Search

Background: Perioperative use of immune checkpoint blockade (ICB) improves survival in patients with early-stage cancer. Treatment-related adverse events (AEs), frequently involve the endocrine system which may increase perioperative complications and affect quality of life. Objective: We conducted a meta-analysis to elucidate the impact of adding ICB to conventional neoadjuvant/adjuvant therapy on the incidence of endocrine AEs. Design: A systematic review and meta-analysis of randomize-controlled trials (RCTs). Data sources and methods: A systematic search of PubMed, Embase, Web of Science, and Cochrane library was performed for RCTs comparing groups with and without the addition of ICB to conventional perioperative therapy in patients with cancer. Outcomes included all-grade and grade 3–5 thyroiditis, hyperthyroidism, hypothyroidism, adrenal insufficiency, hypophysitis, type 1 diabetes mellitus, and hyperglycemia. The odds ratios (ORs) of all-grade and grade 3–5 endocrine were pooled using the random-effect model meta-analysis. Results: Twenty-four RCTs comprising 12,199 patients were identified for meta-analysis. The addition of ICB was associated with higher incidence of thyroiditis [all grade: OR = 3.53 (95% confidence interval (CI): 1.88–6.64)], hyperthyroidism [all-grade: 7.18 (4.30–12.01); grade 3–5: 3.93 (1.21–12.82)], hypothyroidism [all-grade: 5.39 (3.68–7.90); grade 3–5: 3.63 (1.18–11.11)], adrenal insufficiency [all-grade: 3.82 (1.88–7.79); grade 3–5: 5.91 (2.36–14.82)], hypophysitis [all-grade: 10.29 (4.97–21.3); grade 3–5: 5.80 (1.99–16.92)], and type 1 diabetes mellitus [all-grade: 2.24 (1.06–4.74); grade 3–5: 3.49 (1.21–10.08)]. The cumulative incidence of each grade 3–5 endocrine AE was low (

Deep Learning (DL) techniques are being used in various critical applications like self-driving cars. DL techniques such as Deep Neural Networks (DNN), Deep Reinforcement Learning (DRL), Federated Learning (FL), and Transfer Learning (TL) are prone to adversarial attacks, which can make the DL techniques perform poorly. Developing such attacks and their countermeasures is the prerequisite for making artificial intelligence techniques robust, secure, and deployable. Previous survey papers only focused on one or two techniques and are outdated. They do not discuss application domains, datasets, and testbeds in detail. There is also a need to discuss the commonalities and differences among DL techniques. In this paper, we comprehensively discussed the attacks and defenses in four popular DL models, including DNN, DRL, FL, and TL. We also highlighted the application domains, datasets, metrics, and testbeds in these fields. One of our key contributions is to discuss the commonalities and differences among these DL techniques. Insights, lessons, and future research directions are also highlighted in detail.

Abstract??In this study, nanoparticle penetration was measured with a wide range of filter media using silver nanoparticles from 3?nm to 20?nm at three different face velocities in order to define nanoparticle filtration characteristics of commercial fibrous filter media. The silver particles were generated by heating a pure silver powder source via an electric furnace with a temperature of 870?C, which was found to be the optimal temperature for generating an adequate amount of silver nanoparticles for the size range specified above. After size classification using a nano-DMA, the particle counts were measured by an Ultrafine Condensation Particle Counter (UCPC) both upstream and downstream of the test filter to determine the nanoparticle penetration for each specific particle size. Particle sampling time continued long enough to detect more than 105counts at the upstream and 10 counts at the downstream sampling point so that 99.99% efficiency can be detected with the high efficiency filter. The results show a very high uniformity with small error bars for all filter media tested in this study. The particle penetration decreases continuously down to 3?nm as expected from the classical filtration theory, and together with a companion modeling paper by?Wang et al. in this same issue, we found no significant evidence of nanoparticle thermal rebound down to 3?nm. [ABSTRACT FROM AUTHOR]