CT-based intratumoral and peritumoral deep transfer learning features prediction of lymph node metastasis in non-small cell lung cancer.

BACKGROUND: The main metastatic route for lung cancer is lymph node metastasis, and studies have shown that non-small cell lung cancer (NSCLC) has a high risk of lymph node infiltration. OBJECTIVE: This study aimed to compare the performance of handcrafted radiomics (HR) features and deep transfer learning (DTL) features in Computed Tomography (CT) of intratumoral and peritumoral regions in predicting the metastatic status of NSCLC lymph nodes in different machine learning classifier models. METHODS: We retrospectively collected data of 199 patients with pathologically confirmed NSCLC. All patients were divided into training (n = 159) and validation (n = 40) cohorts, respectively. The best HR and DTL features in the intratumoral and peritumoral regions were extracted and selected, respectively. Support Vector Machine (SVM), k-Nearest Neighbors (KNN), Light Gradient Boosting Machine (Light GBM), Multilayer Perceptron (MLP), and Logistic Regression (LR) models were constructed, and the performance of the models was evaluated. RESULTS: Among the five models in the training and validation cohorts, the LR classifier model performed best in terms of HR and DTL features. The AUCs of the training cohort were 0.841 (95% CI: 0.776-0.907) and 0.955 (95% CI: 0.926-0.983), and the AUCs of the validation cohort were 0.812 (95% CI: 0.677-0.948) and 0.893 (95% CI: 0.795-0.991), respectively. The DTL signature was superior to the handcrafted radiomics signature. CONCLUSIONS: Compared with the radiomics signature, the DTL signature constructed based on intratumoral and peritumoral areas in CT can better predict NSCLC lymph node metastasis.

Differential diagnosis of Crohn's disease and intestinal tuberculosis based on ATR-FTIR spectroscopy combined with machine learning.

BACKGROUND: Crohn's disease (CD) is often misdiagnosed as intestinal tuberculosis (ITB). However, the treatment and prognosis of these two diseases are dramatically different. Therefore, it is important to develop a method to identify CD and ITB with high accuracy, specificity, and speed. AIM: To develop a method to identify CD and ITB with high accuracy, specificity, and speed. METHODS: A total of 72 paraffin wax-embedded tissue sections were pathologically and clinically diagnosed as CD or ITB. Paraffin wax-embedded tissue sections were attached to a metal coating and measured using attenuated total reflectance fourier transform infrared spectroscopy at mid-infrared wavelengths combined with XGBoost for differential diagnosis. RESULTS: The results showed that the paraffin wax-embedded specimens of CD and ITB were significantly different in their spectral signals at 1074 cm-1 and 1234 cm-1 bands, and the differential diagnosis model based on spectral characteristics combined with machine learning showed accuracy, specificity, and sensitivity of 91.84%, 92.59%, and 90.90%, respectively, for the differential diagnosis of CD and ITB. CONCLUSION: Information on the mid-infrared region can reveal the different histological components of CD and ITB at the molecular level, and spectral analysis combined with machine learning to establish a diagnostic model is expected to become a new method for the differential diagnosis of CD and ITB.

Circulating cell-free DNA-based methylation pattern in plasma for early diagnosis of esophagus cancer.

With the increased awareness of early tumor detection, the importance of detecting and diagnosing esophageal cancer in its early stages has been underscored. Studies have consistently demonstrated the crucial role of methylation levels in circulating cell-free DNA (cfDNA) in identifying and diagnosing early-stage cancer. cfDNA methylation pertains to the methylation state within the genomic scope of cfDNA and is strongly associated with cancer development and progression. Several research teams have delved into the potential application of cfDNA methylation in identifying early-stage esophageal cancer and have achieved promising outcomes. Recent research supports the high sensitivity and specificity of cfDNA methylation in early esophageal cancer diagnosis, providing a more accurate and efficient approach for early detection and improved clinical management. Accordingly, this review aims to present an overview of methylation-based cfDNA research with a focus on the latest developments in the early detection of esophageal cancer. Additionally, this review summarizes advanced analytical technologies for cfDNA methylation that have significantly benefited from recent advancements in separation and detection techniques, such as methylated DNA immunoprecipitation sequencing (MeDIP-seq). Recent findings suggest that biomarkers based on cfDNA methylation may soon find successful applications in the early detection of esophageal cancer. However, large-scale prospective clinical trials are required to identify the potential of these biomarkers.

Vocal learning-associated convergent evolution in mammalian proteins and regulatory elements.

Vocal production learning ("vocal learning") is a convergently evolved trait in vertebrates. To identify brain genomic elements associated with mammalian vocal learning, we integrated genomic, anatomical, and neurophysiological data from the Egyptian fruit bat (Rousettus aegyptiacus) with analyses of the genomes of 215 placental mammals. First, we identified a set of proteins evolving more slowly in vocal learners. Then, we discovered a vocal motor cortical region in the Egyptian fruit bat, an emergent vocal learner, and leveraged that knowledge to identify active cis-regulatory elements in the motor cortex of vocal learners. Machine learning methods applied to motor cortex open chromatin revealed 50 enhancers robustly associated with vocal learning whose activity tended to be lower in vocal learners. Our research implicates convergent losses of motor cortex regulatory elements in mammalian vocal learning evolution.

Development and validation of a prognostic model based on clinical laboratory biomarkers to predict admission to ICU in Omicron variant-infected hospitalized patients complicated with myocardial injury.

Aims: The aim of this study was to develop and validate a prognostic model based on clinical laboratory biomarkers for the early identification of high-risk patients who require intensive care unit (ICU) admission among those hospitalized with the Omicron variant of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and complicated with myocardial injury (MI). Methods: This single-center study enrolled 263 hospitalized patients with confirmed Omicron variant infection and concurrent MI. The patients were randomly divided into training and validation cohorts. Relevant variables were collected upon admission, and the least absolute shrinkage and selection operator (LASSO) was used to select candidate variables for constructing a Cox regression prognostic model. The model's performance was evaluated in both training and validating cohorts based on discrimination, calibration, and net benefit. Results: Of the 263 eligible patients, 210 were non-ICU patients and 53 were ICU patients. The prognostic model was built using four selected predictors: white blood cell (WBC) count, procalcitonin (PCT) level, C-reactive protein (CRP) level, and blood urea nitrogen (BUN) level. The model showed good discriminative ability in both the training cohort (concordance index: 0.802, 95% CI: 0.716-0.888) and the validation cohort (concordance index: 0.799, 95% CI: 0.681-0.917). For calibration, the predicted probabilities and observed proportions were highly consistent, indicating the model's reliability in predicting outcomes. In the 21-day decision curve analysis, the model had a positive net benefit for threshold probability ranges of 0.2 to 0.8 in the training cohort and nearly 0.2 to 1 in the validation cohort. Conclusion: In this study, we developed a clinically practical model with high discrimination, calibration, and net benefit. It may help to early identify severe and critical cases among Omicron variant-infected hospitalized patients with MI.

Adsorption Behavior and Kinetics of 1,4-Dioxane by Carbon Aerogel.

1,4-dioxane is a potential carcinogen in water and is difficult to deal with due to its robust cycloether bond and complete miscibility with water. To remove 1,4-dioxane in an economically viable and environmentally friendly way, a series of carbon aerogels were synthesized as adsorbents for 1,4-dioxane. The experiment results showed that adsorption performances were closely related to the preparation conditions of carbon aerogels, such as the molar ratio, heating rate, pyrolysis temperature and residence time, which were carefully controlled. Scanning electron microscope analysis revealed the presence of a three-dimensional porous network structure in carbon aerogels. Brunauer-Emmett-Teller analysis results demonstrated an increase in specific surface area (673.89 m2/g) and total pore volume after carbonization, with an increase in mesoporous porosity and a decrease in microporosity. When considering each variable individually, the highest specific surface area of prepared carbon aerogels was achieved at a pyrolysis temperature of 800 °C, a holding time of 1 h, and a heating rate of 2 °C/min. Under optimal experimental conditions, the adsorption removal of 1,4-dioxane by carbon aerogels exceeded 95%, following quasi-second-order kinetics and Langmuir isothermal adsorption isotherms, indicating that monolayer adsorption on the surface of carbon aerogels occurred. The maximum adsorption capacity obtained was 67.28 mg/g at a temperature of 318 K, which was attributed to the presence of a large proportion of mesopores and abundant micropores simultaneously in carbon aerogels. Furthermore, with the interference of chlorinated solvents such as trichloroethylene (TCE), the removal efficiency of 1,4-dioxane had no obvious inhibition effect. Regeneration experiments showed that after five continuous cycles, the carbon aerogels still kept a comparable adsorption capacity, which illustrates its potential application in 1,4-dioxane-polluted water purification.

Sparks of function by de novo protein design.

Information in proteins flows from sequence to structure to function, with each step causally driven by the preceding one. Protein design is founded on inverting this process: specify a desired function, design a structure executing this function, and find a sequence that folds into this structure. This 'central dogma' underlies nearly all de novo protein-design efforts. Our ability to accomplish these tasks depends on our understanding of protein folding and function and our ability to capture this understanding in computational methods. In recent years, deep learning-derived approaches for efficient and accurate structure modeling and enrichment of successful designs have enabled progression beyond the design of protein structures and towards the design of functional proteins. We examine these advances in the broader context of classical de novo protein design and consider implications for future challenges to come, including fundamental capabilities such as sequence and structure co-design and conformational control considering flexibility, and functional objectives such as antibody and enzyme design.

Dietary Bacillus cereus LS2 protects juvenile sea cucumber Apostichopus japonicus against Vibrio splendidus infection.

This study aimed to investigate the effects of Bacillus cereus LS2 on the growth performance, innate immunity, intestinal microbiota, and disease resistance of sea cucumber Apostichopus japonicus. After feeding with LS2 for 30 days, results showed that dietary with LS2 had a significant improvement in the growth rate and immune parameters (including total coelomocytes counts, phagocytosis, respiratory burst, and immune-related enzymes) of juvenile sea cucumbers. Subsequently, transcriptome sequencing and qRT-PCR verification were performed to analyze the potential mechanism of LS2 diet and thus improve the immune response of A. japonicus. GO and KEGG pathway analysis indicated that LS2 can primarily activate the "Lectins" and "complement and coagulation cascades" pathways to modulate the innate immunity of the sea cucumbers. Furthermore, 16S rRNA sequencing was used to analyze the intestinal microbial composition of sea cucumbers after dietary with LS2. Results showed that Proteobacteria, Actinobacteria, Firmicutes, and Bacteroidetes were the most prevalent phyla in A. japonicus intestinal microbiota. The abundance of Actinobacteria (46.20%) and Bacteroidetes (12.80%) were significantly higher in the LS2 group, whereas the relative abundance of Proteobacteria (49.98%) and Firmicutes (14.97%) were higher in the control group. The LDA scores of Nocardiaceae and Rhodococcus were also the highest taxa after the dietary administration of LS2, indicating that Actinobacteria phylum played a pivotal role in the intestinal microbial function of A. japonicus. Overall, these results suggested that feeding with Bacillus LS2 may be beneficial for A. japonicus farming.

Porphyrin Supramolecular Nanoassembly/C3N4 Nanosheet S-Scheme Heterojunctions for Selective Photocatalytic CO2 Reduction toward CO.

The photocatalytic reduction of CO2 with H2O into valuable chemicals is a sustainable carbon-neutral technology for renewable energy; however, the photocatalytic activity and product selectivity remain challenging. Herein, an S-scheme heterojunction photocatalyst with superior CO2 photoreduction performance─porous C3N4 (CN) nanosheets anchored with zinc(II) tetra(4-cyanophenyl)porphyrin (ZnTP) nanoassemblies (denoted as ZnTP/CN)─was designed and prepared via a simple self-assembly process. The constructed ZnTP/CN heterojunction had rich accessible active sites, improved CO2 absorption capacity, and high charge carrier separation efficiency caused by the S-scheme heterojunction. As a result, the obtained ZnTP/CN catalyst exhibited considerable activity for photocatalytic CO2 reduction, yielding CO with a generation rate of 19.4 µmol g-1·h-1 and a high selectivity of 95.8%, which is much higher than that of pristine CN nanosheets (4.53 µmol g-1·h-1, 57.4%). In addition, theoretical calculations and in situ Fourier transform infrared spectra demonstrated that the Zn sites in the porphyrin unit favor CO2 activation and *COOH formation as well as CO desorption, thereby affording a high CO selectivity. This work provides insight into the design and fabrication of efficient S-scheme heterostructure photocatalysts for solar energy storage.

Targeting LncRNA LLNLR-299G3.1 with antisense oligonucleotide inhibits malignancy of esophageal squamous cell carcinoma cells in vitro and in vivo.

Accumulating evidence has indicated that long non-coding RNAs (lncRNAs) play critical roles in the development and progression of cancers, including esophageal squamous cell carcinoma (ESCC). However, the mechanisms of lncRNAs in ESCC are still incompletely understood and therapeutic attempts for in vivo targeting cancer-associated lncRNA remain a challenge. By RNA-sequencing analysis, we identified that LLNLR-299G3.1 was a novel ESCC-associated lncRNA. LLNLR-299G3.1 was up-regulated in ESCC tissues and cells and promoted ESCC cell proliferation and invasion. Silencing of LLNLR-299G3.1 with ASO (antisense oligonucleotide) resulted in opposite effects. Mechanistically, LLNLR-299G3.1 bound to cancer-associated RNA binding proteins and regulated the expression of cancer-related genes, including OSM, TNFRSF4, HRH3, and SSTR3. ChIRP-seq (chromatin isolation by RNA purification and sequencing) revealed that these genes contained enriched chromatin binding sites for LLNLR-299G3.1. Rescue experiments confirmed that the effects of LLNLR-299G3.1 on ESCC cell proliferation were dependent on interaction with HRH3 and TNFRSF4. Therapeutically, intravenous delivery of placental chondroitin sulfate A binding peptide-coated nanoparticles containing antisense oligonucleotide (pICSA-BP-ANPs) strongly inhibited ESCC tumor growth and significantly improved animal survival in vivo. Overall, our results suggest that LLNLR-299G3.1 promotes ESCC malignancy through regulating gene-chromatin interactions and targeting ESCC by pICSA-BP-ANPs may be an effective strategy for the treatment of lncRNA-associated ESCC.

Triggering High Capacity and Superior Reversibility of Manganese Oxides Cathode via Magnesium Modulation for Zn//MnO2 Batteries.

Aqueous zinc-ion batteries (ZIBs) have attracted extensive attention in recent years because of its high volumetric energy density, the abundance of zinc resources, and safety. However, ZIBs still suffer from poor reversibility and sluggish kinetics derived from the unstable cathodic structure and the strong electrostatic interactions between bivalent Zn2+ and cathodes. Herein, magnesium doping into layered manganese dioxide (Mg-MnO2 ) via a simple hydrothermal method as cathode materials for ZIBs is proposed. The interconnected nanoflakes of Mg-MnO2 possess a larger specific surface area compared to pristine δ-MnO2 , providing more electroactive sites and boosting the capacity of batteries. The ion diffusion coefficients of Mg-MnO2 can be enhanced due to the improved electrical conductivity by doped cations and oxygen vacancies in MnO2 lattices. The assembled Zn//Mg-MnO2 battery delivers a high specific capacity of 370 mAh g-1 at a current density of 0.6 A g-1 . Furthermore, the reaction mechanism confirms that Zn2+ insertion occurred after a few cycles of activation reactions. Most important, the reversible redox reaction between Zn2+ and MnOOH is found after several charge-discharge processes, promoting capacity and stability. It believes that this systematic research enlightens the design of high-performance of ZIBs and facilitates the practical application of Zn//MnO2 batteries.

Analysis of pulmonary artery variation based on 3D reconstruction of CT angiography.

Objective: The aim of this study is to acquire pulmonary CT (Computed tomography) angiographic data for the purpose of creating a three-dimensional reconstruction. Additionally, we aim to analyze the features and deviations of the branches in both pulmonary lobes. This information is intended to serve as a more comprehensive and detailed reference for medical professionals when conducting preoperative evaluations and devising surgical plans. Method: Between August 2019 and December 2021, 420 patients were selected from the thoracic surgery department at the First Hospital of Jilin University, and underwent pulmonary 64 channel contrast enhanced CT examinations (Philips ICT 256). The images were acquired at a 1.5 mm slice thickness, and the DCM files that complied with DICOM (Digital Imaging and Communications in Medicine) standards were analysed for 3D (three dimensional) reconstruction using Mimics 22.0 software. The reconstructed pulmonary artery models were assessed by attending chest surgeons and radiologists with over 10 years of clinical experience. The two-dimensional image planes, as well as the coronary and sagittal planes, were utilized to evaluate the arteries. The study analyzed the characteristics and variations of the branches and courses of pulmonary arteries in each lobe of the lungs, with the exception of the subsegmental arterial system. Two chest surgeons and two radiologists with professional titles-all of whom had over a decade of clinical experience-jointly evaluated the 3D models of the pulmonary artery and similarly assessed the characteristics and variations of the branches and courses in each lobe of the lungs. Results: Significant variations were observed in the left superior pulmonary artery across the 420 subjects studied. In the left upper lobe, the blood supply of 4 arteries accounted for 50.5% (n = 212), while the blood supply of 2 arteries in the left lower lobe was the most common, accounting for 79.5% (n = 334). The greatest variation in the right pulmonary artery was observed in the branch supply of the right upper lobe mediastinal artery. In the majority of cases (77.9%), there were two arteries present, which was the most common configuration observed accounting for 64% (n = 269). In the right inferior lobe of the lung, there were typically 2-4 arteries, with 2 arteries being the most common configuration (observed in 79% of cases, n = 332). Conclusion: The three-dimensional reconstruction of pulmonary artery CT angiography enables clear observation of the branches and distribution of the pulmonary artery while also highlighting any variations. This technique holds significant clinical value for preoperative assessments regarding lesions and blood vessels.

Development and validation of a prognostic model based on immune variables to early predict severe cases of SARS-CoV-2 Omicron variant infection.

Background: The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) Omicron variant has prevailed globally since November 2021. The extremely high transmissibility and occult manifestations were notable, but the severity and mortality associated with the Omicron variant and subvariants cannot be ignored, especially for immunocompromised populations. However, no prognostic model for specially predicting the severity of the Omicron variant infection is available yet. In this study, we aim to develop and validate a prognostic model based on immune variables to early recognize potentially severe cases of Omicron variant-infected patients. Methods: This was a single-center prognostic study involving patients with SARS-CoV-2 Omicron variant infection. Eligible patients were randomly divided into the training and validation cohorts. Variables were collected immediately after admission. Candidate variables were selected by three variable-selecting methods and were used to construct Cox regression as the prognostic model. Discrimination, calibration, and net benefit of the model were evaluated in both training and validation cohorts. Results: Six hundred eighty-nine of the involved 2,645 patients were eligible, consisting of 630 non-ICU cases and 59 ICU cases. Six predictors were finally selected to establish the prognostic model: age, neutrophils, lymphocytes, procalcitonin, IL-2, and IL-10. For discrimination, concordance indexes in the training and validation cohorts were 0.822 (95% CI: 0.748-0.896) and 0.853 (95% CI: 0.769-0.942). For calibration, predicted probabilities and observed proportions displayed high agreements. In the 21-day decision curve analysis, the threshold probability ranges with positive net benefit were 0~1 and nearly 0~0.75 in the training and validation cohorts, correspondingly. Conclusions: This model had satisfactory high discrimination, calibration, and net benefit. It can be used to early recognize potentially severe cases of Omicron variant-infected patients so that they can be treated timely and rationally to reduce the severity and mortality of Omicron variant infection.

Microbiome and function alterations in the gastric mucosa of asymptomatic patients with Helicobacter pylori infection.

BACKGROUND: Most patients with Helicobacter pylori (H. pylori) infection have no clinical symptoms, numerous studies reported the gastric microbiome in H. pylori-infected patients, but asymptomatic patients have not been distinguished. How the microbiome and function changes in asymptomatic patients with H. pylori infection remains poorly understood. METHODS: A total of 29 patients were divided into H. pylori-infected asymptomatic group (10 patients), H. pylori-infected symptomatic group (11 patients) and H. pylori-uninfected group (8 patients). Gastric mucosa specimens were taken for histopathological examination, special staining, and 16 S rDNA sequencing. High-throughput results were evaluated by community composition analysis, indicator species analysis, alpha diversity analysis, beta diversity analysis, and function prediction. RESULTS: The gastric microbiota composition at phylum and genus level of H. pylori-infected asymptomatic patients were similar with H. pylori-infected symptomatic group, but different from H. pylori-uninfected patients. The diversity and richness of gastric microbial community declined significantly in H. pylori-infected asymptomatic group comparing with H. pylori-uninfected group. Sphingomonas may be an indicator between symptomatic and asymptomatic patients with H. pylori infection, the AUC value of Sphingomonas is 0.79. Interactions between species increased and altered notably after H. pylori infection. More genera were affected by Helicobacter in H. pylori-infected asymptomatic patients. The function condition changed significantly in asymptomatic patients with H. pylori infection, there was no difference comparing with symptomatic ones. Amino acid metabolism and lipid metabolism strengthened but carbohydrate metabolism remained constant after H. pylori infection. The metabolism of fatty acid and bile acid was disturbed after infection with H. pylori. CONCLUSION: The gastric microbiota composition and function mode changed significantly after H. pylori infection regardless of the presence of clinical symptoms, there was no difference between H. pylori-infected asymptomatic and symptomatic patients. The difference in gastric microbiota composition and interactions between species might be responsible for presence of digestive symptoms.

SARS-CoV-2 Omicron XBB subvariants exhibit enhanced fusogenicity and substantial immune evasion in elderly population, but high sensitivity to pan-coronavirus fusion inhibitors.

Numerous emerging severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) Omicron subvariants have shown significant immune evasion capacity and caused a large number of infections, as well as vaccine-breakthrough infections, especially in elderly populations. Recently emerged Omicron XBB was derived from the BA.2 lineage, but bears a distinct mutant profile in its spike (S) protein. In this study, we found that Omicron XBB S protein drove more efficient membrane-fusion kinetics on human lung-derived cells (Calu-3). Considering the high susceptibility of the elderly to the current Omicron pandemic, we performed a comprehensive neutralization assessment of elderly convalescent or vaccine sera against XBB infection. We found that the sera from elderly convalescent patients who experienced with BA.2 infection or breakthrough infection potently inhibited BA.2 infection, but showed significantly reduced efficacy against XBB. Moreover, recently emerged XBB.1.5 subvariant also showed more significant resistance to the convalescent sera of BA.2- or BA.5-infected elderly. On the other hand, we found that the pan-CoV fusion inhibitors EK1 and EK1C4 can potently block either XBB-S- or XBB.1.5-S-mediated fusion process and viral entry. Moreover, EK1 fusion inhibitor showed potent synergism when combined with convalescent sera of BA.2- or BA.5-infected patients against XBB and XBB.1.5 infection, further indicating that EK1-based pan-CoV fusion inhibitors are promising candidates for development as clinical antiviral agents to combat the Omicron XBB subvariants.

N-Fluorobenzamide-Directed Formal [4+2] Cycloaddition Reaction with Maleic Anhydride: Access to Fluorescent Aminonaphthalic Anhydrides.

We have developed a formal [4+2] cycloaddition reaction of N-fluorobenzamides and maleic anhydride in the presence of CuI and LiOH, and a series of fluorescent 1-amino-2,3-naphthalic anhydrides were produced in good yields. This reaction proceeded via a multistep process involving nitrogen-centered radical generation, 1,5-hydrogen atom transfer, and benzylic radical addition to the amide carbonyl oxygen to generate an N-(tert-butyl) isobenzofuran-1(3H)-imine intermediate, which isomerized to an N-(tert-butyl) isobenzofuran-1-amine via deprotonation and protonation with the aid of LiOH; finally, the amine underwent a [4+2] cycloaddition reaction with maleic anhydride to give the 1-amino-2,3-naphthalic anhydride product upon dehydrating aromatization. Notably, the corresponding naphthalic anhydride products could be transformed into a diverse array of naphthalimides. Both the naphthalic anhydrides and the naphthalimides exhibited similar fluorescent features.

Analysis of the influencing factors in the long-term survival of esophageal cancer.

Background: To analyze the prognosis and diagnostic value of relevant hematological indexes on the survival status of patients with esophageal squamous cell carcinoma after radical surgery. Methods: This study included 206 patients with esophageal cancer who underwent surgical R0 resection. The data, including the basic information, preoperative blood routine, albumin, fibrinogen, surgery-related information, postoperative pathology, and overall survival, of the patients were compared. Results: The survival and death groups showed a significant difference in overall survival (OS), the degree of differentiation, depth of infiltration, pathological stage, vascular infiltration, nerve infiltration, fibrinogen, white blood cell, neutrophils, platelet, and platelet hematocrit (P<0.05). Tumor located in the middle thorax, larger lesion length, deeper invasion, later pathological stage, vascular infiltration, nerve infiltration, lymph node metastasis, cardiovascular disease, and higher smoking grade were risk factors for poor prognosis of esophageal squamous cell carcinoma (ESCC) (P<0.05). Cardiovascular disease, lower differentiation, tumor located in the middle thorax, and nerve infiltration were independent risk factors for the reduction of survival time of patients with ESCC (P<0.05). Conclusions: History of cardiovascular disease, tumor located in the middle chest, poorly differentiated esophageal squamous cell carcinoma, visible nerve cancer invasion, hematocrit (HCT), mean erythrocyte hemoglobin concentration (MCHC), and hemoglobin (HB) are independent risk factors for the long-term survival of patients with ESCC.

Geometric Deep Learning for Structure-Based Ligand Design.

A pervasive challenge in drug design is determining how to expand a ligand-a small molecule that binds to a target biomolecule-in order to improve various properties of the ligand. Adding single chemical groups, known as fragments, is important for lead optimization tasks, and adding multiple fragments is critical for fragment-based drug design. We have developed a comprehensive framework that uses machine learning and three-dimensional protein-ligand structures to address this challenge. Our method, FRAME, iteratively determines where on a ligand to add fragments, selects fragments to add, and predicts the geometry of the added fragments. On a comprehensive benchmark, FRAME consistently improves predicted affinity and selectivity relative to the initial ligand, while generating molecules with more drug-like chemical properties than docking-based methods currently in widespread use. FRAME learns to accurately describe molecular interactions despite being given no prior information on such interactions. The resulting framework for quality molecular hypothesis generation can be easily incorporated into the workflows of medicinal chemists for diverse tasks, including lead optimization, fragment-based drug discovery, and de novo drug design.