资源环境科技发展态势分析平台

Screening mammography aims to identify breast cancer at earlier stages of the disease, when treatment can be more successful(1). Despite the existence of screening programmes worldwide, the interpretation of mammograms is affected by high rates of false positives and false negatives(2). Here we present an artificial intelligence (AI) system that is capable of surpassing human experts in breast cancer prediction. To assess its performance in the clinical setting, we curated a large representative dataset from the UK and a large enriched dataset from the USA. We show an absolute reduction of 5.7% and 1.2% (USA and UK) in false positives and 9.4% and 2.7% in false negatives. We provide evidence of the ability of the system to generalize from the UK to the USA. In an independent study of six radiologists, the AI system outperformed all of the human readers: the area under the receiver operating characteristic curve (AUC-ROC) for the AI system was greater than the AUC-ROC for the average radiologist by an absolute margin of 11.5%. We ran a simulation in which the AI system participated in the double-reading process that is used in the UK, and found that the AI system maintained non-inferior performance and reduced the workload of the second reader by 88%. This robust assessment of the AI system paves the way for clinical trials to improve the accuracy and efficiency of breast cancer screening.

Traditional metallic alloys are mixtures of elements in which the atoms of minority species tend to be distributed randomly if they are below their solubility limit, or to form secondary phases if they are above it. The concept of multiple-principal-element alloys has recently expanded this view, as these materials are single-phase solid solutions of generally equiatomic mixtures of metallic elements. This group of materials has received much interest owing to their enhanced mechanical properties(1-5). They are usually called medium-entropy alloys in ternary systems and high-entropy alloys in quaternary or quinary systems, alluding to their high degree of configurational entropy. However, the question has remained as to how random these solid solutions actually are, with the influence of short-range order being suggested in computational simulations but not seen experimentally(6,7). Here we report the observation, using energy-filtered transmission electron microscopy, of structural features attributable to short-range order in the CrCoNi medium-entropy alloy. Increasing amounts of such order give rise to both higher stacking-fault energy and hardness. These findings suggest that the degree of local ordering at the nanometre scale can be tailored through thermomechanical processing, providing a new avenue for tuning the mechanical properties of medium- and high-entropy alloys.

Metal alloys consisting of three or more major elemental components show enhanced mechanical properties, which are now shown to be correlated with short-range order observed with electron microscopy.

Halide perovskite materials have promising performance characteristics for low-cost optoelectronic applications. Photovoltaic devices fabricated from perovskite absorbers have reached power conversion efficiencies above 25 per cent in single-junction devices and 28 per cent in tandem devices(1,2). This strong performance (albeit below the practical limits of about 30 per cent and 35 per cent, respectively(3)) is surprising in thin films processed from solution at low-temperature, a method that generally produces abundant crystalline defects(4). Although point defects often induce only shallow electronic states in the perovskite bandgap that do not affect performance(5), perovskite devices still have many states deep within the bandgap that trap charge carriers and cause them to recombine non-radiatively. These deep trap states thus induce local variations in photoluminescence and limit the device performance(6). The origin and distribution of these trap states are unknown, but they have been associated with light-induced halide segregation in mixed-halide perovskite compositions(7) and with local strain(8), both of which make devices less stable(9). Here we use photoemission electron microscopy to image the trap distribution in state-of-the-art halide perovskite films. Instead of a relatively uniform distribution within regions of poor photoluminescence efficiency, we observe discrete, nanoscale trap clusters. By correlating microscopy measurements with scanning electron analytical techniques, we find that these trap clusters appear at the interfaces between crystallographically and compositionally distinct entities. Finally, by generating time-resolved photoemission sequences of the photo-excited carrier trapping process(10,11), we reveal a hole-trapping character with the kinetics limited by diffusion of holes to the local trap clusters. Our approach shows that managing structure and composition on the nanoscale will be essential for optimal performance of halide perovskite devices.

ABBV-744, a selective inhibitor of the BD2 domains of BET family proteins, is effective against prostate cancer in mouse xenograft models, with lower toxicities than the dual-bromodomain BET inhibitor ABBV-075.

Proteins of the bromodomain and extra-terminal (BET) domain family are epigenetic readers that bind acetylated histones through their bromodomains to regulate gene transcription. Dual-bromodomain BET inhibitors (DbBi) that bind with similar affinities to the first (BD1) and second (BD2) bromodomains of BRD2, BRD3, BRD4 and BRDt have displayed modest clinical activity in monotherapy cancer trials. A reduced number of thrombocytes in the blood (thrombocytopenia) as well as symptoms of gastrointestinal toxicity are dose-limiting adverse events for some types of DbBi(1-5). Given that similar haematological and gastrointestinal defects were observed after genetic silencing of Brd4 in mice(6), the platelet and gastrointestinal toxicities may represent on-target activities associated with BET inhibition. The two individual bromodomains in BET family proteins may have distinct functions(7-9) and different cellular phenotypes after pharmacological inhibition of one or both bromodomains have been reported(10,11), suggesting that selectively targeting one of the bromodomains may result in a different efficacy and tolerability profile compared with DbBi. Available compounds that are selective to individual domains lack sufficient potency and the pharmacokinetics properties that are required for in vivo efficacy and tolerability assessment(10-13). Here we carried out a medicinal chemistry campaign that led to the discovery of ABBV-744, a highly potent and selective inhibitor of the BD2 domain of BET family proteins with drug-like properties. In contrast to the broad range of cell growth inhibition induced by DbBi, the antiproliferative activity of ABBV-744 was largely, but not exclusively, restricted to cell lines of acute myeloid leukaemia and prostate cancer that expressed the full-length androgen receptor (AR). ABBV-744 retained robust activity in prostate cancer xenografts, and showed fewer platelet and gastrointestinal toxicities than the DbBi ABBV-075(14). Analyses of RNA expression and chromatin immunoprecipitation followed by sequencing revealed that ABBV-744 displaced BRD4 from AR-containing super-enhancers and inhibited AR-dependent transcription, with less impact on global transcription compared with ABBV-075. These results underscore the potential value of selectively targeting the BD2 domain of BET family proteins for cancer therapy.