Identification and validation an anoikis-related gene signature for clinical diagnosis, prognosis and treatment of patients with hepatocellular carcinoma

Identification and validation an anoikis-related gene signature for clinical diagnosis, prognosis and treatment of patients with hepatocellular carcinoma | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Identification and validation an anoikis-related gene signature for clinical diagnosis, prognosis and treatment of patients with hepatocellular carcinoma Haochen Jiang, Tao Wang, Suyin Li, Xiangxue Pan, Weifeng Tan, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4580896/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Hepatocellular carcinoma (HCC) is the most widespread malignancy in the universe, with low early diagnosis rates and high mortality. Therefore, early detection and treatment are critical to improving patients' life. Anoikis is one of the modes of cell death, and resistance to anoikis arising by aggressive tumor cells has been considered a pivotal element in cancer proliferation, while rarely have studies focused on the relationship between HCC and anoikis. Methods Anoikis-related genes were gathered from the GeneCards and MSigDB, and the R software of "limma” and the WGCNA were employed to select anoikis-related differentially expressed genes (ARDEGs). Patients from three independent cohorts (TCGA-LIHC, ICGC, and GSE14520) were classified by Nonnegative Matrix Factorization (NMF) to analyze the overall survival (OS), copy number variation (CNV), tumor microenvironment (TME), and biological characteristics of different HCC clusters. We then rely on the expression of prognostic anoikis-related differentially expressed genes (PARDEGs) to build the signature by the least absolute shrinkage and selection operator (LASSO) regression analysis, then patients were assigned into two risk groups. The study of enrichment pathways, immune microenvironment, clinicopathologic feature stratification, nomogram, tumor mutation burden (TMB), and drug prediction related to the signature was performed. More importantly, the mRNA level of the critical genes was verified at the HCC tissue level. Results HCC patients were randomly segmented into four clusters based on the PARDEGs. The result showed that clusterC2 had the worst survival time and clinical performance. Four PARDEGs, including CD24, SKP2, E2F1, and NDRG1, were selected for conducting a risk model. This risk model was significantly validated by different datasets (TCGA-LIHC, ICGC, and GSE14520) to distinguish the survival status of other HCC patients. Analysis such as the receiver operating characteristic (ROC) analyses, concordance index(C-index), and nomogram indicated that the model had excellent sensitivity and specificity. Drug response and immunotherapy also manifested differently in two risk HCC patients. Conclusion A model constructed with four PARDEGs helps to improve the detection rate of early HCC, long-term prognostic stratification of HCC patients, and postoperative personalized monitoring and treatment plan development, reflecting the medical concept of early screening, early diagnosis, early and precise therapy of HCC. Hepatocellular carcinoma anoikis signature immunotherapy tumor microenvironment Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 Introduction Hepatocellular carcinoma (HCC) remains the third most frequently diagnosed malignancy and is the second leading cause of cancer death worldwide due to its heterogeneity, drug resistance, high postoperative recurrence, and metastasis ( 1 ). China is a major liver malignancy country, where the incidence of new patients and deaths occupied about half of the global patients each year ( 2 ). Studies show more than half HCC patients are diagnosed with advanced liver cancer and systemic therapy is the only treatment option that provides a survival benefit ( 3 ). The FDA approved sorafenib in 2007 as the only systemic therapy available for advanced hepatocellular carcinoma. However, the overall survival of patients treated was less than 8 years ( 4 ). It wasn't until 2017 that regorafenib was approved for second-line treatment of advanced HCC breaking the stalemate ( 5 ). Recently, the FDA approved the marketing of the first dual immune combination therapy for the first-line treatment of patients with unresectable hepatocellular carcinoma, bringing new options and hope ( 6 ). However, the median survival of targeted combination immunotherapy rarely exceeds 2 years, and some patients may develop a certain degree of drug resistance, which seriously affects the treatment effect ( 7 ). Despite great advancements in diagnosis, screening, and treatments, clinical outcomes for patients with HCC were still unsatisfactory due to its poor five survival status ( 8 ). Hence, it is of paramount value to precisely individualize the treatment of patients and to understand the biological mechanisms behind it. Cell death is a vital phenomenon and an irreversible life process of cells. It frequently occurs in normal tissues and is necessary for maintaining cell function and morphology. Cell death can be broadly classified into active death-programmed death, apoptosis and cell necrosis ( 9 ). In 1994, a particular shape of programmed apoptosis induced through separating cells from the extracellular matrix (ECM) and other cells was first named anoikis ( 10 ). Anoikis plays an instrumental role in organism development, tissue self-homeostasis, disease genesis, and tumor metastasis, among others ( 11 ). When cells lose their attachment to the ECM, the pro-apoptotic protein BMF, which is tethered to the cytoskeleton, dissociates from the kinesin light chain two and translocates to the mitochondria promoting the anoikis ( 12 ). Resistance to anoikis is a feature of tumor metastasis, which enables tumor cells to spread to other distant organs through the circulatory system. After the tumor cells are separated from the extracellular matrix adhesion and intercellular contact, they can survive through paracrine autocrine and paracrine mechanisms to resist apoptosis and regain the ability to adhere to spread, metastasis, and invasion ( 13 ). The doctrine of anoikis has proposed new explanatory methods and research concepts for the metastasis of tumors and provided new avenues for treating malignant tumors. The tumor microenvironment (TME) is a mixture of immune cells, stromal cells and blood vessels that serves as the "soil" for tumor growth. Different cells in the tumor microenvironment have other functions, which determine the extreme heterogeneity of tumors ( 14 ). Besides, the interdependent and antagonistic relationship between tumor cells and TME can determine the process and outcome of malignant tumors. Few studies have indicated that TME is intimately linked to anoikis ( 15 ). With the clinical success of tumor immunotherapy, especially Car-T cell immunotherapy technology and immune node therapy, it is crucial to study the mechanism of tumor microenvironment regulation on immune cell function in depth with fundamental research significance ( 16 ). The above shreds of evidence suggested that anoikis played a non-negligible role in the tumor microenvironment. In this context, we aimed to focus on anoikis-related genes to analyze the mode of anoikis in liver cancer and the potential molecular mechanism and drug response of apoptosis in different modes of regulation, which is significant for exploring effective therapies for human malignancies. Material and Methods Data processing We extracted the transcriptome expression information of liver hepatocellular carcinoma (LIHC) with affiliated clinical information from The Cancer Genome Atlas(TCGA; https://portal.gdc.cancer.gov ) and the gene expression omnibus (GEO) database ( http://www.ncbi.nlm.nih.gov/geo ). To obtain accurate results, we excluded HCC patients with a survival time of fewer than 30 days, and finally, 342 tumor samples were chosen in the study from the TCGA database and 221 tumor samples from the GSE14520-GPL3971. In addition,230 HCC samples from the International Cancer Genome Consortium (ICGC) ( https://www.icgc.gov/ ) were also included (Supplement Table S1 ). Copy number variation (CNV) and somatic mutation data of HCC patients in TCGA database were derived from the UCSC Xena database(( https://xena.ucsc.edu/ ), and the chromosome location altered by CNV in the anoikis-related genes was plotted using R software of the "Rcircos” Identification of prognostic anoikis-related differentially expressed genes (PARDEGs) A total amount of 794 anoikis-related genes were obtained from the GeneCards ( https://www.genecards.org/Search/Keyword?queryString=anoikis ) and the Molecular signatures database (MSigDB), and genes with a correlation score > 0.4 were picked out for follow-up analysis (Supplement Table S2). To find ARDEGs, we first obtain DEGs between HCC patients and standard samples in the TCGA-LIHC cohort using the R software of “limma” with false discovery rate (FDR) 1, then the weighted correlation network analysis (WGCNA) method was employed to find out gene modules highly correlation with tumor process; finally, the overlapping genes of the three common modules were defined as anoikis-related differentially expressed genes. At last, univariate Cox regression ( P < 0.01) was adopted to identify PARDEGs. NMF, GSVA, ssGSEA and PCA Nonnegative matrix factorization (NMF) is widely used in disease typing by extracting bio-correlation coefficients of data within the gene expression matrix, organizing genes and samples, and capturing the internal structural features of the data to group samples and belongs to a kind of unsupervised learning which is similar to principal component analysis (PCA) ( 17 ). The PCA was employed to distinguish dimensions between different subtypes better. Gene Set Variation Analysis with the gene set (GSVA) (c2. cp. kegg. v7.5.1) was employed to inquiry the biological mechanisms behind each HCC subtype. The infiltration of 23 immune cells and their distribution in different HCC clusters were detected through the single sample gene set enrichment analysis (ssGSEA). Establishment of prognostic anoikis-related differentially expressed genes Signature We build the risk model relying on the expression of 40 prognostic anoikis-related differentially expressed genes. First, the least absolute shrinkage and selection operator (LASSO) was employed to further narrow the volume of predictive anoikis-related genes. Then, We performed a linear combination of regression coefficients (β) and expression levels of prognostic anoikis-related genes in the multivariate Cox regression model to ascertain the risk score. The calculation equation was described below: Risk scores=∑Coef(i)∗Exp(i). Confirmation of the protein and mRNA level of the four prognostic anoikis-related genes The Human Protein Atlas database (HPA, https://www.proteinatlas.org/ ) was applied to quantify the protein expression in HCC samples and standard samples. A total of eight HCC tumor samples with their paratumor tissues were collected from Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine. Meanwhile, the experiment was conducted with the permission of the ethics committee of Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine (No.2018-630-59-01). Quantitative real-time PCR analysis (qRT-PCR) Human HCC cell lines (HepG2) and normal hepatocytes(L02) were acquired from the Cell Bank of the Chinese Academy of Sciences, Shanghai, China and clustered as described before. A total mRNA was isolated from eight HCC tumor samples with their paratumor tissues and cell lines using TRIzol reagent (Bioteke Corporation, RP40002), and synthesized into cDNA with a Reverse Transcription Master kit (Vazyme, R222-01). QRT-PCR was performed by using ChamQ SYBR qPCR Master Mix (Vazyme, Q311-02). The expression in adjusted tissues normalized the mRNA levels. The sequences of primer used in this study are displayed in Supplement Table S3. Further verification of the risk model We classified HCC patients in two independent databases (the ICGC and GSE14520 datasets) into two risk groups based on the risk score. To further validate the internal reliability of the model, we divided HCC patients in the TCGA database into the training set and the test set in a 7:3 ratio. The Kaplan-Meier analysis, receiver operating characteristic curves (ROC), and Decision curve analysis (DCA) were used to analyze the reliability of the risk model. Drug sensitivity and immunotherapy The R software of pRRophetic was employed to count the half inhibitory concentrations of anti-tumor medicines frequently used to treat HCC concentration (IC50) values. T cell exclusion, T cell dysregulation, MSI (microsatellite instability), and tumor immune dysfunction and exclusion (TIDE) scores were obtained from the TIDE online database (TIDE; http://tide.dfci.harvard.edu/ ) to assess the efficacy of immunotherapy in two different risk HCC patients. We also obtained the Immunophenoscore (IPS) of HCC patients in The Cancer Immunome Atlas ( https://tcia.at/ ). Statistical analysis We utilize R software (version 4.1.2) and its accompanying packages to do the statistical analysis. Unless specifically mentioned, p < 0.05 was designed as a statistically significant difference. Results Determination of anoikis-related differentially expressed genes (ARDEGs) The transcriptome expression data of the TCGA-LIHC was compared by the R software of “limma” with |log2FC|>1 and FDR < 0.05. A total of 5760 DEGs were screened from 50 standard samples and 375 HCC patients, among which 318 were downregulated, and 5442 were upregulated (Fig. 1 A, E). Next, “WGCNA" data packets were used to understand better which genes were significantly involved in the development of liver malignancies. We take the first β that reaches 0.9 as the best β to construct gene modules, and finally, the above condition is satisfied when β = 10 (Fig. 1 B), so ten gene modules were thus formed (Fig. 1 C). As shown in Fig. 2 C and 1 D, genes in cyan color had the highest correlation in HCC patients(cor = 0.53). Therefore, we defined the cyan module as the hub module for subsequent analysis. The “Venn” package was used to overlay the DEGs in the TCGA-LIHC, cyan module, and anoikis-related genes to find the hub anoikis-related genes most closely associated with tumor development. Finally, 56 genes were selected as anoikis-related differentially expressed genes (Fig. 1 E). Further, Spearman’s analyze was performed to have a better recognition of the role of anoikis-related genes in HCC(Fig. 1 F-G). In addition, the PCA showed that 56 anoikis-related genes could be well distinguished from tumors and standard samples (Fig. 1 H). The landscape of genetic variation of prognostic anoikis-related differentially expressed genes (PARDEGs) We used univariate Cox regression( P < 0.01) to identify PARDEGs, and finally, 40 PARGs were included (Fig. 2 A). The omnidirectional regulatory network revealed the connections' landscape, correlations of 40 PARDEGs, and their prognostic values, and we found that all of them were strongly associated with a poor prognosis for HCC(Fig. 2 B). Figure 2 C depicts that only 48 of the 364 samples were mutated, a mutation rate of 13.19%, and missense mutation was the most common kind of mutation, which indicated that PARGs might be stable in HCC samples. Figure 2 D shows the position of the copy number variation (CNV) alterations of PARDEGs on chromosomes. As shown in Fig. 2 E, the CNV amplification frequency was generally different among the PARGs, most of them were observed as the gain in the copy number, and DAP3 performed as the highest CNV amplification, while SFN was the highest deletion frequency gene. In addition, the expression of 40 PARDEGs between HCC samples and standard samples was displayed in Fig. 2 F-G, and we found all of them were overexpressed in tumor tissues compared to normal ones, suggesting the expression of PARDEGs may have a heavy link to the prognosis of HCC. Construction of clusters and biological functions mediated by PARDEGs To further clarify the expression of PARDEGs, consensus clustering of the NMF algorithm was employed to divide HCC clusters depending on the expression of PARDEGs. At last, k = 4 was selected to minimize distractions between different clusters. Thus, four HCC subtypes: clusterC1( n = 55), clusterC2 ( n = 42), clusterC3( n = 76) and clusterC4( n = 169) were identified. The consensus map showed a nonnegative matrix factorization classification (Fig. 3 A). Notably, the KM curve analysis showed cluster C4 had the longest overall survival (OS), disease-free survival (DFS), disease-specific survival (DSS), and progression-free survival (PFS), while cluster C2 had the shortest (Fig. 3 B-E). To further verify the accuracy of the grouping, we used the same approach to cluster the HCC patients in the TCGC dataset and found the exact similar result as in the TCGA database (Fig. 3 F-G). As illustrated in the heatmap (Fig. 3 H), the expression of PARDEGs in cluster C2 was obviously highest among cluster C1, cluster C3, and cluster C4; not surprisingly, patients in cluster C2 had a pronounced linked to advanced clinical characteristics among those in cluster C1, cluster C3, and cluster C4. Since the prognosis based on the expression of PARDEGs differed between these four HCC subtypes, we performed the GSVA to understand better the bio-behavior regulated by PARDEGs and found that metabolic pathways like fatty acid metabolism and bile acid metabolism were significantly enrichment in cluster C4 (Fig. 3 I). This potentially explains well the different prognoses among the four HCC subtypes. Then We experimentally analyzed immune cell infiltrations in four HCC clusters by ssGSEA and found significant differences between the tumor-infiltrating immune cells of the four HCC subtypes (Supplementary Fig. 1). These pieces, as mentioned earlier, of evidence suggest a significant difference between the biological processes and tumor immune microenvironment of four HCC subtypes mediated by PARDEGs. Building the risk model base on the PARDEGs Based on LASSO analysis and multivariate Cox ratio hazard regression analysis, four PARDEGs (E2F1, SKP2, NDRG1, and CD24) were ultimately screened to set the risk model (Fig. 4 A-B). The equation for calculating the risk score was shown below: Risk score=(E2F1×0.0172464245772007 + SKP2×0.0507068315282839 + NDRG1×0.00920961642769134 + CD24×0.0038494255274358) (Fig. 4 C). We also withdrew the expression data of four PARDEGs in the ICGC and GSE14520 databases and used the same algorithm to categorize patients into high-risk and low-risk groups. The risk profile, survival status, and heatmap of the expression of the four genes in high- and low-risk patients were presented in Fig. 4 D-F. Kaplan-Meier survival analysis showed that high-risk patients in these three databases had much worse survival condition than low-risk patients (Fig. 4 G-I). moreover, the TCGA-LIHC cohort with 342 HCC patients was irregularly assigned into a training cohort with 206 patients and a testing cohort with 136 in a 7:3 ratio. Survival and the receiver operating characteristic (ROC) curves demonstrated that the predictive risk model had excellent intra-group stability and potential for development (Supplementary Fig. 2A-D). Moreover, KM survival analysis suggested that higher risk scores were heavily associated with worse survival indicators survival indicators such as DFS, DSS and PFS of HCC patients (Supplement Fig. 2 E-G). Validation and biological analysis of four PARDEGs To validate the precision of the model, first, we performed an analysis of the relationship between the expression level of these four genes and survival time and found that the higher the gene expression, the shorter the survival time of HCC patients, which suggested that these four genes were substantially related to poor prognosis of HCC(Fig. 5 A-D). We then analyzed the immunohistochemical staining of these four genes between HCC patients and healthy individuals through the HPA online database. The results showed that the protein level of four genes were noticeably enhanced in tumor tissues (Fig. 5 E-H). In addition, Expression of four PARDEGs were validated in GEPIA(Fig. 6 A), GSE14520(Fig. 6 B), GSE76427(Fig. 6 C) and GSE54236(Fig. 6 D). Finally, we detected the levels of mRNA in four PARDEGs in human tumor tissues (Fig. 6 E-H) and human HCC cell lines through qRT-PCR experiments and found the same results. To explore the biological significance behind this, we used GSVA analysis and found that these four genes were closely associated with numerous tumor pathways. Notably, all genes were negatively associated with the PPAR signaling pathway, which highlights the critical role of the PPAR signaling pathway in antitumor and provides some perspectives for future tumor therapy (Fig. 6 I). Functional enrichment analysis of the risk model A total of 968 DEGs were picked out by the R software of “limma” with false discovery rate (FDR) = 0.05 and logFCfilter = 1, of which 157 genes were down-regulated and 811 genes were upregulated in the high-risk HCC patients. Gene Ontology (GO) enrichment profiling showed that organelle fission and DNA replication pathways were enriched in DEGs (Fig. 7 A-B). Kyoto Encyclopedia of Genes and Genomes (KEGG) profiling indicated that the Cell cycle and MicroRNAs in cancer were enhanced in the high-risk group (Fig. 7 C). Besides, gene sets enrichment analysis (GSEA) showed that meiotic pathways were enriched in the high-risk group (Fig. 7 D, F), while fatty acid pathways were upregulated in the low-risk group (Fig. 7 E, G). Clinical stratification and clinical correlation analysis based on the model Considering the heterogeneity of HCC, we classified HCC patients into 18 subtypes according to their pervasive clinicopathology characteristics. The results showed that the prognostic model we constructed can distinguish the clinical information very well (Fig. 8 ). In addition, we investigated the relationship between the risk score and clinical characteristics and found that male patients aged ≤ 65 had a significantly increased risk of HCC (Fig. 9 A, F). Furthermore, clinical characteristics such as tumor history, tumor type, and HCC clusters were also studied in our research (Fig. 9 B-E). Our findings also found that risk score was strongly associated with higher tumor grade, stage, and a worse survival time (Fig. 9 G-K). We also present the clinical features, immune infiltration, and prognostic gene expression differences between HCC patients expressed through a heatmap (Fig. 9 M). To validate the general applicability of the model, we analyzed the clinical characteristics of HCC patients in both the ICGC and GSE14520 databases (Supplementary Fig. 3 and Fig. 4 ). Finally, we compare our constructed model with the existing published models ( 18 – 21 ) and find that our constructed model has better stability and prediction ability (Supplementary Figure Fig. 5 ). Establishment of the nomogram clinical characteristics in three databases For the sake of precision treatment, we developed a nomogram that includes age, gender, tumor grade, stage, tumor history, tumor type, tumor vascular, and risk score in the TCGA(Fig. 10 A). The calibration plot for 1, 3, and 5 years show that we have developed nomination charts with low inaccuracies in three databases (Fig. 10 B, E, H). The ROC curves show that the predictive performance of our constructed nomograms is much higher than that of clinical features such as gender, age, class, and stage (Fig. 10 C, F, I). ROC curves and AUC values are generally used to appraise the accuracy of prognostic models, but ROC curves only consider the specificity and sensitivity of the model and pursue accuracy, while Decision curve analysis (DCA) because it compensates for the overfitting of ROC curves, is increasingly used in clinical analysis ( 22 ). We could find that the predictive power of our constructed column line graphs and risk models was much higher than other clinical indicators (Fig. 10 D, G, J). This result was also extended to the ICGC (Supplementary Fig. 6) and GSE14520 dataset (Supplementary Fig. 7). Thus, three independent databases validated the general applicability of the accuracy of the risk model. Immune landscape between two different risk groups Figure 11A shows the immune cell bubble plots derived from different platform algorithms. We then explore the relationship between the immune cells and the risk score. As shown in Fig. 11B, we found the risk score was positively corrected with neutrophil, B cells, APC co-inhibition, and Tumor Infiltrating Lymphocytes (TIL), while negatively corrected with activated dendritic cells (aDCs), cytolytic activity, Mast cells, NK cells, and Type II IFN response. Immunogenomic analysis of more than 1,000 tumor samples from 33 TCGA cancer types has been studied, identifying six immune subtypes, including wound healing(C1), IFN-dominant(C2), inflammatory(C3), lymphocyte depletion(C4), immune resting(C5), and TGF-β dominant(C6). We found that the immune types in the high-risk group were mainly C1 and C2, while C3 was primarily present in the low-risk group (Fig. 11C). In addition, the relationship between microsatellite instability (MSI) and risk scores were also within the scope of our study (Fig. 11D). The RNA-based stemness scores (RNAss) were used to measure the correlation of the risk score with tumor stem cells, and found that the higher the risk score, the stronger the degree of hepatocytes and the lower the degree of differentiation (R = 0.31, p < 0.001) (Fig. 11E). The ESTIMATE algorithm was used to analyze the immune score, the stromal score, and the ESTIMATE score between the low-and high-risk group. The results showed significant differences in stromal score and ESTIMATE score, while no differences in the immune score (Fig. 11F). Interestingly, the TIDE score and T cell dysfunction were lower in high-risk HCC patients than in the low-risk HCC patients, representing that high-risk HCC patients would benefit more from immunotherapy than those in the low-risk group; however, the T cell exclusion was higher in the high-risk group than in the low-risk group(Fig. 11G).These above results may also be the cause of the different prognoses of high- and low-risk patients, which may also provide guidance for immunotherapy in HCC patients. Tumor mutational Burden and Risk-Related Drug-Sensitivity Prediction An oncoplot showed the topmost 20 somatic mutations in two risk groups in the TCGA-LIHC cohort (Fig. 12 A-B). The outcome revealed mutations in the TP53 gene were significantly associated with the prognosis of HCC patients. As shown in Fig. 12 C-D, the survival situation of HCC patients with higher TMB and risk scores was extremely terrible. Given the irreplaceable role of sorafenib and TACE in the treatment of HCC, we then analyzed the expression of four PARDEGs with different responses to Sorafenib treatment from GSE109211 and found that the expression of four PARDEGs were increased in non-responder patients than responder patients (Fig. 12 E). In additionally, risk scores were significantly higher in non-responders than responders to Sorafenib treatment (Fig. 12 F) ( P < 0.001). The same results were observed in transcatheter arterial chemoembolization (TACE) treatment (Fig. 12 G, H), which suggested that Sorafenib and TACE treatments can be of great benefits to lower risk score patients. We also listed the effects of common antineoplastic drugs in two risk groups (Fig. 13 A-C). In light of the importance of immunotherapy in liver cancer therapy, the potential benefit of immunotherapy with PD1 and CTLA4 for HCC patients based on TCGA expression data was calculated via the TCIA online website. The result showed the low-risk group outperformed the higher-risk group when treated with PD1 or CTLA4 (Fig. 13 D-G). Compared to the low-risk group, most immune checkpoints’ expressions of the high-risk group were higher, which suggested that patients in the low-risk group may be better prospects for immunotherapy (Fig. 13 H). Discussion Liver malignancies remain a global health threat, with several incidences expected to exceed one million cases in 2025 ( 23 ). According to epidemiological studies, the annual incidence of HCC in China is 0.03%, which is much higher than that in Japan, Europe, and United States ( < = 0.01%) ( 24 ). Low rates of early diagnosis and 5-year survival, and high rates of surgical recurrence and metastasis are the leading causes of death in HCC patients ( 25 ). According to the study, the diagnosis rate of BCLC stage 0-A patients in China is only 33%, and the 5-year survival rate for patients with progressive HCC is only 11.7%-14.1%, while the diagnosis rate of BCLC stage 0-A in Japan and Taiwan can reach 70% ( 26 ). Therefore, there is an urgent need to further improve the early screening and diagnosis of HCC in China. Anoikis is necessary for the organism's defense, which prevents the shed cells from implanting and growing into untimely locations ( 27 ). However, tumor cells, especially some malignant cells prone to distal metastasis, have extreme resistance to anoikis and, thus, complete metastasis ( 28 ). This suggests that resistance to anoikis may be a prerequisite for the survival of cancer cells during the cycle ( 29 ). Therefore, activation of anoikis can inhibit tumor metastasis ( 30 ). However, research on the progress of anoikis in HCC is still at an embryonic stage. With the rapid development of artificial intelligence, various mathematical algorithms are also widely used in the massive liver cancer data, providing a more convenient method for diagnosing and analyzing liver cancer ( 31 ). 502 anoikis-related genes were included in our study, and by combining multiple algorithms, we finally obtained 56 anoikis-related differentially expressed genes. Then 40 prognostic anoikis-related genes were screened through univariate Cox regression analysis. Based on the expression of these 40 predictive anoikis-related genes, 342 patients in TCGA-LIHC dataset with hepatocellular carcinoma were assigned into 4 HCC subtypes with different overall survival times through the NMF algorithm. To verify the accuracy of the classification, HCC patients in the ICGC dataset were also classified into four types using the same methodology. GSVA analysis showed that patients in clusterC4 were prominently enriched in fatty acid metabolism and bile acid metabolism; in addition, immune cell infiltration analysis showed that ClusterC2 was lower in T cells CD4 memory resting, T cells CD4 memory activated and T cells follicular helper while highest in eosinophils compared to other HCC subtypes. This could explain why ClusterC4 had the best prognosis while clutserC2 had the worst. Next, we found 4 of the 40 prognostic anoikis-related genes by LASSO analysis to construct a predictive model, which was CD24, SKP2, NDRG1, and E2F1. E2F1 is essential in preventing tumorigenesis by strictly controlling the cell cycle, maintaining genomic integrity, and responding to replication stress and DNA damage ( 32 ). Previous studies have found that overexpression of E2F1 has been shown to promote cancers in various organs ( 33 , 34 ). Peng et al. ( 35 ) reported that CAMK2N1 inhibits the proliferation, invasion, and migration of HCC cells by suppressing E2F1-mediated cell cycle signaling. Lee et al. ( 36 ) found that SKP2 expression was significantly upregulated in HCC patients and, more importantly, that overexpression of SKP2 was significantly associated with poor prognosis and advanced clinical features. Yang et al. ( 37 ) found that CD24 was highly expressed in hepatic tumor cell lines and in patients with hepatocellular carcinoma and that overexpression of CD24 was associated with a high invasive and metastatic potential of HCC, suggesting that overexpression of CD24 is an independent risk factor for poor prognosis in patients with hepatocellular carcinoma. NDRG1 has been included in numerous pieces of literature as one of the crucial genes in the prognostic model of HCC ( 38 , 39 ). More importantly, we verified the protein level expression through the HPA public database. On the other hand, we validated our findings at the level of tumor tissues by qRT-PCR. Survival analysis and AUC values showed that the risk score can significantly differentiate between favorable and adverse HCC patients. To further explore its biological significance, we performed GSVA and GSEA analysis and found that fatty acid metabolism plays a pivotal role in tumorigenesis. In our study, we found that all four PARDEGs utilized to build the risk model were negatively correlated with the PPAR signaling pathway, which laterally verified the accuracy of our experiments. Abnormal lipid distribution is a prominent alteration in tumor patients, where abnormal lipid distribution helps moisten tumor cells and promotes tumor proliferation and infiltration. Therefore, inhibition of lipid transport in tumor cells has an invaluable role in minimizing tumor drug resistance and improving immune efficacy ( 40 , 41 ). In addition, tumor cells take advantage of lipid metabolism to regulate the state of stromal cells and the activity of immune cells to establish a tumor microenvironment that creates therapeutic resistance and stimulates cancer recurrence. ( 42 ). PPARα signaling pathway can improve lipid metabolism disorders and accelerate lipid differentiation by increasing fatty acid β-oxidation ( 43 ). These results suggested that developing anti-hepatocellular carcinoma chemotherapeutic agents targeting PPAR may be a breakthrough in treating liver cancer. AFP remains a common and important indicator for diagnosis and efficacy testing of hepatocellular carcinoma, but lacks sensitivity and specificity ( 44 ). The sensitivity and specificity of the model constructed in this study were significantly higher than those of clinical indicators, which can significantly increase the detection rate of early liver cancer and reduce the death rate. It is clear the importance of patient stratification for improving HCC treatment outcomes, so K-M survival curves between different patients according to clinicopathological characteristics and risks were presented. The results showed that high risk scores exhibited a poor prognosis in patients with different clinical characteristics Sorafenib remains the drug of choice for advanced HCC patients, however, greater drug resistance poses a challenge for clinicians ( 45 ). The combination of immunotherapy and molecularly targeted drugs have recently brought a new light to patients with advanced HCC ( 46 ). However, the highly heterogeneous nature of liver cancer shaped the complex tumor microenvironment and led to heterogeneity within and between tumor foci, thus limiting the implementation of precision medicine ( 47 ). Therefore, there was an urgent need for clinical practice to precisely analyze the molecular biomarkers of the tumor microenvironment (TME) occurring in HCC and to develop individualized treatment strategies ( 48 ). This study found that high-risk score patients benefited less from sorafenib and TACE treatment, while patients with high-risk score had significantly higher levels of immune checkpoint expression and lower TIDE scores, suggesting that immunotherapy may benefit patients in high-risk score. In addition, we compared several drugs commonly used in HCC chemotherapy and found that the patients in low-risk score had better efficacy than patients in high-risk score when treated with sorafenib, erlotinib, and sunitinib antitumor. To the best of our knowledge, this was the first prognostic model constructed based on the anoikis-related genes with tumor mutational load, tumor microenvironment, immunotherapy, and drug sensitivity, which provided some support for precision and individualized treatment of liver cancer and provided an alternative way for clinicians to screen for early-stage HCC. However, there are some unavoidable flaws in our study. First, underlying studies on how anoikis-related genes affected HCC were not explored further in depth in our research. In addition, the study's findings need to be further explored in terms of which countries only are appropriate for HCC patients. More importantly, although these four prognostic genes were validated in human tissues, the sample size was small, and there was a lack of real-world HCC cohort validation, so the results of this study have a long way to go. Conclusion In a nutshell, the model based on anoikis-related genes provides new insights and perspectives for researchers to explore tumor mechanisms and new drug development by assessing the prognosis of HCC patients and guiding clinicians in drug treatment. Declarations Author contributions Haochen Jiang and Tao Wang designed this study.Tao Wang and Suyin Li drafted the manuscript. Xuehua Sun, Xiangxue Pan, and Weifeng Tan conducted the experiment. All authors participated in the review of the manuscript and approved the submitted version. Ethics statement The experimental protocol was allowed by the ethics committee of Shuguang Hospital, Shanghai University of Traditional Chinese Medicine Fundings This study was supported by the National Natural Science Foundation of China (No.82074336 to X Sun) and the Program of Shanghai 2020 Science and Technology Innovation Action Plan (No.20S21901600 to X Sun). Acknowledgments The authors would like to express gratitude to all those who quietly helped create the public database. Conflict of interest All authors declare no conflict of interest. Data availability All data in the manuscript is free to download online. The related data were downloaded from the GEO (https://www.ncbi.nlm.nih.gov/geo/), TCGA (https://dcc.icgc.org/), ICGC (https://dcc.icgc.org/projects-/LIRI-JP), HPA (https://www.proteinatlas.org/) and TIDE(http://tide.dfci.harvard.edu/). References Rumgay H, Arnold M, Ferlay J, Lesi O, Cabasag CJ, Vignat J, et al. Global burden of primary liver cancer in 2020 and predictions to 2040. J Hepatol. 2022;77(6):1598–606. 10.1016/j.jhep.2022.08.021 . Yang J, Pan G, Guan L, Liu Z, Wu Y, Liu Z, et al. The burden of primary liver cancer caused by specific etiologies from 1990 to 2019 at the global, regional, and national levels. Cancer Med. 2022;11(5):1357–70. 10.1002/cam4.4530 . Sperandio RC, Pestana RC, Miyamura BV, Kaseb AO. Hepatocellular Carcinoma Immunotherapy. Annu Rev Med. 2022;73:267–78. 10.1146/annurev-med-042220-021121 . Finn RS, Qin S, Ikeda M, Galle PR, Ducreux M, Kim TY, et al. Atezolizumab plus Bevacizumab in Unresectable Hepatocellular Carcinoma. N Engl J Med. 2020;380(20):1894–905. 10.1056/NEJMoa1915745 . Llovet JM, Castet F, Heikenwalder M, Maini MK, Mazzaferro V, Pinato DJ, et al. Immunotherapies for hepatocellular carcinoma. Nat Rev Clin Oncol. 2022;19(3):151–72. 10.1038/s41571-021-00573-2 . Yang C, Zhang H, Zhang L, Zhu AX, Bernards R, Qin W, et al. Evolving therapeutic landscape of advanced hepatocellular carcinoma. Nat Rev Gastroenterol Hepatol. 2022. 10.1038/s41575-022-00704-9 . Cheng AL, Hsu C, Chan SL, Choo SP, Kudo M. Challenges of combination therapy with immune checkpoint inhibitors for hepatocellular carcinoma. J Hepatol. 2020;72(2):307–19. 10.1016/j.jhep.2019.09.025 . Llovet JM, Kelley RK, Villanueva A, Singal AG, Pikarsky E, Roayaie S, et al. Hepatocellular carcinoma. Nat Rev Dis Primers. 2021;7(1):6. 10.1038/s41572-020-00240-3 . Magnani L, Colantuoni M, Mortellaro A, Gasdermins. New Therapeutic Targets in Host Defense, Inflammatory Diseases, and Cancer. Front Immunol. 2022;13:898298. 10.3389/fimmu.2022.898298 . Frisch SM, Francis H. Disruption of epithelial cell-matrix interactions induces apoptosis. J Cell Biol. 1994;124(4):619–26. 10.1083/jcb.124.4.619 . Mason JA, Hagel KR, Hawk MA, Schafer ZT. Metabolism during ECM Detachment: Achilles Heel of Cancer Cells? Trends cancer (2017). 3(7):475–481. 10.1016/j.trecan.2017.04.009 . Delgado M, Tesfaigzi Y. Is BMF central for anoikis and autophagy? Autophagy. 2014;10(1):168–9. 10.4161/auto.26759 . Wang J, Luo Z, Lin L, Sui X, Yu L, Xu C, et al. Anoikis-Associated Lung Cancer Metastasis: Mech Ther Cancers(Basel). 2022;14(19):4791. 10.3390/cancers14194791 . Li F, Kitajima S, Kohno S, Yoshida A, Tange S, Sasaki S, et al. Retinoblastoma Inactivation Induces a Protumoral Microenvironment via Enhanced CCL2 Secretion. Cancer Res. 2019;79(15):3903–15. 10.1158/0008-5472.can-18-3604 . Chen J, Li K, Chen J, Wang X, Ling R, Cheng M, et al. Aberrant translation regulated by METTL1/WDR4-mediated tRNA N7-methylguanosine modification drives head and neck squamous cell carcinoma progression. Cancer Commun (Lond). 2022;42(3):223–44. 10.1002/cac2.12273 . Wilson JL, Nägele T, Linke M, Demel F, Fritsch SD, Mayr HK, et al. Inverse Data-Driven Modeling and Multiomics Analysis Reveals Phgdh as a Metabolic Checkpoint of Macrophage Polarization and Proliferation. Cell Rep. 2020;30(5):1542–e15527. 10.1016/j.celrep.2020.01.011 . Gay CM, Stewart CA, Park EM, Diao L, Groves SM, Heeke S, et al. Patterns of transcription factor programs and immune pathway activation define four major subtypes of SCLC with distinct therapeutic vulnerabilities. Cancer Cell. 2021;39(3):346–e3607. 10.1016/j.ccell.2020.12.014 . Dong W, Xie Y, Huang H. Prognostic Value of Cancer-Associated Fibroblast-Related Gene Signatures in Hepatocellular Carcinoma. Front Endocrinol (Lausanne). 2022;13:884777. 10.3389/fendo.2022.884777 . Long S, Chen Y, Wang Y, Yao Y, Xiao S, Fu K. Identification of Ferroptosis-related molecular model and immune subtypes of hepatocellular carcinoma for individual therapy. Cancer Med. 2023;12(2):2134–47. 10.1002/cam4.5032 . Rao G, Pan H, Sheng X, Liu J. Prognostic Value of Stem Cell Index-Related Characteristics in Primary Hepatocellular Carcinoma. Contrast Media Mol Imaging. 2022;2022:2672033. 10.1155/2022/2672033 . Regmi P, He ZQ, Lia T, Paudyal A, Li FY. N7-Methylguanosine Genes Related Prognostic Biomarker in Hepatocellular Carcinoma. Front Genet. 2022;13:918983. 10.3389/fgene.2022.918983 . Huang YQ, Liang CH, He L, Tian J, Liang CS, Chen X, et al. Development and Validation of a Radiomics Nomogram for Preoperative Prediction of Lymph Node Metastasis in Colorectal Cancer. J Clin Oncol. 2016;34(18):2157–64. 10.1200/jco.2015.65.9128 . Huang DQ, Mathurin P, Cortez-Pinto H, Loomba R. Global epidemiology of alcohol-associated cirrhosis and HCC: trends, projections, and risk factors. Nat Rev Gastroenterol Hepatol. 2022;18:1–13. 10.1038/s41575-022-00688-6 . Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin. 2021;71(3):209–49. 10.3322/caac.21660 . Dasgupta P, Henshaw C, Youlden DR, Clark PJ, Aitken JF, Baade PD. Global Trends in Incidence Rates of Primary Adult Liver Cancers: A Systematic Review and Meta-Analysis. Front Oncol. 2020;10:171. 10.3389/fonc.2020.00171 . Zeng H, Chen W, Zheng R, Zhang S, Ji JS, Zou X, et al. Changing cancer survival in China during 2003-15: a pooled analysis of 17 population-based cancer registries. Lancet Glob Health. 2018;6(5):e555–67. 10.1016/s2214-109x . (18)30127-x. Gilmore AP, Anoikis. Cell Death Differ. 2005;2:1473–7. 10.1038/sj.cdd.4401723 . Jin L, Chun J, Pan C, Kumar A, Zhang G, Ha Y, et al. The PLAG1-GDH1 Axis Promotes Anoikis Resistance and Tumor Metastasis through CamKK2-AMPK Signaling in LKB1-Deficient Lung Cancer. Mol Cell. 2018;69(1):87–e997. 10.1016/j.molcel.2017.11.025 . Kakavandi E, Shahbahrami R, Goudarzi H, Eslami G, Faghihloo E. Anoikis resistance and oncoviruses. J Cell Biochem. 2018;119(3):2484–91. 10.1002/jcb.26363 . Dolinschek R, Hingerl J, Benge A, Zafira C, Schüren E, Ehmoser EK, et al. Constitutive activation of integrin αvβ3 contributes to anoikis resistance of ovarian cancer cells. Mol Oncol. 2021;15(2):503–22. 10.1002/1878-0261.12845 . Zhang X, Wang Z, Tang W, Wang X, Liu R, Bao H, et al. Ultrasensitive and affordable assay for early detection of primary liver cancer using plasma cell-free DNA fragmentomics. Hepatology. 2022;76(3):317–29. 10.1002/hep.32308 . Fouad S, Hauton D, D'Angiolella V. E2F1: Cause and Consequence of DNA Replication Stress. Front Mol Biosci. 2020;7:599332. 10.3389/fmolb.2020.599332 . Conner EA, Lemmer ER, Omori M, Wirth PJ, Factor VM, Thorgeirsson SS. Dual functions of E2F-1 in a transgenic mouse model of liver carcinogenesis. Oncogene. 2000;19(44):5054–62. 10.1038/sj.onc.1203885 . Zheng X, Huang M, Xing L, Yang R, Wang X, Jiang R, et al. The circRNA circSEPT9 mediated by E2F1 and EIF4A3 facilitates the carcinogenesis and development of triple-negative breast cancer. Mol Cancer. 2020;19(1):73. 10.1186/s12943-020-01183-9 . Peng JM, Tseng RH, Shih TC, Hsieh SY. CAMK2N1 suppresses hepatoma growth through inhibiting E2F1-mediated cell-cycle signaling. Cancer Lett. 2021;497:66–76. 10.1016/j.canlet.2020.10.017 . Lee SW, Li CF, Jin G, Cai Z, Han F, Chan CH, et al. Skp2-dependent ubiquitination and activation of LKB1 is essential for cancer cell survival under energy stress. Mol Cell. 2015;57(6):1022–33. 10.1016/j.molcel.2015.01.015 . Yang XR, Xu Y, Yu B, Zhou J, Li JC, Qiu SJ, et al. CD24 is a novel predictor for poor prognosis of hepatocellular carcinoma after surgery. Clin Cancer Res. 2009;15(17):5518–27. 10.1158/1078-0432.ccr-09-0151 . Zeng F, Zhang Y, Han X, Zeng M, Gao Y, Weng J. Employing hypoxia characterization to predict tumor immune microenvironment, treatment sensitivity and prognosis in hepatocellular carcinoma. Comput Struct Biotechnol J. 2021;19:2775–89. 10.1016/j.csbj.2021.03.033 . Cheng J, Xie HY, Xu X, Wu J, Wei X, Su R, et al. NDRG1 as a biomarker for metastasis, recurrence and of poor prognosis in hepatocellular carcinoma. Cancer Lett. 2011;310(1):35–45. 10.1016/j.canlet.2011.06.001 . Röhrig F, Schulze A. The multifaceted roles of fatty acid synthesis in cancer. Nat Rev Cancer. 2016;16(11):732–49. 10.1038/nrc.2016.89 . Hall Z, Chiarugi D, Charidemou E, Leslie J, Scott E, Pellegrinet L, et al. Lipid Remodeling in Hepatocyte Proliferation and Hepatocellular Carcinoma. Hepatology. 2021;73(3):1028–44. Menendez JA, Lupu R. Fatty acid synthase and the lipogenic phenotype in cancer pathogenesis. Nat Rev Cancer. 2007;7(10):763–77. 10.1038/nrc2222 . Li J, Huang Q, Long X, Zhang J, Huang X, Aa J, et al. CD147 reprograms fatty acid metabolism in hepatocellular carcinoma cells through Akt/mTOR/SREBP1c and P38/PPARα pathways. J Hepatol. 2015;63(6):1378–89. 10.1016/j. jhep.2015.07.039. Reig M, Forner A, Rimola J, Ferrer-Fàbrega J, Burrel M, Garcia-Criado Á, et al. BCLC strategy for prognosis prediction and treatment recommendation: The 2022 update. J Hepatol. 2022;76(3):681–93. 10.1016/j.jhep.2021.11.018 . Su GL, Altayar O, O'Shea R, Shah R, Estfan B, Wenzell C. al.AGA Clinical Practice Guideline on Systemic Therapy for Hepatocellular Carcinoma. Gastroenterology. 2022;162(3):920–34. 10.1053/j.gastro.2021.12.276 . Xu F, Jin T, Zhu Y, Dai C. Immune checkpoint therapy in liver cancer. J Exp Clin Cancer Res. 2018;37(1):110. 10.1186/s13046-018-0777-4 . Hiraoka A, Kumada T, Tada T, Hirooka M, Kuriyama K, Tani J, et al. Atezolizumab plus bevacizumab treatment for unresectable hepatocellular carcinoma: Early clinical experience. Cancer Rep (Hoboken). 2022;5(2):e1464. 10.1002/cnr2.1464 . Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144(5):646–74. 10.1016/j.cell.2011.02.013 . Additional Declarations No competing interests reported. Supplementary Files TableS1.docx TableS2.docx TableS3.docx SupplementaryMaterial1.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4580896","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":320903497,"identity":"54d144f7-d2b7-4593-890a-931d086e9b23","order_by":0,"name":"Haochen Jiang","email":"","orcid":"","institution":"Shuguang Hospital, Shanghai University of Traditional Chinese Medicine","correspondingAuthor":false,"prefix":"","firstName":"Haochen","middleName":"","lastName":"Jiang","suffix":""},{"id":320903498,"identity":"56828737-412d-42cc-aec6-ee4684486ef7","order_by":1,"name":"Tao Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA5ElEQVRIie3QMUsDMRTA8RcC6RLaNQU5v8KDQtZ+lRcOMlkpuNxQMKD0hrZ09WM4OuaWTOnezYhfwG4VBL1dMefmkN/8/iTvARTFPyRG+1MiFLfz59cuUbPKJ2PpZ5iWYwJva0wx5JNKkZ6mt6pPrvT05Z4P+Jj0pAj1NXPBNsYJmLQbyuziPBLaG87uwtE8XYCKh8fMK50jwsDWXNijiQJQLTKJqsETfrKtkHpp1nxIYpnrj2wepNQwLJGRQ5/MUIlaUQwyu8tlux29nz9EhYp3p3Ozqibt7vfkG/m38aIoiuJHXydVSWdY9U40AAAAAElFTkSuQmCC","orcid":"","institution":"Shuguang Hospital, Shanghai University of Traditional Chinese Medicine","correspondingAuthor":true,"prefix":"","firstName":"Tao","middleName":"","lastName":"Wang","suffix":""},{"id":320903499,"identity":"3803df0f-c357-4061-8da2-8bc58cbaf5fc","order_by":2,"name":"Suyin Li","email":"","orcid":"","institution":"Shuguang Hospital, Shanghai University of Traditional Chinese Medicine","correspondingAuthor":false,"prefix":"","firstName":"Suyin","middleName":"","lastName":"Li","suffix":""},{"id":320903500,"identity":"4e981dcd-11ff-4b90-a0d2-0d5ab7d00a2b","order_by":3,"name":"Xiangxue Pan","email":"","orcid":"","institution":"Shanghai hospital of Traditional Chinese medicine","correspondingAuthor":false,"prefix":"","firstName":"Xiangxue","middleName":"","lastName":"Pan","suffix":""},{"id":320903501,"identity":"b58db192-7b8f-4457-99cf-fea85753e1b7","order_by":4,"name":"Weifeng Tan","email":"","orcid":"","institution":"Tongji Hospital of Tongji university, Shanghai hospital of Traditional Chinese medicine","correspondingAuthor":false,"prefix":"","firstName":"Weifeng","middleName":"","lastName":"Tan","suffix":""},{"id":320903502,"identity":"72deacac-6929-48f2-a223-2c2c4af0a08b","order_by":5,"name":"Xuehua Sun","email":"","orcid":"","institution":"Shuguang Hospital, Shanghai University of Traditional Chinese Medicine","correspondingAuthor":false,"prefix":"","firstName":"Xuehua","middleName":"","lastName":"Sun","suffix":""}],"badges":[],"createdAt":"2024-06-14 09:11:32","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4580896/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4580896/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":59515718,"identity":"7463f230-bfe1-49f4-b420-b021b28663a9","added_by":"auto","created_at":"2024-07-02 17:33:59","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":738389,"visible":true,"origin":"","legend":"\u003cp\u003eRecognition of anoikis-related differentially expressed genes (ARDEGs). (A) The Volcano plot shows differentially expressed genes (DEGs) in normal and tumor tissues. (B) \u0026nbsp;The determination of soft-thresholding power. (C) Relationship between gene modules and traits. (D) Identification of the modules most relevant to tumor progression. (E) The Venn diagram shows 56 ARDEGs. (F) Correlation network between 56 ARDEGs. (G) Heatmap of 56 ARDEGs correlation. (H) The Principal component analysis (PCA) of 56 ARDEGs in HCC tumor and standard samples.\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-4580896/v1/839a90540c7fe2df8289f6b3.png"},{"id":59517185,"identity":"e995bb12-0e11-483d-9018-32faf6e695cd","added_by":"auto","created_at":"2024-07-02 17:49:59","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1070402,"visible":true,"origin":"","legend":"\u003cp\u003eGenetic variation of prognostic anoikis-related differentially expressed genes (PARDGEs) in hepatocellular carcinoma (HCC). (A) PARDEGs after uniCox regression analysis. (B) Regulatory network of PARDEGs. (C) Waterfall plot of PARDEGs. (D) The location of copy number variation (CNV) alterations of PARDEGs on chromosomes (E) Frequency of CNV of PARDEGs. (F) The boxplot of expressions of PARDEGs between HCC and normal samples. (H) The heatmap of PARDEGs differs in HCC and normal samples. ***\u003cem\u003ep\u003c/em\u003e\u0026lt;0.001.\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-4580896/v1/90b8f9e0b68cee0b9378bdcc.png"},{"id":59517187,"identity":"86caf9d8-498b-4875-b8e2-df1855981a5e","added_by":"auto","created_at":"2024-07-02 17:50:00","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1110704,"visible":true,"origin":"","legend":"\u003cp\u003eConstruction of clusters by nonnegative matrix factorization (NMF) algorithm. (A) NMF rank survey for k=2-10. (B) The heat map shows that hepatocellular carcinoma (HCC) was classified into four HCC clusters in TCGA-LIHC. (C-F) The Kaplan–Meier survival analysis of OS, DFS, DSS and PFS among four HCC clusters in TCGA-LIHC. (G) The heatmap shows that HCC was classified into four HCC clusters in ICGC. (H) The Kaplan–Meier survival analysis of OS among four HCC clusters in ICGC. (I) Heatmap of four HCC clusters with clinicopathological features and expressions of prognostic anoikis-related differentially expressed genes (PARDEGs). (J) Underlying biological mechanisms of four HCC clusters. *\u003cem\u003ep\u003c/em\u003e\u0026lt;0.05, **\u003cem\u003ep\u003c/em\u003e\u0026lt;0.01, ***\u003cem\u003ep\u003c/em\u003e\u0026lt;0.001.\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-4580896/v1/d05c8aead5ac38ca50fad032.png"},{"id":59516844,"identity":"8f0bb5d9-e40f-4128-8669-d130274b0395","added_by":"auto","created_at":"2024-07-02 17:41:59","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":656627,"visible":true,"origin":"","legend":"\u003cp\u003eEstablishment and confirmation of prognostic anoikis-related differentially expressed genes (PARDEGs) in multi cohorts. (A-B) Diagram of the dynamic process of filtering variables by least absolute shrinkage and selection operator (LASSO). (C) Signature Coefficient for PARDEGs. (D) The PARDEGs score distribution, survival status, and gene expression of patients within the TCGA (D), ICGC (E), and GSE14520(F) datasets are shown by ranked dot scatter plots and heatmap. (G-I) Kaplan–Meier analysis between the PARDEGs score groups in the TCGA, ICGC, and GSE14520 dataset, respectively.\u003c/p\u003e","description":"","filename":"Figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-4580896/v1/ed9e13c07fdf74824c2b329b.png"},{"id":59515724,"identity":"008249fa-90bd-40d2-9a10-d92a98bb86de","added_by":"auto","created_at":"2024-07-02 17:33:59","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":1126308,"visible":true,"origin":"","legend":"\u003cp\u003eImmunohistochemistry and biological analysis of four PARDEGs (A-D) Overall survival of CD24(A), E2F1(B), NDRG1(C), and SKP2(D). Immunohistochemistry of CD24(E), E2F1(F), NDRG1(G), and SKP2(H). (I) Biological analysis of four PARDEGs.\u003c/p\u003e","description":"","filename":"Figure5.png","url":"https://assets-eu.researchsquare.com/files/rs-4580896/v1/3ebe0a3c7e7d292cceff9471.png"},{"id":59515731,"identity":"972b1e1e-742e-4460-a3e6-867348c0399e","added_by":"auto","created_at":"2024-07-02 17:34:01","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":383814,"visible":true,"origin":"","legend":"\u003cp\u003eValidation of four PARDEGs with independent datasets and experiments.\u003c/p\u003e\n\u003cp\u003e(A) The mRNA levels of four PARDEGs from the GEPIA, (B) GSE14520, (C) GSE76427 and (D) GSE54236. The transcript levels of four PARDEGs in HCC samples(E) and HCC cell lines(F).\u003c/p\u003e","description":"","filename":"Figure6.png","url":"https://assets-eu.researchsquare.com/files/rs-4580896/v1/93e6c39f47640b92bb42e1bb.png"},{"id":59515729,"identity":"56f1e73c-044e-4e8f-9120-4e77f34f52c6","added_by":"auto","created_at":"2024-07-02 17:34:01","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":927706,"visible":true,"origin":"","legend":"\u003cp\u003eFunctional enrichment analysis of the risk model. (A-B) GO analysis of DEGs between high and low risk groups. (C) KEGG analysis of DEGs between high and low risk groups. (D-G) GSEA analysis of DEGs between high- (D, F) and low-risk groups (E, G).\u003c/p\u003e","description":"","filename":"Figure7.png","url":"https://assets-eu.researchsquare.com/files/rs-4580896/v1/3a41bdbdf9bbd7ab0f8c60b9.png"},{"id":59515734,"identity":"54181c05-1e0b-4929-b7cf-678b2b6415ac","added_by":"auto","created_at":"2024-07-02 17:34:02","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":548293,"visible":true,"origin":"","legend":"\u003cp\u003ePrognostic Relationship between prognostic anoikis-related differentially expressed genes (PARDEGs) and clinical parameters in the TCGA dataset. (A-I) Overall survival of KM analysis between PARDEGs and age(A), cancer history(B), tumor free or not(C), gender(D), grade(E), TMN (F, H) and stage(G).\u003c/p\u003e","description":"","filename":"Figure8.png","url":"https://assets-eu.researchsquare.com/files/rs-4580896/v1/951a2cdf56de1201ff2b2853.png"},{"id":59515727,"identity":"cdaa6f40-66c7-416a-a50e-2f6d2b9db571","added_by":"auto","created_at":"2024-07-02 17:34:00","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":630922,"visible":true,"origin":"","legend":"\u003cp\u003eRisk score profile of prognostic anoikis-related differentially expressed genes (PARDEGs) with different clinical characteristics in TCGA. (A-L) Age, cancer history, type, cluster, gender, gender, TMN, stage, and vascular tumor. (M) Heatmap of correlation between signature and clinical features.\u003c/p\u003e","description":"","filename":"Figure9.png","url":"https://assets-eu.researchsquare.com/files/rs-4580896/v1/5581b9bd1751534ec3386f9d.png"},{"id":59515721,"identity":"da022e93-4105-48b2-aff5-1f65b8148364","added_by":"auto","created_at":"2024-07-02 17:33:59","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":330448,"visible":true,"origin":"","legend":"\u003cp\u003ePrognostic value of prognostic anoikis-related differentially expressed genes (PARDEGs) in TCGA, ICGC, and GSE14520. (A) Nomogram based on PARDEGs scores and clinical characteristics: type, history, vascular, gender, age, stage, and grade in the TCGA dataset. (B, E, H) Calibration curves for 1-year, 3-year, and 5-year in TCGA(B), ICGC(E) and GSE14520(H) datasets. (C, F, I) Area under curve (AUC) values with nomogram, risk score and clinical characteristics in in TCGA(C), ICGC(F) and GSE14520 (I) datasets (D, G, J) Decision curve analysis (DCA) plot of the model in TCGA(D), ICGC(G) and GSE14520 (J) datasets.\u003c/p\u003e","description":"","filename":"Figure10.png","url":"https://assets-eu.researchsquare.com/files/rs-4580896/v1/dfcde76faf49b41e2fb1b5b3.png"},{"id":59516851,"identity":"8ace8ef9-2e36-4aed-9420-89f18c055f91","added_by":"auto","created_at":"2024-07-02 17:42:01","extension":"png","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":602689,"visible":true,"origin":"","legend":"\u003cp\u003eImmune landscape in two risk groups. (A) Immune cells bubble chart between two risk groups. (B) Analysis of immune function in two risk groups. (C) Type of immunization in two risk groups. (D) Relationship between the risk score and MSI (microsatellite instability). (E) Correlation analysis of the risk score with RNA-based stemness scores (RNAss) (F) Box line plot form of the difference in stromal score, immune score, and estimate score in two risk groups. (G) Relationship between the risk score and tumor immune dysfunction and exclusion (TIDE), Dysfunction and Exclusion.\u003c/p\u003e","description":"","filename":"Figure11.png","url":"https://assets-eu.researchsquare.com/files/rs-4580896/v1/c27c4a04e517b6716b1c3fd5.png"},{"id":59517186,"identity":"ecdc648c-be0f-4672-8b99-1e6d65466988","added_by":"auto","created_at":"2024-07-02 17:49:59","extension":"png","order_by":12,"title":"Figure 12","display":"","copyAsset":false,"role":"figure","size":584723,"visible":true,"origin":"","legend":"\u003cp\u003eTumor mutational burden (TMB), Sorafenib, TACE and risk scores. (A) Common mutation frequency genes in the high (A) and low risk (B) groups. (C) Survival analysis of H-TMB and L-TMB. (D) Survival analysis of TMB and the risk score combinations. (E) Differential expression of PARDEGs with different responses to Sorafenib treatment. (F) Differences in the risk score between responder and non-responder HCC patients with Sorafenib treatment. (G) Differential expression of PARDEGs with different responses to TACE treatment (H) Differences in the risk score between responder and non-responder HCC patients with TACE treatment. *\u003cem\u003ep\u003c/em\u003e\u0026lt;0.05, **\u003cem\u003ep\u003c/em\u003e\u0026lt;0.01, ***\u003cem\u003ep\u003c/em\u003e\u0026lt;0.001.\u003c/p\u003e","description":"","filename":"Figure12.png","url":"https://assets-eu.researchsquare.com/files/rs-4580896/v1/9718829b83ed62c47e0f0eb8.png"},{"id":59515726,"identity":"fa6a6a9e-b933-4dba-b106-4c4b535a5072","added_by":"auto","created_at":"2024-07-02 17:34:00","extension":"png","order_by":13,"title":"Figure 13","display":"","copyAsset":false,"role":"figure","size":434363,"visible":true,"origin":"","legend":"\u003cp\u003eDrug efficacy. (A-C) Differences in common anticancer drugs in two risk groups. (D-G) the immunophenoscore of two risk patients based on the TCIA database. (H) Distribution of 32 immune checkpoints.\u003c/p\u003e","description":"","filename":"Figure13.png","url":"https://assets-eu.researchsquare.com/files/rs-4580896/v1/a38f9fcd645f7bea1c956600.png"},{"id":67364875,"identity":"f32fe8ba-2972-45d5-a354-df9973da3f3e","added_by":"auto","created_at":"2024-10-24 07:02:08","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":10189861,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4580896/v1/8bea58d9-469c-4bd8-ba85-eaf1e630981e.pdf"},{"id":59517188,"identity":"ef757fdc-1ea2-4510-adb8-c3e4f07f8485","added_by":"auto","created_at":"2024-07-02 17:50:01","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":13051,"visible":true,"origin":"","legend":"","description":"","filename":"TableS1.docx","url":"https://assets-eu.researchsquare.com/files/rs-4580896/v1/81959aaf43f105ae150eea14.docx"},{"id":59516845,"identity":"e801a8b5-7ed2-4ccf-8d0e-86ce3adbe7fe","added_by":"auto","created_at":"2024-07-02 17:41:59","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":58804,"visible":true,"origin":"","legend":"","description":"","filename":"TableS2.docx","url":"https://assets-eu.researchsquare.com/files/rs-4580896/v1/410736fe0377be79ff615ca9.docx"},{"id":59516842,"identity":"176c0334-b3ea-4620-8ab1-fcd5f2760a38","added_by":"auto","created_at":"2024-07-02 17:41:59","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":11091,"visible":true,"origin":"","legend":"","description":"","filename":"TableS3.docx","url":"https://assets-eu.researchsquare.com/files/rs-4580896/v1/cbef582d70b8ce8768bddad8.docx"},{"id":59516848,"identity":"8f4663f6-1481-4954-ae93-bb5312361129","added_by":"auto","created_at":"2024-07-02 17:42:01","extension":"docx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":3047070,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterial1.docx","url":"https://assets-eu.researchsquare.com/files/rs-4580896/v1/210a604e58f8942771d98ea6.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Identification and validation an anoikis-related gene signature for clinical diagnosis, prognosis and treatment of patients with hepatocellular carcinoma","fulltext":[{"header":"Introduction","content":"\u003cp\u003eHepatocellular carcinoma (HCC) remains the third most frequently diagnosed malignancy and is the second leading cause of cancer death worldwide due to its heterogeneity, drug resistance, high postoperative recurrence, and metastasis (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). China is a major liver malignancy country, where the incidence of new patients and deaths occupied about half of the global patients each year (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e). Studies show more than half HCC patients are diagnosed with advanced liver cancer and systemic therapy is the only treatment option that provides a survival benefit (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e). The FDA approved sorafenib in 2007 as the only systemic therapy available for advanced hepatocellular carcinoma. However, the overall survival of patients treated was less than 8 years (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e). It wasn't until 2017 that regorafenib was approved for second-line treatment of advanced HCC breaking the stalemate (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e). Recently, the FDA approved the marketing of the first dual immune combination therapy for the first-line treatment of patients with unresectable hepatocellular carcinoma, bringing new options and hope (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e). However, the median survival of targeted combination immunotherapy rarely exceeds 2 years, and some patients may develop a certain degree of drug resistance, which seriously affects the treatment effect (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). Despite great advancements in diagnosis, screening, and treatments, clinical outcomes for patients with HCC were still unsatisfactory due to its poor five survival status (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e). Hence, it is of paramount value to precisely individualize the treatment of patients and to understand the biological mechanisms behind it.\u003c/p\u003e \u003cp\u003eCell death is a vital phenomenon and an irreversible life process of cells. It frequently occurs in normal tissues and is necessary for maintaining cell function and morphology. Cell death can be broadly classified into active death-programmed death, apoptosis and cell necrosis (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e). In 1994, a particular shape of programmed apoptosis induced through separating cells from the extracellular matrix (ECM) and other cells was first named anoikis (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e). Anoikis plays an instrumental role in organism development, tissue self-homeostasis, disease genesis, and tumor metastasis, among others (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e). When cells lose their attachment to the ECM, the pro-apoptotic protein BMF, which is tethered to the cytoskeleton, dissociates from the kinesin light chain two and translocates to the mitochondria promoting the anoikis (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e). Resistance to anoikis is a feature of tumor metastasis, which enables tumor cells to spread to other distant organs through the circulatory system. After the tumor cells are separated from the extracellular matrix adhesion and intercellular contact, they can survive through paracrine autocrine and paracrine mechanisms to resist apoptosis and regain the ability to adhere to spread, metastasis, and invasion (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e). The doctrine of anoikis has proposed new explanatory methods and research concepts for the metastasis of tumors and provided new avenues for treating malignant tumors.\u003c/p\u003e \u003cp\u003eThe tumor microenvironment (TME) is a mixture of immune cells, stromal cells and blood vessels that serves as the \"soil\" for tumor growth. Different cells in the tumor microenvironment have other functions, which determine the extreme heterogeneity of tumors (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). Besides, the interdependent and antagonistic relationship between tumor cells and TME can determine the process and outcome of malignant tumors. Few studies have indicated that TME is intimately linked to anoikis (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e). With the clinical success of tumor immunotherapy, especially Car-T cell immunotherapy technology and immune node therapy, it is crucial to study the mechanism of tumor microenvironment regulation on immune cell function in depth with fundamental research significance (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e). The above shreds of evidence suggested that anoikis played a non-negligible role in the tumor microenvironment. In this context, we aimed to focus on anoikis-related genes to analyze the mode of anoikis in liver cancer and the potential molecular mechanism and drug response of apoptosis in different modes of regulation, which is significant for exploring effective therapies for human malignancies.\u003c/p\u003e"},{"header":"Material and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eData processing\u003c/h2\u003e \u003cp\u003eWe extracted the transcriptome expression information of liver hepatocellular carcinoma (LIHC) with affiliated clinical information from The Cancer Genome Atlas(TCGA; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://portal.gdc.cancer.gov\u003c/span\u003e\u003cspan address=\"https://portal.gdc.cancer.gov\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) and the gene expression omnibus (GEO) database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.ncbi.nlm.nih.gov/geo\u003c/span\u003e\u003cspan address=\"http://www.ncbi.nlm.nih.gov/geo\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). To obtain accurate results, we excluded HCC patients with a survival time of fewer than 30 days, and finally, 342 tumor samples were chosen in the study from the TCGA database and 221 tumor samples from the GSE14520-GPL3971. In addition,230 HCC samples from the International Cancer Genome Consortium (ICGC) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.icgc.gov/\u003c/span\u003e\u003cspan address=\"https://www.icgc.gov/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) were also included (Supplement Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). Copy number variation (CNV) and somatic mutation data of HCC patients in TCGA database were derived from the UCSC Xena database((\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://xena.ucsc.edu/\u003c/span\u003e\u003cspan address=\"https://xena.ucsc.edu/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), and the chromosome location altered by CNV in the anoikis-related genes was plotted using R software of the \"Rcircos\u0026rdquo;\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eIdentification of prognostic anoikis-related differentially expressed genes (PARDEGs)\u003c/h2\u003e \u003cp\u003eA total amount of 794 anoikis-related genes were obtained from the GeneCards (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.genecards.org/Search/Keyword?queryString=anoikis\u003c/span\u003e\u003cspan address=\"https://www.genecards.org/Search/Keyword?queryString=anoikis\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) and the Molecular signatures database (MSigDB), and genes with a correlation score\u0026thinsp;\u0026gt;\u0026thinsp;0.4 were picked out for follow-up analysis (Supplement Table S2). To find ARDEGs, we first obtain DEGs between HCC patients and standard samples in the TCGA-LIHC cohort using the R software of \u0026ldquo;limma\u0026rdquo; with false discovery rate (FDR)\u0026thinsp;\u0026lt;\u0026thinsp;0.01 and |log2FC|\u0026gt;1, then the weighted correlation network analysis (WGCNA) method was employed to find out gene modules highly correlation with tumor process; finally, the overlapping genes of the three common modules were defined as anoikis-related differentially expressed genes. At last, univariate Cox regression (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01) was adopted to identify PARDEGs.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eNMF, GSVA, ssGSEA and PCA\u003c/h2\u003e \u003cp\u003eNonnegative matrix factorization (NMF) is widely used in disease typing by extracting bio-correlation coefficients of data within the gene expression matrix, organizing genes and samples, and capturing the internal structural features of the data to group samples and belongs to a kind of unsupervised learning which is similar to principal component analysis (PCA) (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e). The PCA was employed to distinguish dimensions between different subtypes better. Gene Set Variation Analysis with the gene set (GSVA) (c2. cp. kegg. v7.5.1) was employed to inquiry the biological mechanisms behind each HCC subtype. The infiltration of 23 immune cells and their distribution in different HCC clusters were detected through the single sample gene set enrichment analysis (ssGSEA).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eEstablishment of prognostic anoikis-related differentially expressed genes Signature\u003c/h2\u003e \u003cp\u003eWe build the risk model relying on the expression of 40 prognostic anoikis-related differentially expressed genes. First, the least absolute shrinkage and selection operator (LASSO) was employed to further narrow the volume of predictive anoikis-related genes. Then, We performed a linear combination of regression coefficients (β) and expression levels of prognostic anoikis-related genes in the multivariate Cox regression model to ascertain the risk score. The calculation equation was described below: Risk scores=\u0026sum;Coef(i)\u0026lowast;Exp(i).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eConfirmation of the protein and mRNA level of the four prognostic anoikis-related genes\u003c/h2\u003e \u003cp\u003eThe Human Protein Atlas database (HPA, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.proteinatlas.org/\u003c/span\u003e\u003cspan address=\"https://www.proteinatlas.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) was applied to quantify the protein expression in HCC samples and standard samples. A total of eight HCC tumor samples with their paratumor tissues were collected from Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine. Meanwhile, the experiment was conducted with the permission of the ethics committee of Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine (No.2018-630-59-01).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eQuantitative real-time PCR analysis (qRT-PCR)\u003c/h2\u003e \u003cp\u003eHuman HCC cell lines (HepG2) and normal hepatocytes(L02) were acquired from the Cell Bank of the Chinese Academy of Sciences, Shanghai, China and clustered as described before. A total mRNA was isolated from eight HCC tumor samples with their paratumor tissues and cell lines using TRIzol reagent (Bioteke Corporation, RP40002), and synthesized into cDNA with a Reverse Transcription Master kit (Vazyme, R222-01). QRT-PCR was performed by using ChamQ SYBR qPCR Master Mix (Vazyme, Q311-02). The expression in adjusted tissues normalized the mRNA levels. The sequences of primer used in this study are displayed in Supplement Table S3.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eFurther verification of the risk model\u003c/h2\u003e \u003cp\u003eWe classified HCC patients in two independent databases (the ICGC and GSE14520 datasets) into two risk groups based on the risk score. To further validate the internal reliability of the model, we divided HCC patients in the TCGA database into the training set and the test set in a 7:3 ratio. The Kaplan-Meier analysis, receiver operating characteristic curves (ROC), and Decision curve analysis (DCA) were used to analyze the reliability of the risk model.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eDrug sensitivity and immunotherapy\u003c/h2\u003e \u003cp\u003eThe R software of pRRophetic was employed to count the half inhibitory concentrations of anti-tumor medicines frequently used to treat HCC concentration (IC50) values. T cell exclusion, T cell dysregulation, MSI (microsatellite instability), and tumor immune dysfunction and exclusion (TIDE) scores were obtained from the TIDE online database (TIDE; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://tide.dfci.harvard.edu/\u003c/span\u003e\u003cspan address=\"http://tide.dfci.harvard.edu/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) to assess the efficacy of immunotherapy in two different risk HCC patients. We also obtained the Immunophenoscore (IPS) of HCC patients in The Cancer Immunome Atlas (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://tcia.at/\u003c/span\u003e\u003cspan address=\"https://tcia.at/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eWe utilize R software (version 4.1.2) and its accompanying packages to do the statistical analysis. Unless specifically mentioned, p\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was designed as a statistically significant difference.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eDetermination of anoikis-related differentially expressed genes (ARDEGs)\u003c/h2\u003e \u003cp\u003eThe transcriptome expression data of the TCGA-LIHC was compared by the R software of \u0026ldquo;limma\u0026rdquo; with |log2FC|\u0026gt;1 and FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.05. A total of 5760 DEGs were screened from 50 standard samples and 375 HCC patients, among which 318 were downregulated, and 5442 were upregulated (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA, E). Next, \u0026ldquo;WGCNA\" data packets were used to understand better which genes were significantly involved in the development of liver malignancies. We take the first β that reaches 0.9 as the best β to construct gene modules, and finally, the above condition is satisfied when β\u0026thinsp;=\u0026thinsp;10 (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB), so ten gene modules were thus formed (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC). As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC and\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD, genes in cyan color had the highest correlation in HCC patients(cor\u0026thinsp;=\u0026thinsp;0.53). Therefore, we defined the cyan module as the hub module for subsequent analysis. The \u0026ldquo;Venn\u0026rdquo; package was used to overlay the DEGs in the TCGA-LIHC, cyan module, and anoikis-related genes to find the hub anoikis-related genes most closely associated with tumor development. Finally, 56 genes were selected as anoikis-related differentially expressed genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eE). Further, Spearman\u0026rsquo;s analyze was performed to have a better recognition of the role of anoikis-related genes in HCC(Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eF-G). In addition, the PCA showed that 56 anoikis-related genes could be well distinguished from tumors and standard samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eH).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eThe landscape of genetic variation of prognostic anoikis-related differentially expressed genes (PARDEGs)\u003c/h2\u003e \u003cp\u003eWe used univariate Cox regression(\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01) to identify PARDEGs, and finally, 40 PARGs were included (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). The omnidirectional regulatory network revealed the connections' landscape, correlations of 40 PARDEGs, and their prognostic values, and we found that all of them were strongly associated with a poor prognosis for HCC(Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB). Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC depicts that only 48 of the 364 samples were mutated, a mutation rate of 13.19%, and missense mutation was the most common kind of mutation, which indicated that PARGs might be stable in HCC samples. Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD shows the position of the copy number variation (CNV) alterations of PARDEGs on chromosomes. As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eE, the CNV amplification frequency was generally different among the PARGs, most of them were observed as the gain in the copy number, and DAP3 performed as the highest CNV amplification, while SFN was the highest deletion frequency gene. In addition, the expression of 40 PARDEGs between HCC samples and standard samples was displayed in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eF-G, and we found all of them were overexpressed in tumor tissues compared to normal ones, suggesting the expression of PARDEGs may have a heavy link to the prognosis of HCC.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eConstruction of clusters and biological functions mediated by PARDEGs\u003c/h2\u003e \u003cp\u003eTo further clarify the expression of PARDEGs, consensus clustering of the NMF algorithm was employed to divide HCC clusters depending on the expression of PARDEGs. At last, k\u0026thinsp;=\u0026thinsp;4 was selected to minimize distractions between different clusters. Thus, four HCC subtypes: clusterC1(\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;55), clusterC2 (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;42), clusterC3(\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;76) and clusterC4(\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;169) were identified. The consensus map showed a nonnegative matrix factorization classification (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA). Notably, the KM curve analysis showed cluster C4 had the longest overall survival (OS), disease-free survival (DFS), disease-specific survival (DSS), and progression-free survival (PFS), while cluster C2 had the shortest (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB-E). To further verify the accuracy of the grouping, we used the same approach to cluster the HCC patients in the TCGC dataset and found the exact similar result as in the TCGA database (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eF-G). As illustrated in the heatmap (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eH), the expression of PARDEGs in cluster C2 was obviously highest among cluster C1, cluster C3, and cluster C4; not surprisingly, patients in cluster C2 had a pronounced linked to advanced clinical characteristics among those in cluster C1, cluster C3, and cluster C4. Since the prognosis based on the expression of PARDEGs differed between these four HCC subtypes, we performed the GSVA to understand better the bio-behavior regulated by PARDEGs and found that metabolic pathways like fatty acid metabolism and bile acid metabolism were significantly enrichment in cluster C4 (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eI). This potentially explains well the different prognoses among the four HCC subtypes. Then We experimentally analyzed immune cell infiltrations in four HCC clusters by ssGSEA and found significant differences between the tumor-infiltrating immune cells of the four HCC subtypes (Supplementary Fig.\u0026nbsp;1). These pieces, as mentioned earlier, of evidence suggest a significant difference between the biological processes and tumor immune microenvironment of four HCC subtypes mediated by PARDEGs.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eBuilding the risk model base on the PARDEGs\u003c/h2\u003e \u003cp\u003eBased on LASSO analysis and multivariate Cox ratio hazard regression analysis, four PARDEGs (E2F1, SKP2, NDRG1, and CD24) were ultimately screened to set the risk model (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA-B). The equation for calculating the risk score was shown below: Risk score=(E2F1\u0026times;0.0172464245772007\u0026thinsp;+\u0026thinsp;SKP2\u0026times;0.0507068315282839\u0026thinsp;+\u0026thinsp;NDRG1\u0026times;0.00920961642769134\u0026thinsp;+\u0026thinsp;CD24\u0026times;0.0038494255274358) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). We also withdrew the expression data of four PARDEGs in the ICGC and GSE14520 databases and used the same algorithm to categorize patients into high-risk and low-risk groups. The risk profile, survival status, and heatmap of the expression of the four genes in high- and low-risk patients were presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD-F. Kaplan-Meier survival analysis showed that high-risk patients in these three databases had much worse survival condition than low-risk patients (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eG-I). moreover, the TCGA-LIHC cohort with 342 HCC patients was irregularly assigned into a training cohort with 206 patients and a testing cohort with 136 in a 7:3 ratio. Survival and the receiver operating characteristic (ROC) curves demonstrated that the predictive risk model had excellent intra-group stability and potential for development (Supplementary Fig.\u0026nbsp;2A-D). Moreover, KM survival analysis suggested that higher risk scores were heavily associated with worse survival indicators survival indicators such as DFS, DSS and PFS of HCC patients (Supplement Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eE-G).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eValidation and biological analysis of four PARDEGs\u003c/h2\u003e \u003cp\u003eTo validate the precision of the model, first, we performed an analysis of the relationship between the expression level of these four genes and survival time and found that the higher the gene expression, the shorter the survival time of HCC patients, which suggested that these four genes were substantially related to poor prognosis of HCC(Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA-D). We then analyzed the immunohistochemical staining of these four genes between HCC patients and healthy individuals through the HPA online database. The results showed that the protein level of four genes were noticeably enhanced in tumor tissues (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eE-H). In addition, Expression of four PARDEGs were validated in GEPIA(Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA), GSE14520(Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB), GSE76427(Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eC) and GSE54236(Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eD). Finally, we detected the levels of mRNA in four PARDEGs in human tumor tissues (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eE-H) and human HCC cell lines through qRT-PCR experiments and found the same results. To explore the biological significance behind this, we used GSVA analysis and found that these four genes were closely associated with numerous tumor pathways. Notably, all genes were negatively associated with the PPAR signaling pathway, which highlights the critical role of the PPAR signaling pathway in antitumor and provides some perspectives for future tumor therapy (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eI).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eFunctional enrichment analysis of the risk model\u003c/h2\u003e \u003cp\u003eA total of 968 DEGs were picked out by the R software of \u0026ldquo;limma\u0026rdquo; with false discovery rate (FDR)\u0026thinsp;=\u0026thinsp;0.05 and logFCfilter\u0026thinsp;=\u0026thinsp;1, of which 157 genes were down-regulated and 811 genes were upregulated in the high-risk HCC patients. Gene Ontology (GO) enrichment profiling showed that organelle fission and DNA replication pathways were enriched in DEGs (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eA-B). Kyoto Encyclopedia of Genes and Genomes (KEGG) profiling indicated that the Cell cycle and MicroRNAs in cancer were enhanced in the high-risk group (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eC). Besides, gene sets enrichment analysis (GSEA) showed that meiotic pathways were enriched in the high-risk group (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eD, F), while fatty acid pathways were upregulated in the low-risk group (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eE, G).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003eClinical stratification and clinical correlation analysis based on the model\u003c/h2\u003e \u003cp\u003eConsidering the heterogeneity of HCC, we classified HCC patients into 18 subtypes according to their pervasive clinicopathology characteristics. The results showed that the prognostic model we constructed can distinguish the clinical information very well (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e). In addition, we investigated the relationship between the risk score and clinical characteristics and found that male patients aged\u0026thinsp;\u0026le;\u0026thinsp;65 had a significantly increased risk of HCC (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eA, F). Furthermore, clinical characteristics such as tumor history, tumor type, and HCC clusters were also studied in our research (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eB-E). Our findings also found that risk score was strongly associated with higher tumor grade, stage, and a worse survival time (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eG-K). We also present the clinical features, immune infiltration, and prognostic gene expression differences between HCC patients expressed through a heatmap (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eM). To validate the general applicability of the model, we analyzed the clinical characteristics of HCC patients in both the ICGC and GSE14520 databases (Supplementary Fig.\u0026nbsp;3 and Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Finally, we compare our constructed model with the existing published models (\u003cspan additionalcitationids=\"CR19 CR20\" citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e) and find that our constructed model has better stability and prediction ability (Supplementary Figure Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003eEstablishment of the nomogram clinical characteristics in three databases\u003c/h2\u003e \u003cp\u003eFor the sake of precision treatment, we developed a nomogram that includes age, gender, tumor grade, stage, tumor history, tumor type, tumor vascular, and risk score in the TCGA(Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eA). The calibration plot for 1, 3, and 5 years show that we have developed nomination charts with low inaccuracies in three databases (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eB, E, H). The ROC curves show that the predictive performance of our constructed nomograms is much higher than that of clinical features such as gender, age, class, and stage (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eC, F, I). ROC curves and AUC values are generally used to appraise the accuracy of prognostic models, but ROC curves only consider the specificity and sensitivity of the model and pursue accuracy, while Decision curve analysis (DCA) because it compensates for the overfitting of ROC curves, is increasingly used in clinical analysis (\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e). We could find that the predictive power of our constructed column line graphs and risk models was much higher than other clinical indicators (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eD, G, J). This result was also extended to the ICGC (Supplementary Fig.\u0026nbsp;6) and GSE14520 dataset (Supplementary Fig.\u0026nbsp;7). Thus, three independent databases validated the general applicability of the accuracy of the risk model.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003eImmune landscape between two different risk groups\u003c/h2\u003e \u003cp\u003eFigure 11A shows the immune cell bubble plots derived from different platform algorithms. We then explore the relationship between the immune cells and the risk score. As shown in Fig.\u0026nbsp;11B, we found the risk score was positively corrected with neutrophil, B cells, APC co-inhibition, and Tumor Infiltrating Lymphocytes (TIL), while negatively corrected with activated dendritic cells (aDCs), cytolytic activity, Mast cells, NK cells, and Type II IFN response. Immunogenomic analysis of more than 1,000 tumor samples from 33 TCGA cancer types has been studied, identifying six immune subtypes, including wound healing(C1), IFN-dominant(C2), inflammatory(C3), lymphocyte depletion(C4), immune resting(C5), and TGF-β dominant(C6). We found that the immune types in the high-risk group were mainly C1 and C2, while C3 was primarily present in the low-risk group (Fig.\u0026nbsp;11C). In addition, the relationship between microsatellite instability (MSI) and risk scores were also within the scope of our study (Fig.\u0026nbsp;11D). The RNA-based stemness scores (RNAss) were used to measure the correlation of the risk score with tumor stem cells, and found that the higher the risk score, the stronger the degree of hepatocytes and the lower the degree of differentiation (R\u0026thinsp;=\u0026thinsp;0.31, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;11E). The ESTIMATE algorithm was used to analyze the immune score, the stromal score, and the ESTIMATE score between the low-and high-risk group. The results showed significant differences in stromal score and ESTIMATE score, while no differences in the immune score (Fig.\u0026nbsp;11F). Interestingly, the TIDE score and T cell dysfunction were lower in high-risk HCC patients than in the low-risk HCC patients, representing that high-risk HCC patients would benefit more from immunotherapy than those in the low-risk group; however, the T cell exclusion was higher in the high-risk group than in the low-risk group(Fig.\u0026nbsp;11G).These above results may also be the cause of the different prognoses of high- and low-risk patients, which may also provide guidance for immunotherapy in HCC patients.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003eTumor mutational Burden and Risk-Related Drug-Sensitivity Prediction\u003c/h2\u003e \u003cp\u003eAn oncoplot showed the topmost 20 somatic mutations in two risk groups in the TCGA-LIHC cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e12\u003c/span\u003eA-B). The outcome revealed mutations in the TP53 gene were significantly associated with the prognosis of HCC patients. As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e12\u003c/span\u003eC-D, the survival situation of HCC patients with higher TMB and risk scores was extremely terrible. Given the irreplaceable role of sorafenib and TACE in the treatment of HCC, we then analyzed the expression of four PARDEGs with different responses to Sorafenib treatment from GSE109211 and found that the expression of four PARDEGs were increased in non-responder patients than responder patients (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e12\u003c/span\u003eE). In additionally, risk scores were significantly higher in non-responders than responders to Sorafenib treatment (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e12\u003c/span\u003eF) (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). The same results were observed in transcatheter arterial chemoembolization (TACE) treatment (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e12\u003c/span\u003eG, H), which suggested that Sorafenib and TACE treatments can be of great benefits to lower risk score patients. We also listed the effects of common antineoplastic drugs in two risk groups (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e13\u003c/span\u003eA-C). In light of the importance of immunotherapy in liver cancer therapy, the potential benefit of immunotherapy with PD1 and CTLA4 for HCC patients based on TCGA expression data was calculated via the TCIA online website. The result showed the low-risk group outperformed the higher-risk group when treated with PD1 or CTLA4 (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e13\u003c/span\u003eD-G). Compared to the low-risk group, most immune checkpoints\u0026rsquo; expressions of the high-risk group were higher, which suggested that patients in the low-risk group may be better prospects for immunotherapy (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e13\u003c/span\u003eH).\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eLiver malignancies remain a global health threat, with several incidences expected to exceed one million cases in 2025 (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e). According to epidemiological studies, the annual incidence of HCC in China is 0.03%, which is much higher than that in Japan, Europe, and United States (\u0026thinsp;\u0026lt;\u0026thinsp;=\u0026thinsp;0.01%) (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e). Low rates of early diagnosis and 5-year survival, and high rates of surgical recurrence and metastasis are the leading causes of death in HCC patients (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e). According to the study, the diagnosis rate of BCLC stage 0-A patients in China is only 33%, and the 5-year survival rate for patients with progressive HCC is only 11.7%-14.1%, while the diagnosis rate of BCLC stage 0-A in Japan and Taiwan can reach 70% (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e). Therefore, there is an urgent need to further improve the early screening and diagnosis of HCC in China.\u003c/p\u003e \u003cp\u003eAnoikis is necessary for the organism's defense, which prevents the shed cells from implanting and growing into untimely locations (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e). However, tumor cells, especially some malignant cells prone to distal metastasis, have extreme resistance to anoikis and, thus, complete metastasis (\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e). This suggests that resistance to anoikis may be a prerequisite for the survival of cancer cells during the cycle (\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e). Therefore, activation of anoikis can inhibit tumor metastasis (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e). However, research on the progress of anoikis in HCC is still at an embryonic stage. With the rapid development of artificial intelligence, various mathematical algorithms are also widely used in the massive liver cancer data, providing a more convenient method for diagnosing and analyzing liver cancer (\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e502 anoikis-related genes were included in our study, and by combining multiple algorithms, we finally obtained 56 anoikis-related differentially expressed genes. Then 40 prognostic anoikis-related genes were screened through univariate Cox regression analysis. Based on the expression of these 40 predictive anoikis-related genes, 342 patients in TCGA-LIHC dataset with hepatocellular carcinoma were assigned into 4 HCC subtypes with different overall survival times through the NMF algorithm. To verify the accuracy of the classification, HCC patients in the ICGC dataset were also classified into four types using the same methodology. GSVA analysis showed that patients in clusterC4 were prominently enriched in fatty acid metabolism and bile acid metabolism; in addition, immune cell infiltration analysis showed that ClusterC2 was lower in T cells CD4 memory resting, T cells CD4 memory activated and T cells follicular helper while highest in eosinophils compared to other HCC subtypes. This could explain why ClusterC4 had the best prognosis while clutserC2 had the worst.\u003c/p\u003e \u003cp\u003eNext, we found 4 of the 40 prognostic anoikis-related genes by LASSO analysis to construct a predictive model, which was CD24, SKP2, NDRG1, and E2F1. E2F1 is essential in preventing tumorigenesis by strictly controlling the cell cycle, maintaining genomic integrity, and responding to replication stress and DNA damage (\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e). Previous studies have found that overexpression of E2F1 has been shown to promote cancers in various organs (\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e). Peng et al. (\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e) reported that CAMK2N1 inhibits the proliferation, invasion, and migration of HCC cells by suppressing E2F1-mediated cell cycle signaling. Lee et al. (\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e) found that SKP2 expression was significantly upregulated in HCC patients and, more importantly, that overexpression of SKP2 was significantly associated with poor prognosis and advanced clinical features. Yang et al. (\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e) found that CD24 was highly expressed in hepatic tumor cell lines and in patients with hepatocellular carcinoma and that overexpression of CD24 was associated with a high invasive and metastatic potential of HCC, suggesting that overexpression of CD24 is an independent risk factor for poor prognosis in patients with hepatocellular carcinoma. NDRG1 has been included in numerous pieces of literature as one of the crucial genes in the prognostic model of HCC (\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e). More importantly, we verified the protein level expression through the HPA public database. On the other hand, we validated our findings at the level of tumor tissues by qRT-PCR. Survival analysis and AUC values showed that the risk score can significantly differentiate between favorable and adverse HCC patients.\u003c/p\u003e \u003cp\u003eTo further explore its biological significance, we performed GSVA and GSEA analysis and found that fatty acid metabolism plays a pivotal role in tumorigenesis. In our study, we found that all four PARDEGs utilized to build the risk model were negatively correlated with the PPAR signaling pathway, which laterally verified the accuracy of our experiments. Abnormal lipid distribution is a prominent alteration in tumor patients, where abnormal lipid distribution helps moisten tumor cells and promotes tumor proliferation and infiltration. Therefore, inhibition of lipid transport in tumor cells has an invaluable role in minimizing tumor drug resistance and improving immune efficacy (\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e). In addition, tumor cells take advantage of lipid metabolism to regulate the state of stromal cells and the activity of immune cells to establish a tumor microenvironment that creates therapeutic resistance and stimulates cancer recurrence. (\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e). PPARα signaling pathway can improve lipid metabolism disorders and accelerate lipid differentiation by increasing fatty acid β-oxidation (\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e). These results suggested that developing anti-hepatocellular carcinoma chemotherapeutic agents targeting PPAR may be a breakthrough in treating liver cancer.\u003c/p\u003e \u003cp\u003eAFP remains a common and important indicator for diagnosis and efficacy testing of hepatocellular carcinoma, but lacks sensitivity and specificity (\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e). The sensitivity and specificity of the model constructed in this study were significantly higher than those of clinical indicators, which can significantly increase the detection rate of early liver cancer and reduce the death rate. It is clear the importance of patient stratification for improving HCC treatment outcomes, so K-M survival curves between different patients according to clinicopathological characteristics and risks were presented. The results showed that high risk scores exhibited a poor prognosis in patients with different clinical characteristics\u003c/p\u003e \u003cp\u003eSorafenib remains the drug of choice for advanced HCC patients, however, greater drug resistance poses a challenge for clinicians (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e). The combination of immunotherapy and molecularly targeted drugs have recently brought a new light to patients with advanced HCC (\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e). However, the highly heterogeneous nature of liver cancer shaped the complex tumor microenvironment and led to heterogeneity within and between tumor foci, thus limiting the implementation of precision medicine (\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e). Therefore, there was an urgent need for clinical practice to precisely analyze the molecular biomarkers of the tumor microenvironment (TME) occurring in HCC and to develop individualized treatment strategies (\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e). This study found that high-risk score patients benefited less from sorafenib and TACE treatment, while patients with high-risk score had significantly higher levels of immune checkpoint expression and lower TIDE scores, suggesting that immunotherapy may benefit patients in high-risk score. In addition, we compared several drugs commonly used in HCC chemotherapy and found that the patients in low-risk score had better efficacy than patients in high-risk score when treated with sorafenib, erlotinib, and sunitinib antitumor.\u003c/p\u003e \u003cp\u003eTo the best of our knowledge, this was the first prognostic model constructed based on the anoikis-related genes with tumor mutational load, tumor microenvironment, immunotherapy, and drug sensitivity, which provided some support for precision and individualized treatment of liver cancer and provided an alternative way for clinicians to screen for early-stage HCC.\u003c/p\u003e \u003cp\u003eHowever, there are some unavoidable flaws in our study. First, underlying studies on how anoikis-related genes affected HCC were not explored further in depth in our research. In addition, the study's findings need to be further explored in terms of which countries only are appropriate for HCC patients. More importantly, although these four prognostic genes were validated in human tissues, the sample size was small, and there was a lack of real-world HCC cohort validation, so the results of this study have a long way to go.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn a nutshell, the model based on anoikis-related genes provides new insights and perspectives for researchers to explore tumor mechanisms and new drug development by assessing the prognosis of HCC patients and guiding clinicians in drug treatment.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHaochen Jiang and Tao Wang designed this study.Tao Wang and Suyin Li drafted the manuscript. Xuehua Sun, Xiangxue Pan, and Weifeng Tan conducted the experiment.\u0026nbsp;All authors participated in the review of the manuscript and approved the submitted version.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe experimental protocol was allowed by the ethics committee of Shuguang Hospital, Shanghai University of Traditional Chinese Medicine\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFundings\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;This study was supported by the National Natural Science Foundation of China (No.82074336 to X Sun) and the Program of Shanghai 2020 Science and Technology Innovation Action Plan (No.20S21901600 to X Sun).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors would like to express gratitude to all those who quietly helped create the public database.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors declare no conflict of interest.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data in the manuscript is free to download online. The related data were downloaded from the GEO (https://www.ncbi.nlm.nih.gov/geo/), TCGA (https://dcc.icgc.org/), ICGC (https://dcc.icgc.org/projects-/LIRI-JP), HPA (https://www.proteinatlas.org/) and TIDE(http://tide.dfci.harvard.edu/).\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eRumgay H, Arnold M, Ferlay J, Lesi O, Cabasag CJ, Vignat J, et al. Global burden of primary liver cancer in 2020 and predictions to 2040. J Hepatol. 2022;77(6):1598\u0026ndash;606. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jhep.2022.08.021\u003c/span\u003e\u003cspan address=\"10.1016/j.jhep.2022.08.021\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang J, Pan G, Guan L, Liu Z, Wu Y, Liu Z, et al. The burden of primary liver cancer caused by specific etiologies from 1990 to 2019 at the global, regional, and national levels. Cancer Med. 2022;11(5):1357\u0026ndash;70. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/cam4.4530\u003c/span\u003e\u003cspan address=\"10.1002/cam4.4530\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSperandio RC, Pestana RC, Miyamura BV, Kaseb AO. Hepatocellular Carcinoma Immunotherapy. Annu Rev Med. 2022;73:267\u0026ndash;78. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1146/annurev-med-042220-021121\u003c/span\u003e\u003cspan address=\"10.1146/annurev-med-042220-021121\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFinn RS, Qin S, Ikeda M, Galle PR, Ducreux M, Kim TY, et al. Atezolizumab plus Bevacizumab in Unresectable Hepatocellular Carcinoma. N Engl J Med. 2020;380(20):1894\u0026ndash;905. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1056/NEJMoa1915745\u003c/span\u003e\u003cspan address=\"10.1056/NEJMoa1915745\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLlovet JM, Castet F, Heikenwalder M, Maini MK, Mazzaferro V, Pinato DJ, et al. Immunotherapies for hepatocellular carcinoma. Nat Rev Clin Oncol. 2022;19(3):151\u0026ndash;72. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41571-021-00573-2\u003c/span\u003e\u003cspan address=\"10.1038/s41571-021-00573-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang C, Zhang H, Zhang L, Zhu AX, Bernards R, Qin W, et al. Evolving therapeutic landscape of advanced hepatocellular carcinoma. Nat Rev Gastroenterol Hepatol. 2022. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41575-022-00704-9\u003c/span\u003e\u003cspan address=\"10.1038/s41575-022-00704-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCheng AL, Hsu C, Chan SL, Choo SP, Kudo M. Challenges of combination therapy with immune checkpoint inhibitors for hepatocellular carcinoma. J Hepatol. 2020;72(2):307\u0026ndash;19. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jhep.2019.09.025\u003c/span\u003e\u003cspan address=\"10.1016/j.jhep.2019.09.025\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLlovet JM, Kelley RK, Villanueva A, Singal AG, Pikarsky E, Roayaie S, et al. Hepatocellular carcinoma. Nat Rev Dis Primers. 2021;7(1):6. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41572-020-00240-3\u003c/span\u003e\u003cspan address=\"10.1038/s41572-020-00240-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMagnani L, Colantuoni M, Mortellaro A, Gasdermins. New Therapeutic Targets in Host Defense, Inflammatory Diseases, and Cancer. Front Immunol. 2022;13:898298. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fimmu.2022.898298\u003c/span\u003e\u003cspan address=\"10.3389/fimmu.2022.898298\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFrisch SM, Francis H. Disruption of epithelial cell-matrix interactions induces apoptosis. J Cell Biol. 1994;124(4):619\u0026ndash;26. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1083/jcb.124.4.619\u003c/span\u003e\u003cspan address=\"10.1083/jcb.124.4.619\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMason JA, Hagel KR, Hawk MA, Schafer ZT. Metabolism during ECM Detachment: Achilles Heel of Cancer Cells? \u003cem\u003eTrends cancer\u003c/em\u003e (2017). 3(7):475\u0026ndash;481. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.trecan.2017.04.009\u003c/span\u003e\u003cspan address=\"10.1016/j.trecan.2017.04.009\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDelgado M, Tesfaigzi Y. Is BMF central for anoikis and autophagy? Autophagy. 2014;10(1):168\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.4161/auto.26759\u003c/span\u003e\u003cspan address=\"10.4161/auto.26759\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang J, Luo Z, Lin L, Sui X, Yu L, Xu C, et al. Anoikis-Associated Lung Cancer Metastasis: Mech Ther Cancers(Basel). 2022;14(19):4791. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/cancers14194791\u003c/span\u003e\u003cspan address=\"10.3390/cancers14194791\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi F, Kitajima S, Kohno S, Yoshida A, Tange S, Sasaki S, et al. Retinoblastoma Inactivation Induces a Protumoral Microenvironment via Enhanced CCL2 Secretion. Cancer Res. 2019;79(15):3903\u0026ndash;15. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1158/0008-5472.can-18-3604\u003c/span\u003e\u003cspan address=\"10.1158/0008-5472.can-18-3604\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen J, Li K, Chen J, Wang X, Ling R, Cheng M, et al. Aberrant translation regulated by METTL1/WDR4-mediated tRNA N7-methylguanosine modification drives head and neck squamous cell carcinoma progression. Cancer Commun (Lond). 2022;42(3):223\u0026ndash;44. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/cac2.12273\u003c/span\u003e\u003cspan address=\"10.1002/cac2.12273\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWilson JL, N\u0026auml;gele T, Linke M, Demel F, Fritsch SD, Mayr HK, et al. Inverse Data-Driven Modeling and Multiomics Analysis Reveals Phgdh as a Metabolic Checkpoint of Macrophage Polarization and Proliferation. Cell Rep. 2020;30(5):1542\u0026ndash;e15527. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.celrep.2020.01.011\u003c/span\u003e\u003cspan address=\"10.1016/j.celrep.2020.01.011\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGay CM, Stewart CA, Park EM, Diao L, Groves SM, Heeke S, et al. Patterns of transcription factor programs and immune pathway activation define four major subtypes of SCLC with distinct therapeutic vulnerabilities. Cancer Cell. 2021;39(3):346\u0026ndash;e3607. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ccell.2020.12.014\u003c/span\u003e\u003cspan address=\"10.1016/j.ccell.2020.12.014\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDong W, Xie Y, Huang H. Prognostic Value of Cancer-Associated Fibroblast-Related Gene Signatures in Hepatocellular Carcinoma. Front Endocrinol (Lausanne). 2022;13:884777. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fendo.2022.884777\u003c/span\u003e\u003cspan address=\"10.3389/fendo.2022.884777\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLong S, Chen Y, Wang Y, Yao Y, Xiao S, Fu K. Identification of Ferroptosis-related molecular model and immune subtypes of hepatocellular carcinoma for individual therapy. Cancer Med. 2023;12(2):2134\u0026ndash;47. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/cam4.5032\u003c/span\u003e\u003cspan address=\"10.1002/cam4.5032\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRao G, Pan H, Sheng X, Liu J. Prognostic Value of Stem Cell Index-Related Characteristics in Primary Hepatocellular Carcinoma. Contrast Media Mol Imaging. 2022;2022:2672033. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1155/2022/2672033\u003c/span\u003e\u003cspan address=\"10.1155/2022/2672033\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRegmi P, He ZQ, Lia T, Paudyal A, Li FY. N7-Methylguanosine Genes Related Prognostic Biomarker in Hepatocellular Carcinoma. Front Genet. 2022;13:918983. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fgene.2022.918983\u003c/span\u003e\u003cspan address=\"10.3389/fgene.2022.918983\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang YQ, Liang CH, He L, Tian J, Liang CS, Chen X, et al. Development and Validation of a Radiomics Nomogram for Preoperative Prediction of Lymph Node Metastasis in Colorectal Cancer. J Clin Oncol. 2016;34(18):2157\u0026ndash;64. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1200/jco.2015.65.9128\u003c/span\u003e\u003cspan address=\"10.1200/jco.2015.65.9128\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang DQ, Mathurin P, Cortez-Pinto H, Loomba R. Global epidemiology of alcohol-associated cirrhosis and HCC: trends, projections, and risk factors. Nat Rev Gastroenterol Hepatol. 2022;18:1\u0026ndash;13. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41575-022-00688-6\u003c/span\u003e\u003cspan address=\"10.1038/s41575-022-00688-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin. 2021;71(3):209\u0026ndash;49. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3322/caac.21660\u003c/span\u003e\u003cspan address=\"10.3322/caac.21660\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDasgupta P, Henshaw C, Youlden DR, Clark PJ, Aitken JF, Baade PD. Global Trends in Incidence Rates of Primary Adult Liver Cancers: A Systematic Review and Meta-Analysis. Front Oncol. 2020;10:171. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fonc.2020.00171\u003c/span\u003e\u003cspan address=\"10.3389/fonc.2020.00171\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZeng H, Chen W, Zheng R, Zhang S, Ji JS, Zou X, et al. Changing cancer survival in China during 2003-15: a pooled analysis of 17 population-based cancer registries. Lancet Glob Health. 2018;6(5):e555\u0026ndash;67. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/s2214-109x\u003c/span\u003e\u003cspan address=\"10.1016/s2214-109x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. (18)30127-x.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGilmore AP, Anoikis. Cell Death Differ. 2005;2:1473\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/sj.cdd.4401723\u003c/span\u003e\u003cspan address=\"10.1038/sj.cdd.4401723\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJin L, Chun J, Pan C, Kumar A, Zhang G, Ha Y, et al. The PLAG1-GDH1 Axis Promotes Anoikis Resistance and Tumor Metastasis through CamKK2-AMPK Signaling in LKB1-Deficient Lung Cancer. Mol Cell. 2018;69(1):87\u0026ndash;e997. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.molcel.2017.11.025\u003c/span\u003e\u003cspan address=\"10.1016/j.molcel.2017.11.025\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKakavandi E, Shahbahrami R, Goudarzi H, Eslami G, Faghihloo E. Anoikis resistance and oncoviruses. J Cell Biochem. 2018;119(3):2484\u0026ndash;91. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/jcb.26363\u003c/span\u003e\u003cspan address=\"10.1002/jcb.26363\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDolinschek R, Hingerl J, Benge A, Zafira C, Sch\u0026uuml;ren E, Ehmoser EK, et al. Constitutive activation of integrin αvβ3 contributes to anoikis resistance of ovarian cancer cells. Mol Oncol. 2021;15(2):503\u0026ndash;22. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/1878-0261.12845\u003c/span\u003e\u003cspan address=\"10.1002/1878-0261.12845\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang X, Wang Z, Tang W, Wang X, Liu R, Bao H, et al. Ultrasensitive and affordable assay for early detection of primary liver cancer using plasma cell-free DNA fragmentomics. Hepatology. 2022;76(3):317\u0026ndash;29. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/hep.32308\u003c/span\u003e\u003cspan address=\"10.1002/hep.32308\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFouad S, Hauton D, D'Angiolella V. E2F1: Cause and Consequence of DNA Replication Stress. Front Mol Biosci. 2020;7:599332. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fmolb.2020.599332\u003c/span\u003e\u003cspan address=\"10.3389/fmolb.2020.599332\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eConner EA, Lemmer ER, Omori M, Wirth PJ, Factor VM, Thorgeirsson SS. Dual functions of E2F-1 in a transgenic mouse model of liver carcinogenesis. Oncogene. 2000;19(44):5054\u0026ndash;62. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/sj.onc.1203885\u003c/span\u003e\u003cspan address=\"10.1038/sj.onc.1203885\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZheng X, Huang M, Xing L, Yang R, Wang X, Jiang R, et al. The circRNA circSEPT9 mediated by E2F1 and EIF4A3 facilitates the carcinogenesis and development of triple-negative breast cancer. Mol Cancer. 2020;19(1):73. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12943-020-01183-9\u003c/span\u003e\u003cspan address=\"10.1186/s12943-020-01183-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePeng JM, Tseng RH, Shih TC, Hsieh SY. CAMK2N1 suppresses hepatoma growth through inhibiting E2F1-mediated cell-cycle signaling. Cancer Lett. 2021;497:66\u0026ndash;76. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.canlet.2020.10.017\u003c/span\u003e\u003cspan address=\"10.1016/j.canlet.2020.10.017\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee SW, Li CF, Jin G, Cai Z, Han F, Chan CH, et al. Skp2-dependent ubiquitination and activation of LKB1 is essential for cancer cell survival under energy stress. Mol Cell. 2015;57(6):1022\u0026ndash;33. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.molcel.2015.01.015\u003c/span\u003e\u003cspan address=\"10.1016/j.molcel.2015.01.015\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang XR, Xu Y, Yu B, Zhou J, Li JC, Qiu SJ, et al. CD24 is a novel predictor for poor prognosis of hepatocellular carcinoma after surgery. Clin Cancer Res. 2009;15(17):5518\u0026ndash;27. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1158/1078-0432.ccr-09-0151\u003c/span\u003e\u003cspan address=\"10.1158/1078-0432.ccr-09-0151\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZeng F, Zhang Y, Han X, Zeng M, Gao Y, Weng J. Employing hypoxia characterization to predict tumor immune microenvironment, treatment sensitivity and prognosis in hepatocellular carcinoma. Comput Struct Biotechnol J. 2021;19:2775\u0026ndash;89. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.csbj.2021.03.033\u003c/span\u003e\u003cspan address=\"10.1016/j.csbj.2021.03.033\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCheng J, Xie HY, Xu X, Wu J, Wei X, Su R, et al. NDRG1 as a biomarker for metastasis, recurrence and of poor prognosis in hepatocellular carcinoma. Cancer Lett. 2011;310(1):35\u0026ndash;45. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.canlet.2011.06.001\u003c/span\u003e\u003cspan address=\"10.1016/j.canlet.2011.06.001\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eR\u0026ouml;hrig F, Schulze A. The multifaceted roles of fatty acid synthesis in cancer. Nat Rev Cancer. 2016;16(11):732\u0026ndash;49. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/nrc.2016.89\u003c/span\u003e\u003cspan address=\"10.1038/nrc.2016.89\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHall Z, Chiarugi D, Charidemou E, Leslie J, Scott E, Pellegrinet L, et al. Lipid Remodeling in Hepatocyte Proliferation and Hepatocellular Carcinoma. Hepatology. 2021;73(3):1028\u0026ndash;44.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMenendez JA, Lupu R. Fatty acid synthase and the lipogenic phenotype in cancer pathogenesis. Nat Rev Cancer. 2007;7(10):763\u0026ndash;77. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/nrc2222\u003c/span\u003e\u003cspan address=\"10.1038/nrc2222\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi J, Huang Q, Long X, Zhang J, Huang X, Aa J, et al. CD147 reprograms fatty acid metabolism in hepatocellular carcinoma cells through Akt/mTOR/SREBP1c and P38/PPARα pathways. J Hepatol. 2015;63(6):1378\u0026ndash;89. 10.1016/j. jhep.2015.07.039.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eReig M, Forner A, Rimola J, Ferrer-F\u0026agrave;brega J, Burrel M, Garcia-Criado \u0026Aacute;, et al. BCLC strategy for prognosis prediction and treatment recommendation: The 2022 update. J Hepatol. 2022;76(3):681\u0026ndash;93. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jhep.2021.11.018\u003c/span\u003e\u003cspan address=\"10.1016/j.jhep.2021.11.018\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSu GL, Altayar O, O'Shea R, Shah R, Estfan B, Wenzell C. al.AGA Clinical Practice Guideline on Systemic Therapy for Hepatocellular Carcinoma. Gastroenterology. 2022;162(3):920\u0026ndash;34. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1053/j.gastro.2021.12.276\u003c/span\u003e\u003cspan address=\"10.1053/j.gastro.2021.12.276\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu F, Jin T, Zhu Y, Dai C. Immune checkpoint therapy in liver cancer. J Exp Clin Cancer Res. 2018;37(1):110. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s13046-018-0777-4\u003c/span\u003e\u003cspan address=\"10.1186/s13046-018-0777-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHiraoka A, Kumada T, Tada T, Hirooka M, Kuriyama K, Tani J, et al. Atezolizumab plus bevacizumab treatment for unresectable hepatocellular carcinoma: Early clinical experience. Cancer Rep (Hoboken). 2022;5(2):e1464. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/cnr2.1464\u003c/span\u003e\u003cspan address=\"10.1002/cnr2.1464\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144(5):646\u0026ndash;74. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cell.2011.02.013\u003c/span\u003e\u003cspan address=\"10.1016/j.cell.2011.02.013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Hepatocellular carcinoma, anoikis, signature, immunotherapy, tumor microenvironment","lastPublishedDoi":"10.21203/rs.3.rs-4580896/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4580896/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eHepatocellular carcinoma (HCC) is the most widespread malignancy in the universe, with low early diagnosis rates and high mortality. Therefore, early detection and treatment are critical to improving patients' life. Anoikis is one of the modes of cell death, and resistance to anoikis arising by aggressive tumor cells has been considered a pivotal element in cancer proliferation, while rarely have studies focused on the relationship between HCC and anoikis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAnoikis-related genes were gathered from the GeneCards and MSigDB, and the R software of \"limma” and the WGCNA were employed to select anoikis-related differentially expressed genes (ARDEGs). Patients from three independent cohorts (TCGA-LIHC, ICGC, and GSE14520) were classified by Nonnegative Matrix Factorization (NMF) to analyze the overall survival (OS), copy number variation (CNV), tumor microenvironment (TME), and biological characteristics of different HCC clusters. We then rely on the expression of prognostic anoikis-related differentially expressed genes (PARDEGs) to build the signature by the least absolute shrinkage and selection operator (LASSO) regression analysis, then patients were assigned into two risk groups. The study of enrichment pathways, immune microenvironment, clinicopathologic feature stratification, nomogram, tumor mutation burden (TMB), and drug prediction related to the signature was performed. More importantly, the mRNA level of the critical genes was verified at the HCC tissue level.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHCC patients were randomly segmented into four clusters based on the PARDEGs. The result showed that clusterC2 had the worst survival time and clinical performance. Four PARDEGs, including CD24, SKP2, E2F1, and NDRG1, were selected for conducting a risk model. This risk model was significantly validated by different datasets (TCGA-LIHC, ICGC, and GSE14520) to distinguish the survival status of other HCC patients. Analysis such as the receiver operating characteristic (ROC) analyses, concordance index(C-index), and nomogram indicated that the model had excellent sensitivity and specificity. Drug response and immunotherapy also manifested differently in two risk HCC patients.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA model constructed with four PARDEGs helps to improve the detection rate of early HCC, long-term prognostic stratification of HCC patients, and postoperative personalized monitoring and treatment plan development, reflecting the medical concept of early screening, early diagnosis, early and precise therapy of HCC.\u003c/p\u003e","manuscriptTitle":"Identification and validation an anoikis-related gene signature for clinical diagnosis, prognosis and treatment of patients with hepatocellular carcinoma","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-07-02 17:33:54","doi":"10.21203/rs.3.rs-4580896/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"74ccf853-7332-4440-aa33-6a8d6e1878b6","owner":[],"postedDate":"July 2nd, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-10-24T06:53:51+00:00","versionOfRecord":[],"versionCreatedAt":"2024-07-02 17:33:54","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4580896","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4580896","identity":"rs-4580896","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}