Explainable AI Method for Tinnitus Diagnosis via Neighbor-Augmented Knowledge Graph and Traditional Chinese Medicine: Development and Validation Study.

BACKGROUND: Tinnitus diagnosis poses a challenge in otolaryngology owing to an extremely complex pathogenesis, lack of effective objectification methods, and factor-affected diagnosis. There is currently a lack of explainable auxiliary diagnostic tools for tinnitus in clinical practice. OBJECTIVE: This study aims to develop a diagnostic model using an explainable artificial intelligence (AI) method to address the issue of low accuracy in tinnitus diagnosis. METHODS: In this study, a knowledge graph-based tinnitus diagnostic method was developed by combining clinical medical knowledge with electronic medical records. Electronic medical record data from 1267 patients were integrated with traditional Chinese clinical medical knowledge to construct a tinnitus knowledge graph. Subsequently, weights were introduced, which measured patient similarity in the knowledge graph based on mutual information values. Finally, a collaborative neighbor algorithm was proposed, which scored patient similarity to obtain the recommended diagnosis. We conducted 2 group experiments and 1 case derivation to explore the effectiveness of our models and compared the models with state-of-the-art graph algorithms and other explainable machine learning models. RESULTS: The experimental results indicate that the method achieved 99.4% accuracy, 98.5% sensitivity, 99.6% specificity, 98.7% precision, 98.6% F1-score, and 99% area under the receiver operating characteristic curve for the inference of 5 tinnitus subtypes among 253 test patients. Additionally, it demonstrated good interpretability. The topological structure of knowledge graphs provides transparency that can explain the reasons for the similarity between patients. CONCLUSIONS: This method provides doctors with a reliable and explainable diagnostic tool that is expected to improve tinnitus diagnosis accuracy.

Combined Hydroxyethyl Starch Luteolin Nanocrystals for Effective Anti-Hyperuricemia Effect in Mice Model.

Introduction: Although flavonoid compounds exhibit various pharmacological activities, their clinical applications are restricted by low oral bioavailability owing to their poor solubility. Nanocrystals (NCs) represent an excellent strategy for enhancing the oral bioavailability of flavonoids. Hydroxyethyl starch (HES), a biomaterial compound used as a plasma expander, could be an ideal stabilizer material for preparing flavonoid NCs. Methods: HES was used to stabilize flavonoid nanocrystals (NCs), using luteolin (LUT) as a model drug. After full characterization, the freeze-drying and storage stability, solubility, intestinal absorption, pharmacokinetics, and in vivo anti-hyperuricemic effect of the optimized HES-stabilized LUT NCs (LUT-HES NCs) were investigated. Results: Uniformed LUT-HES NCs were prepared with mean particle size of 191.1±16.8 nm, zeta potential of about -23 mV, drug encapsulation efficiency of 98.52 ± 1.01%, and drug loading of 49.26 ± 0.50%. The freeze-dried LUT-HES NCs powder showed good re-dispersibility and storage stability for 9 months. Notably, compared with the coarse drug, LUT-HES NCs exhibited improved saturation solubility (7.49 times), increased drug dissolution rate, enhanced Caco-2 cellular uptake (2.78 times) and oral bioavailability (Fr=355.7%). Pharmacodynamic studies showed that LUT-HES NCs remarkably lowered serum uric acid levels by 69.93% and ameliorated renal damage in hyperuricemic mice. Conclusion: HES is a potential stabilizer for poorly soluble flavonoid NCs and provides a promising strategy for the clinical application of these compounds. LUT-HES NCs may be an alternative or complementary strategy for hyperuricemia treatment.

Ultrahigh-response hydrogen sensor based on PdO/NiO co-doped In2O3 nanotubes.

Hydrogen can be regarded as an ideal type of secondary energy considering its potential for achieving renewable and sustainable development due to water being its sole combustion product and its possible production by solar energy-based water electrolysis. Monitoring the presence and concentration of hydrogen during production, transportation, and application requires a hydrogen gas sensor with high response, high selectivity, and fast response and recovery times. In an attempt to meet these requirements, NiO and PdO are used in the co-doping of In2O3 nanotubes by subsequent electrospinning and impregnation under UV irradiation. The fabricated hydrogen gas sensor demonstrates an ultrahigh response of 487.52, a fast response time of 1 s and high selectivity at an operating temperature of 160 °C, which characteristics are superior to reported monometal-doped hydrogen sensors. The remarkable gas sensing performance could be attributed to the synergistic effect of the resistance modulation, the chemical sensitization of PdO, and the catalytic effect of NiO. This study demonstrates that co-doping of PdO and NiO on In2O3 nanotubes is an effective way to improve hydrogen sensing characteristics more effectively than doping with PdO or NiO alone, and provides a potential application for the fast and accurate detection of hydrogen.

Polymer Microchannel and Micromold Surface Polishing for Rapid, Low-Quantity Polydimethylsiloxane and Thermoplastic Microfluidic Device Fabrication.

Polymer-based micromolding has been proposed as an alternative to SU-8 micromolding for microfluidic chip fabrication. However, surface defects such as milling marks may result in rough microchannels and micromolds, limiting microfluidic device performance. Therefore, we use chemical and mechanical methods for polishing polymer microchannels and micromolds. In addition, we evaluated their performance in terms of removing the machining (milling) marks on polymer microchannel and micromold surfaces. For chemical polishing, we use solvent evaporation to polish the sample surfaces. For mechanical polishing, wool felt polishing bits with an abrasive agent were employed to polish the sample surfaces. Chemical polishing reduced surface roughness from 0.38 µm (0 min, after milling) to 0.13 µm after 6 min of evaporation time. Mechanical polishing reduced surface roughness from 0.38 to 0.165 µm (optimal pressing length: 0.3 mm). As polishing causes abrasion, we evaluated sample geometry loss after polishing. Mechanically and chemically polished micromolds had optimal micromold distortion percentages of 1.01% ± 0.76% and 1.10% ± 0.80%, respectively. Compared to chemical polishing, mechanical polishing could better maintain the geometric integrity since it is locally polished by computer numerical control (CNC) miller. Using these surface polishing methods with optimized parameters, polymer micromolds and microchannels can be rapidly produced for polydimethylsiloxane (PDMS) casting and thermoplastic hot embossing. In addition, low-quantity (15 times) polymer microchannel replication is demonstrated in this paper.

CircRNA_000864 Upregulates B-cell Translocation Gene 2 Expression and Represses Migration and Invasion in Pancreatic Cancer Cells by Binding to miR-361-3p.

BACKGROUND: Pancreatic cancer is a fatal disease with a very poor prognosis due to its characteristic insidious symptoms, early metastasis, and chemoresistance. Circular RNAs (circRNAs) are involved in the development of pancreatic cancer. AIM: Hence, the aim of this study is to elucidate the mechanism of circRNA_000864 in regulating BTG2 expression in pancreatic cancer by binding to miR-361-3p. METHODS: CircRNA_000864, miR-361-3p, and BTG2 expression patterns in the pancreatic cancer tissues were detected by RT-qPCR. Correlation among circRNA_000864, miR-361-3p, and BTG2 was evaluated by RNA-pull down assay, RNA Immunoprecipitation assay, and dual-luciferase reporter gene assay. After ectopic expression and depletion experiments, 5-ethynyl-2'-deoxyuridine assay, Transwell assay, and flow cytometry were employed to assess the cell proliferation, migration and invasion, cell cycle, and apoptosis. Xenotransplantation of nude mice was conducted to detect the effects of circRNA_000864, miR-361-3p, and BTG2 on tumor growth. RESULTS: CircRNA_000864 and BTG2 were poorly expressed, and miR-361-3p was highly expressed in the pancreatic cancer tissues. CircRNA_000864 bound to miR-361-3p could target BTG2. Cell proliferation, migration, and invasion were inhibited, and the cell cycle arrest and apoptosis were stimulated after overexpression of circRNA_000864 or BTG2 or downregulation of miR-361-3p. Overexpression of circRNA_000864 or downregulation of miR-361-3p also decreased the tumor growth in vivo. CONCLUSIONS: Conjointly, our findings elicited that the overexpression of circRNA_000864 could promote BTG2 expression to inhibit pancreatic cancer development by binding to miR-361-3p, which represents an appealing therapeutic target for the treatment of pancreatic cancer.