Rapid Plasma Proteome Profiling via Nanoparticle Protein Corona and Direct Infusion Mass Spectrometry.

Noninvasive detection of protein biomarkers in plasma is crucial for clinical purposes. Liquid chromatography-mass spectrometry (LC-MS) is the gold standard technique for plasma proteome analysis, but despite recent advances, it remains limited by throughput, cost, and coverage. Here, we introduce a new hybrid method that integrates direct infusion shotgun proteome analysis (DISPA) with nanoparticle (NP) protein corona enrichment for high-throughput and efficient plasma proteomic profiling. We realized over 280 protein identifications in 1.4 min collection time, which enables a potential throughput of approximately 1000 samples daily. The identified proteins are involved in valuable pathways, and 44 of the proteins are FDA-approved biomarkers. The robustness and quantitative accuracy of this method were evaluated across multiple NPs and concentrations with a mean coefficient of variation of 17%. Moreover, different protein corona profiles were observed among various NPs based on their distinct surface modifications, and all NP protein profiles exhibited deeper coverage and better quantification than neat plasma. Our streamlined workflow merges coverage and throughput with precise quantification, leveraging both DISPA and NP protein corona enrichment. This underscores the significant potential of DISPA when paired with NP sample preparation techniques for plasma proteome studies.

The Future of Proteomics is Up in the Air: Can Ion Mobility Replace Liquid Chromatography for High Throughput Proteomics?

The coevolution of liquid chromatography (LC) with mass spectrometry (MS) has shaped contemporary proteomics. LC hyphenated to MS now enables quantification of more than 10,000 proteins in a single injection, a number that likely represents most proteins in specific human cells or tissues. Separations by ion mobility spectrometry (IMS) have recently emerged to complement LC and further improve the depth of proteomics. Given the theoretical advantages in speed and robustness of IMS in comparison to LC, we envision that ongoing improvements to IMS paired with MS may eventually make LC obsolete, especially when combined with targeted or simplified analyses, such as rapid clinical proteomics analysis of defined biomarker panels. In this perspective, we describe the need for faster analysis that might drive this transition, the current state of direct infusion proteomics, and discuss some technical challenges that must be overcome to fully complete the transition to entirely gas phase proteomics.

Label-Free Quantification from Direct Infusion Shotgun Proteome Analysis (DISPA-LFQ) with CsoDIAq Software.

Large-scale proteome analysis requires rapid and high-throughput analytical methods. We recently reported a new paradigm in proteome analysis where direct infusion and ion mobility are used instead of liquid chromatography (LC) to achieve rapid and high-throughput proteome analysis. Here, we introduce an improved direct infusion shotgun proteome analysis protocol including label-free quantification (DISPA-LFQ) using CsoDIAq software. With CsoDIAq analysis of DISPA data, we can now identify up to ∼2000 proteins from the HeLa and 293T proteomes, and with DISPA-LFQ, we can quantify ∼1000 proteins from no more than 1 µg of sample within minutes. The identified proteins are involved in numerous valuable pathways including central carbon metabolism, nucleic acid replication and transport, protein synthesis, and endocytosis. Together with a high-throughput sample preparation method in a 96-well plate, we further demonstrate the utility of this technology for performing high-throughput drug analysis in human 293T cells. The total time for data collection from a whole 96-well plate is approximately 8 h. We conclude that the DISPA-LFQ strategy presents a valuable tool for fast identification and quantification of proteins in complex mixtures, which will power a high-throughput proteomic era of drug screening, biomarker discovery, and clinical analysis.

A Hybrid Missing Data Imputation Method for Batch Process Monitoring Dataset.

Batch process monitoring datasets usually contain missing data, which decreases the performance of data-driven modeling for fault identification and optimal control. Many methods have been proposed to impute missing data; however, they do not fulfill the need for data quality, especially in sensor datasets with different types of missing data. We propose a hybrid missing data imputation method for batch process monitoring datasets with multi-type missing data. In this method, the missing data is first classified into five categories based on the continuous missing duration and the number of variables missing simultaneously. Then, different categories of missing data are step-by-step imputed considering their unique characteristics. A combination of three single-dimensional interpolation models is employed to impute transient isolated missing values. An iterative imputation based on a multivariate regression model is designed for imputing long-term missing variables, and a combination model based on single-dimensional interpolation and multivariate regression is proposed for imputing short-term missing variables. The Long Short-Term Memory (LSTM) model is utilized to impute both short-term and long-term missing samples. Finally, a series of experiments for different categories of missing data were conducted based on a real-world batch process monitoring dataset. The results demonstrate that the proposed method achieves higher imputation accuracy than other comparative methods.

FedRAD: Heterogeneous Federated Learning via Relational Adaptive Distillation.

As the development of the Internet of Things (IoT) continues, Federated Learning (FL) is gaining popularity as a distributed machine learning framework that does not compromise the data privacy of each participant. However, the data held by enterprises and factories in the IoT often have different distribution properties (Non-IID), leading to poor results in their federated learning. This problem causes clients to forget about global knowledge during their local training phase and then tends to slow convergence and degrades accuracy. In this work, we propose a method named FedRAD, which is based on relational knowledge distillation that further enhances the mining of high-quality global knowledge by local models from a higher-dimensional perspective during their local training phase to better retain global knowledge and avoid forgetting. At the same time, we devise an entropy-wise adaptive weights module (EWAW) to better regulate the proportion of loss in single-sample knowledge distillation versus relational knowledge distillation so that students can weigh losses based on predicted entropy and learn global knowledge more effectively. A series of experiments on CIFAR10 and CIFAR100 show that FedRAD has better performance in terms of convergence speed and classification accuracy compared to other advanced FL methods.

An Optimization Method of Production-Distribution in Multi-Value-Chain.

Value chain collaboration management is an effective means for enterprises to reduce costs and increase efficiency to enhance competitiveness. Vertical and horizontal collaboration have received much attention, but the current collaboration model combining the two is weak in terms of task assignment and node collaboration constraints in the whole production-distribution process. Therefore, in the enterprise dynamic alliance, this paper models the MVC (multi-value-chain) collaboration process for the optimization needs of the MVC collaboration network in production-distribution and other aspects. Then a MVC collaboration network optimization model is constructed with the lowest total production-distribution cost as the optimization objective and with the delivery cycle and task quantity as the constraints. For the high-dimensional characteristics of the decision space in the multi-task, multi-production end, multi-distribution end, and multi-level inventory production-distribution scenario, a genetic algorithm is used to solve the MVC collaboration network optimization model and solve the problem of difficult collaboration of MVC collaboration network nodes by adjusting the constraints among genes. In view of the multi-level characteristics of the production-distribution scenario, two chromosome coding methods are proposed: staged coding and integrated coding. Moreover, an algorithm ERGA (enhanced roulette genetic algorithm) is proposed with enhanced elite retention based on a SGA (simple genetic algorithm). The comparative experiment results of SGA, SEGA (strengthen elitist genetic algorithm), ERGA, and the analysis of the population evolution process show that ERGA is superior to SGA and SEGA in terms of time cost and optimization results through the reasonable combination of coding methods and selection operators. Furthermore, ERGA has higher generality and can be adapted to solve MVC collaboration network optimization models in different production-distribution environments.

An Unsupervised Learning-Based Regional Deformable Model for Automated Multi-Organ Contour Propagation.

The aim of this study is to evaluate a regional deformable model based on a deep unsupervised learning model for automatic contour propagation in breast cone-beam computed tomography-guided adaptive radiation therapy. A deep unsupervised learning model was introduced to map breast's tumor bed, clinical target volume, heart, left lung, right lung, and spinal cord from planning computed tomography to cone-beam CT. To improve the traditional image registration method's performance, we used a regional deformable framework based on the narrow-band mapping, which can mitigate the effect of the image artifacts on the cone-beam CT. We retrospectively selected 373 anonymized cone-beam CT volumes from 111 patients with breast cancer. The cone-beam CTs are divided into three sets. 311 / 20 / 42 cone-beam CT images were used for training, validating, and testing. The manual contour was used as reference for the testing set. We compared the results between the reference and the model prediction for evaluating the performance. The mean Dice between manual reference segmentations and the model predicted segmentations for breast tumor bed, clinical target volume, heart, left lung, right lung, and spinal cord were 0.78 ± 0.09, 0.90 ± 0.03, 0.88 ± 0.04, 0.94 ± 0.03, 0.95 ± 0.02, and 0.77 ± 0.07, respectively. The results demonstrated a good agreement between the reference and the proposed contours. The proposed deep learning-based regional deformable model technique can automatically propagate contours for breast cancer adaptive radiotherapy. Deep learning in contour propagation was promising, but further investigation was warranted.

Laser Desorption/Ionization Mass Spectrometry Imaging: A New Tool to See through Nanoscale Particles in Biological Systems.

Understanding the fate of nanoscale particles (NPs) in biological systems is significant with the increasing risk for human exposure. Recent research endeavors in laser desorption/ionization mass spectrometry imaging (LDI-MSI) have enriched the toolbox for evaluation of NPs' behavior in biological tissues, especially in aspects including sub-organ bio-distribution, clearance, quantification and surface chemistry variation analysis. In recognition of the potential for advancement in LDI MSI, this concept provides a brief overview of recent research works in LDI MSI for NPs, illustrates new applications that demonstrate the superiority of this technique, and highlights a series of perspectives and directions to move the field forward.

Distributed Supervision Model for Enterprise Data Asset Trading Based on Blockchain Multi-Channel in Industry Alliance.

Compared with traditional physical commodities, data are intangible and easy to leak, and the related trading process has problems, such as complex participating roles, lengthy information flow, poor supervisory coverage and difficult information traceability. To handle these problems, we construct a distributed supervision model for data trading based on blockchain, and conduct multi-party hierarchical and multi-dimensional supervision of the whole process of data trading through collaborative supervision before the event, at present and after the event. First, the characteristics of information flow in the data trading process are analyzed, and the main subject and key supervision information in the data trading process are sorted out and refined. Secondly, combined with the actual business process of data trading supervision, a multi-channel structure of distributed supervision is proposed by adopting an access-verification-traceability strategy. Finally, under the logical framework of the supervision model, the on-chain hierarchical structure and the data hybrid storage method of "on-chain + off-chain" are designed, and multi-supervisor-oriented hierarchical supervision and post-event traceability are realized through smart contracts. The results show that the constructed blockchain-based distributed supervision model of data trading can effectively isolate and protect sensitive and private information between data trading, so as to realize the whole process, multi-subject and differentiated supervision of key information of data trading, and provide an effective and feasible method for the controllable and safe supervision of data trading.

Semi-Supervised Instance-Segmentation Model for Feature Transfer Based on Category Attention.

In the task of image instance segmentation, semi-supervised instance segmentation algorithms have received constant research attention over recent years. Among these algorithms, algorithms based on transfer learning are better than algorithms based on pseudo-label generation in terms of segmentation performance, but they can not make full use of the relevant characteristics of source tasks. To improve the accuracy of these algorithms, this work proposes a semi-supervised instance segmentation model AFT-Mask (attention-based feature transfer Mask R-CNN) based on category attention. The AFT-Mask model takes the result of object-classification prediction as "attention" to improve the performance of the feature-transfer module. In detail, we designed a migration-optimization module for connecting feature migration and classification prediction to enhance segmentation-prediction accuracy. To verify the validity of the AFT-Mask model, experiments were conducted on two types of datasets. Experimental results show that the AFT-Mask model can achieve effective knowledge transfer and improve the performance of the benchmark model on semi-supervised instance segmentation.

Intelligent Reflecting Surface-Assisted Physical Layer Key Generation with Deep Learning in MIMO Systems.

Physical layer secret key generation (PLKG) is a promising technology for establishing effective secret keys. Current works for PLKG mostly study key generation schemes in ideal communication environments with little or even no signal interference. In terms of this issue, exploiting the reconfigurable intelligent reflecting surface (IRS) to assist PLKG has caused an increasing interest. Most IRS-assisted PLKG schemes focus on the single-input-single-output (SISO), which is limited in future communications with multi-input-multi-output (MIMO). However, MIMO could bring a serious overhead of channel reciprocity extraction. To fill the gap, this paper proposes a novel low-overhead IRS-assisted PLKG scheme with deep learning in the MIMO communications environments. We first combine the direct channel and the reflecting channel established by the IRS to construct the channel response function, and we propose a theoretically optimal interaction matrix to approach the optimal achievable rate. Then we design a channel reciprocity-learning neural network with an IRS introduced (IRS-CRNet), which is exploited to extract the channel reciprocity in time division duplexing (TDD) systems. Moreover, a PLKG scheme based on the IRS-CRNet is proposed. Final simulation results verify the performance of the PLKG scheme based on the IRS-CRNet in terms of key generation rate, key error rate and randomness.

Development and Validation of a Deep Learning CT Signature to Predict Survival and Chemotherapy Benefit in Gastric Cancer: A Multicenter, Retrospective Study.

OBJECTIVE: We aimed to develop a deep learning-based signature to predict prognosis and benefit from adjuvant chemotherapy using preoperative computed tomography (CT) images. BACKGROUND: Current staging methods do not accurately predict the risk of disease relapse for patients with gastric cancer. METHODS: We proposed a novel deep neural network (S-net) to construct a CT signature for predicting disease-free survival (DFS) and overall survival in a training cohort of 457 patients, and independently tested it in an external validation cohort of 1158 patients. An integrated nomogram was constructed to demonstrate the added value of the imaging signature to established clinicopathologic factors for individualized survival prediction. Prediction performance was assessed with respect to discrimination, calibration, and clinical usefulness. RESULTS: The DeLIS was associated with DFS and overall survival in the overall validation cohort and among subgroups defined by clinicopathologic variables, and remained an independent prognostic factor in multivariable analysis (P< 0.001). Integrating the imaging signature and clinicopathologic factors improved prediction performance, with C-indices: 0.792-0.802 versus 0.719-0.724, and net reclassification improvement 10.1%-28.3%. Adjuvant chemotherapy was associated with improved DFS in stage II patients with high-DeLIS [hazard ratio = 0.362 (95% confidence interval 0.149-0.882)] and stage III patients with high- and intermediate-DeLIS [hazard ratio = 0.611 (0.442-0.843); 0.633 (0.433-0.925)]. On the other hand, adjuvant chemotherapy did not affect survival for patients with low-DeLIS, suggesting a predictive effect (Pinteraction = 0.048, 0.016 for DFS in stage II and III disease). CONCLUSIONS: The proposed imaging signature improved prognostic prediction and could help identify patients most likely to benefit from adjuvant chemotherapy in gastric cancer.

Mass Spectrometry Imaging Reveals In Situ Behaviors of Multiple Components in Aerosol Particles.

The inhalation of atmospheric particles is deleterious to human health. However, as a complex mixture, tracing the behaviors of multiple components from real aerosol particles is crucial but unachievable by the existing methods. Here, taking advantage of the intrinsic fingerprints of elemental carbon (EC) and organic carbon (OC) in carbonaceous aerosol (CA) upon laser irradiation, we proposed a label-free mass spectrometry imaging method to visualize and quantify the deposition, translocation and component variation of CA in organs. With this method, the heterogeneous deposition, clearance and release behavior of CA in lung, that more OC was released in parenchyma and OC was cleared faster than EC, was observed. The translocation of CA to extrapulmonary organs including kidney, liver, spleen and even brain was also verified and quantified. By comparing the ratio of OC to EC, an organ-specific release behavior of OC from CA during circulation was revealed. In orthotopic lung and liver tumor, OC was found to penetrate more into tumor foci than EC. This technique provides deeper information for understanding the systemic health effects of aerosol particles.

Ultrafast Photocatalytic Reaction Screening by Mass Spectrometry.

Here we report a semiconductor-assisted laser desorption/ionization mass spectrometry (SA-LDI MS) platform to monitor photocatalytic reactions online and apply it for ultrafast reaction screening. In this method, we use photocatalytic nanomaterials as the substrate for LDI and then initiate and monitor the reactions simultaneously. The features of our method include the following: (i) It has a reaction acceleration effect: only seconds are needed in our interfacial reactions vs hours in conventional bulk phase. (ii) The reaction trend in our system agrees with that in bulk phase. (iii) By adding a stable analogue of reactant as internal standard, a quantification of the reaction can be achieved. (iv) The sensitivity is high: for 500 amol of reactant, the photocatalytic reaction can still be initiated and detected. This platform has advantages in ultrafast reaction screening (e.g., screening of nine catalysts requires 24 h by the UPLC-MS system but only 10 min by SA-LDI MS). Furthermore, the high specificity of MS enables the screening of catalytic selectivity of A-TiO2 nanoparticles for a methyl red (MR) and acid yellow (AY) mixture, whose absorption wavelengths are overlapped and thus cannot be discriminated by conventional optical methods. Furthermore, by using SA-LDI MS, we also monitored reductive debrominations during the degradation process of polybrominated diphenyl ethers (PBDEs), which is a type of important pollutant that is difficult to degrade and detect in liquid phase, and the photocatalytic reduction of CO2. Overall, SA-LDI MS realizes ultrafast photocatalytic reaction screening for the first time and provides practical analytical value in the field of catalyst screening.

Prognostic value and nomograms of proximal margin distance in gastric cancer with radical distal gastrectomy.

OBJECTIVE: The proximal margin (PM) distance for distal gastrectomy (DG) of gastric cancer (GC) remains controversial. This study investigated the prognostic value of PM distance for survival outcomes, and aimed to combine clinicopathologic variables associated with survival outcomes after DG with different PM distance for GC into a prediction nomogram. METHODS: Patients who underwent radical DG from June 2004 to June 2014 at Department of General Surgery, Nanfang Hospital, Southern Medical University were included. The first endpoints of the prognostic value of PM distance (assessed in 0.5 cm increments) for disease-free survival (DFS) and overall survival (OS) were assessed. Multivariate analysis by Cox proportional hazards regression was performed using the training set, and the nomogram was constructed, patients were chronologically assigned to the training set for dates from June 1, 2004 to January 30, 2012 (n=493) and to the validation set from February 1, 2012 to June 30, 2014 (n=211). RESULTS: Among 704 patients with pTNM stage I, pTNM stage II, T1-2, T3-4, N0, differentiated type, tumor size ≤5.0 cm, a PM of (2.1-5.0) cmvs. PM≤2.0 cm showed a statistically significant difference in DFS and OS, while a PM>5.0 cm was not associated with any further improvement in DFS and OSvs. a PM of 2.1-5.0 cm. In patients with pTNM stage III, N1, N2-3, undifferentiated type, tumor size >5.0 cm, the PM distance was not significantly correlated with DFS and OS between patients with a PM of (2.1-5.0) cm and a PM≤2 cm, or between patients with a PM >5.0 cm and a PM of (2.1-5.0) cm, so there were no significant differences across the three PM groups. In the training set, the C-indexes of DFS and OS, were 0.721 and 0.735, respectively, and in the validation set, the C-indexes of DFS and OS, were 0.752 and 0.751, respectively. CONCLUSIONS: It is necessary to obtain not less than 2.0 cm of PM distance in early-stage disease, while PM distance was not associated with long-term survival in later and more aggressive stages of disease because more advanced GC is a systemic disease. Different types of patients should be considered for removal of an individualized PM distance intra-operatively. We developed a universally applicable prediction model for accurately determining the 1-year, 3-year and 5-year DFS and OS of GC patients according to their preoperative clinicopathologic characteristics and PM distance.

Direct identification of forensic body fluids by MALDI-MS.

The rapid identification of human body fluids is meaningful for forensic casework. However, current methods suffer from several limitations such as poor sensitivity, time consumption and big sample consumption. Herein, we developed a mass spectrometry method to distinguish human body fluids (blood, semen, urine, sweat, and saliva) based on small molecular regions with no pretreatment, microliter sample consumption and high throughput. A highly sensitive and high salt-tolerance matrix N-(1-naphthyl)ethylenediamine dihydrochloride (NEDC) was used to efficiently detect metabolites in complex humoral environment. Some characteristic small metabolic molecules such as heme, hemin, creatinine, phosphate acid, uric acid, citric acid and lactic acid were identified and served as potential biomarkers to differentiate different body fluid types. Further principal component analysis (PCA) was performed to cluster the body fluid samples and three principal components allowed 75% clustering of all body fluid types. Blind testing revealed that nine out of ten unknown body fluid samples could be correctly classified into their corresponding group. This novel method can efficiently differentiate five body fluids with minimal interferences due to the storage time (less than 12 months) and carrier materials (cotton, fabric and tissue). The whole process from sampling to recording of mass spectra of body fluids can be finished in less than 10 minutes. We believe that this developed strategy has significant implications for rapid and effective human body fluid screening in forensic casework.

ImmunoScore Signature: A Prognostic and Predictive Tool in Gastric Cancer.

OBJECTIVE: We postulated that the ImmunoScore (IS) could markedly improve the prediction of postsurgical survival and chemotherapeutic benefits in gastric cancer (GC). SUMMARY BACKGROUND DATA: A prediction model for GC patients was developed using data from 879 consecutive patients. METHODS: The expression of 27 immune features was detected in 251 specimens by using immunohistochemistry, and a 5-feature-based ISGC was then constructed using the LASSO Cox regression model. Testing and validation cohorts were included to validate the model. RESULTS: Using the LASSO model, we established an ISGC classifier based on 5 features: CD3invasive margin (IM), CD3center of tumor (CT), CD8IM, CD45ROCT, and CD66bIM. Significant differences were found between the high-ISGC and low-ISGC patients in the training cohort in 5-year disease-free survival (45.0% vs. 4.4%, respectively; P <0.001) and 5-year overall survival (48.8% vs. 6.7%, respectively; P <0.001). Multivariate analysis revealed that the ISGC classifier was an independent prognostic factor. A combination of ISGC and tumor, node, and metastasis (TNM) had better prognostic value than TNM stage alone. Further analysis revealed that stage II and III GC patients with high-ISGC exhibited a favorable response to adjuvant chemotherapy. Finally, we constructed 2 nomograms to predict which patients with stages II and III GC might benefit from adjuvant chemotherapy after surgery. CONCLUSIONS: The ISGC classifier could effectively predict recurrence and survival of GC, and complemented the prognostic value of the TNM staging system. Moreover, the ISGC might be a useful predictive tool to identify stage II and III GC patients who would benefit from adjuvant chemotherapy.

1.4 min Plasma Proteome Profiling via Nanoparticle Protein Corona and Direct Infusion Mass Spectrometry.

Non-invasive detection of protein biomarkers in plasma is crucial for clinical purposes. Liquid chromatography mass spectrometry (LC-MS) is the gold standard technique for plasma proteome analysis, but despite recent advances, it remains limited by throughput, cost, and coverage. Here, we introduce a new hybrid method, which integrates direct infusion shotgun proteome analysis (DISPA) with nanoparticle (NP) protein coronas enrichment for high throughput and efficient plasma proteomic profiling. We realized over 280 protein identifications in 1.4 minutes collection time, which enables a potential throughput of approximately 1,000 samples daily. The identified proteins are involved in valuable pathways and 44 of the proteins are FDA approved biomarkers. The robustness and quantitative accuracy of this method were evaluated across multiple NPs and concentrations with a mean coefficient of variation at 17%. Moreover, different protein corona profiles were observed among various nanoparticles based on their distinct surface modifications, and all NP protein profiles exhibited deeper coverage and better quantification than neat plasma. Our streamlined workflow merges coverage and throughput with precise quantification, leveraging both DISPA and NP protein corona enrichments. This underscores the significant potential of DISPA when paired with NP sample preparation techniques for plasma proteome studies.

Morphological diversity of cancer cells predicts prognosis across tumor types.

BACKGROUND: Intratumor heterogeneity drives disease progression and treatment resistance, which can lead to poor patient outcomes. Here, we present a computational approach for quantification of cancer cell diversity in routine hematoxylin-eosin-stained histopathology images. METHODS: We analyzed publicly available digitized whole-slide hematoxylin-eosin images for 2000 patients. Four tumor types were included: lung, head and neck, colon, and rectal cancers, representing major histology subtypes (adenocarcinomas and squamous cell carcinomas). We performed single-cell analysis on hematoxylin-eosin images and trained a deep convolutional autoencoder to automatically learn feature representations of individual cancer nuclei. We then computed features of intranuclear variability and internuclear diversity to quantify tumor heterogeneity. Finally, we used these features to build a machine-learning model to predict patient prognosis. RESULTS: A total of 68 million cancer cells were segmented and analyzed for nuclear image features. We discovered multiple morphological subtypes of cancer cells (range = 15-20) that co-exist within the same tumor, each with distinct phenotypic characteristics. Moreover, we showed that a higher morphological diversity is associated with chromosome instability and genomic aneuploidy. A machine-learning model based on morphological diversity demonstrated independent prognostic values across tumor types (hazard ratio range = 1.62-3.23, P < .035) in validation cohorts and further improved prognostication when combined with clinical risk factors. CONCLUSIONS: Our study provides a practical approach for quantifying intratumor heterogeneity based on routine histopathology images. The cancer cell diversity score can be used to refine risk stratification and inform personalized treatment strategies.

Talk-to-Edit: Fine-Grained 2D and 3D Facial Editing via Dialog.

Facial editing is to manipulate the facial attributes of a given face image. Nowadays, with the development of generative models, users can easily generate 2D and 3D facial images with high fidelity and 3D-aware consistency. However, existing works are incapable of delivering a continuous and fine-grained editing mode (e.g., editing a slightly smiling face to a big laughing one) with natural interactions with users. In this work, we propose Talk-to-Edit, an interactive facial editing framework that performs fine-grained attribute manipulation through dialog between the user and the system. Our key insight is to model a continual "semantic field" in the GAN latent space. 1) Unlike previous works that regard the editing as traversing straight lines in the latent space, here the fine-grained editing is formulated as finding a curving trajectory that respects fine-grained attribute landscape on the semantic field. 2) The curvature at each step is location-specific and determined by the input image as well as the users' language requests. 3) To engage the users in a meaningful dialog, our system generates language feedback by considering both the user request and the current state of the semantic field. We demonstrate the effectiveness of our proposed framework on both 2D and 3D-aware generative models. We term the semantic field for the 3D-aware models as "tri-plane" flow, as it corresponds to the changes not only in the color space but also in the density space. We also contribute CelebA-Dialog, a visual-language facial editing dataset to facilitate large-scale study. Specifically, each image has manually annotated fine-grained attribute annotations as well as template-based textual descriptions in natural language. Extensive quantitative and qualitative experiments demonstrate the superiority of our framework in terms of 1) the smoothness of fine-grained editing, 2) the identity/attribute preservation, and 3) the visual photorealism and dialog fluency. Notably, the user study validates that our overall system is consistently favored by around 80% of the participants.