A novel liver-function-indicators-based prognosis signature for patients with hepatocellular carcinoma treated with anti-programmed cell death-1 therapy.

BACKGROUND: The liver function reserve has a significant impact on the therapeutic effects of anti-programmed cell death-1 (PD-1) for hepatocellular carcinoma (HCC). This study aimed to comprehensively evaluate the ability of liver-function-based indicators to predict prognosis and construct a novel prognostic score for HCC patients with anti-PD-1 immunotherapy. METHODS: Between July 2018 and January 2020, patients diagnosed with HCC who received anti-PD-1 treatment were screened for inclusion in the study. The valuable prognostic liver-function-based indicators were selected using Cox proportional hazards regression analysis to build a novel liver-function-indicators-based signature (LFIS). Concordance index (C-index), the area under the receiver operating characteristic (ROC) curve (AUC), and Kaplan-Meier survival curves were utilized to access the predictive performance of LFIS. RESULTS: A total of 434 HCC patients who received anti-PD-1 treatment were included in the study. The LFIS, based on alkaline phosphatase-to-albumin ratio index, Child-Pugh score, platelet-albumin score, aspartate aminotransferase-to-lymphocyte ratio index, and gamma-glutamyl transpeptidase-to-lymphocyte ratio index, was constructed and identified as an independent risk factor for patient survival. The C-index of LFIS for overall survival (OS) was 0.692, which was higher than the other single liver-function-based indicator. The AUC of LFIS at 6-, 12-, 18-, and 24-month were 0.74, 0.714, 0.747, and 0.865 for OS, respectively. Patients in the higher-risk LFIS group were associated with both worse OS and PFS. An online and easy-to-use calculator was further constructed for better application of the LFIS signature. CONCLUSION: The LFIS score had an excellent prognosis prediction ability superior to every single liver-function-based indicator for anti-PD-1 treatment in HCC patients. It is a reliable, easy-to-use tool to stratify risk for OS and PFS in HCC patients who received anti-PD-1 treatment.

Joint Learning of Neural Transfer and Architecture Adaptation for Image Recognition.

Current state-of-the-art visual recognition systems usually rely on the following pipeline: 1) pretraining a neural network on a large-scale data set (e.g., ImageNet) and 2) finetuning the network weights on a smaller, task-specific data set. Such a pipeline assumes that the sole weight adaptation is able to transfer the network capability from one domain to another domain based on a strong assumption that a fixed architecture is appropriate for all domains. However, each domain with a distinct recognition target may need different levels/paths of feature hierarchy, where some neurons may become redundant, and some others are reactivated to form new network structures. In this work, we prove that dynamically adapting network architectures tailored for each domain task along with weight finetuning benefits in both efficiency and effectiveness, compared to the existing image recognition pipeline that only tunes the weights regardless of the architecture. Our method can be easily generalized to an unsupervised paradigm by replacing supernet training with self-supervised learning in the source domain tasks and performing linear evaluation in the downstream tasks. This further improves the search efficiency of our method. Moreover, we also provide principled and empirical analysis to explain why our approach works by investigating the ineffectiveness of existing neural architecture search. We find that preserving the joint distribution of the network architecture and weights is of importance. This analysis not only benefits image recognition but also provides insights for crafting neural networks. Experiments on five representative image recognition tasks, such as person re-identification, age estimation, gender recognition, image classification, and unsupervised domain adaptation, demonstrate the effectiveness of our method.

Towards a three-component model of fan loyalty: a case study of Chinese youth.

The term "fan loyalty" refers to the loyalty felt and expressed by a fan towards the object of his/her fanaticism in both everyday and academic discourses. However, much of the literature on fan loyalty has paid little attention to the topic from the perspective of youth pop culture. The present study explored the meaning of fan loyalty in the context of China. Data were collected by the method of in-depth interviews with 16 young Chinese people aged between 19 and 25 years who currently or once were pop fans. The results indicated that fan loyalty entails three components: involvement, satisfaction, and affiliation. These three components regulate the process of fan loyalty development, which can be divided into four stages: inception, upgrade, zenith, and decline. This model provides a conceptual explanation of why and how young Chinese fans are loyal to their favorite stars. The implications of the findings are discussed.