard double reading with arbitration in a sample of 249,402 mammograms from a representative screening population. A commercial AI system replaced the first reader (Scenario 1: Integrated AIfirst), the second reader (Scenario 2: Integrated AIsecond), or both readers for triaging of low- and high-risk cases (Integrated AItriage). AI threshold values were partly chosen based on previous validation and fixing screen-read volume reduction at approximately 50% across scenarios. Detection accuracy measures were calculated. Compared with standard double reading, Integrated AIfirst showed no evidence of a difference in accuracy metrics except for a higher arbitration rate (+0.99%; P < .001). Integrated AIsecond had lower sensitivity (-1.58%; P < 0.001), negative predictive value (NPV) (- 0.01%; P < .001) and recall rate (< 0.06%; P = 0.04), but a higher positive predictive value (PPV) (+0.03%; P < .001) and arbitration rate (+1.22%; P < .001). Integrated AItriage achieved higher sensitivity (+1.33%; P < .001), PPV (+0.36%; P = .03), and NPV (+0.01%; P < .001) but lower arbitration rate (-0.88%; P < .001). Replacing one or both readers with AI seems feasible, however, the site of application in the workflow can have clinically relevant effects on accuracy and workload. ©RSNA, 2024.

Education in Focused Assessment With Sonography for Trauma Using Immersive Virtual Reality: A Prospective, Interventional Cohort Study and Non-inferiority Analysis With a Historical Control.

OBJECTIVE: Focused assessment with sonography for trauma (FAST) is a valuable ultrasound procedure in emergency settings, and there is a need for evidence-based education in FAST to ensure competencies. Immersive virtual reality (IVR) is a progressive training modality gaining traction in the field of ultrasound training. IVR holds several economic and practical advantages to the common instructor-led FAST courses using screen-based simulation (SBS). METHODS: This prospective, interventional cohort study investigated whether training FAST using IVR unsupervised and out-of-hospital was non-inferior to a historical control group training at a 90 min SBS course in terms of developing FAST competencies in novices. Competencies were assessed in both groups using the same post-training simulation-based FAST test with validity evidence, and a non-inferiority margin of 2 points was chosen. RESULTS: A total of 27 medical students attended the IVR course, and 27 junior doctors attended the SBS course. The IVR group trained for a median time of 117 min and scored a mean 14.2 ± 2.0 points, compared with a mean 13.7 ± 2.5 points in the SBS group. As the lower bound of the 95% confidence interval at 13.6 was within the range of the non-inferiority margin (11.7-13.7 points), training FAST in IVR for a median of 117 min was found non-inferior to training at a 90 min SBS course. No significant correlation was found between time spent in IVR and test scores. CONCLUSION: Within the limitations of the use of a historical control group, the results suggest that IVR could be an alternative to SBS FAST training and suitable for unsupervised, out-of-hospital courses in basic FAST competencies.

Using real-life data to model the impact of increasing BCG vaccination coverage and scar prevalence on all-cause infant mortality.

PURPOSE: Estimating the potential impact on infant mortality of increasing Bacille Calmette-Guérin (BCG) vaccination coverage and BCG scar prevalence. METHODS: Guinea-Bissau Health and Demographic Surveillance System data on BCG vaccination coverage, scar status, and all-cause mortality were used for this study. Mortality risk (MR) by scar status was assessed in Cox models providing adjusted mortality rate ratios (aMRRs). Distributions were fitted for survival, vaccination coverage, and scar prevalence. Models for 12-month mortality were calculated. We utilized World Bank data on birth rates and mortality rates to assess the potential global impact of optimizing BCG vaccination programs. RESULTS: BCG coverage was 81% and scar prevalence 42% among 2-month-old infants, and the 1- to 12-month scar/no scar aMRR was 0.40 (0.22, 0.76). Modeling 2-month 99% vaccination coverage with 95% developing scars would change the 1- to 12-month MR by -8% (-21%, +12%). Globally, the reduction in the MR between 1- and 12-month would be -14% (-14%, -15%), corresponding to -208,075 (-214,453, -204,023) fewer infant deaths/year. CONCLUSIONS: We confirmed previous observations: having a BCG scar markedly reduces infant MR. Increasing current global 2-month BCG vaccination coverage from 76% to 99%, and scar prevalence among vaccinated infants from 52% to 95% might reduce global infant mortality by >200,000 deaths/year. Thus, optimizing BCG vaccination programs to focus on increasing early BCG vaccination coverage and the overall scar prevalence would have major public health benefits.