Effect of modification methods on the physical properties and immunomodulatory activity of particulate ß-glucan.

ß-Glucan is an immunoenhancing agent whose biological activities are linked to molecular structure. On that basis, the polysaccharide can be physiochemically modified to produce valuable functional materials. This study investigated the physical properties and immunostimulatory activity of modified ß-glucan. Alkali-treated ß-glucan had a distinct shape and smaller particle size than untreated ß-glucan. The reduced particle size was conducive to the stability of the suspension because the ß-glucan appeared to be completely dissolved by this treatment, forming an amorphous mass. Furthermore, alkali treatment improved the immunostimulating activity of ß-glucan, whereas exposure of macrophages to heat-treated ß-glucan decreased their immune activity. ß-Glucan with reduced particle size by wet-grinding also displayed immunomodulatory activities. These results suggested that the particle size of ß-glucan is a key factor in ß-glucan-induced immune responses of macrophages. Thus, the modification of the ß-glucan particle size provides new opportunities for developing immunoenhancing nutraceuticals or pharmacological therapies in the future.

Current technologies for heavy metal removal from food and environmental resources.

Exposure to heavy metals in water and food poses a significant threat to human well-being, necessitating the efficient removal of these contaminants. The process of urban development exacerbates heavy metal pollution, thereby increasing risks to both human health and ecosystems. Heavy metals have the capacity to enter the food chain, undergo bioaccumulation and magnify, ultimately resulting in adverse effects on human health. Therefore, implementing effective pollution control measures and adopting sustainable practices are crucial for mitigating exposure and associated health risks. Various innovative approaches, including adsorption, ion exchange, and electrochemical technology, are currently being actively investigated to cope with the issue of heavy metal contamination. These innovative methods offer benefits such as efficient recycling, cost-effectiveness and environmental friendliness. In this review, we summarize recent advances for removing heavy metals from water, soil and food, providing valuable guidance for environmental engineers and researchers seeking to address contamination challenges.

Effects of Icariin and Its Metabolites on GPCR Regulation and MK-801-Induced Schizophrenia-Like Behaviors in Mice.

Icariin, a major bioactive compound found in the Epimedium genus, has been reported to exert protective effects against neurodegenerative disorders. In the current study, we aimed to investigate the regulatory effect of icariin and its active metabolites (icariside II and icaritin) against prime G-protein-coupled receptor targets, considering their association with neuronal disorders. Icariside II exhibited selective agonist activity towards the dopamine D3 receptor (D3R), with half-maximal effective concentrations of 13.29 µM. Additionally, they effectively inhibited the specific binding of radioligands to D3R. Molecular docking analysis revealed that icariside II potentially exerts its agonistic effect through hydrogen-bonding interaction with Asp110 of the D3R, accompanied by negative binding energy. Conversely, icaritin demonstrated selective antagonist effects on the muscarinic acetylcholine M2 receptor (M2R). Radioligand binding assay and molecular docking analysis identified icaritin as an orthosteric ligand for M2R. Furthermore, all three compounds, icariin and its two metabolites, successfully mitigated MK-801-induced schizophrenia-like symptoms, including deficits in prepulse inhibition and social interaction, in mice. In summary, these findings highlight the potential of icariin and its metabolites as promising lead structures for the discovery of new drugs targeting cognitive and neurodegenerative disorders.

Effects of Interfacial Electron Transport on Field Electron Emission from Carbon Nanotube Paste Emitters.

Field electron emission from carbon nanotubes (CNT) is preceded by the transport of electrons from the cathode metal to emission sites. Specifically, a supporting layer indispensable for adhesion of CNT paste emitters onto the cathode metal would impose a potential barrier, depending on its work function and interfacial electron transport behaviors. In this paper, we investigated the supporting layer of silicon carbide and nickel nanoparticles reacted onto a Kovar alloy (Fe-Ni-Co) cathode substrate, which has been adopted for reliable CNT paste emitters. The X-ray diffraction, X-ray photoelectron spectroscopy, ultraviolet photoelectron spectroscopy, and electrical conductivity measurements showed that the reaction of silicon carbide and nickel nanoparticles on the Kovar metal strongly depends upon the post-vacuum-annealing conditions and can be classified into two procedures of a diffusion-induced reaction (DIR) and a diffusion-limited reaction (DLR). The prolonged annealing at 750 °C for 5 h before the main annealing of the CNT paste emitters at 800 °C for 5 min led to the DIR that has enhanced the Ni silicide phase and a lower potential barrier for the interfacial electron transport, resulting in increased and weakly temperature-dependent field electron emission from the CNT paste emitters. On the other hand, the DLR with only the main anneal of the CNT paste emitters at 800 °C for 5 min gave rise to a higher potential barrier for the electron transport and so lower and strongly temperature-dependent field electron emission. From the results of the interfacial electron transport for the DIR and DLR mechanisms in the CNT paste emitters, we concluded that the ambient temperature dependency of field electron emission from CNT tips in the moderate range of up to 400 °C, still controversial, is mainly attributed to the supporting layer of the CNT emitter rather than its intrinsic electron emission.

Neuromodulatory feasibility of a current limiter-based tDCS device: a resting-state electroencephalography study.

Recently, we introduced a current limiter-based novel transcranial direct-current stimulation (tDCS) device that does not generate significant tDCS-induced electrical artifacts, thereby facilitating simultaneous electroencephalography (EEG) measurement during tDCS application. In this study, we investigated the neuromodulatory effect of the tDCS device using resting-state EEG data measured during tDCS application in terms of EEG power spectral densities (PSD) and brain network indices (clustering coefficient and path length). Resting-state EEG data were recorded from 10 healthy subjects during both eyes-open (EO) and eyes-closed (EC) states for each of five different conditions (baseline, sham, post-sham, tDCS, and post-tDCS). In the tDCS condition, tDCS was applied for 12 min with a current intensity of 1.5 mA, whereas tDCS was applied only for the first 30 s in the sham condition. EEG PSD and brain network indices were computed for the alpha frequency band most closely associated with resting-state EEG. Both alpha PSD and network indices were found to significantly increase during and after tDCS application compared to those of the baseline condition in the EO state, but not in the EC state owing to the ceiling effect. Our results demonstrate the neuromodulatory effect of the tDCS device that does not generate significant tDCS-induced electrical artifacts, thereby allowing simultaneous measurement of electrical brain activity. We expect our novel tDCS device to be practically useful in exploring the impact of tDCS on neuromodulation more precisely using ongoing EEG data simultaneously measured during tDCS application.

Machine Learning-Based Models for Prediction of Critical Illness at Community, Paramedic, and Hospital Stages.

Overcrowding of emergency department (ED) has put a strain on national healthcare systems and adversely affected the clinical outcomes of critically ill patients. Early identification of critically ill patients prior to ED visits can help induce optimal patient flow and allocate medical resources effectively. This study aims to develop ML-based models for predicting critical illness in the community, paramedic, and hospital stages using Korean National Emergency Department Information System (NEDIS) data. Random forest and light gradient boosting machine (LightGBM) were applied to develop predictive models. The predictive model performance based on AUROC in community stage, paramedic stage, and hospital stage was estimated to be 0.870 (95% CI: 0.869-0.871), 0.897 (95% CI: 0.896-0.898), and 0.950 (95% CI: 0.949-0.950) in random forest and 0.877 (95% CI: 0.876-0.878), 0.899 (95% CI: 0.898-0.900), and 0.950 (95% CI: 0.950-0.951) in LightGBM, respectively. The ML models showed high performance in predicting critical illness using variables available at each stage, which can be helpful in guiding patients to appropriate hospitals according to their severity of illness. Furthermore, a simulation model can be developed for proper allocation of limited medical resources.

Phytochemical profiling of Symplocos tanakana Nakai and S. sawafutagi Nagam. leaf and identification of their antioxidant and anti-diabetic potential.

Symplocos sp. contains various phytochemicals and is used as a folk remedy for treatment of diseases such as enteritis, malaria, and leprosy. In this study, we discovered that 70% ethanol extracts of Symplocos sawafutagi Nagam. and S. tanakana Nakai leaves have antioxidant and anti-diabetic effects. The components in the extracts were profiled using high-performance liquid chromatography coupled to electrospray ionization and quadrupole time-of-flight mass spectrometry; quercetin-3-O-(6''-O-galloyl)-ß-d-galactopyranoside (6) and tellimagrandin II (7) were the main phenolic compounds. They acted as strong antioxidants with excellent radical scavenging activity and as inhibitors of non-enzymatic advanced glycation end-products (AGEs) formation. Mass fragmentation analysis demonstrated that compounds 6 and 7 could form mono- or di-methylglyoxal adducts via reaction with methylglyoxal, which is a reactive carbonyl intermediate and an important precursor of AGEs. In addition, compound 7 effectively inhibited the binding between AGE2 and receptor for AGEs as well as the activity of α-glucosidase. Enzyme kinetic study revealed that compound 7 acts as a competitive inhibitor of α-glucosidase, through interaction with the active site of the enzyme. Therefore, compounds 6 and 7, the major constituents of S. sawafutagi and S. tanakana leaves, are promising for developing drugs for preventing or treating diseases caused by aging and excessive sugar consumption.

Association of ultra-early diffusion-weighted magnetic resonance imaging with neurological outcomes after out-of-hospital cardiac arrest.

BACKGROUND: This study aimed to investigate the association between ultra-early (within 6 h after return of spontaneous circulation [ROSC]) brain diffusion-weighted magnetic resonance imaging (DW-MRI) and neurological outcomes in comatose survivors after out-of-hospital cardiac arrest. METHODS: We conducted a registry-based observational study from May 2018 to February 2022 at a Chungnam national university hospital in Daejeon, Korea. Presence of high-signal intensity (HSI) (PHSI) was defined as a HSI on DW-MRI with corresponding hypoattenuation on the apparent diffusion coefficient map irrespective of volume after hypoxic ischemic brain injury; absence of HSI was defined as AHSI. The primary outcome was the dichotomized cerebral performance category (CPC) at 6 months, defined as good (CPC 1-2) or poor (CPC 3-5). RESULTS: Of the 110 patients (30 women [27.3%]; median (interquartile range [IQR]) age, 58 [38-69] years), 48 (43.6%) had a good neurological outcome, time from ROSC to MRI scan was 2.8 h (IQR 2.0-4.0 h), and the PHSI on DW-MRI was observed in 46 (41.8%) patients. No patients in the PHSI group had a good neurological outcome compared with 48 (75%) patients in the AHSI group. In the AHSI group, cerebrospinal fluid (CSF) neuron-specific enolase (NSE) levels were significantly lower in the group with good neurological outcome compared to the group with poor neurological outcome (20.1 [14.4-30.7] ng/mL vs. 84.3 [32.4-167.0] ng/mL, P < 0.001). The area under the curve for PHSI on DW-MRI was 0.87 (95% confidence interval [CI] 0.80-0.93), and the specificity and sensitivity for predicting a poor neurological outcome were 100% (95% CI 91.2%-100%) and 74.2% (95% CI 62.0-83.5%), respectively. A higher sensitivity was observed when CSF NSE levels were combined (88.7% [95% CI 77.1-95.1%]; 100% specificity). CONCLUSIONS: In this cohort study, PHSI findings on ultra-early DW-MRI were associated with poor neurological outcomes 6 months following the cardiac arrest. The combined CSF NSE levels showed higher sensitivity at 100% specificity than on DW-MRI alone. Prospective multicenter studies are required to confirm these results.

Light-FER: A Lightweight Facial Emotion Recognition System on Edge Devices.

Facial emotion recognition (FER) systems are imperative in recent advanced artificial intelligence (AI) applications to realize better human-computer interactions. Most deep learning-based FER systems have issues with low accuracy and high resource requirements, especially when deployed on edge devices with limited computing resources and memory. To tackle these problems, a lightweight FER system, called Light-FER, is proposed in this paper, which is obtained from the Xception model through model compression. First, pruning is performed during the network training to remove the less important connections within the architecture of Xception. Second, the model is quantized to half-precision format, which could significantly reduce its memory consumption. Third, different deep learning compilers performing several advanced optimization techniques are benchmarked to further accelerate the inference speed of the FER system. Lastly, to experimentally demonstrate the objectives of the proposed system on edge devices, Light-FER is deployed on NVIDIA Jetson Nano.

Deep-Learning Algorithm and Concomitant Biomarker Identification for NSCLC Prediction Using Multi-Omics Data Integration.

Early diagnosis of lung cancer to increase the survival rate, which is currently at a low range of mid-30%, remains a critical need. Despite this, multi-omics data have rarely been applied to non-small-cell lung cancer (NSCLC) diagnosis. We developed a multi-omics data-affinitive artificial intelligence algorithm based on the graph convolutional network that integrates mRNA expression, DNA methylation, and DNA sequencing data. This NSCLC prediction model achieved a 93.7% macro F1-score, indicating that values for false positives and negatives were substantially low, which is desirable for accurate classification. Gene ontology enrichment and pathway analysis of features revealed that two major subtypes of NSCLC, lung adenocarcinoma and lung squamous cell carcinoma, have both specific and common GO biological processes. Numerous biomarkers (i.e., microRNA, long non-coding RNA, differentially methylated regions) were newly identified, whereas some biomarkers were consistent with previous findings in NSCLC (e.g., SPRR1B). Thus, using multi-omics data integration, we developed a promising cancer prediction algorithm.

A Propensity Score-Matched Comparison of In-Hospital Mortality between Dedicated Regional Trauma Centers and Emergency Medical Centers in the Republic of Korea.

Background: In the Republic of Korea, a trauma care system was not created until 2012, at which point regional trauma centers (RTCs) were established nationwide. In accordance with the national emergency care system and legislation, regional and local emergency medical centers (EMCs) also treat patients presenting with trauma. The aim of the present study was to assess whether treatment in RTCs is truly associated with better patient outcomes than that in EMCs by means of propensity score-matched comparisons and to identify populations that would benefit from treatment in RTCs. Methods: This study analyzed the data of patients with consecutive emergency visits between January 1, 2018, and December 31, 2018, collected in the National Emergency Department Information System registry. Data from RTCs, designated regional EMCs, or local EMCs were included; data from smaller emergency departments were excluded because, in Korea, dedicated RTCs are established only in hospitals with regional or local EMCs. Propensity scores for treatment in RTCs or EMCs were estimated by logistic regression using linear terms. Mortality rates in RTCs and EMCs were compared between the matched samples. Results: The in-hospital mortality rates in the matched cases treated in RTCs and EMCs were 1.4% and 1.6%, respectively. The odds ratio for in-hospital mortality in RTCs over EMCs was 0.984 (95% confidence interval: 0.813-1.191). Among the subgroups evaluated, the subgroup of patients with injuries involving the chest or lower limbs showed a significant difference in the in-hospital mortality rate. Conclusion: There was no significant difference in the overall severity-adjusted mortality rate between patients treated in RTCs and EMCs. Treatment in an RTC might benefit those with injuries involving the chest or lower limbs.

Influence of the Level of Emergency Medical Facility on the Short-Term Treatment Results of Cardiac Arrest: Out-of-Hospital Cardiac Arrest and Interhospital Transfer.

Objective: This study aimed to elucidate whether direct transport of out-of-hospital cardiac arrest (OHCA) patients to higher-level emergency medical centres (EMCs) would result in better survival compared to resuscitation in smaller local emergency departments (EDs) and subsequent transfer. Methods: This study was a retrospective population-based analysis of cases registered in the national database of 2019. This study investigated the immediate results of cardiopulmonary resuscitation for OHCA compared between EMCs and EDs and the results of therapeutic temperature management (TTM) compared between the patients directly transported from the field and those transferred from other hospitals. In-hospital mortality was compared using multivariate logistic regression. Results: From the population dataset, 11,493 OHCA patients were extracted. (8,912 in the EMC group vs. 2,581 in the ED group). Multivariate logistic regression revealed that the odds for ED mortality were lower with treatment in EDs than with treatment in EMCs. (odds ratio 0.712 (95% confidence interval (CI): 0.638-0.796)). From the study dataset, 1,798 patients who received TTM were extracted. (1,164 in the direct visit group vs. 634 in the transferred group). Multivariate regression analysis showed that the odds ratio for overall mortality was 1.411 (95% CI: 0.809-2.446) in the transferred group. (p = 0.220). Conclusion: The immediate outcome of OHCA patients who were transported to EDs was not inferior to that of EMCs. Therefore, it would be acceptable to transport OHCA patients to the nearest emergency facilities rather than to the specialized centres in distant areas.

Phytoestrogen Coumestrol Selectively Inhibits Monoamine Oxidase-A and Amyloid ß Self-Aggregation.

Pueraria lobata leaves contain a variety of phytoestrogens, including flavonoids, isoflavonoids, and coumestan derivatives. In this study, we aimed to identify the active ingredients of P. lobata leaves and to elucidate their function in monoamine oxidase (MAO) activation and Aß self-aggregation using in vitro and in silico approaches. To the best of our knowledge, this is the first study to elucidate coumestrol as a selective and competitive MAO-A inhibitor. We identified that coumestrol, a coumestan-derivative, exhibited a selective inhibitory effect against MAO-A (IC50 = 1.99 ± 0.68 µM), a key target protein for depression. In a kinetics analysis with 0.5 µg MAO-A, 40-160 µM substrate, and 25 °C reaction conditions, coumestrol acts as a competitive MAO-A inhibitor with an inhibition constant of 1.32 µM. During an in silico molecular docking analysis, coumestrol formed hydrogen bonds with FAD and pi-pi bonds with hydrophobic residues at the active site of the enzyme. Moreover, based on thioflavin-T-based fluorometric assays, we elucidated that coumestrol effectively prevented self-aggregation of amyloid beta (Aß), which induces an inflammatory response in the central nervous system (CNS) and is a major cause of Alzheimer's disease (AD). Therefore, coumestrol could be used as a CNS drug to prevent diseases such as depression and AD by the inhibition of MAO-A and Aß self-aggregation.

Induction of apoptosis through inactivation of ROS-dependent PI3K/Akt signaling pathway by platycodin D in human bladder urothelial carcinoma cells.

Platycodin D (PD) is a triterpenoid saponin, a major bioactive constituent of the roots of Platycodon grandiflorum, which is well known for possessing various pharmacological properties. However, the anti-cancer mechanism of PD in bladder cancer cells remains poorly understood. In the current study, we investigated the effect of PD on the growth of human bladder urothelial carcinoma cells. PD treatment significantly reduced the cell survival of bladder cancer cells associated with induction of apoptosis and DNA damage. PD inhibited the expression of inhibitor of apoptosis family members, activated caspases, and induced cleavage of poly (ADP-ribose) polymerase. PD also increased the release of cytochrome c into the cytoplasm by disrupting the mitochondrial membrane potential while upregulating the expression ratio of Bax to Bcl-2. The PD-mediated anti-proliferative effect was significantly inhibited by pre-treatment with a pancaspase inhibitor, but not by an inhibitor of necroptosis. Moreover, PD suppressed the phosphoinositide 3-kinase (PI3K)/Akt/mammalian target of rapamycin (mTOR) signaling pathway, and the apoptosis-inducing effect of PD was further enhanced by a PI3K inhibitor. In addition, PD increased the accumulation of reactive oxygen species (ROS), whereas N-acetyl cysteine (NAC), an ROS inhibitor, significantly attenuated the growth inhibition and inactivation of the PI3K/Akt/mTOR signaling caused by PD. Furthermore, NAC significantly suppressed apoptosis, DNA damage, and decreased cell viability induced by PD treatment. Collectively, our findings indicated that PD blocked the growth of bladder urothelial carcinoma cells by inducing ROS-mediated inactivation of the PI3K/Akt/mTOR signaling.

The preventive effect of loganin on oxidative stress-induced cellular damage in human keratinocyte HaCaT cells.

Loganin is a type of iridoid glycosides isolated from Corni fructus and is known to have various pharmacological properties, but studies on its antioxidant activity are still lacking. Therefore, in this study, the preventive effect of loganin on oxidative stress-mediated cellular damage in human keratinocyte HaCaT cells was investigated. Our results show that loganin pretreatment in a non-toxic concentration range significantly improved cell survival in hydrogen peroxide (H2O2)-treated HaCaT cells, which was associated with inhibition of cell cycle arrest at the G2/M phase and induction of apoptosis. H2O2-induced DNA damage and reactive oxygen species (ROS) generation were also greatly reduced in the presence of loganin. Moreover, H2O2 treatment enhanced the cytoplasmic release of cytochrome c, upregulation of the Bax/Bcl-2 ratio and degradation of cleavage of poly (ADP-ribose) polymerase, whereas loganin remarkably suppressed these changes. In addition, loganin obviously attenuated H2O2-induced autophagy while inhibiting the increased accumulation of autophagosome proteins, including as microtubule-associated protein 1 light chain 3-II and Beclin-1, and p62, an autophagy substrate protein, in H2O2-treated cells. In conclusion, our current results suggests that loganin could protect HaCaT keratinocytes from H2O2-induced cellular injury by inhibiting mitochondrial dysfunction, autophagy and apoptosis. This finding indicates the applicability of loganin in the prevention and treatment of skin diseases caused by oxidative damage.

Induction of apoptotic cell death in human bladder cancer cells by ethanol extract of Zanthoxylum schinifolium leaf, through ROS-dependent inactivation of the PI3K/Akt signaling pathway.

BACKGROUND/OBJECTIVES: Zanthoxylum schinifolium is traditionally used as a spice for cooking in East Asian countries. This study was undertaken to evaluate the anti-proliferative potential of ethanol extracts of Z. schinifolium leaves (EEZS) against human bladder cancer T24 cells. MATERIALS/METHODS: Subsequent to measuring the cytotoxicity of EEZS, the anti-cancer activity was measured by assessing apoptosis induction, reactive oxygen species (ROS) generation, and mitochondrial membrane potential (MMP). In addition, we determined the underlying mechanism of EEZS-induced apoptosis through various assays, including Western blot analysis. RESULTS: EEZS treatment concentration-dependently inhibited T24 cell survival, which is associated with apoptosis induction. Exposure to EEZS induced the expression of Fas and Fas-ligand, activated caspases, and subsequently resulted to cleavage of poly (ADP-ribose) polymerase. EEZS also enhanced the expression of cytochrome c in the cytoplasm by suppressing MMP, following increase in the ratio of Bax:Bcl-2 expression and truncation of Bid. However, EEZS-mediated growth inhibition and apoptosis were significantly diminished by a pan-caspase inhibitor. Moreover, EEZS inhibited activation of the phosphoinositide 3-kinase (PI3K)/Akt pathway, and the apoptosis-inducing potential of EEZS was promoted in the presence of PI3K/Akt inhibitor. In addition, EEZS enhanced the production of ROS, whereas N-acetyl cysteine (NAC), a ROS scavenger, markedly suppressed growth inhibition and inactivation of the PI3K/Akt signaling pathway induced by EEZS. Furthermore, NAC significantly attenuated the EEZS-induced apoptosis and reduction of cell viability. CONCLUSIONS: Taken together, our results indicate that exposure to EEZS exhibits anti-cancer activity in T24 bladder cancer cells through ROS-dependent induction of apoptosis and inactivation of the PI3K/Akt signaling pathway.

Automatic Cancer Cell Taxonomy Using an Ensemble of Deep Neural Networks.

Microscopic image-based analysis has been intensively performed for pathological studies and diagnosis of diseases. However, mis-authentication of cell lines due to misjudgments by pathologists has been recognized as a serious problem. To address this problem, we propose a deep-learning-based approach for the automatic taxonomy of cancer cell types. A total of 889 bright-field microscopic images of four cancer cell lines were acquired using a benchtop microscope. Individual cells were further segmented and augmented to increase the image dataset. Afterward, deep transfer learning was adopted to accelerate the classification of cancer types. Experiments revealed that the deep-learning-based methods outperformed traditional machine-learning-based methods. Moreover, the Wilcoxon signed-rank test showed that deep ensemble approaches outperformed individual deep-learning-based models (p < 0.001) and were in effect to achieve the classification accuracy up to 97.735%. Additional investigation with the Wilcoxon signed-rank test was conducted to consider various network design choices, such as the type of optimizer, type of learning rate scheduler, degree of fine-tuning, and use of data augmentation. Finally, it was found that the using data augmentation and updating all the weights of a network during fine-tuning improve the overall performance of individual convolutional neural network models.

Verification of De-Identification Techniques for Personal Information Using Tree-Based Methods with Shapley Values.

With the development of big data and cloud computing technologies, the importance of pseudonym information has grown. However, the tools for verifying whether the de-identification methodology is correctly applied to ensure data confidentiality and usability are insufficient. This paper proposes a verification of de-identification techniques for personal healthcare information by considering data confidentiality and usability. Data are generated and preprocessed by considering the actual statistical data, personal information datasets, and de-identification datasets based on medical data to represent the de-identification technique as a numeric dataset. Five tree-based regression models (i.e., decision tree, random forest, gradient boosting machine, extreme gradient boosting, and light gradient boosting machine) are constructed using the de-identification dataset to effectively discover nonlinear relationships between dependent and independent variables in numerical datasets. Then, the most effective model is selected from personal information data in which pseudonym processing is essential for data utilization. The Shapley additive explanation, an explainable artificial intelligence technique, is applied to the most effective model to establish pseudonym processing policies and machine learning to present a machine-learning process that selects an appropriate de-identification methodology.

A Novel Multistage Transfer Learning for Ultrasound Breast Cancer Image Classification.

Breast cancer diagnosis is one of the many areas that has taken advantage of artificial intelligence to achieve better performance, despite the fact that the availability of a large medical image dataset remains a challenge. Transfer learning (TL) is a phenomenon that enables deep learning algorithms to overcome the issue of shortage of training data in constructing an efficient model by transferring knowledge from a given source task to a target task. However, in most cases, ImageNet (natural images) pre-trained models that do not include medical images, are utilized for transfer learning to medical images. Considering the utilization of microscopic cancer cell line images that can be acquired in large amount, we argue that learning from both natural and medical datasets improves performance in ultrasound breast cancer image classification. The proposed multistage transfer learning (MSTL) algorithm was implemented using three pre-trained models: EfficientNetB2, InceptionV3, and ResNet50 with three optimizers: Adam, Adagrad, and stochastic gradient de-scent (SGD). Dataset sizes of 20,400 cancer cell images, 200 ultrasound images from Mendeley and 400 ultrasound images from the MT-Small-Dataset were used. ResNet50-Adagrad-based MSTL achieved a test accuracy of 99 ± 0.612% on the Mendeley dataset and 98.7 ± 1.1% on the MT-Small-Dataset, averaging over 5-fold cross validation. A p-value of 0.01191 was achieved when comparing MSTL against ImageNet based TL for the Mendeley dataset. The result is a significant improvement in the performance of artificial intelligence methods for ultrasound breast cancer classification compared to state-of-the-art methods and could remarkably improve the early diagnosis of breast cancer in young women.