Deep learning-assisted diagnosis of benign and malignant parotid tumors based on contrast-enhanced CT: a multicenter study.

OBJECTIVES: To develop deep learning-assisted diagnosis models based on CT images to facilitate radiologists in differentiating benign and malignant parotid tumors. METHODS: Data from 573 patients with histopathologically confirmed parotid tumors from center 1 (training set: n = 269; internal-testing set: n = 116) and center 2 (external-testing set: n = 188) were retrospectively collected. Six deep learning models (MobileNet V3, ShuffleNet V2, Inception V3, DenseNet 121, ResNet 50, and VGG 19) based on arterial-phase CT images, and a baseline support vector machine (SVM) model integrating clinical-radiological features with handcrafted radiomics signatures were constructed. The performance of senior and junior radiologists with and without optimal model assistance was compared. The net reclassification index (NRI) and integrated discrimination improvement (IDI) were calculated to evaluate the clinical benefit of using the optimal model. RESULTS: MobileNet V3 had the best predictive performance, with sensitivity increases of 0.111 and 0.207 (p < 0.05) in the internal- and external-testing sets, respectively, relative to the SVM model. Clinical benefit and overall efficiency of junior radiologist were significantly improved with model assistance; for the internal- and external-testing sets, respectively, the AUCs improved by 0.128 and 0.102 (p < 0.05), the sensitivity improved by 0.194 and 0.120 (p < 0.05), the NRIs were 0.257 and 0.205 (p < 0.001), and the IDIs were 0.316 and 0.252 (p < 0.001). CONCLUSIONS: The developed deep learning models can assist radiologists in achieving higher diagnostic performance and hopefully provide more valuable information for clinical decision-making in patients with parotid tumors. KEY POINTS: • The developed deep learning models outperformed the traditional SVM model in predicting benign and malignant parotid tumors. • Junior radiologist can obtain greater clinical benefits with assistance from the optimal deep learning model. • The clinical decision-making process can be accelerated in patients with parotid tumors using the established deep learning model.

Bacterial transformation of lignin: key enzymes and high-value products.

Lignin, a natural organic polymer that is recyclable and inexpensive, serves as one of the most abundant green resources in nature. With the increasing consumption of fossil fuels and the deterioration of the environment, the development and utilization of renewable resources have attracted considerable attention. Therefore, the effective and comprehensive utilization of lignin has become an important global research topic, with the goal of environmental protection and economic development. This review focused on the bacteria and enzymes that can bio-transform lignin, focusing on the main ways that lignin can be utilized to produce high-value chemical products. Bacillus has demonstrated the most prominent effect on lignin degradation, with 89% lignin degradation by Bacillus cereus. Furthermore, several bacterial enzymes were discussed that can act on lignin, with the main enzymes consisting of dye-decolorizing peroxidases and laccase. Finally, low-molecular-weight lignin compounds were converted into value-added products through specific reaction pathways. These bacteria and enzymes may become potential candidates for efficient lignin degradation in the future, providing a method for lignin high-value conversion. In addition, the bacterial metabolic pathways convert lignin-derived aromatics into intermediates through the "biological funnel", achieving the biosynthesis of value-added products. The utilization of this "biological funnel" of aromatic compounds may address the heterogeneous issue of the aromatic products obtained via lignin depolymerization. This may also simplify the separation of downstream target products and provide avenues for the commercial application of lignin conversion into high-value products.

Integrating genomics, molecular docking, and protein expression to explore new perspectives on polystyrene biodegradation.

Faced with the escalating challenge of global plastic pollution, this study specifically addresses the research gap in the biodegradation of polystyrene (PS). A PS-degrading bacterial strain was isolated from the gut of Tenebrio molitor, and genomics, molecular docking, and proteomics were employed to thoroughly investigate the biodegradation mechanisms of Pseudomonas putida H-01 against PS. Using scanning electron microscopy (SEM), Fourier-transform infrared spectroscopy (ATR-FTIR), and contact angle analysis, significant morphological and structural changes in the PS films under the influence of the H-01 strain were observed. The study revealed several potential degradation genes and ten enzymes that were specifically upregulated in the PS degradation environment. Additionally, a novel protein with laccase-like activity, LacQ1, was purified from this strain for the first time, and its crucial role in the PS degradation process was confirmed. Through molecular docking and molecular dynamics (MD) simulations, the interactions between the enzymes and PS were detailed, elucidating the binding and catalytic mechanisms of the degradative enzymes with the substrate. These findings have deepened our understanding of PS degradation.

The Effect of Magnetic Resonance Imaging Based Radiomics Models in Discriminating stage I-II and III-IVa Nasopharyngeal Carcinoma.

BACKGROUND: Nasopharyngeal carcinoma (NPC) is a common tumor in China. Accurate stages of NPC are crucial for treatment. We therefore aim to develop radiomics models for discriminating early-stage (I-II) and advanced-stage (III-IVa) NPC based on MR images. METHODS: 329 NPC patients were enrolled and randomly divided into a training cohort (n = 229) and a validation cohort (n = 100). Features were extracted based on axial contrast-enhanced T1-weighted images (CE-T1WI), T1WI, and T2-weighted images (T2WI). Least absolute shrinkage and selection operator (LASSO) was used to build radiomics signatures. Seven radiomics models were constructed with logistic regression. The AUC value was used to assess classification performance. The DeLong test was used to compare the AUCs of different radiomics models and visual assessment. RESULTS: Models A, B, C, D, E, F, and G were constructed with 13, 9, 7, 9, 10, 7, and 6 features, respectively. All radiomics models showed better classification performance than that of visual assessment. Model A (CE-T1WI + T1WI + T2WI) showed the best classification performance (AUC: 0.847) in the training cohort. CE-T1WI showed the greatest significance for staging NPC. CONCLUSION: Radiomics models can effectively distinguish early-stage from advanced-stage NPC patients, and Model A (CE-T1WI + T1WI + T2WI) showed the best classification performance.

A Combined Model Integrating Radiomics and Deep Learning Based on Contrast-Enhanced CT for Preoperative Staging of Laryngeal Carcinoma.

RATIONALE AND OBJECTIVES: Accurate staging of laryngeal carcinoma can inform appropriate treatment decision-making. We developed a radiomics model, a deep learning (DL) model, and a combined model (incorporating radiomics features and DL features) based on the venous-phase CT images and explored the performance of these models in stratifying patients with laryngeal carcinoma into stage I-II and stage III-IV, and also compared these models with radiologists. MATERIALS AND METHODS: Three hundreds and nineteen patients with pathologically confirmed laryngeal carcinoma were randomly divided into a training set (n = 223) and a test set (n = 96). In the training set, the radiomics features with inter- and intraclass correlation coefficients (ICCs)> 0.75 were screened by Spearman correlation analysis and recursive feature elimination (RFE); then support vector machine (SVM) classifier was applied to develop the radiomics model. The DL model was built using ResNet 18 by the cropped 2D regions of interest (ROIs) in the maximum tumor ROI slices and the last fully connected layer of this network served as the DL feature extractor. Finally, a combined model was developed by pooling the radiomics features and extracted DL features to predict the staging. RESULTS: The area under the curves (AUCs) for radiomics model, DL model, and combined model in the test set were 0.704 (95% confidence interval [CI]: 0.588-0.820), 0.724 (95% CI: 0.613-0.835), and 0.849 (95% CI: 0.755-0.943), respectively. The combined model outperformed the radiomics model and the DL model in discriminating stage I-II from stage III-IV (p = 0.031 and p = 0.020, respectively). Only the combined model performed significantly better than radiologists (p < 0.050 for both). CONCLUSION: The combined model can help tailor the therapeutic strategy for laryngeal carcinoma patients by enabling more accurate preoperative staging.

Altered connectivity in the cognitive control-related prefrontal cortex in Parkinson's disease with rapid eye movement sleep behavior disorder.

Rapid eye movement sleep behavior disorder (RBD) frequently occurs in Parkinson's disease (PD), however, the exact pathophysiological mechanism is not clear. The prefrontal cortex (PFC), especially ventrolateral prefrontal cortex (VLPFC), dorsolateral prefrontal cortex (DLPFC), and inferior frontal gyrus (IFG) which may play roles by regulating cognitive control processes. The purpose of this study was to investigate whether there is abnormal functional connectivity (FC) maps and volume changes in PD with RBD(PD-RBD). We recruited 20 PD-RBD, 20 PD without RBD (PD-nRBD), and 20 normal controls (NC). We utilized resting-state functional Magnetic Resonance Imaging (rs-MRI) to explore FC changes based on regions of interest (VLPFC, DLPFC, and IFG), and used voxel-based morphology technology to analyze whole-brain volumes by 3D-T1 structural MRI. Except the REM sleep behavioral disorders questionnaire (RBDSQ), the PD-RBD showed lower visuospatial/executive and attention scores than the NC group. The RBDSQ scores were significantly positively correlated with zFC of right DLPFC to bilateral posterior cingulate cortex (PCC) (P = 0.0362, R = 0.4708, AlphaSim corrected) and also significantly positively correlated with zFC of left VLPFC to right inferior temporal (P = 0.0157, R = 0.5323, AlphaSim corrected) in PD-RBD group. Furthermore, abnormal correlations with zFC values were also found in some cognitive subdomains in PD-RBD group. The study may suggest that in PD-RBD patients, the presence of RBD may be related to the abnormal FC of VLPFC and DLPFC, meanwhile, the abnormal FC of DLPFC and IFG may be related to the mechanisms of cognitive impairment.

Multiphasic CT-Based Radiomics Analysis for the Differentiation of Benign and Malignant Parotid Tumors.

Objective: This study aims to investigate the value of machine learning models based on clinical-radiological features and multiphasic CT radiomics features in the differentiation of benign parotid tumors (BPTs) and malignant parotid tumors (MPTs). Methods: This retrospective study included 312 patients (205 cases of BPTs and 107 cases of MPTs) who underwent multiphasic enhanced CT examinations, which were randomly divided into training (N = 218) and test (N = 94) sets. The radiomics features were extracted from the plain, arterial, and venous phases. The synthetic minority oversampling technique was used to balance minority class samples in the training set. Feature selection methods were done using the least absolute shrinkage and selection operator (LASSO), mutual information (MI), and recursive feature extraction (RFE). Two machine learning classifiers, support vector machine (SVM), and logistic regression (LR), were then combined in pairs with three feature selection methods to build different radiomics models. Meanwhile, the prediction performances of different radiomics models based on single phase (plain, arterial, and venous phase) and multiphase (three-phase combination) were compared to determine which model construction method and phase were more discriminative. In addition, clinical models based on clinical-radiological features and combined models integrating radiomics features and clinical-radiological features were established. The prediction performances of the different models were evaluated by the area under the receiver operating characteristic (ROC) curve (AUC) and the drawing of calibration curves. Results: Among the 24 established radiomics models composed of four different phases, three feature selection methods, and two machine learning classifiers, the LASSO-SVM model based on a three-phase combination had the optimal prediction performance with AUC (0.936 [95% CI = 0.866, 0.976]), sensitivity (0.78), specificity (0.90), and accuracy (0.86) in the test set, and its prediction performance was significantly better than with the clinical model based on LR (AUC = 0.781, p = 0.012). In the test set, the combined model based on LR had a lower AUC than the optimal radiomics model (AUC = 0.933 vs. 0.936), but no statistically significant difference (p = 0.888). Conclusion: Multiphasic CT-based radiomics analysis showed a machine learning model based on clinical-radiological features and radiomics features has the potential to provide a valuable tool for discriminating benign from malignant parotid tumors.

MRI radiomics-based machine learning model integrated with clinic-radiological features for preoperative differentiation of sinonasal inverted papilloma and malignant sinonasal tumors.

Objective: To explore the best MRI radiomics-based machine learning model for differentiation of sinonasal inverted papilloma (SNIP) and malignant sinonasal tumor (MST), and investigate whether the combination of radiomics features and clinic-radiological features can produce a superior diagnostic performance. Methods: The database of 247 patients with SNIP (n=106) or MST (n=141) were analyzed. Dataset from scanner A were randomly divided into training set (n=135) and test set 1 (n=58) in a ratio of 7:3, and dataset from scanner B and C were used as an additional independent test set 2 (n=54). Fourteen clinic-radiological features were analyzed by using univariate analysis, and those with significant differences were applied to construct clinical model. Based on the radiomics features extracted from single sequence (T2WI or CE-T1WI) and combined sequence, four commonly used classifiers (logistic regression (LR), support vector machine (SVM), decision tree (DT) and k-nearest neighbor (KNN)) were employed to constitute twelve different machine learning models, and the best-performing one was confirmed as the optimal radiomics model. Furthermore, a combined model incorporated best radiomics feature subsets and clinic-radiological features was developed. The diagnostic performances of these models were assessed by the area under the receiver operating characteristic (ROC) curve (AUC) and the calibration curves. Results: Five clinic-radiological features (age, convoluted cerebriform pattern sign, heterogeneity, adjacent bone involvement and infiltration of surrounding tissue) were considered to be significantly different between the tumor groups (P < 0.05). Among the twelve machine learning models, the T2WI-SVM model exhibited optimal predictive efficacy for classification tasks on the two test sets, with the AUC of 0.878 and 0.914, respectively. For three types of diagnostic models, the combined model achieved highest AUC of 0.912 (95%CI: 0.807-0.970) and 0.927 (95%CI: 0.823-0.980) for differentiation of SNIP and MST in test 1 and test 2 sets, which performed prominently better than clinical model (P=0.011, 0.005), but not significantly different from the optimal radiomics model (P=0.100, 0.452). Conclusion: The machine learning model based on T2WI sequence and SVM classifier achieved best performance in differentiation of SNIP and MST, and the combination of radiomics features and clinic-radiological features significantly improved the diagnostic capability of the model.

Mechanisms of BPA Degradation and Toxicity Resistance in Rhodococcus equi.

Bisphenol A (BPA) pollution poses an increasingly serious problem. BPA has been detected in a variety of environmental media and human tissues. Microbial degradation is an effective method of environmental BPA remediation. However, BPA is also biotoxic to microorganisms. In this study, Rhodococcus equi DSSKP-R-001 (R-001) was used to degrade BPA, and the effects of BPA on the growth metabolism, gene expression patterns, and toxicity-resistance mechanisms of Rhodococcus equi were analyzed. The results showed that R-001 degraded 51.2% of 5 mg/L BPA and that 40 mg/L BPA was the maximum BPA concentration tolerated by strain R-001. Cytochrome P450 monooxygenase and multicopper oxidases played key roles in BPA degradation. However, BPA was toxic to strain R-001, exhibiting nonlinear inhibitory effects on the growth and metabolism of this bacterium. R-001 bacterial biomass, total protein content, and ATP content exhibited V-shaped trends as BPA concentration increased. The toxic effects of BPA included the downregulation of R-001 genes related to glycolysis/gluconeogenesis, pentose phosphate metabolism, and glyoxylate and dicarboxylate metabolism. Genes involved in aspects of the BPA-resistance response, such as base excision repair, osmoprotectant transport, iron-complex transport, and some energy metabolisms, were upregulated to mitigate the loss of energy associated with BPA exposure. This study helped to clarify the bacterial mechanisms involved in BPA biodegradation and toxicity resistance, and our results provide a theoretical basis for the application of strain R-001 in BPA pollution treatments.

Cell cycle-dependent regulation of a human DNA helicase that localizes in DNA damage foci.

Mutational studies of human DNA helicase B (HDHB) have suggested that its activity is critical for the G1/S transition of the cell cycle, but the nature of its role remains unknown. In this study, we show that during G1, ectopically expressed HDHB localizes in nuclear foci induced by DNA damaging agents and that this focal pattern requires active HDHB. During S and G2/M, HDHB localizes primarily in the cytoplasm. A carboxy-terminal domain from HDHB confers cell cycle-dependent localization, but not the focal pattern, to a reporter protein. A cluster of potential cyclin-dependent kinase phosphorylation sites in this domain was modified at the G1/S transition and maintained through G2/M of the cell cycle in vivo, coincident with nuclear export of HDHB. Serine 967 of HDHB was the major site phosphorylated in vivo and in vitro by cyclin-dependent kinases. Mutational analysis demonstrated that phosphorylation of serine 967 is crucial in regulating the subcellular localization of ectopically expressed HDHB. We propose that the helicase of HDHB operates primarily during G1 to process endogenous DNA damage before the G1/S transition, and it is largely sequestered in the cytoplasm during S/G2.

Identification of Anti-EGFR and Anti-ErbB3 Dual Variable Domains Immunoglobulin (DVD-Ig) Proteins with Unique Activities.

Epidermal growth factor receptor (EGFR) and receptor tyrosine-protein kinase 3 (ErbB3) are two well-established targets in cancer therapy. There is significant crosstalk among these two receptors and others. To block signaling from both EGFR and ErbB3, we generated anti-EGFR and anti-ErbB3 DVD-Ig proteins. Two DVD-Ig proteins maintained the functions of the combination of the two parental antibodies. The DVD-Ig proteins inhibit cell signaling and proliferation in A431 and FaDu cells while half DVD-Ig proteins lost proliferation inhibition function. Interestingly, in the presence of ß-Heregulin (HRG), the DVD-Ig proteins show synergies with respect to inhibiting cell proliferation. The DVD-Ig proteins downregulate EGFR protein expression in the presence of HRG, which may be due to receptor internalization. Furthermore, the DVD-Ig proteins remarkably disrupt ß-Heregulin binding to FaDu cells.

Characterization of ABT-806, a Humanized Tumor-Specific Anti-EGFR Monoclonal Antibody.

Despite clinical efficacy, current approved agents targeting EGFR are associated with on-target toxicities as a consequence of disrupting normal EGFR function. MAb 806 is a novel EGFR antibody that selectively targets a tumor-selective epitope suggesting that a mAb 806-based therapeutic would retain antitumor activity without the on-target toxicities associated with EGFR inhibition. To enable clinical development, a humanized variant of mAb 806 designated ABT-806 was generated and is currently in phase 1 trials. We describe the characterization of binding and functional properties of ABT-806 compared with the clinically validated anti-EGFR antibody cetuximab. ABT-806 binds the mutant EGFRvIII with high affinity and, relative to cetuximab, exhibits increased potency against glioblastoma multiforme cell line and patient-derived xenografts expressing this form of the receptor. ABT-806 also inhibits the growth of squamous cell carcinoma xenograft models expressing high levels of wild-type EGFR, associated with inhibition of EGFR signaling, although higher doses of ABT-806 than cetuximab are required for similar activity. ABT-806 enhances in vivo potency of standard-of-care therapies used to treat glioblastoma multiforme and head and neck squamous cell carcinoma. An indium-labeled version of ABT-806, [(111)In]-ABT-806, used to investigate the relationship between dose and receptor occupancy, revealed greater receptor occupancy at lowers doses in an EGFRvIII-expressing model and significant uptake in an orthotopic model. Collectively, these results suggest that ABT-806 may have antitumor activity superior to cetuximab in EGFRvIII-expressing tumors, and similar activity to cetuximab in tumors highly overexpressing wild-type EGFR with reduced toxicity.

Identification of anti-ErbB2 dual variable domain immunoglobulin (DVD-Ig™) proteins with unique activities.

Inhibiting ErbB2 signaling with monoclonal antibodies (mAbs) or small molecules is an established therapeutic strategy in oncology. We have developed anti-ErbB2 Dual Variable Domain Immunoglobulin (DVD-Ig) proteins that capture the function of a combination of two anti-ErbB2 antibodies. In addition, some of the anti-ErbB2 DVD-Ig proteins gain the new functions of enhancing ErbB2 signaling and cell proliferation in N87 cells. We further found that two DVD-Ig proteins, DVD687 and DVD688, have two distinct mechanisms of actions in Calu-3 and N87 cells. DVD687 enhances cell cycle progression while DVD688 induces apoptosis in N87 cells. Using a half DVD687, we found that avidity may play a key role in the agonist activity of DVD687 in N87 cells.

Antitumor activity of a novel bispecific antibody that targets the ErbB2/ErbB3 oncogenic unit and inhibits heregulin-induced activation of ErbB3.

The prevalence of ErbB2 amplification in breast cancer has resulted in the heavy pursuit of ErbB2 as a therapeutic target. Although both the ErbB2 monoclonal antibody trastuzumab and ErbB1/ErbB2 dual kinase inhibitor lapatinib have met with success in the clinic, many patients fail to benefit. In addition, the majority of patients who initially respond will unfortunately ultimately progress on these therapies. Activation of ErbB3, the preferred dimerization partner of ErbB2, plays a key role in driving ErbB2-amplified tumor growth, but we have found that current ErbB2-directed therapies are poor inhibitors of ligand-induced activation. By simulating ErbB3 inhibition in a computational model of ErbB2/ErbB3 receptor signaling, we predicted that a bispecific antibody that docks onto ErbB2 and subsequently binds to ErbB3 and blocks ligand-induced receptor activation would be highly effective in ErbB2-amplified tumors, with superior activity to a monospecific ErbB3 inhibitor. We have developed a bispecific antibody suitable for both large scale production and systemic therapy by generating a single polypeptide fusion protein of two human scFv antibodies linked to modified human serum albumin. The resulting molecule, MM-111, forms a trimeric complex with ErbB2 and ErbB3, effectively inhibiting ErbB3 signaling and showing antitumor activity in preclinical models that is dependent on ErbB2 overexpression. MM-111 can be rationally combined with trastuzumab or lapatinib for increased antitumor activity and may in the future complement existing ErbB2-directed therapies to treat resistant tumors or deter relapse.

Control of cyclin D1 and breast tumorigenesis by the EglN2 prolyl hydroxylase.

2-Oxoglutarate-dependent dioxygenases, including the EglN prolyl hydroxylases that regulate HIF, can be inhibited with drug-like molecules. EglN2 is estrogen inducible in breast carcinoma cells and the lone Drosophila EglN interacts genetically with Cyclin D1. Although EglN2 is a nonessential gene, we found that EglN2 inactivation decreases Cyclin D1 levels and suppresses mammary gland proliferation in vivo. Regulation of Cyclin D1 is a specific attribute of EglN2 among the EglN proteins and is HIF independent. Loss of EglN2 catalytic activity inhibits estrogen-dependent breast cancer tumorigenesis and can be rescued by exogenous Cyclin D1. EglN2 depletion also impairs the fitness of lung, brain, and hematopoietic cancer lines. These findings support the exploration of EglN2 inhibitors as therapeutics for estrogen-dependent breast cancer and other malignancies.

A dominant-negative mutant of human DNA helicase B blocks the onset of chromosomal DNA replication.

A cDNA encoding a human ortholog of mouse DNA helicase B, which may play a role in DNA replication, has been cloned and expressed as a recombinant protein. The predicted human DNA helicase B (HDHB) protein contains conserved helicase motifs (superfamily 1) that are strikingly similar to those of bacterial recD and T4 dda proteins. The HDHB gene is expressed at low levels in liver, spleen, kidney, and brain and at higher levels in testis and thymus. Purified recombinant HDHB hydrolyzed ATP and dATP in the presence of single-stranded DNA, displayed robust 5'-3' DNA helicase activity, and interacted physically and functionally with DNA polymerase alpha-primase. HDHB proteins with mutations in the Walker A or B motif lacked ATPase and helicase activity but retained the ability to interact with DNA polymerase alpha-primase, suggesting that the mutants might be dominant over endogenous HDHB in human cells. When purified HDHB protein was microinjected into the nucleus of cells in early G(1), the mutant proteins inhibited DNA synthesis, whereas the wild type protein had no effect. Injection of wild type or mutant protein into cells at G(1)/S did not prevent DNA synthesis. The results suggest that HDHB function is required for S phase entry.