A Hydrogel-coated Wood Membrane with Intelligent Oil Pollution Detection for Emulsion Separation.

Considering the potential threats posed by oily wastewater to the ecosystem, it is urgently in demand to develop efficient, eco-friendly, and intelligent oil/water separation materials to enhance the safety of the water environment. Herein, an intelligent hydrogel-coated wood (PPT/PPy@DW) membrane with self-healing, self-cleaning, and oil pollution detection performances is fabricated for the controllable separation of oil-in-water (O/W) emulsions and water-in-oil (W/O) emulsions. The PPT/PPy@DW is prepared by loading polypyrrole (PPy) particles on the delignified wood (DW) membranes, further modifying the hydrogel layer as an oil-repellent barrier. The layered porous structure and selective wettability endow PPT/PPy@DW with great separation performance for various O/W emulsions (≥98.69% for separation efficiency and ≈1000 L m-2 h-1 bar-1 for permeance). Notably, the oil pollution degree of PPT/PPy@DW can be monitored in real-time based on the changed voltage generated during O/W emulsion separation, and the oil-polluted PPT/PPy@DW can be self-cleaned by soaking in water to recover its separation performance. The high affinity of PPT/PPy@DW for water makes it effective in trapping water from the mixed surfactant-stabilized W/O emulsions. The prepared eco-friendly and low-cost multifunctional hydrogel wood membrane shows promising potential in on-demand oil/water separation and provides new ideas for the functional improvement of new biomass oil/water separation membrane materials.

Suggested blood donor deferral strategy regarding hepatitis B infections in China.

BACKGROUND: Hepatitis B virus (HBV) reactivity in individual immunologic and nucleic acid tests (NAT) tests does not represent the true infectious status of the blood donor. This study discusses the use of confirmatory tests to determine when deferral of blood donors is appropriate. METHODS: HBsAg or HBV NAT reactive samples were confirmed via a neutralisation test. All the HBsAg reactive but neutralisation test negative samples were subjected to further anti-HBc testing. The receiver operating characteristic curve was used to obtain the best threshold value using signal-to-cut-off ratios of two HBsAg enzyme-linked immunosorbent assay reagents. RESULTS: Of the 780 HBV reactive samples collected, there were 467 HBsAg reactive but HBV DNA negative samples, of which 65 (13.92%) and 402 (86.08%) were neutralisation test positive and negative, respectively. Of the 402, 91 samples (30% of tested samples) were anti-HBc reactive. HBV DNA positive specimens negative by virus neutralisation were >80% HBcAg positive. A screening strategy was proposed for Chinese blood collection agencies. CONCLUSION: These findings suggest that adopting a screening algorithm for deferring HBV reactive blood donors based on HBsAg and NAT testing followed with HBsAg S/CO consideration and HBcAg testing can be both safe and feasible in China.

Perilla frutescens L. alleviates trimethylamine N-oxide-induced apoptosis in the renal tubule by regulating ASK1-JNK phosphorylation.

Trimethylamine N-oxide (TMAO) is associated with overall mortality in patients with chronic kidney disease (CKD). Previous findings suggest that P. frutescens (L.) can alleviate renal injury, but its effects and mechanisms underlying alleviation of TMAO-induced kidney damage remain unclear. In this study, a TMAO injury model, in vivo and in vitro, was established to clarify the effects and mechanisms of P. frutescens in alleviating TMAO-induced kidney injury. The results show that TMAO (60 mM/L) can induce the activation of apoptosis signal-regulating kinase 1 (ASK1)-c-Jun N-terminal kinase (JNK), thus aggravating downstream cell apoptosis in vitro. The study also found that P. frutescens aqueous extract (PFAE) (5 mg/mL) can inhibit TMAO-induced apoptosis by downregulating ASK1-JNK phosphorylation. In the in vivo experiments, it was demonstrated that TMAO can increase the levels of blood urea nitrogen and cystatin C, aggravating renal tubular epithelial apoptosis. The results also show that PFAE can reduce TMAO-induced renal damage by inhibiting ASK1-JNK phosphorylation in vivo. Our findings confirmed that P. frutescens can alleviate TMAO-induced renal tubule apoptosis by regulating ASK1-JNK phosphorylation, indicating that P. frutescens may be an effective treatment for alleviating TMAO damage in CKD.

[Effect of traditional Chinese medicine in attenuating chronic kidney disease and its complications by regulating gut microbiota-derived metabolite trimethylamine N-oxide: a review].

Trimethylamine N-oxide(TMAO), a metabolite of gut microbiota, is closely associated with chronic kidney disease(CKD). It can aggravate the kidney injury and promote the occurrence of complications of CKD mainly by inducing renal fibroblast activation, vascular endothelial inflammation, macrophage foaming, platelet hyperreactivity, and inhibition of reverse cholesterol transport. Thus it is of great significance for clinical treatment of CKD to regulate circulating TMAO and alleviate its induced body damage. Currently, therapeutic strategies for TMAO regulation include dietary structure adjustment, lifestyle intervention, intestinal microflora regulation, and inhibition of intestinal trimethylamine synthesis and liver trimethylamine oxidation. Chinese medicinal herbs have the clinical advantage of multi-component and multi-target effects, and application of traditional Chinese medicine(TCM) to synergistically regulating TMAO and improving CKD via multiple pathways has broad prospects. This study systematically reviewed the clinical relevance and mechanism of TMAO in aggravating CKD renal function deterioration and complication progression. In addition, the effect and mechanism of TCM in improving TMAO-induced kidney injury, cardiovascular disease, hyperlipidemia, thrombosis and osteoporosis were summarized. The results provided a theoretical basis for TCM in attenuating gut microbiota-derived metabolite TMAO and improving CKD, as well as a basis and direction for in-depth clinical development and mechanism research in the future.

On-site detection of As(III) based on silver nanoparticles aggregation mediated by phosphates using surface-enhanced Raman scattering (SERS).

A portable and simple method was developed for on-site selective determination of As(III) based on the SERS signal of As(III)-O vibration. The method relies on the synergistic effect of nanoparticles aggregation and analyte adsorption. Experimental results demonstrated that phosphate replaced the ligands of HH@Ag NPs, which in turn facilitated the adsorption of As(III) on the surface of HH@Ag NPs. The phosphate was introduced as an agglomerating agent to improve the detection ability of the method for As(III). The method shows good selectivity and linear relationship between 5 × 10-8 and 0.8 × 10-6 M, with the detection limit of 1.8 × 10-9 M. The method was applied to actual water samples and successfully detected As(III), indicating that the method could have application potential in actual detection scenarios.

First annual report of Chinese haemovigilance network.

BACKGROUND AND OBJECTIVES: Haemovigilance involves surveillance of the whole chain of blood transfusion with the aim of identifying adverse events and errors and improving outcomes for patients. The Chinese Haemovigilance Network, founded in August 2017, has witnessed a rapid development in the last three years. MATERIALS AND METHODS: Based on the 1,022 cases in 2019, we analysed the adverse reactions (ARs) by blood component, clinical outcome severity and demography of recipients in an effort to publish the first annual Chinese haemovigilance report. RESULTS: The AR rate associated with blood transfusion in 2019 was 0·2% in China. Allergic reactions and FNHTR were the two most common adverse symptoms, accounting for 97·7% of the reports. Two-thirds of the TAD, AHTR and TACO and all of the HTR and DHTR resulted in hospitalization or prolongation of hospitalization. Plasma and AP were usually associated with allergic reaction (81·1%), whereas red cells more commonly cause FNHTR (68·8%) and all the AHTR, HTR, DSTR and DHTR. 84·1% of patients were aged 16 years or over, and the majority of the TAD, AHTR, TACO and HTR involved patients aged 60 and above. The ratio of serious adverse reactions (SARs) was 8·2%. Allergic reaction and FNHTR were top two (85·7%) SARs. The first case related to anti-D immunoglobulin was detected in a DHTR report. CONCLUSION: This report provides the world's first overview of transfusion-related adverse reactions in China. This report is useful for better understanding transfusion risks in China.

Comparison of digital subtraction angiography combined arterial thrombectomy versus simple arterial thrombectomy in the treatment of acute lower limb ischemia.

BACKGROUND: This study aimed to compare the clinical efficacy of digital subtraction angiography (DSA) combined arterial thrombectomy versus simple arterial thrombectomy in the treatment of acute lower limb ischemia (ALI). METHODS: This retrospective cohort study collected the clinical data from 124 patients (128 affected lower limbs) with ALI who underwent emergency surgery from March 2010 to November 2019. Patients were consecutively divided into Group A and Group B. Patients in Group A underwent simple arterial thrombectomy via the Fogarty catheterization. Patients in Group B underwent arterial thrombectomy, and the DSA was performed during the surgery. The differences in the success rate of primary surgery, the second intervention rate, and the amputation/mortality rate within 30-days after surgery were compared. RESULTS: In Group A, 4 of 70 limbs (5.7%) were amputated, 54 of 70 limbs (77.1%) had improved blood flow, 14 of 70 limbs (20.0%) received a second intervention, and 3 of 68 patients (4.4%) died within 30 days. In Group B, 1 of 58 limbs (1.7%) was amputated, 56 of 58 limbs (96.6%) had improved blood flow, 3 of 58 limbs (5.2%) received a second intervention, and 2 of 56 patients (3.5%) died within 30-days. The success rate of primary surgery, the second intervention rate, and the amputation rate of Group B were significantly lower than Group A (P < 0.05). CONCLUSION: Arterial thrombectomy combined with DSA may effectively improve the clinical efficacy of patients with ALI.

The prevalence of human T-cell leukemia virus in blood donors in China.

BACKGROUND: China has not yet incorporated routine human T-lymphotropic virus (HTLV)-1/2 blood donor screening, even though HTLV has been reported in the southeastern coastal region. This study was conducted to investigate the prevalence of HTLV in five major regions across of China. METHODS: From January 2016 to December 2017, blood samples were collected in 20 blood centers located in different regions of China. These samples were screened for HTLV-1/2 antibodies using enzyme-linked immunosorbent assay (ELISA). If the test samples were reactive, the samples were confirmed with a western blot (WB) assay. If the results of WB were indeterminate, the donor was interviewed after a minimum lapse of 8 weeks. All follow-up samples from donors were tested for anti-HTLV-1/2 with ELISA and WB. RESULTS: There were 875,453 donor samples tested for anti-HTLV-1/2 by ELISA. In all, 365 samples tested negative, 22 samples tested positive by WB, and 14 samples with HTLV status undetermined due to being lost to follow-up. The prevalences were 11.09, 5.96, 3.16, 2.88 and 0.98 per 100,000 in Xiamen, Changsha, Beijing, Shenzhen, and Nanjing blood center, respectively. The prevalences were 0 per 100,000 for all 15 other blood centers. There was significant differences in the prevalence of HTLV in different regions of China (p = 0.0011). CONCLUSION: In China, HTLV-1 confirmed positive donors are mainly from southeastern coastal areas. It may be necessary to conduct HTLV screening in these areas to reduce the risk of transfusion-transmitted HTLV.

Data Anomaly Detection for Internet of Vehicles Based on Traffic Cellular Automata and Driving Style.

The data validity of safe driving in the Internet of Vehicles (IoV) is the basis of improving the safety of vehicles. Different from a traditional information systems, the data anomaly analysis of vehicle safety driving faces the diversity of data anomaly and the randomness and subjectivity of the driver's driving behavior. How to combine the characteristics of the IOV data with the driving style analysis to provide effective real-time anomaly data detection has become an important issue in the IOV applications. This paper aims at the critical safety data analysis, considering the large computing cost generated by the real-time anomaly detection of all data in the data package. We preprocess it through the traffic cellular automata model which is built to achieve the ideal abnormal detection effect with limited computing resources. On the basis of this model, the Anomaly Detection based on Driving style (ADD) algorithm is proposed to realize real-time and online detection of anomaly data related to safe driving. Firstly, this paper designs the driving coefficient and proposes a driving style quantization model to represent the driving style of the driver. Then, based on driving style quantization model and vehicle driving state information, a data anomaly detection algorithm is developed by using Gaussian mixture model (GMM). Finally, combining with the application scenarios of multi-vehicle collaboration in the Internet of Vehicles, this paper uses real data sets and simulation data sets to analyze the effectiveness of the proposed ADD algorithm.

Sensitive and Selective Detection of Antibiotic D-Penicillamine Based on a Dual-Mode Probe of Fluorescent Carbon Dots and Gold Nanoparticles.

This paper reported a dual-mode probe for D-penicillamine on the basis of pH-mediated gold nanoparticles aggregation and fluorescence resonance energy transfer (FRET) from carbon dots. D-penicillamine is a zwitterionic compound and has different forms depending on specific pH ranges. The thiol group of D-penicillamine has high affinity towards the surface of gold nanoparticles and can replace other surface ligands. When pH values were close to its isoelectrical point (pH(I)), the D-penicillamine capped gold nanoparticles aggregated through hydrogen bonding or electrostatic interactions, resulting in the releasing of carbon dots from gold nanoparticles. The dual-mode probe consisted of fluorescent carbon dots and gold nanoparticles, and the fluorescence of carbon dots was quenched by the attached gold nanoparticles due to the FRET. Then, the fluorescence can be recovered in presence of D-penicillamine due to the gold nanoparticles aggregation in specific pH range. Under the optimum conditions, the probe has linear response for D-penicillamine in the 0.25-1.5 µM concentration range with a detection limit of 0.085 µM. This method provides a potential application in sensitive detection of D-penicillamine.

Fast batch searching for protein homology based on compression and clustering.

BACKGROUND: In bioinformatics community, many tasks associate with matching a set of protein query sequences in large sequence datasets. To conduct multiple queries in the database, a common used method is to run BLAST on each original querey or on the concatenated queries. It is inefficient since it doesn't exploit the common subsequences shared by queries. RESULTS: We propose a compression and cluster based BLASTP (C2-BLASTP) algorithm to further exploit the joint information among the query sequences and the database. Firstly, the queries and database are compressed in turn by procedures of redundancy analysis, redundancy removal and distinction record. Secondly, the database is clustered according to Hamming distance among the subsequences. To improve the sensitivity and selectivity of sequence alignments, ten groups of reduced amino acid alphabets are used. Following this, the hits finding operator is implemented on the clustered database. Furthermore, an execution database is constructed based on the found potential hits, with the objective of mitigating the effect of increasing scale of the sequence database. Finally, the homology search is performed in the execution database. Experiments on NCBI NR database demonstrate the effectiveness of the proposed C2-BLASTP for batch searching of homology in sequence database. The results are evaluated in terms of homology accuracy, search speed and memory usage. CONCLUSIONS: It can be seen that the C2-BLASTP achieves competitive results as compared with some state-of-the-art methods.

An "on-off-on" fluorescent probe for ascorbic acid based on Cu-ZnCdS quantum dots and α-MnO2 nanorods.

This work established a fluorescence approach for detecting ascorbic acid (AA) based on Cu-ZnCdS quantum dots (Cu-ZnCdS QDs) and α-MnO2 nanorods. Cu-ZnCdS QDs and α-MnO2 nanorods were characterized by high-resolution transmission electron microscopy (HRTEM), fluorescence spectroscopy, inductively coupled plasma optical emission spectroscopy (ICP-OES) and X-ray diffraction (XRD). In the presence of α-MnO2 nanorods, the fluorescence of Cu-ZnCdS QDs was greatly quenched through the inner filter effect (IFE). Subsequently, AA can trigger the decomposition of the α-MnO2 nanorods which can reduce α-MnO2 to Mn2+ and recover the fluorescence. Under optimal conditions, a linear relation was obtained over the range 5.02-401.77 µM with a 31.62 µM detection limit. Through applying the fluorescent sensing system for detecting AA, a satisfactory result is obtained with recoveries ranging from 89.23% to 110.99%.

Transition metal (Co, Ni) nanoparticles wrapped with carbon and their superior catalytic activities for the reversible hydrogen storage of magnesium hydride.

Magnesium hydride (MgH2) exhibits long-term stability and has recently been developed as a safe alternative to store hydrogen in the solid state, due to its high capacity of 7.6 wt% H2 and low cost compared to other metal hydrides. However, the high activation energy and poor kinetics of MgH2 lead to inadequate hydrogen storage properties, resulting in low energy efficiency. Nano-catalysis is deemed to be the most effective strategy in improving the kinetics performance of hydrogen storage materials. In this work, robust and efficient architectures of carbon-wrapped transition metal (Co/C, Ni/C) nanoparticles (8-16 nm) were prepared and used as catalysts in the MgH2 system via ball milling to improve its de/rehydrogenation kinetics. Between the two kinds of nano-catalysts, the Ni/C nanoparticles exhibit a better catalytic efficiency. MgH2 doped with 6% Ni/C (MgH2-6%Ni/C) exhibits a peak dehydrogenation temperature of 275.7 °C, which is 142.7, 54.2 and 32.5 °C lower than that of commercial MgH2, milled MgH2 and MgH2 doped with 6% Co/C (MgH2-6%Co/C), respectively. MgH2 doped with 6% Ni/C can release about 6.1 wt% H2 at 250 °C. More importantly, the dehydrogenated MgH2-6%Ni/C is even able to uptake 5.0 wt% H2 at 100 °C within 20 s. Moreover, a cycling test of MgH2 doped with 8% Ni/C demonstrates its excellent hydrogen absorption/desorption stability with respect to both capacity (up to 6.5 wt%) and kinetics (within 8 min at 275 °C for dehydrogenation and within 10 s at 200 °C for rehydrogenation). Mechanistic research reveals that the in situ formed Mg2Ni and Mg2NiH4 nanoparticles can be regarded as advanced catalytically active species in the MgH2-Ni/C system. Meanwhile, the carbon attached around the surface of transition metal nanoparticles can successfully inhibit the aggregation of the catalysts and achieve the steadily, prompting de/rehydrogenation during the subsequent cycling process. The intrinsic catalytic effects and the uniform distributions of Mg2Ni and Mg2NiH4 result in a favorable catalytic efficiency and cycling stability. Nano-catalysts with this kind of morphology can also be applied to other metal hydrides to improve their kinetics performance and cycling stability.

Evaluation of three-dimensional printing for laparoscopic partial nephrectomy of renal tumors: a preliminary report.

OBJECTIVES: To investigate the impact of three-dimensional (3D) printing on the surgical planning, potential of training and patients' comprehension of minimally invasive surgery for renal tumors. METHODS: Patients of a T1N0M0 single renal tumor and indicated for laparoscopic partial nephrectomy were selected. CT data were sent for post-processing and output to the 3D printer to create kidney models with tumor. By presenting to experienced laparoscopic urologists and patients, respectively, the models' realism, effectiveness for surgical planning and training, and patients' comprehension of disease and procedure were evaluated with plotted questionnaires (10-point rating scales, 1-not at all useful/not at all realistic/poor, 10-very useful/very realistic/excellent). The size of resected tumors was compared with that on the models. RESULTS: Ten kidney models of such patients were fabricated successfully. The overall effectiveness in surgical planning and training (7.8 ± 0.7-8.0 ± 1.1), and realism (6.0 ± 0.6-7.8 ± 1.0) were reached by four invited urologists. Intraoperative correlation was advocated by the two performing urologists. Patients were fascinated with the demonstration of a tactile "diseased organ" (average ≥ 9.0). The size deviation was 3.4 ± 1.3 mm. CONCLUSIONS: Generating kidney models of T1N0M0 tumors with 3D printing are feasible with refinements to be performed. Face and content validity was obtained when those models were presented to experienced urologists for making practical planning and training. Understandings of the disease and procedure from patients were well appreciated with this novel technology.

Evaluation of an individual-donation nucleic acid amplification testing algorithm for detecting hepatitis B virus infection in Chinese blood donors.

BACKGROUND: This multicenter study was performed to evaluate the efficiency of a multiplex individual-donation nucleic acid amplification technology (ID-NAT) and discriminatory testing algorithm for detecting hepatitis B virus (HBV) infection in Chinese blood donors. STUDY DESIGN AND METHODS: A total of 1,205,796 hepatitis B surface antigen (HBsAg)-nonreactive donations from 10 blood centers were tested by ID-NAT using the Ultrio assay. Multiplex Ultrio-reactive donations were tested in the discriminatory tests as well as in quantitative polymerase chain reaction (qPCR) and in supplemental electrochemiluminescence immunoassays for HBsAg, hepatitis B surface antibody (anti-HBs), hepatitis B e antigen, and antibody to hepatitis B core antigen (anti-HBc). Meanwhile, a control group of 4317 Ultrio-nonreactive donations was tested for anti-HBc and anti-HBs. RESULTS: Of all donations, 2033 (0.17%) were reactive in the multiplex Ultrio assay. Among 1776 further tested samples, 548 (30.9%) were HBV discriminatory assay (dHBV)-reactive, while 1214 (68.4%) were nonreactive. Of 472 Ultrio+ and dHBV+ samples 86.2% were qPCR positive compared to 15.0% in 1046 Ultrio+ and dHBV- samples. The proportion of anti-HBc+ and anti-HBs- (potentially infectious) donations was higher in 409 Ultrio+ and dHBV+ than in 1028 Ultrio+ and dHBV- samples (51.3% vs. 31.1%, p < 0.001). The yield rate of Ultrio+, dHBV+, and qPCR+ donations was estimated at 1 in 2500, but at 1 in 1100 when all supplemental tests were taken into account assuming that 44% of detected donations by Ultrio were false reactive. CONCLUSIONS: A quarter of HBsAg-negative Ultrio+ and dHBV- donations in China are likely given by potentially infectious low-viral-load occult carriers. Although this has no implication for blood safety, the testing algorithm needs to be redesigned to more efficiently discriminate between true and false NAT reactivity.

Joule Heating-Assisted Crude Oil Purification by a Poly(pyrrole)-Modified Microfibril Cellulose Membrane.

Using membrane materials to purify viscous watery oil from industrial production processes and accidental oil spills is of great importance but still challenging. Based on the excellent electrical conductivity and electric-thermal conversion of poly(pyrrole) (PPy), a hydrophobic PPy-modified micro-fibrillated cellulose membrane (P-CP) was successfully prepared. The size of the P-CP membrane can be customized to meet specific requirements. In this research, the membrane diameter is capable of reaching 24 cm. By applying a voltage ranging from 0 to 12 V, the surface temperature of the P-CP membrane can be elevated to roughly 120 °C. After 10 cycles of heating and cooling under 12 V voltage, the electric-thermal curves, surface hydrophobicity, and pore structure of P-CP membrane can remain stable, which suggests remarkable electric-thermal stability and reliability despite prolonged operation. The P-CP membrane shows good linearity between voltage and current (R2 = 0.997) and easy temperature control from room temperature to ∼120 °C at low supply voltage (0-12 V). Under the condition of 12 V power supply and self-gravity, the separation flux of the P-CP membrane for water-in-oil (W/O) emulsions (kerosene, diesel) is 2-3 times higher than that at room temperature, and the separation efficiency is also improved. Importantly, the P-CP membrane shows excellent separation performance for high viscosity water-in-crude oil emulsions, with a separation flux of 40 L m-2 h-1 by gravity. Compared to the situation without electricity, the separation flux of water-in-crude oil emulsion has increased four-fold. The joule heating of the P-CP membrane expands its service time and application scenarios, demonstrating its great application prospects in actual viscous oil-water emulsion separation.

Structural evolution of metatitanic acid and iron removal during hydrolysis of industrial TiOSO4 solution.

The ferrous ion content of metatitanic acid affected the whiteness, purity and applications of TiO2, controlled by the hydrolysis conditions and metatitanic acid structure. The structural evolution of metatitanic acid and ferrous ion removal was investigated by hydrolyzing the industrial TiOSO4 solution. The hydrolysis degree was conformed to Boltzmann model with good fitting. TiO2 content of metatitanic acid gradually increased as hydrolysis proceeding due to its stronger compact structure and weaker colloidal property, caused by the aggregation and adjustment of the precipitated particles. At lower TiOSO4 concentration, the crystal size increased significantly, lattice strain decreased, and average particle size constantly adjusted and reduced. The micropores and mesopores were mainly formed by aggregating and stacking of primary agglomerate particles, bonded and filled with sulfate and hydroxyl. The ferrous ion content decreased linearly with the increase of TiO2 content, and reducing moisture content of metatitanic acid was an effective way to reduce Fe content. This would save more water and energy consumption, help to improve the clean production level of TiO2.

Cross-attention multi-branch CNN using DCE-MRI to classify breast cancer molecular subtypes.

Purpose: The aim of this study is to improve the accuracy of classifying luminal or non-luminal subtypes of breast cancer by using computer algorithms based on DCE-MRI, and to validate the diagnostic efficacy of the model by considering the patient's age of menarche and nodule size. Methods: DCE-MRI images of patients with non-specific invasive breast cancer admitted to the Second Affiliated Hospital of Dalian Medical University were collected. There were 160 cases in total, with 84 cases of luminal type (luminal A and luminal B and 76 cases of non-luminal type (HER 2 overexpressing and triple negative). Patients were grouped according to thresholds of nodule sizes of 20 mm and age at menarche of 14 years. A cross-attention multi-branch net CAMBNET) was proposed based on the dataset to predict the molecular subtypes of breast cancer. Diagnostic performance was assessed by accuracy, sensitivity, specificity, F1 and area under the ROC curve (AUC). And the model is visualized with Grad-CAM. Results: Several classical deep learning models were included for diagnostic performance comparison. Using 5-fold cross-validation on the test dataset, all the results of CAMBNET are significantly higher than the compared deep learning models. The average prediction recall, accuracy, precision, and AUC for luminal and non-luminal types of the dataset were 89.11%, 88.44%, 88.52%, and 96.10%, respectively. For patients with tumor size <20 mm, the CAMBNET had AUC of 83.45% and ACC of 90.29% for detecting triple-negative breast cancer. When classifying luminal from non-luminal subtypes for patients with age at menarche years, our CAMBNET model achieved an ACC of 92.37%, precision of 92.42%, recall of 93.33%, F1of 92.33%, and AUC of 99.95%. Conclusions: The CAMBNET can be applied in molecular subtype classification of breasts. For patients with menarche at 14 years old, our model can yield more accurate results when classifying luminal and non-luminal subtypes. For patients with tumor sizes ≤20 mm, our model can yield more accurate result in detecting triple-negative breast cancer to improve patient prognosis and survival.