Effects of a comprehensive brain computed tomography deep learning model on radiologist detection accuracy.

OBJECTIVES: Non-contrast computed tomography of the brain (NCCTB) is commonly used to detect intracranial pathology but is subject to interpretation errors. Machine learning can augment clinical decision-making and improve NCCTB scan interpretation. This retrospective detection accuracy study assessed the performance of radiologists assisted by a deep learning model and compared the standalone performance of the model with that of unassisted radiologists. METHODS: A deep learning model was trained on 212,484 NCCTB scans drawn from a private radiology group in Australia. Scans from inpatient, outpatient, and emergency settings were included. Scan inclusion criteria were age ≥ 18 years and series slice thickness ≤ 1.5 mm. Thirty-two radiologists reviewed 2848 scans with and without the assistance of the deep learning system and rated their confidence in the presence of each finding using a 7-point scale. Differences in AUC and Matthews correlation coefficient (MCC) were calculated using a ground-truth gold standard. RESULTS: The model demonstrated an average area under the receiver operating characteristic curve (AUC) of 0.93 across 144 NCCTB findings and significantly improved radiologist interpretation performance. Assisted and unassisted radiologists demonstrated an average AUC of 0.79 and 0.73 across 22 grouped parent findings and 0.72 and 0.68 across 189 child findings, respectively. When assisted by the model, radiologist AUC was significantly improved for 91 findings (158 findings were non-inferior), and reading time was significantly reduced. CONCLUSIONS: The assistance of a comprehensive deep learning model significantly improved radiologist detection accuracy across a wide range of clinical findings and demonstrated the potential to improve NCCTB interpretation. CLINICAL RELEVANCE STATEMENT: This study evaluated a comprehensive CT brain deep learning model, which performed strongly, improved the performance of radiologists, and reduced interpretation time. The model may reduce errors, improve efficiency, facilitate triage, and better enable the delivery of timely patient care. KEY POINTS: • This study demonstrated that the use of a comprehensive deep learning system assisted radiologists in the detection of a wide range of abnormalities on non-contrast brain computed tomography scans. • The deep learning model demonstrated an average area under the receiver operating characteristic curve of 0.93 across 144 findings and significantly improved radiologist interpretation performance. • The assistance of the comprehensive deep learning model significantly reduced the time required for radiologists to interpret computed tomography scans of the brain.

Evaluating Starter Feeding on Ruminal Function in Yak Calves: Combined 16S rRNA Sequencing and Metabolomics.

For young ruminants, starter feeding can effectively facilitate the growth and development of rumen in ruminants, but the development of rumen is an important physiological challenge as it remains unclear for the mechanism of starter feeding stimulating. In this study, we performed an analysis of ruminal microbiota and their metabolites in yak calves to explore how the ruminal microbiota and their metabolites stimulate the ruminal function. This study associated 16S rRNA sequencing with liquid chromatography-mass spectrometry (LC-MS)-based metabolomics to evaluate the effects of starter feeding on ruminal microbiota diversity and metabolites in yak calves. We designed the experiment using 20 yak calves that were assigned equally into 2 groups, based on feeding milk replacer; the control (RA) group was fed with alfalfa hay while the treatment (RAS) group was fed with alfalfa hay and starter. After the experiment, we investigated the ruminal microbiota and metabolites through 16S rRNA sequencing and LC-MS-based metabolomics. During the preweaning period, the RAS group significantly promoted the growth performance and ruminal development in yak calves, including increases in body weight, chest girth, and development of rumen (P < 0.05). The RAS group increased the relative abundance of Bacteroidota, Proteobacteria, Chloroflexi, Synergistota, and Spirochaetota and decreased the abundance of Firmicutes, Desulfobacterota, Actinobacteriota, and Actinobacteriota at the phylum level (P < 0.05). At the genus level, the ruminal content of the RAS group was significantly enriched for Rikenellaceae_RC9_gut_group and Ruminococcus, while depleted for Prevotella, Christensenellaceae_R-7_group, and NK4A214_group (P < 0.05). A total of 37 metabolites were identified between the RA group and the RAS group, of which 15 metabolites were upregulated and 22 metabolites were downregulated compared with the RA group. Metabolic pathway analyses indicated that upregulated the metabolites of the RAS group yak calves were related to carbohydrate metabolism, ubiquinone, and other terpenoid-quinone biosynthesis, while the downregulated metabolic pathway was relevant to xenobiotic biodegradation, metabolism, and nucleotide metabolism. In summary, starter feeding before weaning significantly increased the dry matter intake and body weight of yak calves, changed the diversity and abundance of ruminal microbiota, and positively regulated the good development of ruminal morphology and function, providing an important basis for high-quality cultivation and the nutritional level of nutrition of yak calves in the Qinghai Tibet plateau. This study is based on the availability of 16S rRNA sequencing and LC-MS-based metabolomics in clarifying the function of starter feeding in the yak calves.

Acoustic-optical phonon up-conversion and hot-phonon bottleneck in lead-halide perovskites.

The hot-phonon bottleneck effect in lead-halide perovskites (APbX3) prolongs the cooling period of hot charge carriers, an effect that could be used in the next-generation photovoltaics devices. Using ultrafast optical characterization and first-principle calculations, four kinds of lead-halide perovskites (A=FA+/MA+/Cs+, X=I-/Br-) are compared in this study to reveal the carrier-phonon dynamics within. Here we show a stronger phonon bottleneck effect in hybrid perovskites than in their inorganic counterparts. Compared with the caesium-based system, a 10 times slower carrier-phonon relaxation rate is observed in FAPbI3. The up-conversion of low-energy phonons is proposed to be responsible for the bottleneck effect. The presence of organic cations introduces overlapping phonon branches and facilitates the up-transition of low-energy modes. The blocking of phonon propagation associated with an ultralow thermal conductivity of the material also increases the overall up-conversion efficiency. This result also suggests a new and general method for achieving long-lived hot carriers in materials.