Explore the vast range of titles available.

Pancreatic ductal adenocarcinoma (PDAC), the most deadly solid malignancy, is typically detected late and at an inoperable stage. Early or incidental detection is associated with prolonged survival, but screening asymptomatic individuals for PDAC using a single test remains unfeasible due to the low prevalence and potential harms of false positives. Non-contrast computed tomography (CT), routinely performed for clinical indications, offers the potential for large-scale screening, however, identification of PDAC using non-contrast CT has long been considered impossible. Here, we develop a deep learning approach, pancreatic cancer detection with artificial intelligence (PANDA), that can detect and classify pancreatic lesions with high accuracy via non-contrast CT. PANDA is trained on a dataset of 3,208 patients from a single center. PANDA achieves an area under the receiver operating characteristic curve (AUC) of 0.986–0.996 for lesion detection in a multicenter validation involving 6,239 patients across 10 centers, outperforms the mean radiologist performance by 34.1% in sensitivity and 6.3% in specificity for PDAC identification, and achieves a sensitivity of 92.9% and specificity of 99.9% for lesion detection in a real-world multi-scenario validation consisting of 20,530 consecutive patients. Notably, PANDA utilized with non-contrast CT shows non-inferiority to radiology reports (using contrast-enhanced CT) in the differentiation of common pancreatic lesion subtypes. PANDA could potentially serve as a new tool for large-scale pancreatic cancer screening. A deep learning model provides high accuracy in detecting pancreatic lesions in multicenter data, outperforming radiology specialists.

With the advancement of SiC wafers toward 12 inches and innovations in laser cutting technology, new demands have emerged for SiC grinding techniques—namely, high efficiency, low loss, and low wear ratio. This paper investigates electrochemical-assisted grinding of SiC using a grinding wheel–SiC pair model system, examining the effects of electrolyte type, concentration, voltage, load, and rotational speed on wear behavior. Experimental results reveal that NaCl is the most effective electrolyte among the six candidates tested. In the NaCl system, wear behavior is strongly influenced by the interplay between voltage and rotational speed. At a constant voltage of 3 V, increasing the rotational speed to 600 rpm produces a wear area of 1911.93 μm2, while at a higher voltage of 7 V with a lower speed of 200 rpm, the wear area reaches 1301.96 μm2, indicating that optimal material removal requires synergistic matching of electrical and mechanical parameters. At 2 wt% NaCl, a sudden change in wear behavior occurs at 6–7 min, indicating a dynamic balance between oxide formation and mechanical removal. Rotational speed shows a turning point at 600 rpm, where the wear mechanism shifts significantly, marking the transition to a synergistically enhanced regime. EDS analysis confirms that Na2SO4 increases surface oxygen content by 54.4% compared to deionized water, demonstrating enhanced electrochemical oxidation. The optimal parameter window for synergistic removal is identified as 1–2 wt% NaCl, 5–7 V, 600 rpm, and 100–150 g. This study provides quantitative insights into the synergistic removal mechanism of SiC, offering a theoretical foundation for developing efficient, low-loss electrochemical grinding technologies.

The global rise in steatotic liver disease poses a significant public health challenge. While non-contrast computed tomography scans hold promise for opportunistic detection of steatotic liver disease, their potential for staging and risk assessment remains underexplored. Here we present a multimodal AI model trained on a large dataset, comprising of (n=968) histopathologically and (n=1103) radiologically confirmed cases, validated against both histology (n=660) and MRI-PDFF (n=375) gold standards, demonstrating high accuracy in detecting mild to severe steatosis (AUC: 0.904–0.929) and clinically significant fibrosis (AUC: 0.824–0.888). Furthermore, integrating the model into the standard clinical pathway improves primary risk screening in a retrospective patient cohort (n=1192), identifying 36% more patients at risk of fibrosis progression. Using Cox proportional hazard model, we observe that the intermediate-high risk patients identified by the optimized clinical pathway exhibits a significantly higher incidence of cirrhosis (hazard ratio: 5.54: 2.69–11.42), showcasing the model’s potential for early detection and management of steatotic liver disease. This study presents MAOSS, a multimodal AI model that repurposes non-contrast CT scans and leverages clinical features to detect and stage liver steatosis and fibrosis. Here the authors show MAOSS accurately stratifies cirrhosis progression risk when embedded into the standard clinical workflow, enabling scalable, opportunistic screening for early intervention of steatotic liver disease.

The evolutionary loss of ednrb2 in specific vertebrate lineages, such as mammals and cypriniform fish, raises fundamental questions about its functional necessity and potential redundancy or synergy with paralogous endothelin receptors in pigment cell development. In teleosts possessing both ednrb1a and ednrb2 (e.g., Nile tilapia), their respective and combined roles in regulating neural crest-derived pigment cell precursors remains unresolved. Using CRISPR/Cas9, we generated single and double ednrb mutants to dissect their functions. We demonstrated that ednrb1a and ednrb2 synergistically govern the differentiation and migration of iridophore precursors. While ednrb1a is broadly essential for iridophore development, ednrb2 plays a unique and indispensable role in the colonization of iridophores in the dorsal iris. Double mutants exhibit near-complete iridophore loss; severe depletion of melanophores, xanthophores, and erythrophores; and a striking, fertile, transparent phenotype. Crucially, this iridophore deficiency does not impair systemic guanine synthesis pathways. mRNA rescue experiments confirmed mitfa as a key downstream effector within the Ednrb signaling cascade. This work resolves the synergistic regulation of pigment cell fates by Ednrb receptors and establishes a mechanism for generating transparent ermplasm.

ObjectivesTo explore how people perceive different advice for rotator cuff disease in terms of words/feelings evoked by the advice and treatment needs.SettingWe performed a content analysis of qualitative data collected in a randomised experiment.Participants2028 people with shoulder pain read a vignette describing someone with rotator cuff disease and were randomised to: bursitis label plus guideline-based advice, bursitis label plus treatment recommendation, rotator cuff tear label plus guideline-based advice and rotator cuff tear label plus treatment recommendation. Guideline-based advice included encouragement to stay active and positive prognostic information. Treatment recommendation emphasised that treatment is needed for recovery.Primary and secondary outcomesParticipants answered questions about: (1) words/feelings evoked by the advice; (2) treatments they feel are needed. Two researchers developed coding frameworks to analyse responses.Results1981 (97% of 2039 randomised) responses for each question were analysed. Guideline-based advice (vs treatment recommendation) more often elicited words/feelings of reassurance, having a minor issue, trust in expertise and feeling dismissed, and treatment needs of rest, activity modification, medication, wait and see, exercise and normal movements. Treatment recommendation (vs guideline-based advice) more often elicited words/feelings of needing treatment/investigation, psychological distress and having a serious issue, and treatment needs of injections, surgery, investigations, and to see a doctor.ConclusionsWords/feelings evoked by advice for rotator cuff disease and perceived treatment needs may explain why guideline-based advice reduces perceived need for unnecessary care compared to a treatment recommendation.

The rapid growth in the size and complexity of large language models has im-posed severe challenges on memory management, particularly when these models are deployed on GPUs with limited memory. This thesis introduces a fine-grained paging mechanism that dynamically offloads and reloads tensors at the granularity of individual operations, thereby mitigating out-of-memory (OOM) issues during the inference and prefill phase of transformer-based models. Instead of traditional static, layer-based offloading methods, the proposed approach uses compile-time, simulation-based memory allocation to optimize GPU memory usage, making runtime possible under severe memory constraints.This work is based off of the Einsummable system, a framework that represents tensor computations using Einstein summation notation. Einsummable transforms high-level mathematical specifications into an optimized execution pipeline through a series of intermediate representations, notably the TASKGRAPH and the MEMGRAPH. The TASKGRAPH captures the data dependencies and operational flow of tensor computations, while the MEMGRAPH extends this representation by incorporating detailed memory location information and managing offload-reload operations. The transformation from TASKGRAPH to MEMGRAPH is achieved through a simulated execution process the core of this thesis that relies on two key components: an allocation horizon, which pre-allocates memory for future operations, and an execution horizon, which tracks the simulated execution progress of the computation.A key contribution of this thesis is the design and implementation of specialized memory allocation routines simMalloc, simMallocForceReld, and simMallocOffld. These routines not only allocate memory for tensor outputs but also manage dependencies by inserting offload and reload nodes into the MEMGRAPH whenever GPU memory resources depletes. By leveraging full knowledge of the simulated execution order, our offload-reload heuristic selects tensors for offloading based on their computed reuse distance, thereby deferring memory transfers until they are most convenient. This future-aware strategy mitigates the frequency and impact of memory transfers compared to reactive approaches, enabling a finer control over GPU memory usage.Extensive experimental evaluations were conducted using two configurations of NVIDIA GPUs Tesla P100 and V100 to benchmark the performance of the proposed system against state-of-the-art techniques such as ZERO-Inference. The evaluation focused on the prefill stage of inference in LLAMA models with 7B and 65B parameters, a phase known to be particularly memory-bound. The results demonstrate that the fine-grained paging mechanism supports a broader range of configurations, successfully executing inference tasks across varying batch sizes and sequence lengths. While the finer granularity of tensor-level management introduces some communication overhead due to more frequent offloading and reloading, the overall improvements in memory utilization and reduction in OOM errors outweigh these costs.In summary, this thesis makes a contribution to the field of deep learning by addressing the critical challenge of GPU memory constraints through a fine-grained paging mechanism. Future work will explore further optimizations to reduce communication overhead, overall computation latency, and GPU RAM utilization.

High spectral resolution imaging spectroscopy plays a crucial role in solar observation, regularly serving as a backend instrument for solar telescopes. These instruments find direct application in deriving Doppler velocity from hyperspectral images, offering insights into the dynamic motion of matter on the solar surface. In this study, we present the development of a Fabry–Pérot interferometer (FPI) based imaging spectrometer operating at the Fe I (617.3 nm) wavelength for precise Doppler velocity measurements. The spectrometer features a moderate spectral resolution of λ / Δ λ ≈ 60 , 000 , aiming to balance the imaging signal-to-noise ratio (SNR). The instrument underwent successful observational experiments on the 65-cm Educational Adaptive-Optics Solar Telescope (EAST) at the Shanghai Astronomy Museum. Obtained Doppler velocities were compared with data from the Helioseismic and Magnetic Imager (HMI), the maximum column and row correlation coefficients are 0.9261 and 0.9603, respectively. The estimated cut-off normalized frequency of the power spectral density (PSD) curve for velocity map is approximately 0.4/0.21 times higher than that observed in the HMI data, with potentially higher spatial resolution achievable under better seeing conditions. Based on the estimated imaging SNR levels, the accuracy of velocity measurements is approximately 50 m s −1 .

A process is being developed to recycle zinc-bearing metallurgical dust by reductive sintering. In the present work, the reaction behavior of zinc and iron oxides was studied in different conditions in CO–CO 2 atmosphere, to understand the processes involved and determine the optimal conditions. The results showed that dezincification started to become significant when the coke addition was 9.0 wt.% of the amount of raw material, the corresponding CO content in the gases being 20 vol.%. Iron oxide played an important role in the reduction of ZnO: when the CO content was less than 20 vol.%, ZnO and Fe 3 O 4 reacted to generate ZnFe 2 O 4 , CO 2 being the oxidizer that promoted the conversion of Fe 2+ to Fe 3+ . Increasing the temperature was also conducive to the generation of ZnFe 2 O 4 . The effect of iron oxide on the ZnO reduction gradually weakened when the CO content was increased above 20 vol.%. To realize reduction of ZnO and increase the removal rate of zinc, the atmosphere and temperature should be controlled in the thermodynamic stability region of FeO and Zn, where zinc vaporizes and is removed in elemental form.

Lymph node (LN) metastasis status is one of the most critical prognostic and cancer staging clinical factors for patients with resectable pancreatic ductal adenocarcinoma (PDAC, generally for any types of solid malignant tumors). Pre-operative prediction of LN metastasis from non-invasive CT imaging is highly desired, as it might be directly and conveniently used to guide the follow-up neoadjuvant treatment decision and surgical planning. Most previous studies only use the tumor characteristics in CT imaging alone to implicitly infer LN metastasis. To the best of our knowledge, this is the first work to propose a fully-automated LN segmentation and identification network to directly facilitate the LN metastasis status prediction task for patients with PDAC. Specially, (1) we explore the anatomical spatial context priors of pancreatic LN locations by generating a guiding attention map from related organs and vessels to assist segmentation and infer LN status. As such, LN segmentation is impelled to focus on regions that are anatomically adjacent or plausible with respect to the specific organs and vessels. (2) The metastasized LN identification network is trained to classify the segmented LN instances into positives or negatives by reusing the segmentation network as a pre-trained backbone and padding a new classification head. (3) Importantly, we develop a LN metastasis status prediction network that combines and aggregates the holistic patient-wise diagnosis information of both LN segmentation/identification and deep imaging characteristics by the PDAC tumor region. Extensive quantitative nested five-fold cross-validation is conducted on a discovery dataset of 749 patients with PDAC. External multi-center clinical evaluation is further performed on two other hospitals of 191 total patients. Our multi-staged LN metastasis status prediction network statistically significantly outperforms strong baselines of nnUNet and several other compared methods, including CT-reported LN status, radiomics, and deep learning models.

Alkali metal chlorides emitted from sintering flue gas are easily adsorbed on the surface of activated carbon (AC) in the purification process. In this paper, NaCl particles adsorbing onto AC were simulated by impregnation method, and the size and morphology of NaCl particles were similar to those of NaCl-PM 10 emitted from sintering flue gas. With the adsorption of NaCl particles, 2–10-μm pores of AC were filled, and the specific surface area of AC decreased. But NaCl led to the increase of acidic functional groups on the surface of AC. When 0.75 wt% NaCl was adsorbed, it was beneficial for AC catalytic denitration (de-NOx), because the chemical reaction was strengthened by acidic functional groups, so it showed a certain promotion of de-NOx efficiency. As 1.5 wt% NaCl and 3 wt% NaCl were adsorbed, NaCl had an inhibitory effect on AC de-NOx, which was because the specific surface area of AC decreased, and the prevention of physical adsorption played a major role. As a result, the de-NOx efficiency of AC adsorbed with 3 wt% NaCl decreased from 40.59 to 23.02% at 150 °C. Therefore, the absorption of NaCl fine particles on AC should not exceed 0.75 wt%.