Atypical and variable attention patterns reveal reduced contextual priors in children with autism spectrum disorder.

Accumulating evidence suggests that individuals with autism spectrum disorder (ASD) show impairments in using contextual priors to predict others' actions and make intention inference. Yet less is known about whether and how children with ASD acquire contextual priors during action observation and how contextual priors relate to their action prediction and intention inference. To form proper contextual priors, individuals need to observe the social scenes in a reliable manner and focus on socially relevant information. By employing a data-driven scan path method and areas of interest (AOI)-based analysis, the current study investigated how contextual priors would relate to action prediction and intention understanding in 4-to-9-year-old children with ASD (N = 56) and typically developing (TD) children (N = 50) during free viewing of dynamic social scenes with different intentions. Results showed that children with ASD exhibited higher intra-subject variability when scanning social scenes and reduced attention to socially relevant areas. Moreover, children with high-level action prediction and intention understanding showed lower intra-subject variability and increased attention to socially relevant areas. These findings suggest that altered fixation patterns might restrain children with ASD from acquiring proper contextual priors, which has cascading downstream effects on their action prediction and intention understanding.

Whole-Genome DNA Methylation Profiling of Intrahepatic Cholangiocarcinoma Reveals Prognostic Subtypes with Distinct Biological Drivers.

Intrahepatic cholangiocarcinoma (iCCA) is the second most prevalent primary liver cancer. Although the genetic characterization of iCCA has led to targeted therapies for treating tumors with FGFR2 alterations and IDH1/2 mutations, only a limited number of patients can benefit from these strategies. Epigenomic profiles have emerged as potential diagnostic and prognostic biomarkers for improving the treatment of cancers. In this study, we conducted whole-genome bisulfite sequencing on 331 iCCAs integrated with genetic, transcriptomic, and proteomic analyses, demonstrating the existence of four DNA methylation subtypes of iCCAs (S1-S4) that exhibited unique postoperative clinical outcomes. The S1 group was an IDH1/2 mutation-specific subtype with moderate survival. The S2 subtype was characterized by the lowest methylation level and the highest mutational burden among the four subtypes and displayed upregulation of a gene-expression pattern associated with cell cycle/DNA replication. The S3 group was distinguished by high interpatient heterogeneity of tumor immunity, a gene-expression pattern associated with carbohydrate metabolism, and an enrichment of KRAS alterations. Patients with the S2 and S3 subtypes had the shortest survival among the four subtypes. Tumors in the S4 subtype, which had the best prognosis, showed global methylation levels comparable to normal controls, increased FGFR2 fusions/BAP1 mutations, and the highest copy-number variant burdens. Further integrative and functional analyses identified GBP4 demethylation, which is highly prevalent in the S2 and S3 groups, as an epigenetic oncogenic factor that regulates iCCA proliferation, migration, and invasion. Together, this study identifies prognostic methylome alterations and epigenetic drivers in iCCA. SIGNIFICANCE: Characterization of the DNA methylome of intrahepatic cholangiocarcinoma integrated with genomic, transcriptomic, and proteomic analyses uncovers molecular mechanisms affected by genome-wide DNA methylation alterations, providing a resource for identifying potential therapeutic targets.

[Machine learning algorithms for identifying autism spectrum disorder through eye-tracking in different intention videos]. / ä¸åŒæ„å›¾åœºæ™¯çœ¼åŠ¨æ³¨è§†æ¨¡å¼æœºå™¨å­¦ä¹ ç®—æ³•è¯†åˆ«å­¤ç‹¬ç—‡è°±ç³»éšœç¢çš„ç ”ç©¶.

OBJECTIVES: To investigate the differences in visual perception between children with autism spectrum disorder (ASD) and typically developing (TD) children when watching different intention videos, and to explore the feasibility of machine learning algorithms in objectively distinguishing between ASD children and TD children. METHODS: A total of 58 children with ASD and 50 TD children were enrolled and were asked to watch the videos containing joint intention and non-joint intention, and the gaze duration and frequency in different areas of interest were used as original indicators to construct classifier-based models. The models were evaluated in terms of the indicators such as accuracy, sensitivity, and specificity. RESULTS: When using eight common classifiers, including support vector machine, linear discriminant analysis, decision tree, random forest, and K-nearest neighbors (with K values of 1, 3, 5, and 7), based on the original feature indicators, the highest classification accuracy achieved was 81.90%. A feature reconstruction approach with a decision tree classifier was used to further improve the accuracy of classification, and then the model showed the accuracy of 91.43%, the specificity of 89.80%, and the sensitivity of 92.86%, with an area under the receiver operating characteristic curve of 0.909 (P<0.001). CONCLUSIONS: The machine learning model based on eye-tracking data can accurately distinguish ASD children from TD children, which provides a scientific basis for developing rapid and objective ASD screening tools.

The evolution of atomistic structure and mechanical property of coke in the gasification process with CO2 and H2O at different temperatures: A ReaxFF molecular dynamics study.

CONTEXT: An atomistic coke carbon model was constructed to simulate the structural evolution in the gasification and stretching process. The coke model was placed in a box with different CO2/H2O content to investigate the evolution of the atomistic structure of coke during the gasification. It was found that different atmospheric concentrations had different effects on the structure and reaction sites of the coke model. The CO2 molecules tended to dissolve on the surface of coke and disrupt its surface structure, while H2O molecules were more likely to enter the coke model to disrupt the internal structure. For tensile simulation, it was found that CO2 and H2O had different effects on the tensile resistance of the coke model. Controlling the composition content of the reaction gas can effectively influence the tensile strength of the coke model. By revealing the behavior of coke model at the micro scale, it provides a theoretical basis for the industrial coke application process. METHODS: Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS) was used to conduct the molecular dynamics using the reactive force field (ReaxFF). The atomistic model of coke carbon was constructed using the well-known annealing and quenching method, and its composition is determined according to the element analysis of industrial coke. The structural evolution in the gasification with CO2/H2O and the stretching process were analyzed in detail. Molecular dynamics simulations with reactive force field (ReaxFF-MD) were used to simulate the coke dissolution reaction under CO2/H2O atmosphere and the coke stretching process. The atmosphere ratio was modified to investigate the changes in coke structure under different atmosphere conditions. The Packmol software was used to place gas and coke models into the same box. During the reaction process, the Ovito software was used to perform corresponding visualization analysis on the changes in the atomic structure of coke.

Effect of Different sp2 Bond Contents on the Reactivity and Mechanical Properties of Coke Carbon: A ReaxFF Molecular Dynamics Study.

In this study, ReaxFF-MD was used to construct a large-molecule model of coke containing 3000 atoms, and the sp2 bond content of the model was controlled by changing the heating and cooling rates. The increase of the sp2 bond content led to a significant difference in the reactivity of coke. The presence of the sp2 bond caused the carbon atoms inside the coke to change into a circular structure, making it more difficult for the gaseous atoms to adsorb on the surface of the coke. It significantly reduced the gasification reaction rate of coke in the CO2 and H2O atmospheres. In the tensile simulation experiment, it was found that the stretching process of coke was mainly divided into three stages: an elastic stretching stage, a plastic stretching stage, and a model fracture stage. During the stretching process, the carbon ring structure would undergo a C-C bond fracture while generating carbon chains to resist stress. The results indicated that the presence of sp2 bonds can effectively reduce the phenomenon of excessive local stress on coke to improve its tensile resistance. The method developed in this paper may provide further ideas and platforms for the research on coke performance.

Supramolecular nanoarchitectonics of phenolic-based nanofiller for controlled diffusion of versatile drugs in hydrogels.

Drug-dependent design of hydrogels is currently required for engineering the controlled release of therapeutics, which is a major contributor to the technical challenges relating to the clinical translation of hydrogel-drug systems. Herein, by integrating supramolecular phenolic-based nanofillers (SPFs) into hydrogel microstructures we developed a facile strategy to endow a range of clinically relevant hydrogels with controlled release properties for diverse therapeutic agents. The assembly of multiscale SPF aggregates leads to tunable mesh size and multiple dynamic interactions between SPF aggregates and drugs, which relaxes the available choices of drugs and hydrogels. This simple approach allowed for the controlled release of 12 representative drugs evaluated with 8 commonly used hydrogels. Moreover, the anesthetic drug lidocaine was loaded into SPF-integrated alginate hydrogel and demonstrated sustained release for 14 days in vivo, validating the potential for long-term anesthesia in patients.

Nanostructured particles assembled from natural building blocks for advanced therapies.

Advanced treatments based on immune system manipulation, gene transcription and regulation, specific organ and cell targeting, and/or photon energy conversion have emerged as promising therapeutic strategies against a range of challenging diseases. Naturally derived macromolecules (e.g., proteins, lipids, polysaccharides, and polyphenols) have increasingly found use as fundamental building blocks for nanostructured particles as their advantageous properties, including biocompatibility, biodegradability, inherent bioactivity, and diverse chemical properties make them suitable for advanced therapeutic applications. This review provides a timely and comprehensive summary of the use of a broad range of natural building blocks in the rapidly developing field of advanced therapeutics with insights specific to nanostructured particles. We focus on an up-to-date overview of the assembly of nanostructured particles using natural building blocks and summarize their key scientific and preclinical milestones for advanced therapies, including adoptive cell therapy, immunotherapy, gene therapy, active targeted drug delivery, photoacoustic therapy and imaging, photothermal therapy, and combinational therapy. A cross-comparison of the advantages and disadvantages of different natural building blocks are highlighted to elucidate the key design principles for such bio-derived nanoparticles toward improving their performance and adoption. Current challenges and future research directions are also discussed, which will accelerate our understanding of designing, engineering, and applying nanostructured particles for advanced therapies.

Long noncoding RNA TLNC1 promotes the growth and metastasis of liver cancer via inhibition of p53 signaling.

BACKGROUND: Long non-coding RNAs (lncRNAs) have been demonstrated to play vital roles in cancer development and progression. However, their biological roles and function mechanisms in liver cancer remain largely unknown. METHODS: RNA-seq was performed with clinical hepatoma tissues and paired adjacent normal liver tissues to identify differentially expressed lncRNAs. qPCR was utilized to examine the expression levels of lncRNAs. We studied the function of TLNC1 in cell growth and metastasis of hepatoma with both cell and mouse models. RNA-seq, RNA pull-down coupled with mass spectrometry, RNA immunoprecipitation, dual luciferase reporter assay, and surface plasmon resonance analysis were used to analyze the functional mechanism of TLNC1. RESULTS: Based on the intersection of our own RNA-seq, TCGA RNA-seq, and TCGA survival analysis data, TLNC1 was identified as a potential tumorigenic lncRNA of liver cancer. TLNC1 significantly enhanced the growth and metastasis of hepatoma cells both in vitro and in vivo. TLNC1 exerted its tumorigenic function through interaction with TPR and inducing the TPR-mediated transportation of p53 from nucleus to cytoplasm, thus repressing the transcription of p53 target genes and finally contributing to the progression of liver cancer. CONCLUSIONS: TLNC1 is a promising prognostic factor of liver cancer, and the TLNC1-TPR-p53 axis can serve as a potential therapeutic target for hepatoma treatment.

Predicting Genomic Alterations of Phosphatidylinositol-3 Kinase Signaling in Hepatocellular Carcinoma: A Radiogenomics Study Based on Next-Generation Sequencing and Contrast-Enhanced CT.

BACKGROUND: Exploring the genomic landscape of hepatocellular carcinoma (HCC) provides clues for therapeutic decision-making. Phosphatidylinositol-3 kinase (PI3K) signaling is one of the key pathways regulating HCC aggressiveness, and its genomic alterations have been correlated with sorafenib response. In this study, we aimed to predict somatic mutations of the PI3K signaling pathway in HCC samples through machine-learning-based radiomic analysis. METHODS: HCC patients who underwent next-generation sequencing and preoperative contrast-enhanced CT were recruited from West China Hospital and The Cancer Genome Atlas for model training and validation, respectively. Radiomic features were extracted from volumes of interest (VOIs) covering the tumor (VOItumor) and peritumoral areas (5 mm [VOI5mm], 10 mm [VOI10mm], and 20 mm [VOI20mm] from tumor margin). Factor analysis, logistic regression analysis, least absolute shrinkage and selection operator, and random forest analysis were applied for feature selection and model construction. Model performance was characterized based on the area under the receiver operating characteristic curve (AUC). RESULTS: A total of 132 HCC patients (mean age: 61.1 ± 14.7 years; 108 men) were enrolled. In the training set, the AUCs of radiomic signatures based on single CT phases were moderate (AUC 0.694-0.771). In the external validation set, the radiomic signature based on VOI10mm in arterial phase demonstrated the highest AUC (0.733) among all models. No improvement in model performance was achieved after adding the tumor radiomic features or manually assessed qualitative features. CONCLUSIONS: Machine-learning-based radiomic analysis had potential for characterizing alterations of PI3K signaling in HCC and could help identify potential candidates for sorafenib treatment.

CircNFIB inhibits tumor growth and metastasis through suppressing MEK1/ERK signaling in intrahepatic cholangiocarcinoma.

BACKGROUND: Considerable evidence shows that circular RNAs (circRNAs) play an important role in tumor development. However, their function in intrahepatic cholangiocarcinoma (ICC) metastasis and the underlying mechanisms are incompletely understood. METHODS: circNFIB (hsa_circ_0086376, termed as cNFIB hereafter) was identified in human ICC tissues through circRNAs sequencing. The biological role of cNFIB was determined in vitro and in vivo by gain or loss of functional experiments. Fluorescence in situ hybridization (FISH), RNA immunoprecipitation (RIP) and RNA pull-down assays were conducted to analyze the interaction of cNFIB with dual specificity mitogen-activated protein kinase kinase1 (MEK1). Duolink in situ proximity ligation assay (PLA) and coimmunoprecipitation (co-IP) assay were used to investigate the effects of cNFIB on the interaction between MEK1 and mitogen-activated protein kinase 2 (ERK2). Finally, a series of in vitro and in vivo experiments were performed to explore the influences of cNFIB on the anti-tumor activity of trametinib (a MEK inhibitor). RESULTS: cNFIB was significantly down-regulated in human ICC tissues with postoperative metastases. The loss of cNFIB was highly associated with aggressive characteristics and predicted unfavorable prognosis in ICC patients. Functional studies revealed that cNFIB inhibited the proliferation and metastasis of ICC cells in vitro and in vivo. Mechanistically, cNFIB competitively interacted with MEK1, which induced the dissociation between MEK1 and ERK2, thereby resulting in the suppression of ERK signaling and tumor metastasis. Moreover, we found that ICC cells with high levels of cNFIB held the potential to delay the trametinib resistance. Consistently, in vivo and in vitro studies demonstrated that cotreatment with trametinib and lentivirus vector encoding cNFIB showed greater inhibitory effect than isolated trametinib treatment. CONCLUSIONS: Our findings identified that cNFIB played a key role in ICC growth and metastasis by regulating MEK1/ERK signaling. Given the efficacy of cNFIB modulation on ICC suppression and trametinib sensitivity, cNFIB appears to be a potential therapeutic molecule for ICC treatment.

Integrated proteogenomic analysis revealed the metabolic heterogeneity in noncancerous liver tissues of patients with hepatocellular carcinoma.

Understanding the adjacent liver microenvironment of hepatocellular carcinoma (HCC) with possible metastasis tendency might provide a strategy for risk classification of patients and potential therapies by converting the unique metastasis-inclined microenvironment to a metastasis-averse one. In this study, we performed an integrated proteogenomic analysis to have a comprehensive view on the heterogeneity of hepatic microenvironment contributing to HCC metastasis. Pairing mRNA-protein analysis revealed consistent and discordant mRNA-protein expressions in metabolism regulations and cancer-related pathways, respectively. Proteomic profiling identified three subgroups associated with the recurrence-free survival of patients. These proteomic subgroups demonstrated distinct features in metabolic reprogramming, which was potentially modified by epigenetic alterations. This study raises the point of metabolic heterogeneity in HCC noncancerous tissues and may offer a new perspective on HCC treatment.

Efficacy and Safety of Second-line Treatments in Patients with Advanced Hepatocellular Carcinoma after Sorafenib Failure: A Meta-analysis.

BACKGROUND AND AIMS: In the last decade, several second-line therapies followed by sorafenib in patients with advanced hepatocellular carcinoma (HCC) have been reported. But the outcomes were different from each other. This meta-analysis aimed to evaluate the efficacy and safety of the second-line therapies followed by sorafenib in patients with advanced HCC. METHODS: Embase (1974 to October 2019) and Ovid MEDLINE (1946 to October 2019) were searched for randomized clinical trials on second-line therapies followed by sorafenib in patients with advanced HCC. The quality of each study was assessed by the modified Jadad scale. Statistical analysis was carried out by RevMan5.3 software. Efficacy and safety were analyzed. Efficacy included overall survival (OS), disease control rate, time to progression, and progression-free survival. RESULTS: Eight studies involving 3,173 patients were eligible. No difference in OS was found between the second-line treatment group and the control group (HR=0.87, 95% CI: 0.74-1.01, p=0.06). Disease control rate (relative risk (RR)=1.36, 95% CI: 1.16-1.60, p=0.0002), time to progression (HR=0.64, 95% CI: 0.51-0.81, p=0.0002) and progression-free survival (HR=0.60, 95% CI: 0.46-0.77, p<0.0001) were significantly improved by the second-line therapies. There was a slight difference in adverse events of any grade (RR=1.07, 95% CI: 1.00-1.14, p=0.03) between the two groups. CONCLUSIONS: These second-line therapies followed by sorafenib may potentially improve the prognosis in patients with advanced HCC. Compared with other second-line therapies, regorafenib seemed to be more effective.

Cumulative damage effect of jaundice may be an effective predictor of complications in patients undergoing radical resection of Bismuth type II or above hilar cholangiocarcinoma.

BACKGROUND: There is currently no preoperative risk assessment system for predicting complications after radical resection of hilar cholangiocarcinoma. This study examined the association between the cumulative damage effect of jaundice (CDEJ) and the complications of radical resection of Bismuth II or above hilar cholangiocarcinoma. METHODS: Patients who underwent radical resection of hilar cholangiocarcinoma at the Department of Hepatobiliary Surgery, West China Hospital of Sichuan University, from April 2010 to January 2018 were retrospectively included. RESULTS: Of the 171 included patients, 115 (67.3%) patients experienced complications. Multivariate analysis found that CDEJ [odds ratio (OR) =1.0001, 95% confidence interval (95% CI) =1.000027-1.000239, P=0.014], cholangitis (OR =9.638, 95% CI =2.683-34.622, P=0.001), and preoperative bilirubin (OR =1.006, 95% CI =1.002-1.01, P=0.004) were independently associated with the incidence of complications. CDEJ (OR =1.0001, 95% CI =1.00001-1.00019, P=0.024), age (OR =1.083, 95% CI =1.029-1.14, P=0.002), preoperative bilirubin (OR =1.083, 95% CI =1.029-1.14, P=0.002), and future liver remnant (FLR) (OR =0.963, 95% CI =0.941-0.986, P=0.002) were independently associated with hepatic failure. To predict the incidence of complications, the following criteria were used. For the CDEJ cutoff of 2,151, the area under the receiver operating characteristic curve (AUC) was 0.69 (95% CI =0.615-0.759), the sensitivity was 66.09%, and the specificity was 69.64%. For the preoperative bilirubin cutoff of 111.7 µmol/L, the AUC was 0.65 (95% CI =0.573-0.721), the sensitivity was 84.35%, and the specificity was 42.86%. To predict hepatic failure, the following criteria were used. For the CDEJ cutoff of 3,931.95, the AUC was 0.605 (95% CI =0.582-0.679), the sensitivity was 51.28%, and the specificity was 70.45%. For the preoperative bilirubin cutoff of 115.9 µmol/L, the AUC was 0.638 (95% CI =0.561-0.71), the sensitivity was 92.31%, and the specificity was 32.58%. For the FLR cutoff of 50, the AUC was 0.638 (95% CI =0.515-0.667), the sensitivity was 48.72%, and the specificity was 78.79%. CONCLUSIONS: CDEJ was independently associated with complications and can moderately predict complications after surgical resection of hilar cholangiocarcinoma.

Circular RNA circDLC1 inhibits MMP1-mediated liver cancer progression via interaction with HuR.

Rationale: circular RNAs (circRNAs) have been demonstrated to play a crucial role in cancer progression. KIAA1429, a key component of the m6A methyltransferase complex, has recently been reported to promote hepatocellular carcinoma (HCC) progression by regulating the m6A methylation. The aim of present study is to investigate the role of circular RNAs in KIAA1429-mediated HCC progression. Methods: RNA sequencing (RNA-seq) and methylated RNA immunoprecipitation sequencing (m6A-seq) were utilized to identify KIAA1429-regulated circRNAs. The effects of circDLC1 on proliferation and metastasis of hepatoma cells were examined in vitro and in vivo. RT-qPCR was used to measure the expression of circDLC1 in HCC tissues and hepatoma cells. RNA FISH, RIP assays and biotin-labeled RNA pull-down were used to investigate the downstream effector of circDLC1. The downstream targets of circDLC1 were identified using RNA-seq. Results: Our data demonstrated that circDLC1 was downregulated in HCC tissues and closely relevant to favorable prognosis. Overexpression of circDLC1 inhibited the proliferation and motility of hepatoma cells in vitro and in vivo, while silencing of circDLC1 played the opposite role. Mechanistic investigations revealed that circDLC1 could bind to RNA-binding protein HuR, which subsequently reduced the interaction between HuR and MMP1 mRNAs, and thus inhibited the expression of MMP1, ultimately contributing to inhibition of HCC progression. Conclusion: Our work suggests that circDLC1, a downstream target of KIAA1429, is a promising prognostic marker for HCC patients, and the circDLC1-HuR-MMP1 axis may serve as a potential therapeutic target for HCC treatment.

Overexpression of Mothers Against Decapentaplegic Homolog 7 Activates the Yes-Associated Protein/NOTCH Cascade and Promotes Liver Carcinogenesis in Mice and Humans.

BACKGROUND AND AIMS: Mothers against decapentaplegic homolog (SMAD) 7 is an antagonist of TGF-ß signaling. In the present investigation, we sought to determine the relevance of SMAD7 in liver carcinogenesis using in vitro and in vivo approaches. APPROACH AND RESULTS: We found that SMAD7 is up-regulated in a subset of human HCC samples with poor prognosis. Gene set enrichment analysis revealed that SMAD7 expression correlates with activated yes-associated protein (YAP)/NOTCH pathway and cholangiocellular signature genes in HCCs. These findings were substantiated in human HCC cell lines. In vivo, overexpression of Smad7 alone was unable to initiate HCC development, but it significantly accelerated c-Myc/myeloid cell leukemia 1 (MCL1)-induced mouse HCC formation. Consistent with human HCC data, c-Myc/MCL1/Smad7 liver tumors exhibited an increased cholangiocellular gene expression along with Yap/Notch activation and epithelial-mesenchymal transition (EMT). Intriguingly, blocking of the Notch signaling did not affect c-Myc/MCL1/Smad7-induced hepatocarcinogenesis while preventing cholangiocellular signature expression and EMT, whereas ablation of Yap abolished c-Myc/MCL1/Smad7-driven HCC formation. In mice overexpressing a myristoylated/activated form of AKT, coexpression of SMAD7 accelerated carcinogenesis and switched the phenotype from HCC to intrahepatic cholangiocarcinoma (iCCA) lesions. In human iCCA, SMAD7 expression was robustly up-regulated, especially in the most aggressive tumors, and directly correlated with the levels of YAP/NOTCH targets as well as cholangiocellular and EMT markers. CONCLUSIONS: The present data indicate that SMAD7 contributes to liver carcinogenesis by activating the YAP/NOTCH signaling cascade and inducing a cholangiocellular and EMT signature.

Higher expression of cell division cycle-associated protein 5 predicts poorer survival outcomes in hepatocellular carcinoma.

The upregulation of cell division cycle associated protein 5 (CDCA5) has been observed in various cancer types. However, the prognostic value of CDCA5 and its underlying mechanism contributing to tumorigenesis in hepatocellular carcinoma (HCC) remain poorly understood. We used tissue microarray (TMA) to evaluate the prognosis of 304 HCC samples based on their CDCA5 expression, and analyzed the genomic features correlated with CDCA5 by using dataset from The Cancer Genome Atlas (TCGA). Compared with adjacent normal tissues, increased expression of CDCA5 was found in HCC tissues. Moreover, higher expression of CDCA5 was associated with inferior OS and DFS outcomes in HCC patients. The enrichment plots showed that the gene signatures in cell cycle, DNA replication and p53 pathways were enriched in patients with higher CDCA5 expression. Meanwhile, statistically higher mutations burdens in TP53 could also be observed in CDCA5-high patients. Integrative analysis based on miRNAseq and methylation data demonstrated a potential association between CDCA5 expression and epigenetic changes. In conclusion, our study provided the evidence of CDCA5 as an oncogenic promoter in HCC and the potential function of CDCA5 in affecting tumor microenvironment.

Classification and Prognosis Prediction from Histopathological Images of Hepatocellular Carcinoma by a Fully Automated Pipeline Based on Machine Learning.

OBJECTIVE: The aim of this study was to develop quantitative feature-based models from histopathological images to distinguish hepatocellular carcinoma (HCC) from adjacent normal tissue and predict the prognosis of HCC patients after surgical resection. METHODS: A fully automated pipeline was constructed using computational approaches to analyze the quantitative features of histopathological slides of HCC patients, in which the features were extracted from the hematoxylin and eosin (H&E)-stained whole-slide images of HCC patients from The Cancer Genome Atlas and tissue microarray images from West China Hospital. The extracted features were used to train the statistical models that classify tissue slides and predict patients' survival outcomes by machine-learning methods. RESULTS: A total of 1733 quantitative image features were extracted from each histopathological slide. The diagnostic classifier based on 31 features was able to successfully distinguish HCC from adjacent normal tissues in both the test [area under the receiver operating characteristic curve (AUC) 0.988] and external validation sets (AUC 0.886). The random-forest prognostic model using 46 features was able to significantly stratify patients in each set into longer- or shorter-term survival groups according to their assigned risk scores. Moreover, the prognostic model we constructed showed comparable predicting accuracy as TNM staging systems in predicting patients' survival at different time points after surgery. CONCLUSIONS: Our findings suggest that machine-learning models derived from image features can assist clinicians in HCC diagnosis and its prognosis prediction after hepatectomy.