Enhancing computer image recognition with improved image algorithms.

Advances in computer image recognition have significantly impacted many industries, including healthcare, security and autonomous systems. This paper aims to explore the potential of improving image algorithms to enhance computer image recognition. Specifically, we will focus on regression methods as a means to improve the accuracy and efficiency of identifying images. In this study, we will analyze various regression techniques and their applications in computer image recognition, as well as the resulting performance improvements through detailed examples and data analysis. This paper deals with the problems related to visual image processing in outdoor unstructured environment. Finally, the heterogeneous patterns are converted into the same pattern, and the heterogeneous patterns are extracted from the fusion features of data modes. The simulation results show that the perception ability and recognition ability of outdoor image recognition in complex environment are improved.

Tumor Radiomic Features on Pretreatment MRI to Predict Response to Lenvatinib plus an Anti-PD-1 Antibody in Advanced Hepatocellular Carcinoma: A Multicenter Study.

Introduction: Lenvatinib plus an anti-PD-1 antibody has shown promising antitumor effects in patients with advanced hepatocellular carcinoma (HCC), but with clinical benefit limited to a subset of patients. We developed and validated a radiomic-based model to predict objective response to this combination therapy in advanced HCC patients. Methods: Patients (N = 170) who received first-line combination therapy with lenvatinib plus an anti-PD-1 antibody were retrospectively enrolled from 9 Chinese centers; 124 and 46 into the training and validation cohorts, respectively. Radiomic features were extracted from pretreatment contrast-enhanced MRI. After feature selection, clinicopathologic, radiomic, and clinicopathologic-radiomic models were built using a neural network. The performance of models, incremental predictive value of radiomic features compared with clinicopathologic features and relationship between radiomic features and survivals were assessed. Results: The clinicopathologic model modestly predicted objective response with an AUC of 0.748 (95% CI: 0.656-0.840) and 0.702 (95% CI: 0.547-0.884) in the training and validation cohorts, respectively. The radiomic model predicted response with an AUC of 0.886 (95% CI: 0.815-0.957) and 0.820 (95% CI: 0.648-0.984), respectively, with good calibration and clinical utility. The incremental predictive value of radiomic features to clinicopathologic features was confirmed with a net reclassification index of 47.9% (p < 0.001) and 41.5% (p = 0.025) in the training and validation cohorts, respectively. Furthermore, radiomic features were associated with overall survival and progression-free survival both in the training and validation cohorts, but modified albumin-bilirubin grade and neutrophil-to-lymphocyte ratio were not. Conclusion: Radiomic features extracted from pretreatment MRI can predict individualized objective response to combination therapy with lenvatinib plus an anti-PD-1 antibody in patients with unresectable or advanced HCC, provide incremental predictive value over clinicopathologic features, and are associated with overall survival and progression-free survival after initiation of this combination regimen.

Radiographic and α-fetoprotein response predict pathologic complete response to immunotherapy plus a TKI in hepatocellular carcinoma: a multicenter study.

BACKGROUND: Pathologic complete response (pCR) following preoperative systemic therapy is associated with improved outcomes after subsequent liver transplant/resection in hepatocellular carcinoma (HCC). However, the relationship between radiographic and histopathological response remains unclear. METHODS: We retrospectively examined patients with initially unresectable HCC who received tyrosine kinase inhibitor (TKI) plus anti-programmed death 1 (PD-1) therapy before undergoing liver resection between March 2019 and September 2021 across 7 hospitals in China. Radiographic response was evaluated using mRECIST. A pCR was defined as no viable tumor cells in resected samples. RESULTS: We included 35 eligible patients, of whom 15 (42.9%) achieved pCR after systemic therapy. After a median follow-up of 13.2 months, tumors recurred in 8 non-pCR and 1 pCR patient. Before resection, there were 6 complete responses, 24 partial responses, 4 stable disease cases, and 1 progressive disease case, per mRECIST. Predicting pCR by radiographic response yielded an area under the receiver operating characteristic curve (AUC) of 0.727 (95% CI: 0.558-0.902), with an optimal cutoff value of 80% reduction in the enhanced area in MRI (called major radiographic response), which had a 66.7% sensitivity, 85.0% specificity, and a 77.1% diagnostic accuracy. When radiographic response was combined with α-fetoprotein response, the AUC was 0.926 (95% CI: 0.785-0.999); the optimal cutoff value was 0.446, which had a 91.7% sensitivity, 84.6%, specificity, and an 88.0% diagnostic accuracy. CONCLUSIONS: In patients with unresectable HCC receiving combined TKI/anti-PD 1 therapy, major radiographic response alone or combined with α-fetoprotein response may predict pCR.

Characteristics of different asthma phenotypes associated with cough: a prospective, multicenter survey in China.

BACKGROUND: Asthma is a heterogeneous disease with variable symptoms, which presents with cough either as the sole or predominant symptom with or without wheezing. We compared the clinical and pathophysiological characteristics of cough predominant asthma (CPA), cough variant asthma (CVA) and classic asthma (CA) in order to determine any differential phenotypic traits. METHODS: In 20 clinics across China, a total of 2088 patients were finally recruited, including 327 CVA, 1041 CPA and 720 CA patients. We recorded cough and wheezing visual analogue scale, Leicester cough questionnaire (LCQ) and asthma control test scores. Fractional exhaled nitric oxide (FeNO), induced sputum cell counts, and capsaicin cough challenge were also measured and compared. RESULTS: CPA patients more frequently presented with cough as the initial symptom, and laryngeal symptoms (p < 0.001), had less symptoms related with rhinitis/sinusitis and gastroesophageal reflux (p < 0.05) than CA patients. Comorbidities including rhinitis and gastroesophageal reflux were similar, while the proportion of COPD and bronchiectasis was higher in CA patients. There were no differences in FeNO levels, sputum eosinophil and neutrophil counts, FEV1 (%pred) decreased from CVA to CPA to CA patients (p < 0.001). Cough sensitivity was higher in CVA and CPA compared to CA (p < 0.001), and was positively correlated with LCQ scores. CONCLUSIONS: CVA, CPA and CA can be distinguished by the presence of laryngeal symptoms, cough sensitivity and airflow obstruction. Asthma-associated chronic cough was not associated with airway inflammation or comorbidities in our cohort. Trial registration The Chinese Clinical Trial Registration Center, ChiCTR-POC-17011646, 13 June 2017.

Reversible rearrangement of magnetic nanoparticles in solution studied using time-resolved SAXS method.

Superparamagnetic nanoparticles have broad applications in biology and medicines. Quantitative measurements of magnetic beads in solution are essential in gaining comprehensive understanding of their dynamics and developing applications. Here, using synchrotron X-ray sources combined with well controlled magnetic fields, the results from small-angle X-ray scattering (SAXS) experiments on superparamagnetic particles in solution under the influence of external magnetic fields are reported. The particles mostly remain in monodispersed states and the linear aggregates tend to be aligned with the external magnetic field. After removing the magnetic fields, the superparamagnetic nanoparticles quickly recover to their original states indicating high reversibility of the rearrangement under the control of a magnetic field. The external magnetic field instrument composed of paired permanent magnets is integrated into the SAXS beamline at the Shanghai Synchrotron Radiation Facility providing a platform for studying time-resolved dynamics induced by magnetic fields.

Computer aided protein engineering to enhance the thermo-stability of CXCR1- T4 lysozyme complex.

CXCR1, a member in G-protein coupled receptor (GPCR) family, binds to chemokine interleukin-8 (IL-8) specifically and transduces signals to mediate immune and inflammatory responses. Despite the importance of CXCR1, high-resolution structure determination is hindered by the challenges in crystallization. It has been shown that properly designed mutants with enhanced thermostability, together with fusion partner proteins, can be useful to form crystals for GPCR proteins. In this study, in silico protein design was carried out by using homology modeling and molecular dynamics simulations. To validate the computational modeling results, the thermostability of several mutants and the wild type were measured experimentally. Both computational results and experimental data suggest that the mutant L126W has a significant improvement in the thermostability. This study demonstrated that in silico design can guide protein engineering and potentially facilitate protein crystallography research.

Using Quantitative Seroproteomics to Identify Antibody Biomarkers in Pancreatic Cancer.

Pancreatic cancer is the fourth leading cause of cancer-related deaths in the United States. Less than 6% of patients survive beyond the fifth year due to inadequate early diagnostics and ineffective treatment options. Our laboratory has developed an allogeneic, granulocyte-macrophage colony-stimulating factor (GM-CSF)-secreting pancreatic cancer vaccine (GVAX) that has been tested in phase II clinical trials. Here, we employed a serum antibodies-based SILAC immunoprecipitation (SASI) approach to identify proteins that elicit an antibody response after vaccination. The SASI approach uses immunoprecipitation with patient-derived antibodies that is coupled to quantitative stable isotope-labeled amino acids in cell culture (SILAC). Using mass spectrometric analysis, we identified more than 150 different proteins that induce an antibody response after vaccination. The regulatory subunit 12A of protein phosphatase 1 (MYPT1 or PPP1R12A), regulatory subunit 8 of the 26S proteasome (PSMC5), and the transferrin receptor (TFRC) were shown to be pancreatic cancer-associated antigens recognized by postvaccination antibodies in the sera of patients with favorable disease-free survival after GVAX therapy. We further interrogated these proteins in over 80 GVAX-treated patients' pancreases and uniformly found a significant increase in the expression of MYPT1, PSMC5, and TFRC in neoplastic compared with non-neoplastic pancreatic ductal epithelium. We show that the novel SASI approach can identify antibody targets specifically expressed in patients with improved disease-free survival after cancer vaccine therapy. These targets need further validation to be considered as possible pancreatic cancer biomarkers.

Prediction of structural features and application to outer membrane protein identification.

Protein three-dimensional (3D) structures provide insightful information in many fields of biology. One-dimensional properties derived from 3D structures such as secondary structure, residue solvent accessibility, residue depth and backbone torsion angles are helpful to protein function prediction, fold recognition and ab initio folding. Here, we predict various structural features with the assistance of neural network learning. Based on an independent test dataset, protein secondary structure prediction generates an overall Q3 accuracy of ~80%. Meanwhile, the prediction of relative solvent accessibility obtains the highest mean absolute error of 0.164, and prediction of residue depth achieves the lowest mean absolute error of 0.062. We further improve the outer membrane protein identification by including the predicted structural features in a scoring function using a simple profile-to-profile alignment. The results demonstrate that the accuracy of outer membrane protein identification can be improved by ~3% at a 1% false positive level when structural features are incorporated. Finally, our methods are available as two convenient and easy-to-use programs. One is PSSM-2-Features for predicting secondary structure, relative solvent accessibility, residue depth and backbone torsion angles, the other is PPA-OMP for identifying outer membrane proteins from proteomes.

B7-H3 Promotes Pathogenesis of Autoimmune Disease and Inflammation by Regulating the Activity of Different T Cell Subsets.

B7-H3 is a cell surface molecule in the immunoglobulin superfamily that is frequently upregulated in response to autoantigens and pathogens during host T cell immune responses. However, B7-H3's role in the differential regulation of T cell subsets remains largely unknown. Therefore, we constructed a new B7-H3 deficient mouse strain (B7-H3 KO) and evaluated the functions of B7-H3 in the regulation of Th1, Th2, and Th17 subsets in experimental autoimmune encephalomyelitis (EAE), experimental asthma, and collagen-induced arthritis (CIA); these mouse models were used to predict human immune responses in multiple sclerosis, asthma, and rheumatoid arthritis, respectively. Here, we demonstrate that B7-H3 KO mice have significantly less inflammation, decreased pathogenesis, and limited disease progression in both EAE and CIA mouse models when compared with littermates; these results were accompanied by a decrease in IFN-γ and IL-17 production. In sharp contrast, B7-H3 KO mice developed severe ovalbumin (OVA)-induced asthma with characteristic infiltrations of eosinophils in the lung, increased IL-5 and IL-13 in lavage fluid, and elevated IgE anti-OVA antibodies in the blood. Our results suggest B7-H3 has a costimulatory function on Th1/Th17 but a coinhibitory function on Th2 responses. Our studies reveal that B7-H3 could affect different T cell subsets which have important implications for regulating pathogenesis and disease progression in human autoimmune disease.

Galectin-3 Shapes Antitumor Immune Responses by Suppressing CD8+ T Cells via LAG-3 and Inhibiting Expansion of Plasmacytoid Dendritic Cells.

Galectin-3 is a 31-kDa lectin that modulates T-cell responses through several mechanisms, including apoptosis, T-cell receptor (TCR) cross-linking, and TCR downregulation. We found that patients with pancreatic ductal adenocarcinoma (PDA) who responded to a granulocyte-macrophage colony-stimulating factor-secreting allogeneic PDA vaccine developed neutralizing antibodies to galectin-3 after immunization. We show that galectin-3 binds activated antigen-committed CD8(+) T cells only in the tumor microenvironment. Galectin-3-deficient mice exhibit improved CD8(+) T-cell effector function and increased expression of several inflammatory genes. Galectin-3 binds to LAG-3, and LAG-3 expression is necessary for galectin-3-mediated suppression of CD8(+) T cells in vitro. Lastly, galectin-3-deficient mice have elevated levels of circulating plasmacytoid dendritic cells, which are superior to conventional dendritic cells in activating CD8(+) T cells. Thus, inhibiting galectin-3 in conjunction with CD8(+) T-cell-directed immunotherapies should enhance the tumor-specific immune response.

GPCRserver: an accurate and novel G protein-coupled receptor predictor.

G protein coupled receptors (GPCRs), also known as seven-transmembrane domain receptors, pass through the cellular membrane seven times and play diverse biological roles in the cells such as signaling, transporting of molecules and cell-cell communication. In this work, we develop a web server, namely the GPCRserver, which is capable of identifying GPCRs from genomic sequences, and locating their transmembrane regions. The GPCRserver contains three modules: (1) the Trans-GPCR for the transmembrane region prediction by using sequence evolutionary profiles with the assistance of neural network training, (2) the SSEA-GPCR for identifying GPCRs from genomic data by using secondary structure element alignment, and (3) the PPA-GPCR for identifying GPCRs by using profile-to-profile alignment. Our predictor was strictly benchmarked and showed its favorable performance in the real application. The web server and stand-alone programs are publicly available at .

Prediction of outer membrane proteins by combining the position- and composition-based features of sequence profiles.

Locating the transmembrane regions of outer membrane proteins (OMPs) is highly important for deciphering their biological functions at both molecular and cellular levels. Here, we propose a novel method to predict the transmembrane regions of OMPs by employing the position- and composition-based features of sequence profiles. Furthermore, a simple probability-based prediction model, which is estimated by the secondary structures of structurally known OMPs, is also developed. Considering that these two methods are both effective and well complementary, we integrate them into a method called TransOMP, which is also capable of identifying OMPs. Furthermore, we develop an OMP identification measure I_CScore by considering transmembrane regions by TransOMP and secondary structural topology by SSEA-OMP. Our methods were benchmarked against state-of-the-art methods and assessed in the genome of Escherichia coli. Benchmark results confirmed that our methods were reliable and useful. Meanwhile, we constructed an OMP prediction web server, which can be used for OMP identification, transmembrane region location, and 3D model building.

Tyrosine 23 phosphorylation-dependent cell-surface localization of annexin A2 is required for invasion and metastases of pancreatic cancer.

The aggressiveness of pancreatic ductal adenocarcinoma (PDA) is characterized by its high metastatic potential and lack of effective therapies, which is the result of a lack of understanding of the mechanisms involved in promoting PDA metastases. We identified Annexin A2 (ANXA2), a member of the Annexin family of calcium-dependent phospholipid binding proteins, as a new molecule that promotes PDA invasion and metastases. We found ANXA2 to be a PDA-associated antigen recognized by post-treatment sera of patients who demonstrated prolonged survival following treatment with a PDA-specific vaccine. Cell surface ANXA2 increases with PDA development and progression. Knockdown of ANXA2 expression by RNA interference or blocking with anti-ANXA2 antibodies inhibits in vitro invasion of PDA cells. In addition, post-vaccination patient sera inhibits in vitro invasion of PDA cells, suggesting that therapeutic anti-ANXA2 antibodies are induced by the vaccine. Furthermore, cell-surface localization of ANXA2 is tyrosine 23 phosphorylation-dependent; and tyrosine 23 phosphorylation is required for PDA invasion. We demonstrated that tyrosine 23 phosphorylation resulting in surface expression of ANXA2 is required for TGFß-induced, Rho-mediated epithelial-mesenchymal transition (EMT), linking the cellular function of ANXA2 which was previously shown to be associated with small GTPase-regulated cytoskeletal rearrangements, to the EMT process in PDA. Finally, using mouse PDA models, we showed that shRNA knock-down of ANXA2, a mutation at tyrosine 23, or anti-ANXA2 antibodies, inhibit PDA metastases and prolong mouse survival. Thus, ANXA2 is part of a novel molecular pathway underlying PDA metastases and a new target for development of PDA therapeutics.

Nondominant CD8 T cells are active players in the vaccine-induced antitumor immune response.

We previously reported that CD8(+) T cells are directed predominantly toward the immunodominant Her-2/neu (neu) epitope RNEU(420-429) in nontolerized FVB/N but not tolerized HER-2/neu (neu-N) mice. In this study, we screened overlapping peptides of the entire neu protein and identified six new epitopes recognized by vaccine-induced neu-N-derived T cells. Evaluation of individual nondominant responses by tetramer staining and IFN-γ secretion demonstrate that this repertoire is peripherally tolerized. To address the role that the complete CD8(+) T cell repertoire plays in vaccine-induced antitumor immunity, we created a whole-cell vaccine-expressing neu cDNA that has been mutated at the RNEU(420-429) anchor residue, thereby abrogating activation of immunodominant epitope responses. Studies comparing the mutated and nonmutated vaccines indicate that nondominant CD8(+) T cells can induce antitumor immunity when combined with regulatory T cell-depleting agents in both neu-N and FVB/N mice. Collectively, these studies demonstrate that the neu-directed T cell repertoire is not intrinsically incapable of eradicating tumors. Rather, they are suppressed by mechanisms of peripheral tolerance. Thus, these studies provide new insights into the function of the complete T cell repertoire directed toward a clinically relevant tumor Ag in tumor-bearing hosts.

Ectopic expression of vascular cell adhesion molecule-1 as a new mechanism for tumor immune evasion.

Immune escape is an important reason why the immune system cannot control tumor growth, but how escape variants emerge during immunotherapy remains poorly understood. Here, we identify a new mechanism of tumor immune escape using an in vivo selection strategy. We generated a highly immune-resistant cancer cell line (P3) by subjecting a susceptible cancer cell line (P0/TC-1) to multiple rounds of in vivo immune selection. Microarray analysis of P0 and P3 revealed that vascular cell adhesion molecule-1 (VCAM-1) is up-regulated in the P3-resistant variant. Retroviral transfer of VCAM-1 into P0 significantly increased its resistance against a vaccine-induced immune response. Analysis of tumors showed a dramatic decrease in the number of tumor-infiltrating cluster of differentiation 8(+) (CD8(+)) T cells in the tumors expressing VCAM-1. In vitro transwell migration assays showed that VCAM-1 can promote the migration of CD8(+) T cells through its interaction with the alpha(4)beta(1) integrin. Site-directed mutagenesis of VCAM-1 at amino acid residues required for interaction with alpha(4)beta(1) integrin completely abolished the immune resistance conferred by VCAM-1 in vivo. Surface staining showed that most renal cell carcinomas (RCC) express VCAM-1, whereas an RCC that responded to vaccination was VCAM-1 negative. These data provide evidence that tumor expression of VCAM-1 represents a new mechanism of immune evasion and has important implications for the development of immunotherapy for human RCC.

Diverse CD8+ T-cell responses to renal cell carcinoma antigens in patients treated with an autologous granulocyte-macrophage colony-stimulating factor gene-transduced renal tumor cell vaccine.

A phase I clinical trial with granulocyte-macrophage colony-stimulating factor tumor cell vaccines in patients with metastatic renal cell carcinoma (RCC) showed immune cell infiltration at vaccine sites and delayed-type hypersensitivity (DTH) responses to autologous tumor cells indicative of T-cell immunity. To further characterize RCC T-cell responses and identify relevant RCC-associated antigens, we did a detailed analysis of CD8+ T-cell responses in two vaccinated RCC patients who generated the greatest magnitude of DTH response and also displayed a strong clinical response to vaccination (>90% reduction in metastatic tumor volume). Three separate CD8+ T-cell lines (and subsequent derived clones) derived from patient 24 recognized distinct RCC-associated antigens. One recognized a shared HLA-A*0201-restricted antigen expressed by both renal cancer cells and normal kidney cells. This recognition pattern correlated with a positive DTH test to normal kidney cells despite no evidence of impairment of renal function by the patient's remaining kidney after vaccination. A second line recognized a shared HLA-C7-restricted antigen that was IFN-gamma inducible. A third line recognized a unique HLA-A*0101-restricted RCC antigen derived from a mutated KIAA1440 gene specific to the tumor. In addition, two independent CTL lines and three clones were also generated from patient 26 and they recognized autologous tumor cells restricted through HLA-A*0205, HLA-A/B/C, and HLA-B/C. These results show that paracrine granulocyte-macrophage colony-stimulating factor tumor vaccines may generate a diverse repertoire of tumor-reactive CD8+ T-cell responses and emphasize the importance of polyvalency in the design of cancer immunotherapies.

Mesothelin-specific CD8(+) T cell responses provide evidence of in vivo cross-priming by antigen-presenting cells in vaccinated pancreatic cancer patients.

Tumor-specific CD8(+) T cells can potentially be activated by two distinct mechanisms of major histocompatibility complex class I-restricted antigen presentation as follows: direct presentation by tumor cells themselves or indirect presentation by professional antigen-presenting cells (APCs). However, controversy still exists as to whether indirect presentation (the cross-priming mechanism) can contribute to effective in vivo priming of tumor-specific CD8(+) T cells that are capable of eradicating cancer in patients. A clinical trial of vaccination with granulocyte macrophage-colony stimulating factor-transduced pancreatic cancer lines was designed to test whether cross-presentation by locally recruited APCs can activate pancreatic tumor-specific CD8(+) T cells. Previously, we reported postvaccination delayed-type hypersensitivity (DTH) responses to autologous tumor in 3 out of 14 treated patients. Mesothelin is an antigen demonstrated previously by gene expression profiling to be up-regulated in most pancreatic cancers. We report here the consistent induction of CD8(+) T cell responses to multiple HLA-A2, A3, and A24-restricted mesothelin epitopes exclusively in the three patients with vaccine-induced DTH responses. Importantly, neither of the vaccinating pancreatic cancer cell lines expressed HLA-A2, A3, or A24. These results provide the first direct evidence that CD8 T cell responses can be generated via cross-presentation by an immunotherapy approach designed to recruit APCs to the vaccination site.