Combined Survival Analysis Research Articles

DeepSurv landmarking: a deep learning approach for dynamic survival analysis with longitudinal data

Risk prediction models quantify the probability of an individual experiencing specified events, there has significant attention to utilize longitudinal data during follow-up to dynamically update risk estimates. This paper devotes to a dynamic prediction model that combines survival analysis and deep learning techniques, which integrates the deep learning network-DeepSurv into the landmarking framework, to predict time-to-event outcomes with longitudinal covariates. Precisely, at each landmark time point, new information of the dataset is reconstructed, and the DeepSurv network is employed to capture the non-linear relationship between covariates and hazard rates, thereby give the dynamic prediction of survival functions. Furthermore, the importance of different features at each landmark time point is evaluated and ranked to estimate the impact of covariates on predictive outcomes over time. Performances on simulation experiments demonstrate that the proposed DeepSurv landmarking model outperforms state-of-art landmark methods, especially when the dependencies between covariates and the survival process are unknown. Additionally, the model exhibits remarkable dynamic prediction capabilities in the real heart-valve-transplantation-surgery data.

Volumetric and textural analysis of PET/CT in patients with diffuse large B-cell lymphoma highlights the importance of novel MTVrate feature.

To investigate the prognostic value of clinical, volumetric, and radiomics-based textural parameters in baseline [18F]FDG-PET/CT scans of diffuse large B-cell lymphoma (DLBCL) patients. We retrospectively investigated baseline PET/CT scans and collected clinical data of fifty DLBCL patients. PET images were segmented semiautomatically to determine metabolic tumor volume (MTV), then the largest segmented lymphoma volume of interest (VOI) was used to extract first-, second-, and high-order textural features. A novel value, MTVrate was introduced as the quotient of the largest lesion's volume and total body MTV. Receiver operating characteristics (ROC) analyses were performed and 24-months progression-free survival (PFS) of low- and high-risk cohorts were compared by log-rank analyses. A machine learning algorithm was used to build a prognostic model from the available clinical, volumetric, and textural data based on logistic regression. The area-under-the-curve (AUC) on ROC analysis was the highest of MTVrate at 0.74, followed by lactate-dehydrogenase, MTV, and skewness, with AUCs of 0.68, 0.63, and 0.55, respectively which parameters were also able to differentiate the PFS. A combined survival analysis including MTV and MTVrate identified a subgroup with particularly low PFS at 38%. In the machine learning-based model had an AUC of 0.83 and the highest relative importance was attributed to three textural features and both MTV and MTVrate as important predictors of PFS. Individual evaluation of different biomarkers yielded only limited prognostic data, whereas a machine learning-based combined analysis had higher effectivity. MTVrate had the highest prognostic ability on individual analysis and, combined with MTV, it identified a patient group with particularly poor prognosis.

Analysis of methylation-driven genes for predicting the prognosis of patients with oral squamous cell carcinoma.

Oral squamous cell carcinoma (OSCC) is a highly aggressive malignancy that is characterized by early distant metastasis and poor prognosis. DNA methylation plays an important role in the etiology and pathogenesis of OSCC. This study aimed to identify methylation-driven genes through bioinformatics analysis as potential biomarkers for early diagnosis and prognostic assessment of OSCC. Methylation data, RNA sequencing (RNA-seq) data and clinical prognosis information of OSCC patients were retrieved from The Cancer Genome Atlas (TCGA) database. The R packages MethylMix were employed to analyze the correlation between methylation status and corresponding gene expression in tumor and normal tissues to obtain methylation-driven genes. Univariate Cox regression analysis was developed to further screen methylation-driven genes associated with the prognosis of OSCC patients. Subsequently, multivariate Cox regression analysis was utilized to construct a linear prognostic risk prediction model. Furthermore, a combined survival analysis integrating methylation and gene expression was performed to investigate the prognostic value. A total of 374 differentially expressed methylation-driven genes were identified. Seven methylation-driven genes (BST2, KRT15, ZNF134, NT5E, GSTA7P, NAPRT, and GOLPH3L) were found to be significantly associated with patient prognosis. Additionally, four methylation-driven genes (BST2, KRT15, ZNF134 and NAPRT) were used to construct a linear prognostic risk prediction model for OSCC patients. Furthermore, a combined Kaplan-Meier survival analysis revealed that three methylation-driven genes (ZKSCAN7, MFF, ZNF134) alone can be used as independent prognostic markers or drug targets. Our findings facilitate a better understanding of molecular mechanisms of OSCC and provide potential biomarkers of early diagnosis, precision treatment and prognosis evaluation.

Pattern of Expression of MicroRNA in Patients with Radiation-Induced Bladder Injury

Introduction: Bladder cancer (BC) is sensitive to radiation treatment and a subset of patients experience radiation-induced injuries including shrinkage of bladder due to bladder fibrosis. Methods: This study is a retrospective cohort study. Three Japanese BC patients were randomly selected. Using a microRNA (miRNA) array, comparing their samples with or without radiation-induced injuries, we have checked the clustering of miRNA expression. Results: Hsa-miR-130a, hsa-miR-200c, hsa-miR-141, and hsa-miR-96 were found to be highly expressed (>50 times) in patients with fibrotic bladder shrinkage (FBS) compared to those with intact bladder (IB) function. In patients with FBS, hsa-miR-6835, hsa-miR-4675, hsa-miR-371a, and hsa-miR-6885 were detected to have lesser than half expression to IB patients. We have analyzed the significance of these genes in relation to overall survival of 409 BC patients retrieved from the Cancer Genome Atlas data set. All available cutoff values between the lower and upper quartiles of expression are used for the selected genes, and false discovery rate using the Benjamini-Hochberg method is computed to correct for multiple hypothesis testing. We have run combined survival analysis of the mean expression of these four miRNAs highly expressed in FBS patients. 175 patients with high expression had a longer median survival of 98.47 months than 23.73 months in 233 patients with low expression (hazard ratio [HR]: 0.53; 0.39–0.72, log-rank p value: 7.3e−0.5). Combination analysis of all 8 genes including hsa-miR-6835, hsa-miR-4675, hsa-miR-371a, and hsa-miR-6885 showed the same HR for OS. Target scanning for these miRNAs matched specific cytokines known as an early biomarker to develop radiation-induced fibrosis. Conclusions: BC patients with fibrotic radiation injury have specific miRNA expression profile targeting profibrotic cytokines and these miRNAs possibly render to favorable survival.

Identification and validation of DNA methylation-driven gene PCDHB4 as a novel tumor suppressor for glioblastoma diagnosis and prognosis.

Aberrant DNA methylation is a critical regulator of gene expression in the development and progression of glioblastoma (GBM). However, the impact of methylation-driven gene PCDHB4 changes on GBM occurrence and progression remains unclear. Therefore, this study aimed to identify the PCDHB4 gene for early diagnosis and prognostic evaluation and clarify its functional role in GBM. Methylation-driven gene PCDHB4 was selected for GBM using the multi-omics integration method based on publicly available data sets. The diagnostic capabilities of PCDHB4 methylation and 5-hydroxymethylcytosines were validated in tissue and blood cell-free DNA (cfDNA) samples, respectively. Combined survival analysis of PCDHB4 methylation and immune infiltration cells evaluated the prognostic predictive performance of GBM patients. We identified that the PCDHB4 gene achieved high discriminative capabilities for GBM and normal tissues with an area under the curve value of 0.941. PCDHB4 hypermethylation was observed in cfDNA blood samples from GBM patients. Compared with GBM patients with PCDHB4 hypermethylation level, patients with PCDHB4 hypomethylation level had significantly poorer overall survival (p = 0.035). In addition, GBM patients with PCDHB4 hypermethylation and high infiltration of CD4+ T cell activation level had a favorable survival (p = 0.026). Moreover, we demonstrated that mRNA expression of PCDHB4 was downregulated in GBM tissues and upregulated in GBM cell lines with PCDHB4 demethylation, and PCDHB4 overexpression inhibited GBM cell proliferation and migration. In summary, we discovered a novel methylation-driven gene PCDHB4 for the diagnosis and prognosis of GBM and demonstrated that PCDHB4 is a tumor suppressor in vitro experiments.

Comparison of linear and non-linear machine learning models for time-dependent readmission or mortality prediction among hospitalized heart failure patients

Although many models are available to predict prognosis of heart failure patients, most tools combining survival analysis are based on proportional hazard model. Non-linear machine learning algorithms would overcome the limitation of the time-independent hazard ratio assumption and provide more information in readmission or mortality prediction among heart failure patients.The present study collected the clinical information of 1796 hospitalized heart failure patients surviving during hospitalization in a Chinese clinical center from December 2016 to June 2019. A traditional multivariate Cox regression model and three machine learning survival models were developed in derivation cohort. Uno's concordance index and integrated Brier score in validation cohort were calculated to evaluate the discrimination and calibration of different models. Time-dependent AUC and Brier score curves were plotted to assess the performance of models at different time phases.

Impact of the immune molecular profile of the tumor microenvironment on the prognosis of NSCLC.

The present study aimed to clarify the association between macrophages, tumor neo-vessels and programmed cell death-ligand 1 (PD-L1) in the tumor microenvironment and the clinicopathological features of patients with non-small cell lung cancer (NSCLC), and to explore the prognostic factors of stromal features in NSCLC. To determine this, tissue microarrays containing samples of 92 patients with NSCLC were studied using immunohistochemistry and immunofluorescence. The quantitative data demonstrated that in tumor islets, the number of CD68+ and CD206+ tumor-associated macrophages (TAMs) was 8-348 (median, 131) and 2-220 (median, 52), respectively (P<0.001). In tumor stroma, the number of CD68+ and CD206+ TAMs was 23-412 (median, 169) and 7-358 (median, 81), respectively (P<0.001). The number of CD68+ TAMs in each location of the tumor islets and tumor stroma was significantly higher than that of CD206+ TAMs, and they were significantly correlated (P<0.0001). The quantitative density of CD105 and PD-L1 in tumor tissues was 19-368 (median, 156) and 9-493 (median, 103), respectively. Survival analysis revealed that a high density of CD68+ TAMs in tumor stroma and islets and a high density of CD206+ TAMs and PD-L1 in tumor stroma were associated with worse prognosis (both P<0.05). Collectively, the survival analysis demonstrated that the high-density group was related to a worse prognosis regardless of combined neo-vessels and PD-L1 expression with the CD68+ TAMs in tumor islets and stroma, or CD206+ TAMs in tumor islets and stroma. To the best of our knowledge, the present study was the first to provide a multi-component combined prognostic survival analysis of different types of macrophages in different regions with tumor neo-vessels and PD-L1, which demonstrated the importance of macrophages in tumor stroma.

Comparison of State-of-the-Art Neural Network Survival Models with the Pooled Cohort Equations for Cardiovascular Disease Risk Prediction

BackgroundThe Pooled Cohort Equations (PCEs) are race- and sex-specific Cox proportional hazards (PH)-based models used for 10-year atherosclerotic cardiovascular disease (ASCVD) risk prediction with acceptable discrimination. In recent years, neural network models have gained increasing popularity with their success in image recognition and text classification. Various survival neural network models have been proposed by combining survival analysis and neural network architecture to take advantage of the strengths from both. However, the performance of these survival neural network models compared to each other and to PCEs in ASCVD prediction is unknown.MethodsIn this study, we used 6 cohorts from the Lifetime Risk Pooling Project (with 5 cohorts as training/internal validation and one cohort as external validation) and compared the performance of the PCEs in 10-year ASCVD risk prediction with an all two-way interactions Cox PH model (Cox PH-TWI) and three state-of-the-art neural network survival models including Nnet-survival, Deepsurv, and Cox-nnet. For all the models, we used the same 7 covariates as used in the PCEs. We fitted each of the aforementioned models in white females, white males, black females, and black males, respectively. We evaluated models’ internal and external discrimination power and calibration.ResultsThe training/internal validation sample comprised 23216 individuals. The average age at baseline was 57.8 years old (SD = 9.6); 16% developed ASCVD during average follow-up of 10.50 (SD = 3.02) years. Based on 10 × 10 cross-validation, the method that had the highest C-statistics was Deepsurv (0.7371) for white males, Deepsurv and Cox PH-TWI (0.7972) for white females, PCE (0.6981) for black males, and Deepsurv (0.7886) for black females. In the external validation dataset, Deepsurv (0.7032), Cox-nnet (0.7282), PCE (0.6811), and Deepsurv (0.7316) had the highest C-statistics for white male, white female, black male, and black female population, respectively. Calibration plots showed that in 10 × 10 validation, all models had good calibration in all race and sex groups. In external validation, all models overestimated the risk for 10-year ASCVD.ConclusionsWe demonstrated the use of the state-of-the-art neural network survival models in ASCVD risk prediction. Neural network survival models had similar if not superior discrimination and calibration compared to PCEs.

SR-BI expression regulates the gastric cancer tumor immune microenvironment and is associated with poor prognosis

<b>Aim:</b> Scavenger receptor class B, type I (SR-BI) is an integral plasma membrane protein that has been reported to be overexpressed in various malignancies, such as renal cancer, breast cancer, and prostate cancer, and is an independent prognostic factor. However, the clinical value and expression of SR-BI in GC are unknown. Our research aimed to explore the role of SR-BI in combination with immune markers as a diagnostic and prognostic marker for gastric cancer (GC). <b>Methods:</b> GC tissues, paracancerous tissues, and clinicopathological data of 149 patients were collected. The expression level of SR-BI, Tumor-infiltrating lymphocytes (TILs), and PD-L1 were evaluated by immunohistochemistry (IHC). The associations of the SR-BI staining intensity with clinicopathological features and immune markers were determined by the chi-square test. Univariate and multivariate COX regression analyses were used to evaluate independent prognostic factors. Kaplan–Meier analyses were performed to plot the survival curve. <b>Results:</b> Our results indicated that SR-BI was expressed at higher levels in tumor tissues than in adjacent paracancerous tissues (<i>p</i> < 0.001), and patients with high levels of SR-BI expression had a worse prognosis. Univariate and multivariate analyses revealed that high SR-BI expression was an independent factor for poor prognosis. The chi-square test determined that the expression of SR-BI was negatively correlated with CD4+ T cells and CD8+ T cells (CD4+ T cells, <i>p</i> = 0.013; CD8+ T cells, <i>p</i> = 0.021), and positively correlated with PD-L1 (<i>p</i> = 0.022). Finally, survival analysis revealed that CD4+ T cells were associated with the prognosis of GC patients (<i>p</i> = 0.019), and the combined survival analysis of SR-BI and CD4+ T cells was also statistically significant (<i>p</i> = 0.030). <b>Conclusion:</b> SR-BI is highly expressed in GC tissue and associated with poor prognosis. Moreover, SR-BI can also regulate the GC tumor immune microenvironment.

APOBEC3B and CD274 as Combined Biomarkers for Predicting Response to Immunotherapy in Urothelial Carcinoma of the Bladder.

Immunotherapy has become a promising form of treatment for cancers. There is a need to predict response to immunotherapy accurately. In the UCSC Xena, pan-cancer analysis revealed a positive relationship between APOBEC3B (A3B) and tumor mutational burden (R = 0.28, P < 0.001) and microsatellite instability (R = 0.12, P < 0.05). Naturally, the A3B high expression group had higher tumor mutational burden and microsatellite instability than the low expression group. The bladder cancer (BLCA) cohort in The Cancer Genome Atlas (TCGA) revealed tumor mutational signatures of A3B high and low expression groups. Compared to the low expression group, the high expression group had a higher number of SNPs and mutations. Subsequently, A3B was profiled for immune cell infiltration and immune checkpoints in bladder cancer. The results showed that A3B was positively correlated with most immune cells. Compared with the A3B low expression group, the A3B high expression group had higher expression of immune checkpoints. A3B was positively correlated with CD274 (R = 0.12, P = 0.016). This indicated that the high expression of A3B may have a better response to immunotherapy. Furthermore, data from the IMvigor210 immunotherapy clinical trial was used to confirm the findings of this study. The combined survival analysis of A3B and CD274 showed that the group of patients with high expression of CD274 and A3B was found to have a significantly higher survival rate than the rest of the patient group (P < 0.047). The results demonstrated that A3B has a significant role in immunotherapy. Moreover, the combined biomarkers of A3B and CD274 were more effective in predicting response to immunotherapy in bladder urothelial carcinoma. The findings of this study provide valuable insights for precision medicine.

The Genomic Landscape of Vulvovaginal Squamous Cell Carcinoma (048)

The Genomic Landscape of Vulvovaginal Squamous Cell Carcinoma (048)

Screening of key methylation-driven genes CDO1 in breast cancer based on WGCNA.

With the rapid development of genomics and molecular biology, not only have biochemical indicators been used as tumour markers, but many new molecular markers have emerged. Epigenetic abnormalities are a new type of molecular marker, and DNA methylation is an important part of epigenetics. This study used weighted gene coexpression network analysis (WGCNA) to analyse key methylation-driven genes in breast cancer. The RNA-seq transcriptome data, DNA methylation data, and clinical information data of breast cancer patients were downloaded from The Cancer Genome Atlas (TCGA) database, and the MethylMix R package was used to screen methylation-driven genes in breast cancer. The ClusterProfiler package and enrichplot package in R software were used to further analyse the function and signalling pathway of methylation-driven genes. Through univariate and multivariate Cox regression analyses, methylation-driver genes related to prognostic were obtained, a prognostic model was constructed and prognostic characteristics were analysed. The 17 methylation-driven genes related to prognosis were obtained by the WGCNA method in breast cancer, and the prognostic significance of these methylation-driven genes was determined by transcriptome and methylation combined survival analysis. Analysis of functions and signalling pathways showed that these genes were mainly enriched in biological processes and signalling pathway. Through univariate and multivariate Cox regression analyses, a prognostic model of 5 methylation-driven genes was constructed. The AUC of the receiver operating characteristic (ROC) curve of this model was 0.784, showing that the model had a good prediction effect. Based on WGCNA screening, it was found that only CDO1 was the key methylation-driven gene for prognosis in breast cancer, indicating that CDO1 may be an important indicator of the prognosis of breast cancer patients.

Relationship between ATOH1 and tumor microenvironment in colon adenocarcinoma patients with different microsatellite instability status

BackgroundColon adenocarcinoma (COAD) is one of the major varieties of malignant tumors threatening human health today. Immune checkpoint inhibitors (ICIs) have recently begun to emerge as an effective option for the treatment of COAD patients, but not all patients can benefit from ICI treatment. Previous studies have suggested that ICIs boast significant clinical effects on patients with microsatellite instability-high (MSI-H), while conversely patients with microsatellite-stable/microsatellite instability-low (MSS/MSI-L) have shown limited response.MethodsWe used ATAC-seq, RNA-seq, and mutation data from The Cancer Genome Atlas Colon adenocarcinoma (TCGA-COAD) cohort to perform multi-omics differential analysis on COAD samples with different MSI statuses, then further screened genes by additionally combining these results with survival analysis. We analyzed the effects of the screened genes on the tumor microenvironment and immunogenicity of COAD patients, and subsequently determined their influence on the efficacy of ICIs in COAD patients using a series of predictive indexes.ResultsTwelve genes were screened in the TCGA-COAD cohort, and after the combined survival analysis, we identified ATOH1 as having significant effects. ATOH1 is characterized by high chromatin accessibility, high expression, and high mutation in COAD patients in the MSI-H group. COAD patients with high ATOH1 expression are associated with a better prognosis, unique immune microenvironment, and higher efficacy in ICI treatment. Enrichment analysis showed that COAD patients with high ATOH1 expression displayed significant upregulation in their humoral immunity and other related pathways.ConclusionsWe speculate that ATOH1 may influence the efficacy of ICIs therapy in patients with COAD by affecting the immune microenvironment and immunogenicity of the tumor.

Appraisal of clinicopathological prognosticators in advanced acral lentiginous melanoma with characterization of PD-L1 and CD8/CD4 immunoprofiles.

Acral lentiginous melanoma is the most common subtype of cutaneous melanoma in Asian countries. This study aims to clarify the associations between certain histologic and immunohistochemical parameters, and identify their prognostic values. We assessed several histologic features and conducted immunohistochemical study of programmed cell death ligand 1 (clone 22C3) and CD8/CD4 in 61 Taiwanese patients with Stage III/IV, non-BRAF acral lentiginous melanomas. A total of 41 males and 20 females were included, with a median age of 74years. The majority of tumors occurred at nonungual locations (86.9%), with 'foot' being the most frequently affected site (85.2%). Positive programmed cell death ligand 1 staining (combined positive score≥10) was significantly associated with the status of tumor-infiltrating lymphocytes (P=0.036). Lack of skin ulceration was linked to the immunoexpression of CD8/CD4-high (P=0.004). A superior clinical outcome was found in the tumor-infiltrating lymphocytes-present group (P=0.011), and among which, CD8/CD4-high was significantly correlated with better survival (P<0.001). Combined survival analysis revealed that the PD-L1(-) TIL(+) CD8/CD4-high subgroup was associated with favorable prognosis, and cases with PD-L1(+) TIL(-) showed the worst disease specific survival (P<0.001). In the univariate analysis, lymphovascular invasion (P=0.002), skin ulceration (P=0.002), tumor-infiltrating lymphocytes (P=0.015) and CD8/CD4 status (P<0.001) were significant prognostic factors. At the multivariate level, the statuses of CD8/CD4-low (P<0.001) and lymphovascular invasion (P=0.014) represented the independent poor prognosticators. For advanced, non-BRAF acral lentiginous melanomas, comprehensive assessments of these microscopic traits, along with CD8/CD4 and PD-L1 immunoprofiles, may help guide the clinicians and patients through treatment decisions.

Predicting cardiovascular risk from national administrative databases using a combined survival analysis and deep learning approach

BackgroundMachine learning-based risk prediction models may outperform traditional statistical models in large datasets with many variables, by identifying both novel predictors and the complex interactions between them. This study compared deep learning extensions of survival analysis models with Cox proportional hazards models for predicting cardiovascular disease (CVD) risk in national health administrative datasets.MethodsUsing individual person linkage of administrative datasets, we constructed a cohort of all New Zealanders aged 30–74 who interacted with public health services during 2012. After excluding people with prior CVD, we developed sex-specific deep learning and Cox proportional hazards models to estimate the risk of CVD events within 5 years. Models were compared based on the proportion of explained variance, model calibration and discrimination, and hazard ratios for predictor variables.ResultsFirst CVD events occurred in 61 927 of 2 164 872 people. Within the reference group, the largest hazard ratios estimated by the deep learning models were for tobacco use in women (2.04, 95% CI: 1.99, 2.10) and chronic obstructive pulmonary disease with acute lower respiratory infection in men (1.56, 95% CI: 1.50, 1.62). Other identified predictors (e.g. hypertension, chest pain, diabetes) aligned with current knowledge about CVD risk factors. Deep learning outperformed Cox proportional hazards models on the basis of proportion of explained variance (R2: 0.468 vs 0.425 in women and 0.383 vs 0.348 in men), calibration and discrimination (all P <0.0001).ConclusionsDeep learning extensions of survival analysis models can be applied to large health administrative datasets to derive interpretable CVD risk prediction equations that are more accurate than traditional Cox proportional hazards models.

LncRNA FOXP4-AS1 Promotes Progression of Ewing Sarcoma and Is Associated With Immune Infiltrates.

Ewing sarcoma (ES) is a highly malignant primary bone tumor with poor prognosis. Studies have shown that abnormal expression of lncRNA influences the prognosis of tumor patients. Herein, we established that FOXP4-AS1 was up-regulated in ES and this correlated with poor prognosis. Further analysis illustrated that FOXP4-AS1 down-regulation repression growth, migration, along with invasion of ES. On the contrary, up-regulation of FOXP4-AS1 promoted the growth, migration, as well as invasion of ES. To explore the mechanism of FOXP4-AS1, Spearman correlation analysis was carried out to determine genes that were remarkably linked to FOXP4-AS1 expression. The potential functions and pathways involving FOXP4-AS1 were identified by GO analysis, Hallmark gene set enrichment analysis, GSEA, and GSVA. The subcellular fractionation results illustrated that FOXP4-AS1 was primarily located in the cytoplasm of ES cells. Then a ceRNA network of FOXP4-AS1 was constructed. Analysis of the ceRNA network and GSEA yielded two candidate mRNAs for FOXP4-AS1. Results of the combined survival analysis led us to speculate that FOXP4-AS1 may affect the expression of TMPO by sponging miR-298, thereby regulating the malignant phenotype of ES. Finally, we found that FOXP4-AS1 may modulates the tumor immune microenvironment in an extracellular vesicle-mediated manner. In summary, FOXP4-AS1 correlates with poor prognosis of ES. It promotes the growth, migration, as well as invasion of ES cells and may modulate the tumor immune microenvironment.

A DYNAMIC CREDIT SCORING MODEL BASED ON SURVIVAL GRADIENT BOOSTING DECISION TREE APPROACH

Credit scoring, which is typically transformed into a classification problem, is a powerful tool to manage credit risk since it forecasts the probability of default (PD) of a loan application. However, there is a growing trend of integrating survival analysis into credit scoring to provide a dynamic prediction on PD over time and a clear explanation on censoring. A novel dynamic credit scoring model (i.e., SurvXGBoost) is proposed based on survival gradient boosting decision tree (GBDT) approach. Our proposal, which combines survival analysis and GBDT approach, is expected to enhance predictability relative to statistical survival models. The proposed method is compared with several common benchmark models on a real-world consumer loan dataset. The results of out-of-sample and out-of-time validation indicate that SurvXGBoost outperform the benchmarks in terms of predictability and misclassification cost. The incorporation of macroeconomic variables can further enhance performance of survival models. The proposed SurvXGBoost meanwhile maintains some interpretability since it provides information on feature importance.

Establishment of immune prognostic signature and analysis of prospective molecular mechanisms in childhood osteosarcoma patients.

In pediatric tumors, immunotherapy exhibits less toxicity than chemotherapy and radiation. The current study aims to identify potential immune targets in immune-related genes of C-C motif chemokine ligand genes (CCLs) and C-C motif chemokine receptors (CCRs) in childhood osteosarcoma (OS) and to explore the underlying molecular mechanisms of childhood OS. Firstly, we identified immune-related genes in CCLs and CCRs, these genes were used for functional annotation and interaction analysis. Then, the prognostic value of these genes was evaluated using Kaplan-Meier analysis and multivariate COX regression model. And the potential relationship between risk score and immune infiltrating cells was identified. Finally, gene set enrichment analysis was used to determine the underlying molecular mechanism of OS. Immune-related genes in CCLs and CCRs are inextricably linked. The results of survival analysis of these genes show that CCL5, CCL8, CCR4, and CCR5 are significantly associated with the prognosis of childhood OS. The combined effect survival analysis shows that the co-high expression of these 4 genes has a good prognosis for childhood OS. A prognostic signature model was constructed based on the 4 genes mentioned above, and the result of time-dependent receiver operating characteristic curves showed that this model was a good predictor of childhood OS 3- and 5-year prognosis. In addition, the risk score of the constructed prognostic signature model was closely related to immune infiltration. We also found that CCL5, CCL8, and CCR5 may affect the prognosis of OS through complex regulation among Toll-like receptor signaling pathway, mitogen-activated protein kinase (MAPK) family signaling cascade, and nuclear factor-kappaB pathway, whereas CCR4 affects the prognosis of OS by regulating eukaryotic translation. CCL5, CCL8, CCR4, and CCR5 are potential prognostic markers for the prognosis of childhood OS, and the underlying molecular mechanisms of childhood OS have been identified.

Comprehensive Identification of the Human Secretome as Potential Indicators in Treatment Outcome of HPV-Positive and -Negative Cervical Cancer Patients

Background: The aim of this work was to explore the novel and promising biomarkers for the diagnosis and prognosis of cervical cancer patients. Methods: The secretome of primary cervical tissues was extracted and then determined by the LC-MS/MS assay. The level of screened targets was confirmed using the RT-PCR and ELISA in cervical cancer tissue samples. The median expression level of certain targets was used as a cutoff value to divide the patients into 2 groups, and then the patients were followed up. The predictive abilities of the targets on the prognosis were further studied. Results: LC-MS/MS, together with bioinformatic analysis, demonstrated that totally 95 targets were dysregulated in cervical cancer. Among them, ECM2, KLK6, and MASP1 were increased in cervical cancer in a stage-dependent manner, whereas FGA was negatively associated with the stage of cervical cancers. Overall survival (OS) and disease-free survival (DFS) rates were significantly decreased in the KLK6 high group, whereas little difference was found between the high and low groups of other 3 cases. Univariate analysis of the 5-year OS and DFS revealed a significantly worse outcome for patients with KLK6 high tumors. In multivariate analysis, KLK6 remained a highly significant prognostic marker for OS and DFS. Combined survival analysis of KLK6 expression and the HPV infection revealed that KLK6^highHPV⁽⁻⁾ predicted the most poor OS rate and the KLK6^lowHPV⁽⁺⁾ group showed the best prognosis. Conclusion: Through the secretome analysis, we identified a series of secreted proteins differentially expressed in the clinical cancer, among which KLK6 has the potential to become a promising biomarker for the diagnosis and prognosis of cervical cancer patients.

Infiltration by IL22-Producing T Cells Promotes Neutrophil Recruitment and Predicts Favorable Clinical Outcome in Human Colorectal Cancer.

Immune cell infiltration in colorectal cancer effectively predicts clinical outcome. IL22, produced by immune cells, plays an important role in inflammatory bowel disease, but its relevance in colorectal cancer remains unclear. Here, we addressed the prognostic significance of IL22+ cell infiltration in colorectal cancer and its effects on the composition of tumor microenvironment. Tissue microarrays (TMA) were stained with an IL22-specific mAb, and positive immune cells were counted by expert pathologists. Results were correlated with clinicopathologic data and overall survival (OS). Phenotypes of IL22-producing cells were assessed by flow cytometry on cell suspensions from digested specimens. Chemokine production was evaluated in vitro upon colorectal cancer cell exposure to IL22, and culture supernatants were used to assess neutrophil migration in vitro Evaluation of a testing (n = 425) and a validation TMA (n = 89) revealed that high numbers of IL22 tumor-infiltrating immune cells were associated with improved OS in colorectal cancer. Ex vivo analysis indicated that IL22 was produced by CD4+ and CD8+ polyfunctional T cells, which also produced IL17 and IFNγ. Exposure of colorectal cancer cells to IL22 promoted the release of the neutrophil-recruiting chemokines CXCL1, CXCL2, and CXCL3 and enhanced neutrophil migration in vitro Combined survival analysis revealed that the favorable prognostic significance of IL22 in colorectal cancer relied on the presence of neutrophils and was enhanced by T-cell infiltration. Altogether, colorectal cancer-infiltrating IL22-producing T cells promoted a favorable clinical outcome by recruiting beneficial neutrophils capable of enhancing T-cell responses.