Computed tomography-based 3D convolutional neural network deep learning model for predicting micropapillary or solid growth pattern of invasive lung adenocarcinoma.

PURPOSE: To investigate the value of a computed tomography (CT)-based deep learning (DL) model to predict the presence of micropapillary or solid (M/S) growth pattern in invasive lung adenocarcinoma (ILADC). MATERIALS AND METHODS: From June 2019 to October 2022, 617 patients with ILADC who underwent preoperative chest CT scans in our institution were randomly placed into training and internal validation sets in a 4:1 ratio, and 353 patients with ILADC from another institution were included as an external validation set. Then, a self-paced learning (SPL) 3D Net was used to establish two DL models: model 1 was used to predict the M/S growth pattern in ILADC, and model 2 was used to predict that pattern in ≤ 2-cm-diameter ILADC. RESULTS: For model 1, the training cohort's area under the curve (AUC), accuracy, recall, precision, and F1-score were 0.924, 0.845, 0.851, 0.842, and 0.843; the internal validation cohort's were 0.807, 0.744, 0.756, 0.750, and 0.743; and the external validation cohort's were 0.857, 0.805, 0.804, 0.806, and 0.804, respectively. For model 2, the training cohort's AUC, accuracy, recall, precision, and F1-score were 0.946, 0.858, 0.881,0.844, and 0.851; the internal validation cohort's were 0.869, 0.809, 0.786, 0.794, and 0.790; and the external validation cohort's were 0.831, 0.792, 0.789, 0.790, and 0.790, respectively. The SPL 3D Net model performed better than the ResNet34, ResNet50, ResNeXt50, and DenseNet121 models. CONCLUSION: The CT-based DL model performed well as a noninvasive screening tool capable of reliably detecting and distinguishing the subtypes of ILADC, even in small-sized tumors.

Light-Reinforced Key Intermediate for Anticoking To Boost Highly Durable Methane Dry Reforming over Single Atom Ni Active Sites on CeO2.

Dry reforming of methane (DRM) has been investigated for more than a century; the paramount stumbling block in its industrial application is the inevitable sintering of catalysts and excessive carbon emissions at high temperatures. However, the low-temperature DRM process still suffered from poor reactivity and severe catalyst deactivation from coking. Herein, we proposed a concept that highly durable DRM could be achieved at low temperatures via fabricating the active site integration with light irradiation. The active sites with Ni-O coordination (NiSA/CeO2) and Ni-Ni coordination (NiNP/CeO2) on CeO2, respectively, were successfully constructed to obtain two targeted reaction paths that produced the key intermediate (CH3O*) for anticoking during DRM. In particular, the operando diffuse reflectance infrared Fourier transform spectroscopy coupling with steady-state isotopic transient kinetic analysis (operando DRIFTS-SSITKA) was utilized and successfully tracked the anticoking paths during the DRM process. It was found that the path from CH3* to CH3O* over NiSA/CeO2 was the key path for anticoking. Furthermore, the targeted reaction path from CH3* to CH3O* was reinforced by light irradiation during the DRM process. Hence, the NiSA/CeO2 catalyst exhibits excellent stability with negligible carbon deposition for 230 h under thermo-photo catalytic DRM at a low temperature of 472 °C, while NiNP/CeO2 shows apparent coke deposition behavior after 0.5 h in solely thermal-driven DRM. The findings are vital as they provide critical insights into the simultaneous achievement of low-temperature and anticoking DRM process through distinguishing and directionally regulating the key intermediate species.

Synergy-Compensation Effect of Ferroelectric Polarization and Cationic Vacancy Collaboratively Promoting CO2 Photoreduction.

Photocatalytic CO2 reduction is severely limited by the rapid recombination of photo-generated charges and insufficient reactive sites. Creating electric field and defects are effective strategies to inhibit charge recombination and enrich catalytic sites, respectively. Herein, a coupled strategy of ferroelectric poling and cationic vacancy is developed to achieve high-performance CO2 photoreduction on ferroelectric Bi2 MoO6 , and their interesting synergy-compensation relationship is first disclosed. Corona poling increases the remnant polarization of Bi2 MoO6 to enhance the intrinsic electric field for promoting charge separation, while it decreases the CO2 adsorption. The introduced Mo vacancy (VMo ) facilitates the adsorption and activation of CO2 , and surface charge separation by creating local electric field. Unfortunately, VMo largely reduces the remnant polarization intensity. Coupling poling and VMo not only integrate their advantages, resulting in an approximately sevenfold increased surface charge transfer efficiency, but also compensate for their shortcomings, for example, VMo largely alleviates the negative effects of ferroelectric poling on CO2 adsorption. In the absence of co-catalyst or sacrificial agent, the poled Bi2 MoO6 with VMo exhibits a superior CO2 -to-CO evolution rate of 19.75 µmol g-1 h-1 , ≈8.4 times higher than the Bi2 MoO6 nanosheets. This work provides new ideas for exploring the role of polarization and defects in photocatalysis.

Facet-Dependent SERS Activity of Co3O4.

Surface-enhanced Raman spectroscopy (SERS) is an ultra-sensitive and rapid technique that is able to significantly enhance the Raman signals of analytes absorbed on functional substrates by orders of magnitude. Recently, semiconductor-based SERS substrates have shown rapid progress due to their great cost-effectiveness, stability and biocompatibility. In this work, three types of faceted Co3O4 microcrystals with dominantly exposed {100} facets, {111} facets and co-exposed {100}-{111} facets (denoted as C-100, C-111 and C-both, respectively) are utilized as SERS substrates to detect the rhodamine 6G (R6G) molecule and nucleic acids (adenine and cytosine). C-100 exhibited the highest SERS sensitivity among these samples, and the lowest detection limits (LODs) to R6G and adenine can reach 10-7 M. First-principles density functional theory (DFT) simulations further unveiled a stronger photoinduced charge transfer (PICT) in C-100 than in C-111. This work provides new insights into the facet-dependent SERS for semiconductor materials.

[Research on coronavirus disease 2019 (COVID-19) detection method based on depthwise separable DenseNet in chest X-ray images].

Coronavirus disease 2019 (COVID-19) has spread rapidly around the world. In order to diagnose COVID-19 more quickly, in this paper, a depthwise separable DenseNet was proposed. The paper constructed a deep learning model with 2 905 chest X-ray images as experimental dataset. In order to enhance the contrast, the contrast limited adaptive histogram equalization (CLAHE) algorithm was used to preprocess the X-ray image before network training, then the images were put into the training network and the parameters of the network were adjusted to the optimal. Meanwhile, Leaky ReLU was selected as the activation function. VGG16, ResNet18, ResNet34, DenseNet121 and SDenseNet models were used to compare with the model proposed in this paper. Compared with ResNet34, the proposed classification model of pneumonia had improved 2.0%, 2.3% and 1.5% in accuracy, sensitivity and specificity respectively. Compared with the SDenseNet network without depthwise separable convolution, number of parameters of the proposed model was reduced by 43.9%, but the classification effect did not decrease. It can be found that the proposed DWSDenseNet has a good classification effect on the COVID-19 chest X-ray images dataset. Under the condition of ensuring the accuracy as much as possible, the depthwise separable convolution can effectively reduce number of parameters of the model.

Unprecedented Eighteen-Faceted BiOCl with a Ternary Facet Junction Boosting Cascade Charge Flow and Photo-redox.

Exposure of anisotropic crystal facets allows the directional transfer of photoexcited electrons (e- ) and holes (h+ ), for spatial charge separation. High-index facets with a high density of low-coordinated atoms always serve as reactive catalytic sites. However, preparation of multi-facets or high-index facets is highly challenging for layered bismuth-based photocatalysts. Herein, we report the preparation of unprecedented eighteen-faceted BiOCl with {001} top facets and {102} and {112} oblique facets via a hydrothermal process. Compared to the conventional BiOCl square plates with {001} top facets and {110} lateral facets, the eighteen-faceted BiOCl has highly enhanced photocatalytic activity for H2 evolution and hydroxyl radicals (. OH) production. Theoretical calculations and photodeposition results disclose that the of eighteen-faceted BiOCl has a well-matched {001}/{102}/{112} ternary facet junction, which provides a cascade path for more efficient charge flow than the binary facet junction in BiOCl square plates.

An adaptive compensation algorithm for temperature drift of micro-electro-mechanical systems gyroscopes using a strong tracking Kalman filter.

We present an adaptive algorithm for a system integrated with micro-electro-mechanical systems (MEMS) gyroscopes and a compass to eliminate the influence from the environment, compensate the temperature drift precisely, and improve the accuracy of the MEMS gyroscope. We use a simplified drift model and changing but appropriate model parameters to implement this algorithm. The model of MEMS gyroscope temperature drift is constructed mostly on the basis of the temperature sensitivity of the gyroscope. As the state variables of a strong tracking Kalman filter (STKF), the parameters of the temperature drift model can be calculated to adapt to the environment under the support of the compass. These parameters change intelligently with the environment to maintain the precision of the MEMS gyroscope in the changing temperature. The heading error is less than 0.6° in the static temperature experiment, and also is kept in the range from 5° to -2° in the dynamic outdoor experiment. This demonstrates that the proposed algorithm exhibits strong adaptability to a changing temperature, and performs significantly better than KF and MLR to compensate the temperature drift of a gyroscope and eliminate the influence of temperature variation.

Macrophages in acne vulgaris: mediating phagocytosis, inflammation, scar formation, and therapeutic implications.

Macrophages serve as a pivotal nexus in the pathogenesis of acne vulgaris, orchestrating both the elimination of Cutibacterium acnes (C. acnes) and lipid metabolic regulation while also possessing the capacity to exacerbate inflammation and induce cutaneous scarring. Additionally, recent investigations underscore the therapeutic potential inherent in macrophage modulation and challenge current anti-inflammatory strategies for acne vulgaris. This review distills contemporary advances, specifically examining the dual roles of macrophages, underlying regulatory frameworks, and emergent therapeutic avenues. Such nuanced insights hold the promise of guiding future explorations into the molecular etiology of acne and the development of more efficacious treatment modalities.

Study on Formation and Migration Law of Foam in Fractures and Its Influencing Factors.

This study focuses on the characteristics of foam generation, flow, and plugging in different reservoir fracture environments. Through visual physical model experiments and stone core displacement experiments, we analyze the flow regeneration of foam in a simulated reservoir fracture environment as well as its sealing and sweeping mechanisms. The findings reveal that low permeability reservoirs, with their smaller and more intricate fracture structures, are conducive to the generation of high-strength foam. This is due to the stronger shear effect of these fracture structures on the injected surfactant and gas mixture system, resulting in a denser foam system. Consequently, low permeability reservoirs facilitate a series of mechanisms that enhance the fluid sweep efficiency. Furthermore, the experiments demonstrate that higher reservoir fracture roughness intensifies the shear disturbance effect on the injected fluid. This disturbance aids in foam regeneration, increases the flow resistance of the foam, and helps to plug high permeability channels. As a result, the foam optimizes the injection-production profile and improves the fluid sweep efficiency. Stone core displacement experiments further illustrate that during foam flooding, the foam liquid film encapsulates the gas phase, thereby obstructing fluid channeling through the Jamin effect. This forces the subsequently injected fluid into other low-permeability fractures, overcoming the shielding effect of high-permeability fractures on low-permeability fractures. Consequently, this improves the fluid diversion rate of low permeability fractures, effectively inhibiting fluid cross-flow and enhancing sweep efficiency. These experimental results highlight the advantages of foam flooding in the development of complex reservoirs with low permeability fracture structures, demonstrating its efficacy in inhibiting fluid cross-flow and optimizing the injection-production profile.

Deep Learning Models of Multi-Scale Lesion Perception Attention Networks for Diagnosis and Staging of Pneumoconiosis: A Comparative Study with Radiologists.

Accurate prediction of pneumoconiosis is essential for individualized early prevention and treatment. However, the different manifestations and high heterogeneity among radiologists make it difficult to diagnose and stage pneumoconiosis accurately. Here, based on DR images collected from two centers, a novel deep learning model, namely Multi-scale Lesion-aware Attention Networks (MLANet), is proposed for diagnosis of pneumoconiosis, staging of pneumoconiosis, and screening of stage I pneumoconiosis. A series of indicators including area under the receiver operating characteristic curve, accuracy, recall, precision, and F1 score were used to comprehensively evaluate the performance of the model. The results show that the MLANet model can effectively improve the consistency and efficiency of pneumoconiosis diagnosis. The accuracy of the MLANet model for pneumoconiosis diagnosis on the internal test set, external validation set, and prospective test set reached 97.87%, 98.03%, and 95.40%, respectively, which was close to the level of qualified radiologists. Moreover, the model can effectively screen stage I pneumoconiosis with an accuracy of 97.16%, a recall of 98.25, a precision of 93.42%, and an F1 score of 95.59%, respectively. The built model performs better than the other four classification models. It is expected to be applied in clinical work to realize the automated diagnosis of pneumoconiosis digital chest radiographs, which is of great significance for individualized early prevention and treatment.

A Multi-Drug Concentration Gradient Mixing Chip: A Novel Platform for High-Throughput Drug Combination Screening.

Combinatorial drug therapy has emerged as a critically important strategy in medical research and patient treatment and involves the use of multiple drugs in concert to achieve a synergistic effect. This approach can enhance therapeutic efficacy while simultaneously mitigating adverse side effects. However, the process of identifying optimal drug combinations, including their compositions and dosages, is often a complex, costly, and time-intensive endeavor. To surmount these hurdles, we propose a novel microfluidic device capable of simultaneously generating multiple drug concentration gradients across an interlinked array of culture chambers. This innovative setup allows for the real-time monitoring of live cell responses. With minimal effort, researchers can now explore the concentration-dependent effects of single-agent and combination drug therapies. Taking neural stem cells (NSCs) as a case study, we examined the impacts of various growth factors-epithelial growth factor (EGF), platelet-derived growth factor (PDGF), and fibroblast growth factor (FGF)-on the differentiation of NSCs. Our findings indicate that an overdose of any single growth factor leads to an upsurge in the proportion of differentiated NSCs. Interestingly, the regulatory effects of these growth factors can be modulated by the introduction of additional growth factors, whether singly or in combination. Notably, a reduced concentration of these additional factors resulted in a decreased number of differentiated NSCs. Our results affirm that the successful application of this microfluidic device for the generation of multi-drug concentration gradients has substantial potential to revolutionize drug combination screening. This advancement promises to streamline the process and accelerate the discovery of effective therapeutic drug combinations.

Seed-mediated growth synthesis and tunable narrow-band luminescence of quaternary Ag-In-Ga-S alloyed nanocrystals.

Silver-based I-III-VI-type semiconductor nanocrystals have received extensive attention due to their narrow-band luminescence properties. Herein, we demonstrated a seed-mediated growth of quaternary Ag-In-Ga-S (AIGS) nanocrystals (NCs) with narrow-band luminescence. By conducting partial cation exchange with In3+ and Ga3+ based on Ag2S NCs and controlling the Ag/In feeding ratios (0.25 to 2) of Ag-In-S seeds as well as the inventory of 1-dodecanethiol, we achieved optimized luminescence performance in the synthesized AIGS NCs, characterized by a narrow full width at half maximum of less than 40 nm. Meanwhile, narrow-band luminescent AIGS NCs exhibit a tetragonal AgGaS2 crystal structure and a gradient alloy structure, rather than a core-shell structure. Most importantly, the kinetics decay curves of time-resolved photoluminescence and the ground state bleaching in transient absorption generally agree with each other regarding the lifetime of the second decay component, which indicates that the narrow-band luminescence is due to the slow radiative recombination between trapped electrons and trapped holes located at the edge of the conduction band and the deep silver-related trap states (e.g., silver vacancy), respectively. This study provides new insights into the correlation between the narrow-band luminescence properties and the structural characteristics of AIGS NCs.

Value of CT-Based Deep Learning Model in Differentiating Benign and Malignant Solid Pulmonary Nodules ≤ 8 mm.

RATIONALE AND OBJECTIVES: We examined the effectiveness of computed tomography (CT)-based deep learning (DL) models in differentiating benign and malignant solid pulmonary nodules (SPNs) ≤ 8 mm. MATERIALS AND METHODS: The study patients (n = 719) were divided into internal training, internal validation, and external validation cohorts; all had small SPNs and had undergone preoperative chest CTs and surgical resection. We developed five DL models incorporating features of the nodule and five different peri-nodular regions with the Multiscale Dual Attention Network (MDANet) to differentiate benign and malignant SPNs. We selected the best-performing model, which was then compared to four conventional algorithms (VGG19, ResNet50, ResNeXt50, and DenseNet121). Furthermore, another five DL models were constructed using MDANet to distinguish benign tumors from inflammatory nodules and the one performed best was selected out. RESULTS: Model 4, which incorporated the nodule and 15 mm peri-nodular region, best differentiated benign and malignant SPNs. The model had an area under the curve (AUC), accuracy, recall, precision, and F1-score of 0.730, 0.724, 0.711, 0.705, and 0.707 in the external validation cohort. Model 4 also performed better than the other four conventional algorithms. Model 8, which incorporated the nodule and 10 mm peri-nodular region, was the best model for distinguishing benign tumors from inflammatory nodules. The model had an AUC, accuracy, recall, precision, and F1-score of 0.871, 0.938, 0.863, 0.904, and 0.882 in the external validation cohort. CONCLUSION: The study concludes that CT-based DL models built with MDANet can accurately discriminate among small benign and malignant SPNs, benign tumors and inflammatory nodules.

From Chips-in-Lab to Point-of-Care Live Cell Device: Development of a Microfluidic Device for On-Site Cell Culture and High-Throughput Drug Screening.

Live cell assays provide real-time data of cellular responses. In combination with microfluidics, applications such as automated and high-throughput drug screening on live cells can be accomplished in small devices. However, their application in point-of-care testing (POCT) is limited by the requirement for bulky equipment to maintain optimal cell culture conditions. In this study, we propose a POCT device that allows on-site cell culture and high-throughput drug screening on live cells. We first observe that cell viabilities are substantially affected by liquid evaporation within the microfluidic device, which is intrinsic to the polydimethylsiloxane (PDMS) material due to its hydrophobic nature and nanopatterned surface. The unwanted PDMS-liquid-air interface in the cell culture environment can be eliminated by maintaining a persistent humidity of 95-100% or submerging the whole microfluidic device under water. Our results demonstrate that in the POCT device equipped with a water tank, both primary cells and cell lines can be maintained for up to 1 week without the need for external cell culture equipment. Moreover, this device is powered by a standard alkali battery and can automatically screen over 5000 combinatorial drug conditions for regulating neural stem cell differentiation. By monitoring dynamic variations in fluorescent markers, we determine the optimal doses of platelet-derived growth factor and epidermal growth factor to suppress proinflammatory S100A9-induced neuronal toxicities. Overall, this study presents an opportunity to transform lab-on-a-chip technology from a laboratory-based approach to actual point-of-care devices capable of performing complex experimental procedures on-site and offers significant advancements in the fields of personalized medicine and rapid clinical diagnostics.

Blue-Emitting Cs(Pb,Cd)Br3 Nanocrystals Resistant to Electric Field-Induced Ion Segregation.

High-performance solution-processed perovskite light-emitting diodes (PeLEDs) have emerged as a good alternative to the well-established technology of epitaxially grown AIIIBV semiconductor alloys. Colloidal cesium lead halide perovskite nanocrystals (CsPbX3 NCs) exhibit room-temperature excitonic emission that can be spectrally tuned across the entire visible range by varying the content of different halogens at the X-site. Therefore, they present a promising platform for full color display manufacturing. Engineering of highly efficient PeLEDs based on bromide and iodide perovskite NCs emitting green and red light, respectively, does not face major challenges except low operational stability of the devices. Meanwhile, mixed-halide counterparts demonstrating blue luminescence suffer from the electric field-induced phase separation (ion segregation) phenomenon described by the rearrangement (demixing) of mobile halide ions in the crystal lattice. This phenomenon results in an undesirable temporal redshift of the electroluminescence spectrum. However, to realize spectral tuning and, at the same time, address the issue of ion segregation less mobile Cd2+ ion could be introduced in the lattice at Pb2+-site that leads to the band gap opening. Herein, we report an original synthesis of CsPb0.88Cd0.12Br3 perovskite NCs and study their structural and optical properties, in particular electroluminescence. Multilayer PeLEDs based on the obtained NCs exhibit single-peak emission centered at 485 nm along with no noticeable change in the spectral line shape for 30 min which is a significant improvement of the device performance.

A Stand-Alone Microfluidic Chip for Long-Term Cell Culture.

Live-cell microscopy is crucial for biomedical studies and clinical tests. The technique is, however, limited to few laboratories due to its high cost and bulky size of the necessary culture equipment. In this study, we propose a portable microfluidic-cell-culture system, which is merely 15 cm×11 cm×9 cm in dimension, powered by a conventional alkali battery and costs less than USD 20. For long-term cell culture, a fresh culture medium exposed to 5% CO2 is programmed to be delivered to the culture chamber at defined time intervals. The 37 °C culture temperature is maintained by timely electrifying the ITO glass slide underneath the culture chamber. Our results demonstrate that 3T3 fibroblasts, HepG2 cells, MB-231 cells and tumor spheroids can be well-maintained for more than 48 h on top of the microscope stage and show physical characters (e.g., morphology and mobility) and growth rate on par with the commercial stage-top incubator and the widely adopted CO2 incubator. The proposed portable cell culture device is, therefore, suitable for simple live-cell studies in the lab and cell experiments in the field when samples cannot be shipped.

Comparison of single-postlabeling delay and seven-delay three-dimensional pseudo-continuous arterial spin labeling in the assessment of intracranial atherosclerotic disease.

Background: The assessment of cerebral blood flow (CBF) is crucial in the evaluation of intracranial atherosclerotic disease. This study was performed to compare single postlabeling delay (PLD) 3-dimensional pseudo-continuous arterial spin labeling (3D-pCASL) and 7-delay 3D-pCASL magnetic resonance imaging in patients with intracranial atherosclerotic stenosis. Methods: A total of 26 patients with moderate to severe atherosclerotic stenosis or occlusion of an intracranial artery were prospectively enrolled in the study. Perfusion parameters were obtained in various regions of interest (ROIs), namely CBF for single PLDs of 1,525 ms (CBF1525 ms), 2,025 ms (CBF2025 ms), and 2,525 ms (CBF2525 ms) with 3D-pCASL, as well as arterial transit time (ATT) and transit-corrected CBF (CBFtransit-corrected) for 7-delay 3D-pCASL. The consistency of the perfusion parameters between single-PLD 3D-pCASL and 7-delay 3D-pCASL was investigated, and the relationship between vascular stenosis and perfusion parameters was explored. Results: Bland-Altman plots compared the CBF values derived from single-PLD 3D-pCASL to those from CBFtransit-corrected. ATT significantly correlated with the difference between CBFtransit-corrected and CBF1525 ms, CBF2025 ms, and CBF2525 ms, respectively (P<0.05). Binary logistic regression analysis revealed that the CBFtransit-corrected and ATT correlated with the presence of moderate or more severe stenotic vascular territories (P<0.05). Conclusions: The single-PLD 3D-pCASL and the 7-delay 3D-pCASL showed inconsistencies in the assessment of CBF, and the perfusion parameters generated under the standard single-PLD 3D-pCASL were more affected by ATT. Moreover, CBFtransit-corrected and ATT were consistent with stenotic vascular territories, which is useful in the evaluation of intracranial atherosclerotic disease.

A fluidic platform for mobility evaluation of zebrafish with gene deficiency.

Introduction: Zebrafish is a suitable animal model for molecular genetic tests and drug discovery due to its characteristics including optical transparency, genetic manipulability, genetic similarity to humans, and cost-effectiveness. Mobility of the zebrafish reflects pathological conditions leading to brain disorders, disrupted motor functions, and sensitivity to environmental challenges. However, it remains technologically challenging to quantitively assess zebrafish's mobility in a flowing environment and simultaneously monitor cellular behavior in vivo. Methods: We herein developed a facile fluidic device using mechanical vibration to controllably generate various flow patterns in a droplet housing single zebrafish, which mimics its dynamically flowing habitats. Results: We observe that in the four recirculating flow patterns, there are two equilibrium stagnation positions for zebrafish constrained in the droplet, i.e., the "source" with the outward flow and the "sink" with the inward flow. Wild-type zebrafish, whose mobility remains intact, tend to swim against the flow and fight to stay at the source point. A slight deviation from streamline leads to an increased torque pushing the zebrafish further away, whereas zebrafish with motor neuron dysfunction caused by lipin-1 deficiency are forced to stay in the "sink," where both their head and tail align with the flow direction. Deviation angle from the source point can, therefore, be used to quantify the mobility of zebrafish under flowing environmental conditions. Moreover, in a droplet of comparable size, single zebrafish can be effectively restrained for high-resolution imaging. Conclusion: Using the proposed methodology, zebrafish mobility reflecting pathological symptoms can be quantitively investigated and directly linked to cellular behavior in vivo.

Predicting the efficacy of non-steroidal anti-inflammatory drugs in migraine using deep learning and three-dimensional T1-weighted images.

Deep learning (DL) models based on individual images could contribute to tailored therapies and personalized treatment strategies. We aimed to construct a DL model using individual 3D structural images for predicting the efficacy of non-steroidal anti-inflammatory drugs (NSAIDs) in migraine. A 3D convolutional neural network model was constructed, with ResNet18 as the classification backbone, to link structural images to predict the efficacy of NSAIDs. In total, 111 patients were included and allocated to the training and testing sets in a 4:1 ratio. The prediction accuracies of the ResNet34, ResNet50, ResNeXt50, DenseNet121, and 3D ResNet18 models were 0.65, 0.74, 0.65, 0.70, and 0.78, respectively. This model, based on individual 3D structural images, demonstrated better predictive performance in comparison to conventional models. Our study highlights the feasibility of the DL algorithm based on brain structural images and suggests that it can be applied to predict the efficacy of NSAIDs in migraine treatment.

An MRI-based machine learning radiomics can predict short-term response to neoadjuvant chemotherapy in patients with cervical squamous cell carcinoma: A multicenter study.

BACKGROUND AND PURPOSE: Neoadjuvant chemotherapy (NACT) has become an essential component of the comprehensive treatment of cervical squamous cell carcinoma (CSCC). However, not all patients respond to chemotherapy due to individual differences in sensitivity and tolerance to chemotherapy drugs. Therefore, accurately predicting the sensitivity of CSCC patients to NACT was vital for individual chemotherapy. This study aims to construct a machine learning radiomics model based on magnetic resonance imaging (MRI) to assess its efficacy in predicting NACT susceptibility among CSCC patients. METHODS: This study included 234 patients with CSCC from two hospitals, who were divided into a training set (n = 180), a testing set (n = 20), and an external validation set (n = 34). Manual radiomic features were extracted from transverse section MRI images, and feature selection was performed using the recursive feature elimination (RFE) method. A prediction model was then generated using three machine learning algorithms, namely logistic regression, random forest, and support vector machines (SVM), for predicting NACT susceptibility. The model's performance was assessed based on the area under the receiver operating characteristic curve (AUC), accuracy, and sensitivity. RESULTS: The SVM approach achieves the highest scores on both the testing set and the external validation set. In the testing set and external validation set, the AUC of the model was 0.88 and 0.764, and the accuracy was 0.90 and 0.853, the sensitivity was 0.93 and 0.962, respectively. CONCLUSIONS: Machine learning radiomics models based on MRI images have achieved satisfactory performance in predicting the sensitivity of NACT in CSCC patients with high accuracy and robustness, which has great significance for the treatment and personalized medicine of CSCC patients.