Colorectal cancer screening: The value of early detection and modern challenges.

The screening of colorectal cancer (CRC) is pivotal for both the prevention and treatment of this disease, significantly improving early-stage tumor detection rates. This advancement not only boosts survival rates and quality of life for patients but also reduces the costs associated with treatment. However, the adoption of CRC screening methods faces numerous challenges, including the technical limitations of both noninvasive and invasive methods in terms of sensitivity and specificity. Moreover, socioeconomic factors such as regional disparities, economic conditions, and varying levels of awareness affect screening uptake. The coronavirus disease 2019 pandemic further intensified these cha-llenges, leading to reduced screening participation and increased waiting periods. Additionally, the growing prevalence of early-onset CRC necessitates innovative screening approaches. In response, research into new methodologies, including artificial intelligence-based systems, aims to improve the precision and accessibility of screening. Proactive measures by governments and health organizations to enhance CRC screening efforts are underway, including increased advocacy, improved service delivery, and international cooperation. The role of technological innovation and global health collaboration in advancing CRC screening is undeniable. Technologies such as artificial intelligence and gene sequencing are set to revolutionize CRC screening, making a significant impact on the fight against this disease. Given the rise in early-onset CRC, it is crucial for screening strategies to continually evolve, ensuring their effectiveness and applicability.

Enhancing robustness in video recognition models: Sparse adversarial attacks and beyond.

Recent years have witnessed increasing interest in adversarial attacks on images, while adversarial video attacks have seldom been explored. In this paper, we propose a sparse adversarial attack strategy on videos (DeepSAVA). Our model aims to add a small human-imperceptible perturbation to the key frame of the input video to fool the classifiers. To carry out an effective attack that mirrors real-world scenarios, our algorithm integrates spatial transformation perturbations into the frame. Instead of using the lp norm to gauge the disparity between the perturbed frame and the original frame, we employ the structural similarity index (SSIM), which has been established as a more suitable metric for quantifying image alterations resulting from spatial perturbations. We employ a unified optimisation framework to combine spatial transformation with additive perturbation, thereby attaining a more potent attack. We design an effective and novel optimisation scheme that alternatively utilises Bayesian Optimisation (BO) to identify the most critical frame in a video and stochastic gradient descent (SGD) based optimisation to produce both additive and spatial-transformed perturbations. Doing so enables DeepSAVA to perform a very sparse attack on videos for maintaining human imperceptibility while still achieving state-of-the-art performance in terms of both attack success rate and adversarial transferability. Furthermore, built upon the strong perturbations produced by DeepSAVA, we design a novel adversarial training framework to improve the robustness of video classification models. Our intensive experiments on various types of deep neural networks and video datasets confirm the superiority of DeepSAVA in terms of attacking performance and efficiency. When compared to the baseline techniques, DeepSAVA exhibits the highest level of performance in generating adversarial videos for three distinct video classifiers. Remarkably, it achieves an impressive fooling rate ranging from 99.5% to 100% for the I3D model, with the perturbation of just a single frame. Additionally, DeepSAVA demonstrates favourable transferability across various time series models. The proposed adversarial training strategy is also empirically demonstrated with better performance on training robust video classifiers compared with the state-of-the-art adversarial training with projected gradient descent (PGD) adversary.

Mortality prediction with adaptive feature importance recalibration for peritoneal dialysis patients.

The study aims to develop AICare, an interpretable mortality prediction model, using electronic medical records (EMR) from follow-up visits for end-stage renal disease (ESRD) patients. AICare includes a multichannel feature extraction module and an adaptive feature importance recalibration module. It integrates dynamic records and static features to perform personalized health context representation learning. The dataset encompasses 13,091 visits and demographic data of 656 peritoneal dialysis (PD) patients spanning 12 years. An additional public dataset of 4,789 visits from 1,363 hemodialysis (HD) patients is also considered. AICare outperforms traditional deep learning models in mortality prediction while retaining interpretability. It uncovers mortality-feature relationships and variations in feature importance and provides reference values. An AI-doctor interaction system is developed for visualizing patients' health trajectories and risk indicators.

Interpretable Machine Learning for COVID-19: An Empirical Study on Severity Prediction Task.

The black-box nature of machine learning models hinders the deployment of some high-accuracy medical diagnosis algorithms. It is risky to put one's life in the hands of models that medical researchers do not fully understand or trust. However, through model interpretation, black-box models can promptly reveal significant biomarkers that medical practitioners may have overlooked due to the surge of infected patients in the COVID-19 pandemic. This research leverages a database of 92 patients with confirmed SARS-CoV-2 laboratory tests between 18th January 2020 and 5th March 2020, in Zhuhai, China, to identify biomarkers indicative of infection severity prediction. Through the interpretation of four machine learning models, decision tree, random forests, gradient boosted trees, and neural networks using permutation feature importance, partial dependence plot, individual conditional expectation, accumulated local effects, local interpretable model-agnostic explanations, and Shapley additive explanation, we identify an increase in N-terminal pro-brain natriuretic peptide, C-reaction protein, and lactic dehydrogenase, a decrease in lymphocyte is associated with severe infection and an increased risk of death, which is consistent with recent medical research on COVID-19 and other research using dedicated models. We further validate our methods on a large open dataset with 5644 confirmed patients from the Hospital Israelita Albert Einstein, at São Paulo, Brazil from Kaggle, and unveil leukocytes, eosinophils, and platelets as three indicative biomarkers for COVID-19.

[Spatial and temporal variations of vegetation phenology and its response to urbanization in central Yunnan urban agglomeration, Southwest China]. / æ»‡ä¸­åŸŽå¸‚ç¾¤æ¤è¢«ç‰©å€™æ—¶ç©ºå˜åŒ–åŠå ¶å¯¹åŸŽå¸‚åŒ–çš„å“åº”.

Vegetation phenology is an important sensor that responds to environmental changes. Based on MOD13Q1 EVI data, we used the dynamic threshold method to extract vegetation phenological parameters of the central Yunnan urban agglomeration from 2001 to 2020, namely the start of growing season, the end of growing season, and the length of growing season, aiming to reveal the spatiotemporal variations in vegetation phenology and urban-rural differences. The results showed that vegetation phenology of the central Yunnan urban agglomeration from 2001 to 2020 generally showed a phenomenon of delayed start of growing season, delayed the end of growing season (0.66 days per year), and prolonged growing season. Compared with suburban and rural areas, growing season in urban areas in the past 20 years had started earlier (1.05 days per year), ended later (0.91 days per year), and thus growing season had been prolonged (1.79 days per year). Vegetation phenology showed significant difference on the gradient of urban, suburban, and rural areas. The start and the end of growing season of urban vegetation were the earliest, and the length of growing season was the longest, with the most significant changes in the urban areas and within the range of 0-2 km outward. The start of growing season in urban area was significantly earlier, the end of growing season was significantly delayed, and length of growing season was prolonged significantly with the increase of population density, per capita GDP, and the proportion of built-up area. The sensitivity of different phenological periods of vegetation and their duration to environmental changes varied on the gradient of urban, suburban and rural areas. Population density and proportion of built-up area in the study area played an important role in delaying the end of growing season of vegetation in the central Yunnan urban agglomeration.

Development of a performance measurement system for general practitioners' office in China's primary healthcare.

BACKGROUND: General practitioners are the main providers of primary care services. To better strengthen the important role of general practitioners in primary healthcare services, China is promoting the general practitioners' office system. There is a lack of well-accepted methods to measure the performance of general practitioner offices in China. We thus aim to develop a systematic and operable performance measurement system for evaluating the general practitioner's office. METHODS: We establish an index pool of the performance measurement system of general practitioners' offices by a cross-sectional study and the literature research method and adopt the focus group method to establish the preliminary system. The Delphi method is then used to conduct three rounds of consultation to modify indices, which aims to form the final indicator system. We determine the weight of each index by the analytic hierarchy process method, which together with the final indicator system constitutes the final performance measurement system. Finally, we select three offices from three different cities in Sichuan Province, China, as case offices to conduct the case study, aiming to assess its credibility. RESULTS: Our results show that the first office scored 958.5 points, the second scored 768.1 points, and the third scored 947.7 points, which corresponds to the reality of these three offices, meaning that the performance measurement system is effective and manoeuvrable. CONCLUSIONS: Our study provides support for standardizing the functions of China's general practitioner's office, improving the health service quality of generalists, and providing a theoretical basis for the standardization of the general practitioner's office.

Structural and biochemical insights into small RNA 3' end trimming by Arabidopsis SDN1.

A family of DEDDh 3'→5' exonucleases known as Small RNA Degrading Nucleases (SDNs) initiates the turnover of ARGONAUTE1 (AGO1)-bound microRNAs in Arabidopsis by trimming their 3' ends. Here, we report the crystal structure of Arabidopsis SDN1 (residues 2-300) in complex with a 9 nucleotide single-stranded RNA substrate, revealing that the DEDDh domain forms rigid interactions with the N-terminal domain and binds 4 nucleotides from the 3' end of the RNA via its catalytic pocket. Structural and biochemical results suggest that the SDN1 C-terminal domain adopts an RNA Recognition Motif (RRM) fold and is critical for substrate binding and enzymatic processivity of SDN1. In addition, SDN1 interacts with the AGO1 PAZ domain in an RNA-independent manner in vitro, enabling it to act on AGO1-bound microRNAs. These extensive structural and biochemical studies may shed light on a common 3' end trimming mechanism for 3'→5' exonucleases in the metabolism of small non-coding RNAs.

Evolution of two modes of intrinsic RNA polymerase transcript cleavage.

During gene transcription, the RNA polymerase (Pol) active center can catalyze RNA cleavage. This intrinsic cleavage activity is strong for Pol I and Pol III but very weak for Pol II. The reason for this difference is unclear because the active centers of the polymerases are virtually identical. Here we show that Pol II gains strong cleavage activity when the C-terminal zinc ribbon domain (C-ribbon) of subunit Rpb9 is replaced by its counterpart from the Pol III subunit C11. X-ray analysis shows that the C-ribbon has detached from its site on the Pol II surface and is mobile. Mutagenesis indicates that the C-ribbon transiently inserts into the Pol II pore to complement the active center. This mechanism is also used by transcription factor IIS, a factor that can bind Pol II and induce strong RNA cleavage. Together with published data, our results indicate that Pol I and Pol III contain catalytic C-ribbons that complement the active center, whereas Pol II contains a non-catalytic C-ribbon that is immobilized on the enzyme surface. Evolution of the Pol II system may have rendered mRNA transcript cleavage controllable by the dissociable factor transcription factor IIS to enable promoter-proximal gene regulation and elaborate 3'-processing and transcription termination.