Purpose Coronary outlet resistance is influenced by the quantification and distribution of resting coronary blood flow. It is crucial for a more physiologically accurate estimation of fractional flow reserve (FFR) derived from computed tomography angiography (CTA), referred to as FFRCT. This study presents a physiologically personalized (PP)‐based coronary blood flow model involving the outlet boundary condition (BC) and a standardized outlet truncation strategy to estimate the outlet resistance and FFRCT. Methods In this study, a total of 274 vessels were retrospectively collected from 221 patients who underwent coronary CTA and invasive FFR within 14 days. For FFRCT determination, we have employed a PP‐based outlet BC model involving personalized physiological parameters and left ventricular mass (LVM) to quantify resting coronary blood flow. We evaluated the improvement achieved in the diagnostic performance of FFRCT by using the PP‐based outlet BC model relative to the LVM‐based model, with respect to the invasive FFR. Additionally, in order to evaluate the impact of the outlet truncation strategy on FFRCT, 68 vessels were randomly selected and analyzed independently by two operators, by using two different outlet truncation strategies at 1‐month intervals. Results The per‐vessel diagnostic performance of the PP‐based outlet BC model was improved, based on invasive FFR as reference, compared to the LVM‐based model: (i) accuracy/sensitivity/specificity: 91.2%/90.4%/91.8% versus 86.5%/84.6%/87.6%, for the entire dataset of 274 vessels, (ii) accuracy/sensitivity/specificity: 88.7%/82.4%/90.4% versus 82.4%/ 76.5%/84.0%, for moderately stenosis lesions. The standardized outlet truncation strategy showed good repeatability with the Kappa coefficient of 0.908. Conclusions It has been shown that our PP‐based outlet BC model and standardized outlet truncation strategy can improve the diagnostic performance and repeatability of FFRCT.

No abstract

Leveled reading (LR) aims to automatically classify texts according to different reading capabilities and provide appropriate reading materials to readers. However, most state-of-theart LR methods rely on the availability of copious annotated resources, which prevents their adaptation to low-resource languages like Chinese. In our work, to tackle Chinese LR, we explore to perform different language transfer methods on English-Chinese LR. Specifically, we focus on adversarial training and cross-lingual pre-training method to transfer the LR knowledge learned from annotated data in the rich-resource English language to Chinese. For evaluation, we introduce the agebased standard to align datasets with different leveling standards, and conduct experiments in both zero-shot and few-shot settings. Experiments show that the cross-lingual pre-training method can capture language-invariant features more effectively than adversarial training. We also conduct analysis to propose further improvement in cross-lingual LR.