RESUMEN
BACKGROUND: Multiplexed immunofluorescence (mIF) staining, such as CODEX and MIBI, holds significant clinical value for various fields, such as disease diagnosis, biological research, and drug development. However, these techniques are often hindered by high time and cost requirements. METHODS: Here we present a Multimodal-Attention-based virtual mIF Staining (MAS) system that utilises a deep learning model to extract potential antibody-related features from dual-modal non-antibody-stained fluorescence imaging, specifically autofluorescence (AF) and DAPI imaging. The MAS system simultaneously generates predictions of mIF with multiple survival-associated biomarkers in gastric cancer using self- and multi-attention learning mechanisms. FINDINGS: Experimental results with 180 pathological slides from 94 patients with gastric cancer demonstrate the efficiency and consistent performance of the MAS system in both cancer and noncancer gastric tissues. Furthermore, we showcase the prognostic accuracy of the virtual mIF images of seven gastric cancer related biomarkers, including CD3, CD20, FOXP3, PD1, CD8, CD163, and PD-L1, which is comparable to those obtained from the standard mIF staining. INTERPRETATION: The MAS system rapidly generates reliable multiplexed staining, greatly reducing the cost of mIF and improving clinical workflow. FUNDING: Stanford 2022 HAI Seed Grant; National Institutes of Health 1R01CA256890.
RESUMEN
BACKGROUND: Tertiary lymphoid structures (TLSs) are associated with favorable prognosis and enhanced response to anti-cancer therapy. A digital assessment of TLSs could provide an objective alternative that mitigates variability inherent in manual evaluation. This study aimed to develop and validate a digital gene panel based on biological prior knowledge for assessment of TLSs, and further investigate its associations with survival and multiple anti-cancer therapies. MATERIALS AND METHODS: The present study involved 1,704 patients with gastric cancer from seven cancer centers. TLSs were identified morphologically through hematoxylin-and-eosin staining. We further developed a digital score based on targeted gene expression profiling to assess TLSs status, recorded as gene signature of tertiary lymphoid structures (gsTLS). For enhanced interpretability, we employed the SHapley Additive exPlanation (SHAP) analysis to elucidate its contribution to the prediction. We next evaluated the signature's associations with prognosis, and investigated its predictive accuracy for multiple anti-cancer therapies, including adjuvant chemotherapy and immunotherapy. RESULTS: The gsTLS panel with nine gene features achieved high accuracies in predicting TLSs status in the training, internal and external validation cohorts (area under the curve, range: 0.729-0.791). In multivariable analysis, gsTLS remained an independent predictor of disease-free and overall survival (hazard ratio, range: 0.346-0.743, all P < 0.05) after adjusting for other clinicopathological variables. SHAP analysis highlighted gsTLS as the strongest predictor of TLSs status compared with clinical features. Importantly, patients with high gsTLS (but not those with low gsTLS) exhibited substantial benefits from adjuvant chemotherapy (P < 0.05). Furthermore, we found that the objective response rate to anti-programmed cell death protein 1 (anti-PD-1) immunotherapy was significantly higher in the high-gsTLS group (40.7%) versus the low-gsTLS group (5.6%, P = 0.036), and the diagnosis was independent from Epstein-Barr virus (EBV), tumor mutation burden (TMB), and programmed cell death-ligand 1 (PD-L1) expression. CONCLUSION: The gsTLS digital panel enables accurate assessment of TLSs status, and provides information regarding prognosis and responses to multiple therapies for gastric cancer.
RESUMEN
Introduction: Age-related macular degeneration (AMD) is a prevalent, chronic and progressive retinal degenerative disease characterized by an inflammatory response mediated by activated microglia accumulating in the retina. In this study, we demonstrate the therapeutically effects and the underlying mechanisms of microglial repopulation in the laser-induced choroidal neovascularization (CNV) model of exudative AMD. Methods: The CSF1R inhibitor PLX3397 was used to establish a treatment paradigm for microglial repopulation in the retina. Neovascular leakage and neovascular area were examined by fundus fluorescein angiography (FFA) and immunostaining of whole-mount RPE-choroid-sclera complexes in CNV mice receiving PLX3397. Altered cellular senescence was measured by beta-galactosidase (SA-ß-gal) activity and p16INK4a expression. The effect and mechanisms of repopulated microglia on leukocyte infiltration and the inflammatory response in CNV lesions were analyzed. Results: We showed that ten days of the CSF1R inhibitor PLX3397 treatment followed by 11 days of drug withdrawal was sufficient to stimulate rapid repopulation of the retina with new microglia. Microglial repopulation attenuated pathological choroid neovascularization and dampened cellular senescence in CNV lesions. Repopulating microglia exhibited lower levels of activation markers, enhanced phagocytic function and produced fewer cytokines involved in the immune response, thereby ameliorating leukocyte infiltration and attenuating the inflammatory response in CNV lesions. Discussion: The microglial repopulation described herein are therefore a promising strategy for restricting inflammation and choroidal neovascularization, which are important players in the pathophysiology of AMD.
Asunto(s)
Aminopiridinas , Neovascularización Coroidal , Modelos Animales de Enfermedad , Microglía , Animales , Neovascularización Coroidal/etiología , Neovascularización Coroidal/tratamiento farmacológico , Neovascularización Coroidal/metabolismo , Neovascularización Coroidal/patología , Microglía/metabolismo , Microglía/efectos de los fármacos , Ratones , Aminopiridinas/farmacología , Aminopiridinas/uso terapéutico , Ratones Endogámicos C57BL , Degeneración Macular/patología , Degeneración Macular/metabolismo , Degeneración Macular/tratamiento farmacológico , Inflamación , Receptores de Factor Estimulante de Colonias de Granulocitos y Macrófagos/antagonistas & inhibidores , Receptores de Factor Estimulante de Colonias de Granulocitos y Macrófagos/metabolismo , Pirroles/farmacología , Pirroles/uso terapéutico , Senescencia Celular/efectos de los fármacosRESUMEN
Predicting the occurrence of nonproliferative diabetic retinopathy (NPDR) using biochemical parameters is invasive, which limits large-scale clinical application. Noninvasive retinal oxygen metabolism and hemodynamics of 215 eyes from 73 age-matched healthy subjects, 90 diabetic patients without DR, 40 NPDR, and 12 DR with postpanretinal photocoagulation were measured with a custom-built multimodal retinal imaging device. Diabetic patients underwent biochemical examinations. Two logistic regression models were developed to predict NPDR using retinal and biochemical metrics, respectively. The predictive model 1 using retinal metrics incorporated male gender, insulin treatment condition, diastolic duration, resistance index, and oxygen extraction fraction presented a similar predictive power with model 2 using biochemical metrics incorporated diabetic duration, diastolic blood pressure, and glycated hemoglobin A1c (area under curve: 0.73 vs. 0.70; sensitivity: 76% vs. 68%; specificity: 64% vs. 62%). These results suggest that retinal oxygen metabolic and hemodynamic biomarkers may replace biochemical parameters to predict the occurrence of NPDR .
Asunto(s)
Retinopatía Diabética , Hemodinámica , Oxígeno , Retina , Retinopatía Diabética/diagnóstico , Oxígeno/metabolismo , Retina/diagnóstico por imagen , Retina/metabolismo , Modelos Logísticos , Humanos , Masculino , Femenino , Persona de Mediana Edad , Anciano , Valor Predictivo de las PruebasRESUMEN
BACKGROUNDThe tumor immune microenvironment can provide prognostic and therapeutic information. We aimed to develop noninvasive imaging biomarkers from computed tomography (CT) for comprehensive evaluation of immune context and investigate their associations with prognosis and immunotherapy response in gastric cancer (GC).METHODSThis study involved 2,600 patients with GC from 9 independent cohorts. We developed and validated 2 CT imaging biomarkers (lymphoid radiomics score [LRS] and myeloid radiomics score [MRS]) for evaluating the IHC-derived lymphoid and myeloid immune context respectively, and integrated them into a combined imaging biomarker [LRS/MRS: low(-) or high(+)] with 4 radiomics immune subtypes: 1 (-/-), 2 (+/-), 3 (-/+), and 4 (+/+). We further evaluated the imaging biomarkers' predictive values on prognosis and immunotherapy response.RESULTSThe developed imaging biomarkers (LRS and MRS) had a high accuracy in predicting lymphoid (AUC range: 0.765-0.773) and myeloid (AUC range: 0.736-0.750) immune context. Further, similar to the IHC-derived immune context, 2 imaging biomarkers (HR range: 0.240-0.761 for LRS; 1.301-4.012 for MRS) and the combined biomarker were independent predictors for disease-free and overall survival in the training and all validation cohorts (all P < 0.05). Additionally, patients with high LRS or low MRS may benefit more from immunotherapy (P < 0.001). Further, a highly heterogeneous outcome on objective response ârate was observed in 4 imaging subtypes: 1 (-/-) with 27.3%, 2 (+/-) with 53.3%, 3 (-/+) with 10.2%, and 4 (+/+) with 30.0% (P < 0.0001).CONCLUSIONThe noninvasive imaging biomarkers could accurately evaluate the immune context and provide information regarding prognosis and immunotherapy for GC.
Asunto(s)
Neoplasias Gástricas , Humanos , Neoplasias Gástricas/diagnóstico por imagen , Neoplasias Gástricas/terapia , Radiómica , Inmunoterapia , Tomografía Computarizada por Rayos X , Microambiente Tumoral , Biomarcadores , PronósticoRESUMEN
Deep neural networks (DNNs) extract thousands to millions of task-specific features during model training for inference and decision-making. While visualizing these features is critical for comprehending the learning process and improving the performance of the DNNs, existing visualization techniques work only for classification tasks. For regressions, the feature points lie on a high dimensional continuum having an inherently complex shape, making a meaningful visualization of the features intractable. Given that the majority of deep learning applications are regression-oriented, developing a conceptual framework and computational method to reliably visualize the regression features is of great significance. Here, we introduce a manifold discovery and analysis (MDA) method for DNN feature visualization, which involves learning the manifold topology associated with the output and target labels of a DNN. MDA leverages the acquired topological information to preserve the local geometry of the feature space manifold and provides insightful visualizations of the DNN features, highlighting the appropriateness, generalizability, and adversarial robustness of a DNN. The performance and advantages of the MDA approach compared to the existing methods are demonstrated in different deep learning applications.
RESUMEN
Purpose: To investigate changes in foveal avascular area (FAZ) and retinal vein diameter in patients with retinal vein occlusion (RVO) after intravitreal ranibizumab, and to analyze the correlation between ranibizumab therapy and visual gain. Methods: This retrospective study enrolled 95 eyes of 95 patients who had accepted three consecutive monthly ranibizumab injections, including 50 branch RVOs (BRVOs) and 45 central RVOs (CRVOs). BRVOs were divided into ischemia group (n = 32) and non-ischemia group (n = 18), and CRVOs also had ischemia group (n = 28) and non-ischemia group (n = 17). Comprehensive ophthalmic examinations were performed before the first injection and after 6, 12, and 24 months. The FAZ was manually circumscribed on early-phase images of fundus fluorescein angiography. Retinal vein diameters were measured on fundus photographs. Results: After three injections, the FAZ area was significantly enlarged firstly and then reduced in all ischemic RVOs and the non-ischemic BRVOs (p < 0.05), while the retinal vein diameter was significantly reduced firstly and then increased in all groups except for unobstructed branch veins of non-ischemic BRVOs (p < 0.05). The correlation between the FAZ area and best corrected visual acuity was statistically significant in all CRVOs (non-ischemic, r = 0.372; ischemic, r = 0.286; p < 0.01) and ischemic BRVOs (r = 0.180, p < 0.05). Spearman's correlation analysis revealed that the retinal vein diameter was significantly correlated to the larger FAZ area in obstructed branch veins of ischemic BRVOs (r = -0.31, p < 0.01), inferior temporal branch veins of non-ischemic CRVOs (r = -0.461, p < 0.01) and ischemia CRVO groups (superior temporal branch vein, r = -0.226, p < 0.05; inferior temporal branch vein, r = -0.259, p < 0.01). Conclusion: After three consecutive monthly ranibizumab injections, the FAZ area was enlarged and retinal vein diameter reduced with gradual recovery to near baseline from 12 months. These results suggest that ranibizumab therapy can worsen macular ischemia and prevent visual gain in the short term. It has important significance for the treatment and prognosis of RVO, although the natural course of RVO may also affect ischemia and visual gain.
RESUMEN
Nanoparticles (NPs) disguised in the cell membrane are a new type of biomimetic platform. Due to their ability to simulate the unique biological functions of membrane-derived cells, they have become one of the hotspots of research at home and abroad. The tumor-specific antigen antibody carried by breast cancer cell membranes can modify nanoparticles to have homologous tumor targeting. Therefore, nanoparticles wrapped in cancer cell membranes have been widely used in research on the diagnosis and treatment of breast cancer. This article reviews the current situation, prospects, advantages and limitations of nanoparticles modified by cancer cell membranes in the treatment and diagnosis of breast cancer.
RESUMEN
Deep learning-based diagnosis is becoming an indispensable part of modern healthcare. For high-performance diagnosis, the optimal design of deep neural networks (DNNs) is a prerequisite. Despite its success in image analysis, existing supervised DNNs based on convolutional layers often suffer from their rudimentary feature exploration ability caused by the limited receptive field and biased feature extraction of conventional convolutional neural networks (CNNs), which compromises the network performance. Here, we propose a novel feature exploration network named manifold embedded multilayer perceptron (MLP) mixer (ME-Mixer), which utilizes both supervised and unsupervised features for disease diagnosis. In the proposed approach, a manifold embedding network is employed to extract class-discriminative features; then, two MLP-Mixer-based feature projectors are adopted to encode the extracted features with the global reception field. Our ME-Mixer network is quite general and can be added as a plugin to any existing CNN. Comprehensive evaluations on two medical datasets are performed. The results demonstrate that their approach greatly enhances the classification accuracy in comparison with different configurations of DNNs with acceptable computational complexity.
RESUMEN
The retina is one of the most metabolically active tissues in the body. The dysfunction of oxygen kinetics in the retina is closely related to the disease and has important clinical value. Dynamic imaging and comprehensive analyses of oxygen kinetics in the retina depend on the fusion of structural and functional imaging and high spatiotemporal resolution. But it's currently not clinically available, particularly via a single imaging device. Therefore, this work aims to develop a retinal oxygen kinetics imaging and analysis (ROKIA) technology by integrating dual-wavelength imaging with laser speckle contrast imaging modalities, which achieves structural and functional analysis with high spatial resolution and dynamic measurement, taking both external and lumen vessel diameters into account. The ROKIA systematically evaluated eight vascular metrics, four blood flow metrics, and fifteen oxygenation metrics. The single device scheme overcomes the incompatibility of optical design, harmonizes the field of view and resolution of different modalities, and reduces the difficulty of registration and image processing algorithms. More importantly, many of the metrics (such as oxygen delivery, oxygen metabolism, vessel wall thickness, etc.) derived from the fusion of structural and functional information, are unique to ROKIA. The oxygen kinetic analysis technology proposed in this paper, to our knowledge, is the first demonstration of the vascular metrics, blood flow metrics, and oxygenation metrics via a single system, which will potentially become a powerful tool for disease diagnosis and clinical research.
RESUMEN
Embedding high-dimensional data onto a low-dimensional manifold is of both theoretical and practical value. In this article, we propose to combine deep neural networks (DNN) with mathematics-guided embedding rules for high-dimensional data embedding. We introduce a generic deep embedding network (DEN) framework, which is able to learn a parametric mapping from high-dimensional space to low-dimensional space, guided by well-established objectives such as Kullback-Leibler (KL) divergence minimization. We further propose a recursive strategy, called deep recursive embedding (DRE), to make use of the latent data representations for boosted embedding performance. We exemplify the flexibility of DRE by different architectures and loss functions, and benchmarked our method against the two most popular embedding methods, namely, t-distributed stochastic neighbor embedding (t-SNE) and uniform manifold approximation and projection (UMAP). The proposed DRE method can map out-of-sample data and scale to extremely large datasets. Experiments on a range of public datasets demonstrated improved embedding performance in terms of local and global structure preservation, compared with other state-of-the-art embedding methods. Code is available at https://github.com/tao-aimi/DeepRecursiveEmbedding.
RESUMEN
Ultrasound beamforming is a principal factor in high-quality ultrasound imaging. The conventional delay-and-sum (DAS) beamformer generates images with high computational speed but low spatial resolution; thus, many adaptive beamforming methods have been introduced to improve image qualities. However, these adaptive beamforming methods suffer from high computational complexity, which limits their practical applications. Hence, an advanced beamformer that can overcome spatiotemporal resolution bottlenecks is eagerly awaited. In this paper, we propose a novel deep-learning-based algorithm, called the multiconstrained hybrid generative adversarial network (MC-HGAN) beamformer that rapidly achieves high-quality ultrasound imaging. The MC-HGAN beamformer directly establishes a one-shot mapping between the radio frequency signals and the reconstructed ultrasound images through a hybrid generative adversarial network (GAN) model. Through two specific branches, the hybrid GAN model extracts both radio frequency-based and image-based features and integrates them through a fusion module. We also introduce a multiconstrained training strategy to provide comprehensive guidance for the network by invoking intermediates to co-constrain the training process. Moreover, our beamformer is designed to adapt to various ultrasonic emission modes, which improves its generalizability for clinical applications. We conducted experiments on a variety of datasets scanned by line-scan and plane wave emission modes and evaluated the results with both similarity-based and ultrasound-specific metrics. The comparisons demonstrate that the MC-HGAN beamformer generates ultrasound images whose quality is higher than that of images generated by other deep learning-based methods and shows very high robustness in different clinical datasets. This technology also shows great potential in real-time imaging.
Asunto(s)
Algoritmos , Procesamiento de Imagen Asistido por Computador , Humanos , UltrasonografíaRESUMEN
Recently, the use of portable equipment has attracted much attention in the medical ultrasound field. Handheld ultrasound devices have great potential for improving the convenience of diagnosis, but noise-induced artifacts and low resolution limit their application. To enhance the video quality of handheld ultrasound devices, we propose a low-rank representation multipathway generative adversarial network (LRR MPGAN) with a cascade training strategy. This method can directly generate sequential, high-quality ultrasound video with clear tissue structures and details. In the cascade training process, the network is first trained with plane wave (PW) single-/multiangle video pairs to capture dynamic information and then fine-tuned with handheld/high-end image pairs to extract high-quality single-frame information. In the proposed GAN structure, a multipathway generator is applied to implement the cascade training strategy, which can simultaneously extract dynamic information and synthesize multiframe features. The LRR decomposition channel approach guarantees the fine reconstruction of both global features and local details. In addition, a novel ultrasound loss is added to the conventional mean square error (MSE) loss to acquire ultrasound-specific perceptual features. A comprehensive evaluation is conducted in the experiments, and the results confirm that the proposed method can effectively reconstruct high-quality ultrasound videos for handheld devices. With the aid of the proposed method, handheld ultrasound devices can be used to obtain convincing and convenient diagnoses.
Asunto(s)
Ultrasonografía/instrumentación , Algoritmos , Artefactos , Humanos , Procesamiento de Imagen Asistido por Computador/métodos , Aprendizaje Automático , Redes Neurales de la Computación , Ultrasonografía/métodos , Grabación en VideoRESUMEN
As a widely used imaging modality in the medical field, ultrasound has been applied in community medicine, rural medicine, and even telemedicine in recent years. Therefore, the development of portable ultrasound devices has become a popular research topic. However, the limited size of portable ultrasound devices usually degrades the imaging quality, which reduces the diagnostic reliability. To overcome hardware limitations and improve the image quality of portable ultrasound devices, we propose a novel generative adversarial network (GAN) model to achieve mapping between low-quality ultrasound images and corresponding high-quality images. In contrast to the traditional GAN method, our two-stage GAN that cascades a U-Net network prior to the generator as a front end is built to reconstruct the tissue structure, details, and speckle of the reconstructed image. In the training process, an ultrasound plane-wave imaging (PWI) data-based transfer learning method is introduced to facilitate convergence and to eliminate the influence of deformation caused by respiratory activities during data pair acquisition. A gradual tuning strategy is adopted to obtain better results by the PWI transfer learning process. In addition, a comprehensive loss function is presented to combine texture, structure, and perceptual features. Experiments are conducted using simulated, phantom, and clinical data. Our proposed method is compared to four other algorithms, including traditional gray-level-based methods and learning-based methods. The results confirm that the proposed approach obtains the optimum solution for improving quality and offering useful diagnostic information for portable ultrasound images. This technology is of great significance for providing universal medical care.
Asunto(s)
Procesamiento de Imagen Asistido por Computador/métodos , Aprendizaje Automático , Ultrasonografía/métodos , Algoritmos , Arterias Carótidas/diagnóstico por imagen , Diseño de Equipo , Humanos , Procesamiento de Imagen Asistido por Computador/normas , Fantasmas de Imagen , Ultrasonografía/instrumentaciónRESUMEN
In the medical ultrasound field, ultrafast imaging has recently become a hot topic. However, the diagnostic reliability of ultrafast high-frame rate plane-wave (PW) imaging is reduced by its low-quality images. The medical ultrasound equipment on the market usually adopts the line-scanning mode, which can obtain high-quality images at a very low frame rate. In addition, many proven data-driven ultrasound image processing methods are trained by line-scan images. Since the gray-level distributions of line-scan images and PW images are very different, these gray-level distribution-sensitive methods cannot be generalized to ultrafast ultrasound imaging, which limits further applications. Hence, we propose an ultrasound-transfer generative adversarial network to improve the quality of PW images and extend the existing image processing methods to ultrafast ultrasound imaging by reconstructing PW images into line-scan images. This network adopts a residual dense generator with a self-attention system that fully uses the hierarchical features and generates details from all the relevant physiological information. A projection discriminator and spectral normalization are introduced to increase the discernibility and to maintain a balance between the generator and the discriminator. Moreover, we reorganize the transmit sequence of the transducer array to eliminate the negative influence of human movements and facilitate the convergence of the proposed model. The experimental results are evaluated with five metrics, which confirm the feasibility of the proposed method to obtain a line-scan-quality image with a very high frame rate. This technology could significantly popularize ultrafast medical ultrasound imaging.
Asunto(s)
Procesamiento de Imagen Asistido por Computador/métodos , Redes Neurales de la Computación , Ultrasonografía/métodos , Arterias Carótidas/diagnóstico por imagen , Humanos , Músculo Esquelético/diagnóstico por imagenRESUMEN
In recent years, plane-wave imaging (PWI) has attracted considerable attention because of its high temporal resolution. However, the low spatial resolution of PWI limits its clinical applications, which has inspired various studies on the high spatial resolution reconstruction of PW ultrasound images. Although compounding methods and traditional high spatial resolution reconstruction approaches can improve the image quality, these techniques decrease the temporal resolution. Since learning methods can fully reserve the high temporal resolution of PW ultrasounds, a novel convolutional neural network (CNN) model for the high spatial-temporal resolution reconstruction of PW ultrasound images is proposed in this paper. Considering the multiangle form of PW data, a multichannel model is introduced to produce balanced training. To combine local and contextual information, the multiscale model is adopted. These two innovations constitute our multichannel and multiscale CNN (MMCNN) model. Compared with traditional CNN methods, the proposed model uses a two-stage structure in which a cascading wavelet postprocessing stage is combined with the trained MMCNN model. Cascading wavelet postprocessing aims to preserve speckle information. Furthermore, a feedback system is appended to the iteration process of the network training to solve the overfitting problem and help produce convergence. Based on these improvements, an end-to-end mapping is established between a single-angle B-mode PW image and its corresponding multiangle compounded, high-resolution image. The experiments were conducted on simulated, phantom, and real human data. The advantages of our proposed method were compared with a coherent PW compounding method, a conventional maximum a posteriori-based high spatial resolution reconstruction method, and a 2-D CNN compounding method, and the results verified that our approach is capable of attaining a better temporal resolution and comparable spatial resolution. In clinical usage, the proposed method is equipped to satisfy with many ultrafast imaging applications, which require high spatial-temporal resolution. i.
RESUMEN
Recently, advances in computers and high-speed communication tools have led to enhancements in remote medical consultation research. Laws in some localities require hospitals to encrypt patient information (including images of the patient) before transferring the data over a network. Therefore, developing suitable encryption algorithms is quite important for modern medicine. This paper demonstrates a digital image encryption algorithm based on chaotic mapping, which uses the no-period and no-convergence properties of a chaotic sequence to create image chaos and pixel averaging. Then, the chaotic sequence is used to encrypt the image, thereby improving data security. With this method, the security of data and images can be improved.