Search

Peracetic acid (PAA) oxidation catalyzed by metal-free carbons is promising for advanced water decontamination. Nevertheless, developing reaction-oriented and high-performance carbocatalysts has been limited by the ambiguous understanding of the intrinsic relationship between carbon chemical/molecular structure and PAA transformation behavior. Herein, we comprehensively investigated the PAA activation using a family of well-defined sp2/sp3carbon hybrids from annealed nanodiamonds (ANDs). The activity of ANDs displays a volcano-type trend, with respect to the sp2/sp3ratio. Intriguingly, sp3-C-enriched AND exhibits the best catalytic activity for PAA activation and phenolic oxidation, which is different from persulfate chemistry in which the sp2network normally outperforms sp3hybridization. At the electron-rich sp2-C site, PAA undergoes a reduction reaction to generate a reactive complex (AND-PAA*) and induces an electron-transfer oxidation pathway. At the sp3-C site adjacent to C═O, PAA is oxidized to surface-confined OH* and O* successively, which ultimately evolves into singlet oxygen (1O2) as the primary reactive species. Benefiting from the dual nonradical regimes on sp2/sp3hybrids, AND mediates a sustainable redox recycle with PAA to continuously generate reactive species to attack water contaminants, meanwhile maintaining structural/chemical integrity and exceptional reusability in cyclic runs.

With grid modernization, smart inverters are increasingly used to execute advanced controls for distribution network reliability. However, this also increases the cyber-attack space. This paper focuses on the defense approaches to restore the system to normal operation circumstances in the presence of cyber-attacks. A unique deep reinforcement learning (DRL) method is developed to minimize voltage violations and reduce power losses for impacted feeders. The defense problem is reformulated as a Markov decision-making process to dynamically control DERs while minimizing load shedding. This is achieved via an improved soft actor-critic (SAC)-based DRL algorithm, which can govern DER set points and load-shedding scenarios in discrete and continuous modes via the auto-tune entropy and Gaussian policy features. Numerical comparison results on the modified IEEE 123-node system with other control approaches, such as Volt-VAR (VV), Volt-Watt (VW), and model predictive control (MPC) show that the proposed method can eliminate voltage violations and provide feasible control actions that perform complete mitigation of cyber-threats.

The recent advancements in wireless technology enable connected autonomous vehicles (CAVs) to gather information about their environment by vehicle-to-vehicle (V2V) communication. In this work, we design an information-sharing-based multi-agent reinforcement learning (MARL) framework for CAVs, to take advantage of the extra information when making decisions to improve traffic efficiency and safety. The safe actor-critic algorithm we propose has two new techniques: the truncated $\mathcal {Q}$ -function and safe action mapping. The truncated $\mathcal {Q}$ -function utilizes the shared information from neighboring CAVs such that the joint state and action spaces of the $\mathcal {Q}$ -function do not grow in our algorithm for a large-scale CAV system. We prove the bound of the approximation error between the truncated- $\mathcal {Q}$ and global $Q$ -functions. The safe action mapping provides a provable safety guarantee for both the training and execution based on control barrier functions. Using the CARLA simulator for experiments, we show that our approach improves the CAV system’s efficiency in terms of average velocity and comfort under different CAV ratios and different traffic densities. We also show that our approach avoids the execution of unsafe actions and always maintains a safe distance from other vehicles. We construct an obstacle-at-corner scenario to show that the shared vision can help CAVs to observe obstacles earlier and take action to avoid traffic jams. The experiment video is on https://songyanghan.github.io/cavmarl/.

With the worldwide mobility electrification initiative to reduce air pollution and energy security, more and more for-hire vehicles are being replaced with electric ones. A key difference between gas for-hire vehicles and electric for-hire vehicles (EFHV) is their energy replenishment mechanisms, i.e., refueling or charging, which is reflected in two aspects: (i) much longer charging processes versus much shorter refueling processes and (ii) time-varying electricity prices versus time-invariant gasoline prices during a day. The complicated charging issues (e.g., long charging time and dynamic charging pricing) potentially reduce the daily operation time and profits of EFHVs, and also cause overcrowded charging stations during some off-peak charging pricing periods. Motivated by a set of findings obtained from a data-driven investigation and field studies, in this paper, we design a fairness-aware vehicle displacement system called FairMove to jointly optimize the overall profit efficiency and profit fairness of EFHV drivers by considering both the passenger travel demand and vehicle charging demand. We first formulate the EFHV displacement problem as a Markov decision problem, and then we present a fairness-aware multi-agent actor-critic approach to tackle this problem. More importantly, we implement and evaluate FairMove with real-world streaming data from the Chinese city Shenzhen, including GPS data and transaction data from over 20,100 EFHVs, coupled with the data of 123 charging stations, which constitute, to our knowledge, the largest EFHV network in the world. Extensive experimental results show that our fairness-aware FairMove effectively improves the profit efficiency and profit fairness of the EFHV fleet by 26.9% and 54.8%, respectively. It also improves the charging station utilization fairness by 38.4%.

Social mating systems (e.g., monogamy, polygamy, or polyandry) are relatively stable behavioral strategies developed by environmental in animals, but the genetic imprint of a particular mating system is often incongruent with the social mating system due to extrapair matings. However, the genetics of mating systems remain little understood in rodents. In this study, we investigated the genetic signature of the mating system of 141 (63 females, 78 males) field-captured midday gerbils (Meriones meridianus)—a rodent species commonly found in the Minqin Desert in China—through microsatellite site analyses of genetic structure and relatedness. Seven pairs of highly polymorphic microsatellite loci were selected and were highly polymorphic, with the combined exclusion probability greater than 0.99. The parent pair paternity test by Cervus 3.0 software show that 11 mother–offspring and nine father–offspring relationships were identified in 2018, involving 26 individuals from 10 families. Similarly, 19 mother–offspring and 19 father–offspring relationships were identified in 2019, involving 48 individuals from 18 families. All three types of genetic mating structure were identified—monogamy (19 families), polyandry (4 families), and polygyny (5 families), providing evidence that the genetics underlying mating systems in this species are variable, can be incongruent with behavioral evidence for social mating systems, and could vary based on environmental cues, including degree of perceived or actual predation.

•CART-38 cells are effective against preclinical models of human AML, T-ALL, and MM.•CART-38 cells are at least as effective as current immunotherapies for these diseases.

With the rapid development of cities, heterogeneous urban cyber-physical systems are designed to improve citizens’ experience, e.g., navigation and delivery service. However, the integration of services is not designed for disruptive events, an oversight that has rippling effects on service quality. For example, urban transportation systems consist of multiple transport modes that have complementary characteristics of capacities, speeds, and costs, facilitating smooth passenger transfers by planned schedules. Such integration may experience significantly increased delays during disruptions. Current solutions rely on a substitute service to transport passengers from and to affected areas using ad-hoc schedules and static routes, which are inefficient and do not utilize mobility patterns of mobile systems, e.g., dynamic passenger demand. To coordinate heterogeneous transportation systems under disruptions, we design a service to automatically select and integrate part of three systems (subway, bus, and taxi) using systems’ mobility patterns, e.g., predicted supply and demand. The service is presented in a normal version, eRoute, considering both subway and bus, and in a version taking taxis into account, called enhanced eRoute. We implement and evaluate eRoute with datasets including subway, bus and taxi, and a fare collection system. The data-driven evaluation results show that eRoute improves the ratio of served passengers per time interval by up to 11.5 times and reduces the average traveling time by up to 82.1 percent compared with existing solutions.

Electric vehicles (EVs) are being rapidly adopted due to their economic and societal benefits. Autonomous mobility-on-demand (AMoD) systems also embrace this trend. However, the long charging time and high recharging frequency of EVs pose challenges to efficiently managing EV AMoD systems. The complicated dynamic charging and mobility process of EV AMoD systems makes the demand and supply uncertainties significant when designing vehicle balancing algorithms. In this work, we design a data-driven distributionally robust optimization (DRO) approach to balance EVs for both the mobility service and the charging process. The optimization goal is to minimize the worst-case expected cost under both passenger mobility demand uncertainties and EV supply uncertainties. We then propose a novel distributional uncertainty sets construction algorithm that guarantees the produced parameters are contained in desired confidence regions with a given probability. To solve the proposed DRO AMoD EV balancing problem, we derive an equivalent computationally tractable convex optimization problem. Based on real-world EV data of a taxi system, we show that with our solution the average total balancing cost is reduced by 14.49%, and the average mobility fairness and charging fairness are improved by 15.78% and 34.51%, respectively, compared to solutions that do not consider uncertainties.

In electrochemical advanced oxidation processes (EAOPs), the rate-limiting step is the mass transfer of pollutants to the electrodes due to the limited active surface areas. To this end, we established a three-dimensional (3D) EAOP system by coupling conventional graphite electrodes with dispersed carbon nanotubes (CNTs). The electrodes (particularly the anode) induced electric field spontaneously polarized CNTs into dispersed reactive particle electrodes (CNT-PEs) in the solution, which remarkably promoted electrochemical activation of peroxydisulfate (PDS) to generate surface CNT–PDS* complexes and surface-bound radicals (SBRs). Based on the excited potential (ECNT-PEs) at different positions in the 3D electric field, CNT-PEs were activated into three states. (i) ECNT-PEs< Eorganic, CNT-PEs are chemically inert toward DCP oxidation; (ii) Eorganic< ECNT-PEs< Ewater, CNT-PEs will oxidize DCP via an electron-transfer process (ETP); (iii) ECNT-PEs> Ewater, both CNT–PDS* complexes and the anode will oxidize water to produce SBRs. Thus, DCP could be oxidized by CNT–PDS* complexes via ETP to form polychlorophenols on the CNT surface, causing rapid deactivation of the micro-electrodes. In contrast, SBRs attack DCP directly into chloride ions and hydroxylated products, maintaining the surface cleanliness and activity of CNT-PEs for long-term operations.

The C oilfield is located in the Bohai Bay Basin, a typical strong bottom water reservoir. Oilfield reservoir and oil–water distribution are complex. At present, the C oilfield has entered the high water cut development stage, and it is challenging to stabilize oil and control water. The reservoir with an imperfect well pattern has dominant bottom water ridge channels, uneven oil–water interface uplift, limited water drive sweep range, and low inter-well reservoir production degree. The oil layer between the horizontal section of the production well and the top of the reservoir cannot be effectively developed, and the remaining oil is enriched. Therefore, it is urgent to explore new energy supplement methods to improve inter-well and vertical remaining oil production in the C oilfield. In this study, the displacement medium is optimized through indoor experimental simulation. From the experimental results, the remaining oil between the sand bodies can be used in heavy oil reservoirs, and the residual oil between wells can be significantly utilized in the alternate displacement of gas and foam, and the recovery degree of the reservoir is increased by 12.44%. The remaining oil at the top of the reservoir can be used in the upper reservoir to increase the remaining oil in the top of the reservoir by injecting gas and foam alternately in the new reservoir. The final recovery of the reservoir is increased by 6.00%. This experimental study guides tapping the potential of the remaining oil in the offshore strong bottom water reservoir.

Engine as the core component of mechanical equipment, its operating state directly affects whether the equipment can operate normally. Predicting the engine remaining useful life (RUL) can monitor the health of the engine in real time and formulate a timely and reasonable maintenance plan. Aiming at the engine monitoring data with various and long time span, we propose a direct prediction method of engine RUL based on particle swarm optimization (PSO) optimized multi-layer Long Short-Term Memory (LSTM) in this paper. Firstly, the monitoring data that can well reflect the engine degradation trend is screened out, and the samples are constructed through a sliding time window. Then, a multi-layer LSTM model is constructed to mine the deep-seated features of the samples for predicting the engine RUL. Finally, the hyperparameters of the multi-layer LSTM model are optimized automatically by the PSO algorithm to optimize the performance of the model. The effectiveness of this method is verified by NASA data set. RMSE, MAE and the scoring function are used as evaluation indexes. RMSE and score of the prediction results are 12.35 and 284.1, respectively. It has higher prediction accuracy compared with traditional deep learning and machine learning methods.

Supplemental digital content is available in the text.

Molybdenum disulfide (MoS2) thin films, layered quasi-two dimensional (2D) materials, have excellent optical and electronic properties for its special structure. In this paper, the surface morphologies, crystals structures and optical characteristics of MoS2 and Er3+-doped MoS2 (Er: MoS2) thin film samples fabricated on the quartz substrates by sol-gel method were investigated. The result indicated that annealing made the crystallinity and optical absorption characteristics of MoS2 thin films better. The light absorption and photoluminescence intensity was enhanced by doping Er3+. Moreover, the photoluminescence intensity increased with increasing doping concentration, but appears saturation tendency.

Detection of organic materials or heavy ions using the surface plasmon resonance (SPR) effect of the MoS2/Ag system was studies. We found that the reflectivity of the SPR system was sensitive to the incident angle of light, dielectric constant of the detected samples, wavelength of light, and thickness of the Ag film. It reached an extreme minimal value (EMV) at an extreme angle θ. As the dielectric constant of the samples was increased from 1.0 to 3.25, θ increased, enabling high-resolution detection. θ slightly decreased with the thickness of the Ag film changed from 300 to 700 nm for a given detected sample while the EMV of reflectivity increased. Moreover, the reflectivity curves were overlapped when the wavelength of light and the thickness, d of the Ag film both increased from 400 to 700 nm, showing synchronous effect. The results showed that MoS2 was a suitable material for the detection of a wide range of samples and could replace the prism.

MicroRNA (miRNA) biogenesis is finely controlled by complex layers of post-transcriptional regulators, including RNA-binding proteins (RBPs). Here, we show that an RBP, QKI5, activates the processing of primary miR-124-1 (pri-124-1) during erythropoiesis. QKI5 recognizes a distal QKI response element and recruits Microprocessor through interaction with DGCR8. Furthermore, the recruited Microprocessor is brought to pri-124-1 stem loops by a spatial RNA-RNA interaction between two complementary sequences. Thus, mutations disrupting their base-pairing affect the strength of QKI5 activation. When erythropoiesis proceeds, the concomitant decrease of QKI5 releases Microprocessor from pri-124-1 and reduces mature miR-124 levels to facilitate erythrocyte maturation. Mechanistically, miR-124 targets TAL1 and c-MYB, two transcription factors involved in normal erythropoiesis. Importantly, this QKI5-mediated regulation also gives rise to a unique miRNA signature, which is required for erythroid differentiation. Taken together, these results demonstrate the pivotal role of QKI5 in primary miRNA processing during erythropoiesis and provide new insights into how a distal element on primary transcripts affects miRNA biogenesis.

This paper considers a method of coding the sensor outputs in order to detect stealthy false data injection attacks. An intelligent attacker can design a sequence of data injection to sensors and actuators that pass the state estimator and statistical fault detector, based on knowledge of the system parameters. To stay undetected, the injected data should increase the state estimation errors while keep the estimation residues small. We employ a coding matrix to change the original sensor outputs to increase the estimation residues under intelligent data injection attacks. This is a low-cost method compared with encryption schemes over all sensor measurements in communication networks. We show the conditions of a feasible coding matrix under the assumption that the attacker does not have knowledge of the exact coding matrix. An algorithm is developed to compute a feasible coding matrix, and, we show that in general, multiple feasible coding matrices exist. To defend against attackers who estimates the coding matrix via sensor and actuator measurements, time-varying coding matrices are designed according to the detection requirements. A heuristic algorithm to decide the time length of updating a coding matrix is then proposed.

Although recent evidence has emerged that Alzheimer’s disease (AD) and amnestic mild cognitive impairment (aMCI) patients show both regional brain abnormalities and topological degeneration in brain networks, our understanding of the effects of white matter fiber aberrations on brain network topology in AD and aMCI is still rudimentary. In this study, we investigated the regional volumetric aberrations and the global topological abnormalities in AD and aMCI patients. The results showed a widely distributed atrophy in both gray and white matters in the AD and aMCI groups. In particular, AD patients had weaker connectivity with long fiber length than aMCI and normal control (NC) groups, as assessed by fractional anisotropy (FA). Furthermore, the brain networks of all three groups exhibited prominent economical small-world properties. Interestingly, the topological characteristics estimated from binary brain networks showed no significant group effect, indicating a tendency of preserving an optimal topological architecture in AD and aMCI during degeneration. However, significantly longer characteristic path length was observed in the FA weighted brain networks of AD and aMCI patients, suggesting dysfunctional global integration. Moreover, the abnormality of the characteristic path length was negatively correlated with the clinical ratings of cognitive impairment. Thus, the results therefore suggested that the topological alterations in weighted brain networks of AD are induced by the loss of connectivity with long fiber lengths. Our findings provide new insights into the alterations of the brain network in AD and may indicate the predictive value of the network metrics as biomarkers of disease development.

Thickened pituitary stalks (TPSs) on magnetic resonance imaging (MRI) result from diverse pathologies; therefore, it is essential to make specific diagnoses for clinical decision-making. The diagnoses and indications for surgical biopsies in patients with central diabetes insipidus (CDI) and TPSs are thoroughly discussed in this paper. Thirty-seven patients with CDI and TPSs were retrospectively reviewed. The mean age at the diagnosis of CDI was 29.0 ± 15.9 years (range 8.0–63.3), and the median duration of follow-up was 5.5 ± 2.8 years (range 0.7–13.0). Anterior pituitary hormone deficiencies were documented in 26 (70.3 %) patients. All patients had a TPS on MRI at the diagnosis of CDI, and 21 (56.8 %) patients exhibited radiological changes during the follow-up. Of these 21 patients, 11 exhibited increases in the thickness of the stalk, and two patients exhibited reversals of the TPSs. Involvements of the hypothalamus, pituitary gland, basal ganglia or supersellar, and pineal gland were found in four, three, one, and 1 patient, respectively. Ultimately, clear diagnoses were established in 17 patients who underwent biopsies, nine of whom had germinomas, six of whom had Langerhans cell histiocytosis, one of whom had a granular cell tumor, and one of whom had Erdheim–Chester disease. Patients with CDI and TPSs should submit to periodic clinic follow-ups with serial MRI assessments to establish anterior pituitary deficiencies and to detect radiological progressions that are appropriate for surgical biopsies. Endoscopic-assisted microsurgery via the supraorbital keyhole approach is a good choice for the biopsy of pituitary stalk lesions.

In view of the uncertainty transformer state information, state of transformer hierarchical evaluation model is established. Firstly, index system of transformer state is formed; Secondly, in order to get comprehensive index of state, grey clustering method is used for monomial index clustering analysis; Finally in the basis of improving synthesis rules of conflict evidence of D - S evidence theory , clustering coefficient is used as the basic probability assignment function to evaluate the overall state of transformer. Example analysis results show that the proposed evaluation model can accurately grasp the running state of transformer and provide guidance for transformer condition overhaul.

Fortilin is a 172-amino acid multifunctional protein present in both intra- and extracellular spaces. Although fortilin binds and regulates various cellular proteins, the biological role of extracellular fortilin remains unknown. Here we report that fortilin specifically interacts with TGF-β1 and prevents it from activating the TGF-β1 signaling pathway. In a standard immunoprecipitation-western blot assay, fortilin co-immunoprecipitates TGF-β1 and its isoforms. The modified ELISA assay shows that TGF-β1 remains complexed with fortilin in human serum. Both bio-layer interferometry and surface plasmon resonance (SPR) reveal that fortilin directly bind TGF-β1. The SPR analysis also reveals that fortilin and the TGF-β receptor II (TGFβRII) compete for TGF-β1. Both luciferase and secreted alkaline phosphatase (SEAP) reporter assays show that fortilin prevents TGF-β1 from activating Smad3 binding to Smad-binding element (Figure). Fortilin inhibits the phosphorylation of Smad3 in both quantitative western blot assays and ELISA. Finally, fortilin inhibits TGFβ-1-induced differentiation of C3H10T1/2 mesenchymal progenitor cells to smooth muscle cells. A computer-assisted virtual docking reveals that fortilin occupies the pocket of TGF-β1 that is normally occupied by TGFβRII and that TGF-β1 can bind either fortilin or TGFβRII at any given time. These data support the role of extracellular fortilin as a negative regulator of the TGF-β1 signaling pathway.

Background The effect of acupuncture treatment in cerebral ischaemia is controversial. A study was undertaken to assess its effects in rats with transient middle cerebral artery occlusion (tMCAO) and discuss its mechanisms.Methods 21 Sprague–Dawley rats were divided into three groups. Group 1 underwent tMCAO and subsequently received acupuncture treatment, Group 2 underwent tMCAO without treatment and Group 3 only underwent sham operation. The evolution of diffusion tensor imaging (DTI) features in ischaemic lesions from acute to chronic periods was assessed and the correlations with behavioural tests and histopathological changes were examined.Results tMCAO rats displayed side-specific sensorimotor deficits after occlusion. Behavioural scores of rats in group 1 reduced gradually with acupuncture treatment. No significant difference in lesion size on T2-weighted imaging was found between the two tMCAO groups. Relative apparent diffusion coefficient (rADC) and relative fractional anisotropy (rFA) values in the centre and at the edge of the ischaemic lesions reduced at first and then increased to varying degrees. Only changes in the rFA value at the edge of the ischaemic lesions were significantly different between the two tMCAO groups. A more significant increase in growth-associated protein B-50 and synaptophysin protein expression was found in group 1 than in the other groups. No significant correlations were found between behavioural scores, DTI appearances and immunohistochemical results.Conclusions The acupuncture points applied were effective, and improving neuronal regeneration may have a role in the mechanism of acupuncture treatment of post-stroke paralysis of the limbs in rats. MRI, particularly the fractional anisotropy value of DTI, is an appropriate parameter to evaluate the recovery status.

Background: Autologous chimeric antigen receptor T (CAR-T) cell therapies show significant clinical activity against hematologic malignancies including multiple myeloma (MM). Numerous factors influence the effectiveness of this therapeutic approach. The fitness of a CAR-T cell product is an important factor affecting T cell engraftment and persistence, a pre-requisite for effective CAR-T cell therapy. In an effort to deconstruct the cellular heterogeneity of CAR-T cell products and explore relationships between T cell states within the cellular product, gene expression and CAR-T cell engraftment following adoptive transfer, we performed single cell RNA-sequencing (scRNA-seq) on CAR-T cell products from a phase I trial of BCMA-specific CAR-T cells in relapsed/refractory MM patients that included cohorts receiving a similar CAR-T cell dose with and without preconditioning with cyclophosphamide (NCT02546167, Cohen et al, J Clin Invest 2019).

Garfall: Amgen: Honoraria; CRISPR Therapeutics: Research Funding; GlaxoSmithKline: Honoraria; Janssen: Honoraria, Research Funding; Novartis: Research Funding; Tmunity: Research Funding. Bu: Novartis: Current Employment, Patents & Royalties: Co-inventor on patent applications. Orlando: Novartis: Current Employment. Brogdon: Novartis Institutes for Biomedical Research: Current Employment. Pruteanu-Malinici: Novartis: Current Employment. Cohen: Janssen: Consultancy; Oncopeptides: Consultancy; Genentech/Roche: Consultancy; BMS/Celgene: Consultancy; AstraZeneca: Consultancy; Novartis: Research Funding; Takeda: Consultancy; GlaxoSmithKline: Consultancy, Research Funding.

Background: To explore the relationship between the lymph node status and preoperative computed tomography images texture features in pancreatic cancer. Methods: A total of 155 operable pancreatic cancer patients (104 men, 51 women; mean age 63.8 ± 9.6 years), who had undergone contrast-enhanced computed tomography in the arterial and portal venous phases, were enrolled in this retrospective study. There were 73 patients with lymph node metastases and 82 patients without nodal involvement. Four different data sets, with thin (1.25 mm) and thick (5 mm) slices (at arterial phase and portal venous phase) were analysed. Texture analysis was performed by using MaZda software. A combination of feature selection algorithms was used to determine 30 texture features with the optimal discriminative performance for differentiation between lymph node positive and negative groups. The prediction performance of the selected feature was evaluated by receiver operating characteristic (ROC) curve analysis. Results: There were 10 texture features with significant differences between two groups and significance in ROC analysis were identified. They were WavEnLH_s-2(wavelet energy with rows and columns are filtered with low pass and high pass frequency bands with scale factors 2) from wavelet-based features, 135dr_LngREmph (long run emphasis in 135 direction) and 135dr_Fraction (fraction of image in runs in 135 direction) from run length matrix-based features, and seven variables of sum average from coocurrence matrix-based features (SumAverg). The ideal cutoff value for predicting lymph node metastases was 270 for WavEnLH_s-2 (positive likelihood ratio 2.08). In addition, 135dr_LngREmph and 135dr_Fraction were correlated with the ratio of metastatic to examined lymph nodes. Conclusions: Preoperative computed tomography high order texture features provide a useful imaging signature for the prediction of nodal involvement in pancreatic cancer.

The aim of this study is to identify the key computed tomography (CT) imaging findings and clinical characteristics of pancreatic metastases for its differential diagnosis. CT images and clinical features of 18 patients with 36 histopathologically proven pancreatic metastases were retrospectively reviewed. The primary malignancy included non-small cell lung cancer (NSCLC) (n=7), gastrointestinal carcinoma (n=5), renal cell carcinoma (RCC) (n=3), osteosarcoma (n=1), cardiac sarcomas (n=1), and neuroendocrine ethmoid sinus carcinoma (n=1). Pancreatic metastases were metachronous in 12 patients (ranging from 4 to 72 months). Tumor markers were elevated for 8 patients, of which 7 patients had NSCLC and gastrointestinal carcinoma, and 1 patient had osteosarcoma. Metastases from NSCLC and gastrointestinal carcinoma frequently presented as small well-circumscribed lesions, with homogeneous or rim enhancement, and or local pancreatic infiltration instead of focal mass, mimicking local pancreatitis. Neuroendocrine ethmoid sinus carcinoma affecting the pancreas also exhibited local pancreatic infiltration. Metastases from RCC and cardiac sarcomas had typical characteristics of hypervascular lesions. Osteosarcoma metastasizing to pancreas had special manifestation, that is, cystic lesion with thick wall and calcification.