Prognostic model incorporating immune checkpoint genes to predict the immunotherapy efficacy for lung adenocarcinoma: a cohort study integrating machine learning algorithms.

This study aimed to develop and validate a nomogram based on immune checkpoint genes (ICGs) for predicting prognosis and immune checkpoint blockade (ICB) efficacy in lung adenocarcinoma (LUAD) patients. A total of 385 LUAD patients from the TCGA database and 269 LUAD patients in the combined dataset (GSE41272 + GSE50081) were divided into training and validation cohorts, respectively. Three different machine learning algorithms including random forest (RF), least absolute shrinkage and selection operator (LASSO) logistic regression analysis, and support vector machine (SVM) were employed to select the predictive markers from 82 ICGs to construct the prognostic nomogram. The X-tile software was used to stratify patients into high- and low-risk subgroups based on the nomogram-derived risk scores. Differences in functional enrichment and immune infiltration between the two subgroups were assessed using gene set variation analysis (GSVA) and various algorithms. Additionally, three lung cancer cohorts receiving ICB therapy were utilized to evaluate the ability of the model to predict ICB efficacy in the real world. Five ICGs were identified as predictive markers across all three machine learning algorithms, leading to the construction of a nomogram with strong potential for prognosis prediction in both the training and validation cohorts (all AUC values close to 0.800). The patients were divided into high- (risk score ≥ 185.0) and low-risk subgroups (risk score < 185.0). Compared to the high-risk subgroup, the low-risk subgroup exhibited enrichment in immune activation pathways and increased infiltration of activated immune cells, such as CD8 + T cells and M1 macrophages (P < 0.05). Furthermore, the low-risk subgroup had a greater likelihood of benefiting from ICB therapy and longer progression-free survival (PFS) than did the high-risk subgroup (P < 0.05) in the two cohorts receiving ICB therapy. A nomogram based on ICGs was constructed and validated to aid in predicting prognosis and ICB treatment efficacy in LUAD patients.

Long-term risk of overweight in offspring of Chinese women with gestational diabetes defined by IADPSG's but not by WHO's criteria.

AIMS: To examine long-term risk of overweight in offspring of women with gestational diabetes mellitus (GDM) defined by the International Association of Diabetes and Pregnancy Study Group (IADPSG)'s criteria but not by the 1999 World Health Organization (WHO)'s criteria. METHODS: We followed up 1681 mother-child pairs for 8 years in Tianjin, China. Overweight in children aged 1-5 and 6-8 were respectively defined as body mass index-for-age and -sex above the 2 z-score and 1 z-score curves of the WHO's child growth standards. Logistic regression was performed to obtain odds ratios (ORs) and 95% confidence intervals (CIs) of hyperglycemia indices at oral glucose tolerance test and GDMs defined by different criteria for offspring overweight at different ages. RESULTS: Offspring of women with fasting plasma glucose ≥5.1 mmol/L were at increased risk of overweight at 6-8 years old (OR:1.45, 95% CI: 1.09-1.93). GDM defined by the IADPSG's criteria only was associated with increased risk of childhood overweight at 6-8 years old (1.65, 1.13-2.40), as compared with non-GDM by either of the two sets of criteria. CONCLUSIONS: Newly defined GDM by the IADPSG's criteria increased the risk of offspring overweight aged 6-8 years.

[Facial nerve protection in surgery for rare tumors of the internal auditory canal].

Objective:To explore strategies for preserving facial nerve function during surgeries for rare tumors of the internal auditory canal. Methods:A total of 235 cases of internal auditory canal tumors treated between 2010 and 2023 were included, encompassing vestibular schwannomas, cavernous hemangiomas, meningiomas, and other rare tumors. Various data, including clinical presentations, imaging classifications, and treatment processes, were meticulously analyzed to delineate the characteristics of rare tumors and assess pre-and postoperative facial nerve function. Results:Among all internal auditory canal tumors, vestibular schwannomas accounted for 91.9%. In rare tumors, facial nerve schwannomas constituted 5.3%, cavernous hemangiomas 26.3%, meningiomas 15.8%, and arterial aneurysms 10.5%. Significantly, patients with cavernous hemangiomas displayed pronounced invasion of the facial nerve by the tumor, in contrast to other tumor types where clear boundaries with the facial nerve were maintained. During surgery, individualized approaches and strategies for facial nerve protection were implemented for different tumor types, involving intraoperative dissection, tumor excision, and facial nerve reconstruction. Conclusion:Preservation of the facial nerve is crucial in the surgical management of rare tumors of the internal auditory canal. Accurate preoperative diagnosis, appropriate timing of surgery, selective surgical approaches, and meticulous intraoperative techniques can maximize the protection of facial nerve function. Personalized treatment plans and strategies for facial nerve functional reconstruction are anticipated to enhance surgical success rates, reduce the risk of postoperative facial nerve dysfunction, and ultimately improve the quality of life for patients.

Predictive Model to Identify the Long Time Survivor in Patients with Glioblastoma: A Cohort Study Integrating Machine Learning Algorithms.

We aimed to develop and validate a predictive model for identifying long-term survivors (LTS) among glioblastoma (GB) patients, defined as those with an overall survival (OS) of more than 3 years. A total of 293 GB patients from CGGA and 169 from TCGA database were assigned to training and validation cohort, respectively. The differences in expression of immune checkpoint genes (ICGs) and immune infiltration landscape were compared between LTS and short time survivor (STS) (OS<1.5 years). The differentially expressed genes (DEGs) and weighted gene co-expression network analysis (WGCNA) were used to identify the genes differentially expressed between LTS and STS. Three different machine learning algorithms were employed to select the predictive genes from the overlapping region of DEGs and WGCNA to construct the nomogram. The comparison between LTS and STS revealed that STS exhibited an immune-resistant status, with higher expression of ICGs (P<0.05) and greater infiltration of immune suppression cells compared to LTS (P<0.05). Four genes, namely, OSMR, FMOD, CXCL14, and TIMP1, were identified and incorporated into the nomogram, which possessed good potential in predicting LTS probability among GB patients both in the training (C-index, 0.791; 0.772-0.817) and validation cohort (C-index, 0.770; 0.751-0.806). STS was found to be more likely to exhibit an immune-cold phenotype. The identified predictive genes were used to construct the nomogram with potential to identify LTS among GB patients.

Adverse birth outcomes and childhood overweight at age of 3-8 years in a prospective cohort study in Tianjin, China.

OBJECTIVES: To explore associations between adverse birth outcomes and childhood overweight at 3-8 years of age. DESIGN: A prospective cohort study. SETTING: Six central urban districts of Tianjin, China. PARTICIPANTS: 1681 woman-child pairs. METHODS: 1681 woman-child pairs were followed up for 8 years in Tianjin, China. Demographic and clinical information including birth outcomes was collected longitudinally, commencing from first antenatal care visit till postpartum period. Offspring height and weight were measured at 3-8 years of age. High and low weight/length ratios (WLR) at birth were, respectively, defined as ≥90th and ≤10th gestational week and sex-specific percentiles. Overweight for children at 3-5 and 6-8 years of age were, respectively, defined as body mass index (BMI)-for-age and -sex above the 2 z-score and 1 z-score curves of the WHO's child growth standards. Binary logistic regression analysis was used to obtain ORs and 95% CI with a stepwise backward selection method to select independent predictors. PRIMARY OUTCOMES MEASURES: Childhood overweight. RESULTS: Of 1681 children, 10.7% (n=179) and 27.8% (n=468) developed overweight at 3-5 and 6-8 years of age, respectively. Large for gestational age (LGA) was associated with increased risk of overweight at 3-5 years of age (aOR: 1.86, 95% CI: 1.27 to 2.72) while high WLR at birth was associated with increased risk of overweight at 6-8 years of age (1.82, 1.41 to 2.34). Low WLR at birth was associated with decreased risk of overweight at 6-8 years of age (0.52, 0.30 to 0.90). CONCLUSIONS: LGA and high WLR at birth predicted childhood overweight at 3-5 and 6-8 years of age, respectively. Low WLR at birth was associated with decreased risk of childhood overweight at 6-8 years of age.

Maternal gestational diabetes and childhood adiposity risk from 6 to 8 years of age.

BACKGROUND/OBJECTIVE: Previous studies found conflicting results on the association between maternal gestational diabetes mellitus (GDM) and childhood overweight/obesity. This study was to assess the association between maternal GDM and offspring's adiposity risk from 6 to 8 years of age. METHODS: The present study longitudinally followed 1156 mother-child pairs (578 GDM and 578 non-GDM) at 5.9 ± 1.2 years postpartum and retained 912 mother-child pairs (486 GDM and 426 non-GDM) at 8.3 ± 1.6 years postpartum. Childhood body mass index (BMI), waist circumference, body fat and skinfold were measured using standardized methods. RESULTS: Compared with the counterparts born to mothers with normal glucose during pregnancy, children born to mothers with GDM during pregnancy had higher mean values of adiposity indicators (waist circumference, body fat, subscapular skinfold and suprailiac skinfold) at 5.9 and 8.3 years of age. There was a positive association of maternal GDM with changes of childhood adiposity indicators from the 5.9-year to 8.3-year visit, and ß values were significantly larger than zero: +0.10 (95% CI: 0.02-0.18) for z score of BMI for age, +1.46 (95% CI: 0.70-2.22) cm for waist circumference, +1.78% (95% CI: 1.16%-2.40%) for body fat, +2.40 (95% CI: 1.78-3.01) mm for triceps skinfold, +1.59 (95% CI: 1.10-2.09) mm for subscapular skinfold, and +2.03 (95% CI: 1.35-2.71) mm for suprailiac skinfold, respectively. Maternal GDM was associated with higher risks of childhood overweight/obesity, central obesity, and high body fat (Odd ratios 1.41-1.57 at 5.9 years of age and 1.73-2.03 at 8.3 years of age) compared with the children of mothers without GDM. CONCLUSIONS: Maternal GDM was a risk factor of childhood overweight/obesity at both 5.9 and 8.3 years of age, which was independent from several important confounders including maternal pre-pregnancy BMI, gestational weight gain, children's birth weight and lifestyle factors. This significant and positive association became stronger with age.

Interactive effects of gestational diabetes and high pre-pregnancy body mass index on adverse growth patterns of offspring.

AIMS: To examine the independent and interactive effects of maternal gestational diabetes mellitus (GDM) and high pre-pregnancy body mass index (BMI) on the risk of offspring adverse growth patterns. MATERIALS AND METHODS: One thousand six hundred and eighty one mother-child pairs were followed for 8 years in Tianjin, China. Group-based trajectory modelling was used to identify offspring growth patterns. Logistic regression was performed to obtain odds ratios (ORs) and 95% confidence intervals (CIs) of GDM and high pre-pregnancy BMI for offspring adverse growth patterns. Restricted cubic spline was used to identify cut-off points. Additive interactions and multiplicative interactions were used to test interactive effects between GDM and high pre-pregnancy BMI for adverse growth patterns. RESULTS: Four distinct growth patterns were identified in offspring, including normal growth pattern, persistent lean growth pattern, late obesity growth pattern (LOGP), and persistent obesity growth pattern (POGP). Maternal high pre-pregnancy BMI was associated with LOGP and POGP (adjusted OR, 95% CI: 2.38, 1.74-3.25 & 4.92, 2.26-10.73). GDM greatly enhanced the adjusted OR of high pre-pregnancy BMI for LOGP up to 3.48 (95% CI: 2.25-5.38). Additive interactions and multiplicative interactions between both risk factors were significant for LOGP but not for POGP. CONCLUSIONS: Maternal high pre-pregnancy BMI was associated with increased risk of LOGP and POGP, whereas GDM greatly enhanced the risk of high pre-pregnancy BMI for LOGP.

Serum mannan-binding lectin-associated serine proteases in early pregnancy for gestational diabetes in Chinese pregnant women.

Aims: This study aimed to explore associations of mannan-binding lectin-associated serine protease (MASP) levels in early pregnancy with gestational diabetes mellitus (GDM). We also examined interactions of MASPs and deoxycholic acid (DCA)/glycoursodeoxycholic acid (GUDCA) for the GDM risk and whether the interactive effects if any on the GDM risk were mediated via lysophosphatidylcholine (LPC) 18:0. Materials and methods: A 1:1 case-control study (n = 414) nested in a prospective cohort of pregnant women was conducted in Tianjin, China. Binary conditional logistic regressions were performed to examine associations of MASPs with the GDM risk. Additive interaction measures were used to examine interactions between MASPs and DCA/GUDCA for the GDM risk. Mediation analyses and Sobel tests were used to examine mediation effects of LPC18:0 between the copresence of MASPs and DCA/GUDCA on the GDM risk. Results: High MASP-2 was independently associated with GDM [odds ratio (OR): 2.62, 95% confidence interval (CI): 1.44-4.77], while the effect of high MASP-1 on GDM was attributable to high MASP-2 (P for Sobel test: 0.003). Low DCA markedly increased the OR of high MASP-2 alone from 2.53 (1.10-5.85) up to 10.6 (4.22-26.4), with a significant additive interaction. In addition, high LPC18:0 played a significant mediating role in the links from low DCA to GDM and from the copresence of high MASP-2 and low DCA to GDM (P for Sobel test <0.001) but not in the link from high MASP-2 to GDM. Conclusions: High MASP-1 and MASP-2 in early pregnancy were associated with GDM in Chinese pregnant women. MASP-2 amplifies the risk of low DCA for GDM, which is mediated via LPC18:0.

A high activity mesoporous Pt@KIT-6 nanocomposite for selective hydrogenation of halogenated nitroarenes in a continuous-flow microreactor.

In this study, we designed a Pt@KIT-6 nanocomposite prepared by impregnating platinum nanoparticles on the nanopores of the KIT-6 mesoporous material. This Pt@KIT-6 nanocomposite was used as a catalyst in a micro fixed bed reactor (MFBR) for the continuous-flow hydrogenation of halogenated nitroarenes, which demonstrates three advantages. First, the Pt@KIT-6 nanocomposite has a stable mesoporous nanostructure, which effectively enhances the active site and hydrogen adsorption capacity. The uniformly distributed pore structure and large specific surface area were confirmed by electron microscopy and N2 physisorption, respectively. In addition, the aggregation of the loaded metal was avoided, which facilitated the maintenance of high activity and selectivity. The conversion and selectivity reached 99% within 5.0 minutes at room temperature (20 °C). Furthermore, the continuous-flow microreactor allows precise control and timely transfer of the reaction system, reducing the impact of haloid acids. The activity and selectivity of the Pt@KIT-6 nanocomposite showed virtually no degradation after 24 hours of continuous operation of the entire continuous-flow system. Overall, the Pt@KIT-6 nanocomposite showed good catalysis for the hydrogenation of halogenated nitroarenes in the continuous-flow microreactor. This work provides insights into the rational design of a highly active and selective catalyst for selective hydrogenation systems.

Amphiphilic Particle-Stabilized Nanoliter Droplet Reactors with a Multimodal Portable Reader for Distributive Biomarker Quantification.

Compartmentalization, leveraging microfluidics, enables highly sensitive assays, but the requirement for significant infrastructure for their design, build, and operation limits access. Multimaterial particle-based technologies thermodynamically stabilize monodisperse droplets as individual reaction compartments with simple liquid handling steps, precluding the need for expensive microfluidic equipment. Here, we further improve the accessibility of this lab on a particle technology to resource-limited settings by combining this assay system with a portable multimodal reader, thus enabling nanoliter droplet assays in an accessible platform. We show the utility of this platform in measuring N-terminal propeptide B-type natriuretic peptide (NT-proBNP), a heart failure biomarker, in complex medium and patient samples. We report a limit of detection of ∼0.05 ng/mL and a linear response between 0.2 and 2 ng/mL in spiked plasma samples. We also show that, owing to the plurality of measurements per sample, "swarm" sensing acquires better statistical quantitation with a portable reader. Monte Carlo simulations show the increasing capability of this platform to differentiate between negative and positive samples, i.e., below or above the clinical cutoff for acute heart failure (∼0.1 ng/mL), as a function of the number of particles measured. Our platform measurements correlate with gold standard ELISA measurement in cardiac patient samples, and achieve lower variation in measurement across samples compared to the standard well plate-based ELISA. Thus, we show the capabilities of a cost-effective droplet-reader system in accurately measuring biomarkers in nanoliter droplets for diseases that disproportionately affect underserved communities in resource-limited settings.

Serum Sulfur-Containing Amino Acids and Risk of Maternal Gestational Diabetes and Adverse Growth Patterns in Offspring.

BACKGROUND: To estimate associations of sulfur-containing amino acids (SAAs) in the early trimester of pregnancy and gestational diabetes mellitus (GDM) and estimate associations of maternal SAAs with adverse growth patterns in offspring. METHODS: We established a 1:1 matched case-control study (n = 486) from our cohort of pregnant women, and 401 children were followed up at ages 1 to 8 years. We conducted binary conditional logistic regression to estimate the risk associations of serum SAAs with GDM. Multinomial logistic regression was implemented to explore associations of maternal SAAs with adverse growth patterns in the offspring. RESULTS: High serum methionine and cystine were independently associated with increased GDM risk (OR: 1.92, 95%CI: 1.18-3.13 and 2.69, 1.59-4.53). Conversely, a low level of serum taurine was independently associated with increased GDM risk (2.61, 1.64-4.16). Maternal high cystine and low taurine were also associated with an increased risk of persistent obesity growth pattern (POGP) in offspring (OR: 2.79, 95%CI: 1.09-7.17 and 3.92, 1.11-13.89) and the effect was largely independent of GDM. CONCLUSIONS: High serum methionine, cystine and low serum taurine in the early trimester of pregnancy were associated with a greatly increased risk of GDM. Maternal high cystine and low taurine were associated with elevated risk of offspring POGP, largely independent of GDM.

Circulating hyaluronidase in early pregnancy and increased risk of gestational diabetes in Chinese pregnant women: A nested case control study.

BACKGROUND AND AIMS: To explore association of serum hyaluronidase 1 (HYAL1) level in early pregnancy with gestational diabetes mellitus (GDM), and to examine interactive effects of HYAL1 with ceramides species on GDM risk. MATERIALS AND METHODS: We conducted a 1:1 matched case-control study (n = 414) of pregnant women from 2010 to 2012 in Tianjin, China. Blood samples were collected at the first antenatal care visit (at a median of 10th gestational weeks). Binary conditional logistic regression and restricted cubic spline (RCS) analysis were used to examine full-range risk association between HYAL1 and GDM. Additive interactions and multiplicative interactions were employed to test interactive effects of HYAL1 with ceramides species on GDM risk. RESULTS: Ln HYAL1 was linearly associated with GDM risk and the adjusted OR of HYAL1 ≥ vs. < its median for GDM was significant (1.65, 95%CI: 1.08-2.52). High HYAL1 markedly enhanced the ORs of high ceramide 18:0 for GDM from 2.31 (1.06-5.01) to 6.74 (2.85-16.0), and low ceramide 24:0 from 3.08 (1.33-7.11) to 8.15 (3.03-21.9), with significant additive interactions. CONCLUSIONS: High HYAL1 in early pregnancy may increase the risk of GDM in Chinese women, possibly via enhancing the effects of high ceramide 18:0 and low ceramide 24:0 on GDM risk.

Universal linear intensity transformations using spatially incoherent diffractive processors.

Under spatially coherent light, a diffractive optical network composed of structured surfaces can be designed to perform any arbitrary complex-valued linear transformation between its input and output fields-of-view (FOVs) if the total number (N) of optimizable phase-only diffractive features is ≥~2NiNo, where Ni and No refer to the number of useful pixels at the input and the output FOVs, respectively. Here we report the design of a spatially incoherent diffractive optical processor that can approximate any arbitrary linear transformation in time-averaged intensity between its input and output FOVs. Under spatially incoherent monochromatic light, the spatially varying intensity point spread function (H) of a diffractive network, corresponding to a given, arbitrarily-selected linear intensity transformation, can be written as H(m, n; m', n') = |h(m, n; m', n')|2, where h is the spatially coherent point spread function of the same diffractive network, and (m, n) and (m', n') define the coordinates of the output and input FOVs, respectively. Using numerical simulations and deep learning, supervised through examples of input-output profiles, we demonstrate that a spatially incoherent diffractive network can be trained to all-optically perform any arbitrary linear intensity transformation between its input and output if N ≥ ~2NiNo. We also report the design of spatially incoherent diffractive networks for linear processing of intensity information at multiple illumination wavelengths, operating simultaneously. Finally, we numerically demonstrate a diffractive network design that performs all-optical classification of handwritten digits under spatially incoherent illumination, achieving a test accuracy of >95%. Spatially incoherent diffractive networks will be broadly useful for designing all-optical visual processors that can work under natural light.

IRDC-Net: An Inception Network with a Residual Module and Dilated Convolution for Sign Language Recognition Based on Surface Electromyography.

Deaf and hearing-impaired people always face communication barriers. Non-invasive surface electromyography (sEMG) sensor-based sign language recognition (SLR) technology can help them to better integrate into social life. Since the traditional tandem convolutional neural network (CNN) structure used in most CNN-based studies inadequately captures the features of the input data, we propose a novel inception architecture with a residual module and dilated convolution (IRDC-net) to enlarge the receptive fields and enrich the feature maps, applying it to SLR tasks for the first time. This work first transformed the time domain signal into a time-frequency domain using discrete Fourier transformation. Second, an IRDC-net was constructed to recognize ten Chinese sign language signs. Third, the tandem CNN networks VGG-net and ResNet-18 were compared with our proposed parallel structure network, IRDC-net. Finally, the public dataset Ninapro DB1 was utilized to verify the generalization performance of the IRDC-net. The results showed that after transforming the time domain sEMG signal into the time-frequency domain, the classification accuracy (acc) increased from 84.29% to 91.70% when using the IRDC-net on our sign language dataset. Furthermore, for the time-frequency information of the public dataset Ninapro DB1, the classification accuracy reached 89.82%; this value is higher than that achieved in other recent studies. As such, our findings contribute to research into SLR tasks and to improving deaf and hearing-impaired people's daily lives.

Aromatic Amino Acids and Their Interactions with Gut Microbiota-Related Metabolites for Risk of Gestational Diabetes: A Prospective Nested Case-Control Study in a Chinese Cohort.

INTRODUCTION: The aim of this study was to explore associations of aromatic amino acids (AAA) in early pregnancy with gestational diabetes mellitus (GDM), and whether high AAA and gut microbiota-related metabolites had interactive effects on GDM risk. METHODS: We conducted a 1:1 case-control study (n = 486) nested in a prospective cohort of pregnant women from 2010 to 2012. According to the International Association of Diabetes and Pregnancy Study Group's criteria, 243 women were diagnosed with GDM. Binary conditional logistic regression was performed to examine associations of AAA with GDM risk. Interactions between AAA and gut microbiota-related metabolites for GDM were examined using additive interaction measures. RESULTS: High phenylalanine and tryptophan were associated with increased GDM risk (OR: 1.72, 95% CI: 1.07-2.78 and 1.66, 1.02-2.71). The presence of high trimethylamine (TMA) markedly increased the OR of high phenylalanine alone up to 7.95 (2.79-22.71), while the presence of low glycoursodeoxycholic acid (GUDCA) markedly increased the OR of high tryptophan alone up to 22.88 (5.28-99.26), both with significant additive interactions. Furthermore, high lysophosphatidylcholines (LPC18:0) mediated both interactive effects. CONCLUSIONS: High phenylalanine may have an additive interaction with high TMA, while high tryptophan may have an additive interaction with low GUDCA toward increased risk of GDM, both being mediated via LPC18:0.

MRI-based scoring model to predict massive hemorrhage during dilatation and curettage in patients with cesarean scar pregnancy.

OBJECTIVE: To construct a scoring model based on MRI signs to predict massive hemorrhage during dilatation and curettage in cesarean scar pregnancy (CSP) patients. MATERIALS AND METHODS: The MRIs of CSP patients admitted to a tertiary referral hospital between February 2020 and July 2022 were retrospectively reviewed. The included patients were randomly assigned to the training and validation cohorts. The univariate and multivariate logistic regression analyses were adopted to identify the independent risk factors for massive hemorrhage (the amount of bleeding ≥ 200 ml) during the dilatation and curettage. A scoring model predicting intraoperative massive hemorrhage was established where each positive independent risk factor was assigned 1 point, and the predictive power of this model was evaluated both in the training and validation cohorts via the receiver operating characteristic curve. RESULTS: A total of 187 CSP patients were enrolled, who were divided into the training cohort (31 in 131 patients had massive hemorrhage) and validation cohort (10 in 56 patients had massive hemorrhage). The independent risk factors for intraoperative massive hemorrhage included cesarean section diverticulum area (OR = 6.957, 95% CI 1.993-21.887; P = 0.001), uterine scar thickness (OR = 5.113, 95% CI 2.086-23.829; P = 0.025) and gestational sac diameter (OR = 3.853, 95% CI 1.103-13.530; P = 0.025). A scoring model with a total point of 3 was developed and the CSP patients were divided into low-risk (Total points < 2) and high-risk groups (Total points ≥ 2) for intraoperative massive hemorrhage accordingly. This model possessed high prediction performance both in the training cohort (area under the curve [AUC] = 0.896, 95% CI 0.830-0.942) and validation cohort (AUC = 0.915, 95% CI 0.785-1.000). CONCLUSION: We first constructed a MRI-based scoring model for predicting intraoperative massive hemorrhage in CSP patients, which could help the decision-making of the patients' therapy strategies. Low-risk patients can be cured by D&C alone to reduce the financial burden, while high-risk patients require more adequate preoperative preparation or consideration of changing surgical approaches to reduce bleeding risk.

Data-Class-Specific All-Optical Transformations and Encryption.

Diffractive optical networks provide rich opportunities for visual computing tasks. Here, data-class-specific transformations that are all-optically performed between the input and output fields-of-view (FOVs) of a diffractive network are presented. The visual information of the objects is encoded into the amplitude (A), phase (P), or intensity (I) of the optical field at the input, which is all-optically processed by a data-class-specific diffractive network. At the output, an image sensor-array directly measures the transformed patterns, all-optically encrypted using the transformation matrices preassigned to different data classes, i.e., a separate matrix for each data class. The original input images can be recovered by applying the correct decryption key (the inverse transformation) corresponding to the matching data class, while applying any other key will lead to loss of information. All-optical class-specific transformations covering A → A, I → I, and P → I transformations using various image datasets are numerically demonstrated. The feasibility of this framework is also experimentally validated by fabricating class-specific I → I transformation diffractive networks and is successfully tested at different parts of the electromagnetic spectrum, i.e., 1550 nm and 0.75 mm wavelengths. Data-class-specific all-optical transformations provide a fast and energy-efficient method for image and data encryption, enhancing data security and privacy.

Light-field tomographic fluorescence lifetime imaging microscopy.

Fluorescence lifetime imaging microscopy (FLIM) is a powerful imaging technique that enables the visualization of biological samples at the molecular level by measuring the fluorescence decay rate of fluorescent probes. This provides critical information about molecular interactions, environmental changes, and localization within biological systems. However, creating high-resolution lifetime maps using conventional FLIM systems can be challenging, as it often requires extensive scanning that can significantly lengthen acquisition times. This issue is further compounded in three-dimensional (3D) imaging because it demands additional scanning along the depth axis. To tackle this challenge, we developed a novel computational imaging technique called light field tomographic FLIM (LIFT-FLIM). Our approach allows for the acquisition of volumetric fluorescence lifetime images in a highly data-efficient manner, significantly reducing the number of scanning steps required compared to conventional point-scanning or line-scanning FLIM imagers. Moreover, LIFT-FLIM enables the measurement of high-dimensional data using low-dimensional detectors, which are typically low-cost and feature a higher temporal bandwidth. We demonstrated LIFT-FLIM using a linear single-photon avalanche diode array on various biological systems, showcasing unparalleled single-photon detection sensitivity. Additionally, we expanded the functionality of our method to spectral FLIM and demonstrated its application in high-content multiplexed imaging of lung organoids. LIFT-FLIM has the potential to open up new avenues in both basic and translational biomedical research.

Deep learning-enabled virtual histological staining of biological samples.

Histological staining is the gold standard for tissue examination in clinical pathology and life-science research, which visualizes the tissue and cellular structures using chromatic dyes or fluorescence labels to aid the microscopic assessment of tissue. However, the current histological staining workflow requires tedious sample preparation steps, specialized laboratory infrastructure, and trained histotechnologists, making it expensive, time-consuming, and not accessible in resource-limited settings. Deep learning techniques created new opportunities to revolutionize staining methods by digitally generating histological stains using trained neural networks, providing rapid, cost-effective, and accurate alternatives to standard chemical staining methods. These techniques, broadly referred to as virtual staining, were extensively explored by multiple research groups and demonstrated to be successful in generating various types of histological stains from label-free microscopic images of unstained samples; similar approaches were also used for transforming images of an already stained tissue sample into another type of stain, performing virtual stain-to-stain transformations. In this Review, we provide a comprehensive overview of the recent research advances in deep learning-enabled virtual histological staining techniques. The basic concepts and the typical workflow of virtual staining are introduced, followed by a discussion of representative works and their technical innovations. We also share our perspectives on the future of this emerging field, aiming to inspire readers from diverse scientific fields to further expand the scope of deep learning-enabled virtual histological staining techniques and their applications.