Prognostic Model Research Articles

Weakly supervised deep learning-based prognostic model for AI-assisted prospective clinical investigation of cervical cancer

Weakly supervised deep learning-based prognostic model for AI-assisted prospective clinical investigation of cervical cancer

A ferroptosis-associated prognostic model correlated with immune landscape and radiotherapy response in low-grade gliomas (LGGs)

Despite receiving comprehensive treatment, the prognosis for low-grade gliomas (LGGs) patients varies considerably. Recent studies have focused extensively on ferroptosis, across a range of tumor types. Nevertheless, methodologies to evaluate the efficacy of radiotherapy for LGGs, from the perspective of ferroptosis-related genes (FRGs), remain strikingly rare. In this study, we conducted a retrospective study on the transcriptional profiles of LGG patients from the public databases and a local cohort. An FRG model was developed and validated, exhibits heightened robustness when contrasted with the traditional ssGSEA model. Patients demonstrating higher FRG scores were identified as a high-risk group, displaying a worse prognosis. By incorporating the FRG score alongside other prognosis-associated clinical indicators, we formulated an enhanced nomogram to achieve a higher level of prediction performance. Additionally, among LGG patients receiving radiotherapy, a poorer prognosis was observed in the high-risk group. Further investigation revealed that samples from the high-risk group generally exhibit a TME in an immuno-suppressive state. Collectively, we developed an FRG model and a robust nomogram for LGG prognostication. This study suggests that a high FRG score, indicative of an immunosuppressive TME, could potentially lead to a less favorable prognosis for certain LGG patients receiving radiotherapy.

Development and validation of the post-CAR prognostic index for large B-cell lymphoma patients after CAR-T progression in third or later line treatment

Chimeric antigen receptor (CAR) T-cell therapy fails to achieve durable responses in over 60% of relapsed/refractory (R/R) large B-cell lymphoma (LBCL) patients in the third or later line setting. After CAR-T failure, survival outcomes are heterogeneous and a prognostic model in this patient population is lacking. A training cohort of 216 patients with progressive disease (PD) after CAR-T from 12 Spanish centers was used to develop the Post-CAR Prognostic Index (PC-PI); primary endpoint was overall survival (OS) from CAR-T progression. Validation was performed in an external cohort from three different European centers (n = 204). The prognostic score incorporated five variables, assessed at time of PD to CAR-T: ECOG (> 0), hemoglobin (< 10 g/dL), LDH (≥ 2xULN), number of extranodal sites (> 1) and time from CAR-T to PD (< 4 months). Patients were classified in four risk groups with distinct OS (p-value < 0.05 in all comparisons). In the validation cohort, median OS in the low (31%), intermediate-low (26%), intermediate-high (17%) and high risk (26%) were 15.7, 7.1, 1.8 and 1.0 months, respectively (p < 0.05 in all comparisons). Results were consistent following adjustment for subsequent treatment. In the external cohort, the PC-PI showed a C-statistic of 0.79 (95%CI 0.76–0.82), outperforming IPI and R-IPI. In conclusion, the PC-PI score is a novel tool for OS prediction and could facilitate risk-adapted management of LBCL patients relapsing after CAR T-cells. Additionally, these results will help stratification and interpretation of trials and real-world data incorporating CART-exposed patients.

FDG-PET patterns associate with survival in patients with prion disease.

Prion disease classically presents with rapidly progressive dementia, leading to death within months of diagnosis. Advances in diagnostic testing have improved recognition of patients with atypical presentations and protracted disease courses, raising key questions surrounding the relationship between patterns of neurodegeneration and survival. We assessed the contribution of fluorodeoxyglucose (FDG-PET) imaging for this purpose. FDG-PET were performed in 40 clinic patients with prion disease. FDG-PET images were projected onto latent factors generated in an external dataset to yield patient-specific eigenvalues. Eigenvalues were input into a clustering algorithm to generate data-driven clusters, which were compared by survival time. Median age at FDG-PET was 65.3 years (range 23-85). Median time from FDG-PET to death was 3.7 months (range 0.3-19.0). Four data-driven clusters were generated, termed "Neocortical" (n = 7), "Transitional" (n = 12), "Temporo-parietal" (n = 13), and "Deep nuclei" (n = 6). Deep nuclei and transitional clusters had a shorter survival time than the neocortical cluster. Subsequent analyses suggested that this difference was driven by greater hypometabolism of deep nuclei relative to neocortical areas. FDG-PET-patterns were not associated with demographic (age and sex) or clinical (CSF total-tau, 14-3-3) variables. Greater hypometabolism within deep nuclei relative to neocortical areas associated with more rapid decline in patients with prion disease and vice versa. FDG-PET informs large-scale network physiology and may inform the relationship between spreading pathology and survival in patients with prion disease. Future studies should consider whether FDG-PET may enrich multimodal prion disease prognostication models.

A novel risk score model of lactate metabolism for predicting outcomes and immune signatures in acute myeloid leukemia

Acute myeloid leukemia (AML) is a malignant tumor with high recurrence and refractory rates and low survival rates. Increased glycolysis is characteristic of metabolism in AML blast cells and is also associated with chemotherapy resistance. The purpose of this study was to use gene expression and clinical information from The Cancer Genome Atlas (TCGA) database to identify subtypes of AML associated with lactate metabolism. Two different subtypes linked to lactate metabolism, each with specific immunological features and consequences for prognosis, were identified in this study. Using the TCGA and International Cancer Genome Consortium (GEO) cohorts, a prognostic model composed of genes (LMNA, RETN and HK1) for the prognostic value of the lactate metabolism-related risk score prognostic model was created and validated, suggesting possible therapeutic uses. Additionally, the diagnostic value of the prognostic model genes was explored. LMNA and HK1 were ultimately identified as hub genes, and their roles in AML were determined through immune infiltration, GeneMANIA, GSEA, methylation analysis and single-cell analysis. LMNA was upregulated in AML associating with a poor prognosis while HK1 was downregulated in AML associating with a favorable prognosis. The findings underscore the noteworthy impact of genes linked to lactate metabolism in AML and illustrate the possible therapeutic usefulness of the lactate metabolism-related risk score and the hub lactate metabolism-related genes in guiding AML patients’ treatment choices.

A standards-based software tool to support prognostic modelling: application to exemplar heart failure risk scores

Abstract Introduction Heart failure prognosis is an active research area, with the development of several new risk scores each year. While risk score derivation and evaluation have become common practices, these processes currently lack standardisation. Purpose We aimed to: i) develop a standards-based software tool to support prognostic modelling; and ii) apply the tool to derive, evaluate and compare exemplar heart failure risk models to address a hypothetical question of whether using only non-invasive measurements provides similar prognosis value compared to adding invasive measurements. Methods A software tool was developed based on a prognostic modelling workflow defined using the Common Workflow Language and containing 7 steps: univariate Cox proportional hazard (PH) regression, multivariable Cox PH regression with assessment of proportional hazards and linearity with outcome (Schoenfeld and Martingale residuals), derivation of points per risk factor, risk calculation and stratification, discrimination and calibration assessment. Using data from the PEOPLE heart failure cohort [1], two integer-based risk scores were developed to predict 2-year all-cause mortality. The first model utilised 12 variables available at recruitment informed by prior analysis of the PEOPLE cohort to provide a benchmark for comparison, while the second model is based on 9 non-invasive parameters. The analysis was conducted on cases with complete data (n=781), with no internal or external validation. Results By inputting the dataset and indicating the variables of interest for each model, our software tool automatically generated the patient-specific risk scores, risk strata, c-index, and discrimination and calibration plots. A 12-parameter model based on age, sex, ejection fraction, hypertension, diabetes, atrial fibrillation, ischaemic history, NYHA class, heart rate, systolic blood pressure, creatinine and NT-proBNP yielded reasonable discrimination (c-index=0.77). Graphical assessment of risk strata and calibration (Fig. 1.a and c) revealed good performance overall, apart from the highest risk decile where survival was overestimated. The reduced 9-parameter model excluded creatinine, NT-proBNP, and automatically dropped diabetes status for non-linearity. It showed a substantial drop in discrimination (c-index=0.67), confirmed in graphical assessment of risk strata (Fig. 1.b). Calibration was more variable per decile than seen with the 12-parameter model but was more robust to survival overestimation (Fig. 1.d). Conclusion Our software tool ensures the uniformity and comparability of the two risk models derived by a common statistical process. In this example, we showed how the addition of NT-proBNP, creatinine, and diabetes status improves discrimination for predicting heart failure survival. This tool can be leveraged to test new hypotheses in a standardised and reproducible manner to enhance research in heart failure prognosis.

The role of molecular subtypes and immune infiltration characteristics based on disulfidptosis-related genes in ovarian cancer

Ovarian cancer (OC) is the most fatal, gynecological malignancy. Compared with advanced ovarian cancer, the 5 year survival rate of early ovarian cancer is significantly improved, and predicting early detection and diagnosis is very important to improve the prognosis of OC. Recent research has found a new way of cell death: disulfidptosis. Under glucose starvation, abnormal accumulation of disulfide molecules such as Cystine in SLC7A11 overexpression cells induced disulfide stress to trigger cell death. Studies of disulfidptosis are still in their infancy and its role in ovarian cancer progression is unclear. In this study, we used a public database to detect the expression and mutations of disulfidptosis-related genes in OC. Cluster analysis was performed based on disulfidptosis-related genes, and disulfidptosis differential expression genes were analyzed. A prognostic risk model was constructed using three disulfidptosis-related genes, and the reasons for differences in prognosis were explored through immune infiltration analysis and drug sensitivity analysis. The prognostic characteristics of transcriptome based on disulfidptosis-related genes are closely related to the prognosis of OC patients. Finally, quantitative polymerase chain reaction (RT-qPCR) was used to detect the expression of three prognostic genes in clinical OC samples.Our study establishes a link between disulfidptosis and OC, providing new ideas for personalized and precise treatment of OC.

Expression of lipid-metabolism genes is correlated with immune microenvironment and predicts prognosis of hepatocellular carcinoma

Hepatocellular carcinoma (HCC) is one of the most common malignant tumors. This study was aimed to identify a lipid metabolism-related signature associated with the HCC microenvironment to improve the prognostic prediction of HCC patients. Clinical information and expression profile data were downloaded from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases, including the GEO dataset GSE76427. The gene expression profile of lipid metabolism was downloaded from Molecular Signatures Database (MSigDB) database. The infiltrating immune cells were estimated by the Estimation of Stromal and Immune cells in Malignant Tumor tissues using Expression data (ESTIMATE), MCP-counter, and TIMER algorithms. Functional analysis, including Gene ontology (GO) analysis, Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene set enrichment analysis (GSEA) were performed to elucidate the underlying mechanisms. The prognostic risk model was performed by Least absolute shrinkage and selection operator (LASSO) algorithm and Cox regression analysis. Two distinct subgroups of survival were identified. Better prognosis was associated with high immune score, high abundance of immune infiltrating cells, and high immune status. GO and KEGG analysis showed that differentially expressed genes (DEGs) between the two subgroups were mainly enriched in immune related pathways. GSEA analysis suggested that the expression of lipid metabolism related genes (LMRGs) was related to dysregulation of immune in the high-risk group. Risk models and clinical features based on LMRGs predicted HCC prognosis. This study indicated that the lipid metabolism-related signature was important for the prognosis of HCC. The expression of LMRGs was related to the immune microenvironment of HCC patients and could be used to predict the prognosis of HCC.

Prognostic factors and predictive models for primary pulmonary diffuse large B-cell lymphoma: a population-based analysis

ABSTRACT Background Primary pulmonary diffuse large B-cell lymphoma (PP-DLBCL) is a rare extranodal non-Hodgkin's lymphoma (EN-NHL). Its prognosis as an aggressive lymphoma is abysmal, and predictive models are still lacking. Methods We screened patients diagnosed with PP-DLBCL between 2010 and 2019 from the Surveillance, Epidemiology, and End Results (SEER) database. Then, univariate and multivariate COX regression analyses were used to identify independent risk factors affecting patient prognosis. Finally, a novel nomogram was constructed and the model was evaluated by looking at three dimensions. Results A total of 831 patients were included in this study. Most of the patients were elderly (526 (63.8%)) and female (428 (51.9%)). The included patients were randomized in a 7:3 ratio into a training group (577 (70%)) and a validation group (248 (30%)). We concluded that the independent risk factors of prognosis were age, extrapulmonary metastasis, radiotherapy, chemotherapy, and surgical intervention. The results of receiver operating characteristic curves, calibration curves, and decision curve analysis in the training and validation groups confirmed that the risk prediction nomogram could accurately predict the survival of PP-DLBCL. Conclusion This study is the first large population-based clinical data study on PP-DLBCL. A novel predictive model about prognosis has been developed to help clinical decision-making.

Fibrinogen to pre-albumin ratio is an independent prognostic index for patients with pancreatic ductal adenocarcinoma after radical resection

BackgroundThis study aims to elucidate the significance of the preoperative fibrinogen to pre-albumin ratio (FPR) in predicting the prognosis of pancreatic ductal adenocarcinoma (PDAC), a correlation not extensively explored previously.MethodsA cohort of 563 patients diagnosed with PDAC and subjected to radical surgical resection was examined. We meticulously documented a range of inflammatory markers, clinical-pathological features, and oncological outcomes. The prognostic value of preoperative FPR was assessed using Kaplan–Meier survival analysis and Cox proportional hazards regression modeling. Furthermore, the predictive accuracy of FPR was evaluated through time-dependent receiver operating characteristic (ROC) curves and decision curve analyses (DCA).ResultsThe determined optimal threshold for FPR was 14.77, which facilitated the stratification of patients into groups with low and high FPR levels. Notably, patients in the high FPR cohort exhibited significantly reduced recurrence-free survival (RFS) and overall survival (OS) rates compared to their low FPR counterparts. Multivariate Cox regression analysis underscored FPR as an independent prognostic indicator for both RFS and OS. In comparison to the neutrophil-to-lymphocyte ratio (NLR), FPR demonstrated superior prognostic accuracy and clinical utility.ConclusionThe preoperative fibrinogen to pre-albumin ratio serves as an independent prognostic marker for RFS and OS among PDAC patients undergoing radical resection. Our findings suggest that FPR could be a valuable addition to the current prognostic models, potentially guiding therapeutic decision-making and patient management strategies in PDAC.

Lower early mortality after acute pulmonary embolism in patients with WHO-defined obesity: results from a post-hoc cohort analysis with external validation

Abstract Background The relationship between obesity, defined as a Body Mass Index (BMI) ≥ 30 kg/m2, and mortality in venous thrombo-embolism remains controversial. Objectives To compare early outcomes after pulmonary embolism (PE) between obese patients, and non-obese, non-underweight patients. Methods We performed a post-hoc analysis based on prospectively recorded individual patient data from the multicenter BFC-FRANCE registry. We compared rates of outcomes using multivariable logistic regression or a Cox model for 30-day and 6-month outcomes respectively. We assessed the incremental value of adding BMI information on top of the 30-day European Society of Cardiology (ESC) prognostic algorithm. Results A total of 2,390 patients with a BMI ≥ 18.5 kg/m2 (mean age, 66.9±16.8 years; 1,188 men [49.7%]) were included, 686 patients [28.7%] in the obese group, and 1,704 patients [71.3%] in the non-obese group. Mortality rates were significantly lower in obese patients versus non-obese patients at 30 days (OR, 0.59; 95% CI, 0.35-0.98), and 6 months. Cox-model-derived adjusted curves for all-cause death at 6 months, with hazard ratios (HRs) between obese and non-obese patients, after acute PE, are shown in the Figure. Rates of secondary non-fatal outcomes (including recurrent VTE) did not differ between groups. The addition of the obesity information on top of the ESC prognostic model improved global model fit, and discriminatory and calibration capacities, yielding significant reclassification based on the observed mortality rates with the ESC model as reference. Findings were confirmed in an external validation using 35,796 PE patients from the RIETE registry. Conclusion We present evidence indicating lower early and mid-term mortality after PE in patients classified as obese based on BMI, compared with non-obese, non-underweight patients. BMI should likely be incorporated into algorithms or scoring systems for predicting early mortality following PE.

COADREADx: A comprehensive algorithmic dissection of colorectal cancer unravels salient biomarkers and actionable insights into its discrete progression

Background Colorectal cancer is a common condition with an uncommon burden of disease, heterogeneity in manifestation, and no definitive treatment in the advanced stages. Renewed efforts to unravel the genetic drivers of colorectal cancer progression are paramount. Early-stage detection contributes to the success of cancer therapy and increases the likelihood of a favorable prognosis. Here, we have executed a comprehensive computational workflow aimed at uncovering the discrete stagewise genomic drivers of colorectal cancer progression. Methods Using the TCGA COADREAD expression data and clinical metadata, we constructed stage-specific linear models as well as contrast models to identify stage-salient differentially expressed genes. Stage-salient differentially expressed genes with a significant monotone trend of expression across the stages were identified as progression-significant biomarkers. The stage-salient genes were benchmarked using normals-augmented dataset, and cross-referenced with existing knowledge. The candidate biomarkers were used to construct the feature space for learning an optimal model for the digital screening of early-stage colorectal cancers. The candidate biomarkers were also examined for constructing a prognostic model based on survival analysis. Results Among the biomarkers identified are: CRLF1, CALB2, STAC2, UCHL1, KCNG1 (stage-I salient), KLHL34, LPHN3, GREM2, ADCY5, PLAC2, DMRT3 (stage-II salient), PIGR, HABP2, SLC26A9 (stage-III salient), GABRD, DKK1, DLX3, CST6, HOTAIR (stage-IV salient), and CDH3, KRT80, AADACL2, OTOP2, FAM135B, HSP90AB1 (top linear model genes). In particular the study yielded 31 genes that are progression-significant such as ESM1, DKK1, SPDYC, IGFBP1, BIRC7, NKD1, CXCL13, VGLL1, PLAC1, SPERT, UPK2, and interestingly three members of the LY6G6 family. Significant monotonic linear model genes included HIGD1A, ACADS, PEX26, and SPIB. A feature space of just seven biomarkers, namely ESM1, DHRS7C, OTOP3, AADACL2, LPHN3, GABRD, and LPAR1, was sufficient to optimize a RandomForest model that achieved &gt; 98% balanced accuracy (and performant recall) of cancer vs. normal on external validation. Design of an optimal multivariate model based on survival analysis yielded a prognostic panel of three stage-IV salient genes, namely HOTAIR, GABRD, and DKK1. Based on the above sparse signatures, we have developed COADREADx, a web-server for potentially assisting colorectal cancer screening and patient risk stratification. COADREADx provides uncertainty measures for its predictions and needs clinical validation. It has been deployed for experimental non-commercial use at: https://apalanialab.shinyapps.io/coadreadx/.

Discrimination and calibration are insensitive metrics for comparing the performance of cardiovascular risk prediction models

Abstract Background/Introduction Accurate cardiovascular disease (CVD) risk prediction models can facilitate the identification and cost-effective treatment of high-risk individuals. Discrimination and calibration are the most commonly recommended metrics for evaluating model performance. However, both are population-level measures, and it remains unclear whether these metrics are sufficiently sensitive to assess the performance of models designed for use with individual patients. Purpose Using a real-world dataset, we aimed to compare CVD risk estimates for example individuals (individual-level measures) and discrimination and calibration metrics (population-level measures) from two CVD risk prediction models; one with "age" as the only predictor and one with multiple predictors. Methods We constructed a cohort of almost all New Zealanders without CVD or heart failure, alive and aged 30–79 years in 2014, with follow-up linkage to hospitalisations and mortality until 2018 (N = 2,098,359). We derived two sets of sex-specific Cox regression models to predict 5-year CVD risk. The age-only models included "age" as the only predictor, and the full models used nine pre-defined routinely available CVD risk predictors with interaction terms. We calculated and compared individuals’ 5-year CVD risk from the age-only and full models and assessed model performance using Harrell’s C-statistics and calibration plots. Results Risk estimates for individuals using the two models showed substantial differences. A 65-year-old woman had predicted 5-year risks of 4.9% with the age-only model but 16.1% with the full model; a 74-year-old man had predicted risks of 15.2% and 9.2%, respectively, with age-only and full models (Figure 1). However, the models had similar discrimination and calibration (Figure 2). Age-only models had good discrimination with C-statistics of 0.773 (95% CI: 0.771–0.775) in women and 0.740 (0.738–0.742) in men, and calibrated well in both sexes. The full models had slightly better discrimination (C-statistics: 0.814 [0.812–0.816] in women and 0.769 [0.767–0.771] in men), and very similar calibration to the age-only models in both sexes. Conclusions Age-only models showed good discrimination and were well calibrated, but predicted very different risks in individual case studies compared to models with multiple predictors. This study calls into question the use of discrimination and calibration metrics as the main measures for comparing the performance of different CVD risk prediction models. We recommend that examples of predicted risk for individuals and an assessment of the magnitude and prevalence of cardiovascular predictors should also be considered when comparing the performance of prognostic models.

An ensemble machine learning approach for predicting heart failure risk in acute coronary syndrome patients: heartguardai

Abstract Background Heart failure (HF) constitutes a major post-event morbidity in patients experiencing Acute Coronary Syndrome (ACS), underscoring the imperative for precise risk stratification methodologies. Traditional prognostic models, although beneficial, are limited by their reliance on a narrow scope of variables, thereby overlooking the comprehensive analytical potential afforded by extensive clinical datasets. Aims To develop "HeartGuardAI," an ensemble machine learning model that predicts HF risk in ACS patients. By incorporating a wide array of variables including ejection fraction (EF), glomerular filtration rate (GFR), body mas index(BMI), male sex, this model seeks to offer a more practical tool for clinical decision-making. Methods We queried ACTION-ACS a large-scale multicenter database of 3335 patients at 10-year follow-up. We employed a feature selection process informed by Random Forest Classifier to identify the most predictive variables. An ensemble model combining the strengths of Random Forest and Gradient Boosting methodologies was then developed. The model's predictive performance was validated through Stratified K-Fold cross-validation, focusing on Area Under the Curve (AUC) and confidence intervals to assess its efficacy. The model's predictive performance was validated through Stratified K-Fold cross-validation, focusing on Area Under the Curve (AUC) and confidence intervals to assess its efficacy. Results Demonstrating a mean AUC of 0.76 ± 0.036, "HeartGuardAI" showed a high capability to predict HF risk in post-ACS patients and emphasizes the significance of incorporating multifactorial clinical assessments Conclusions HF risk can be predicted by a handy tool developed with advanced machine learning methods. Future studies are needed to prospectively validate this algorithm.Graphic 1Graphic 2

Single-cell and bulk transcriptomic datasets enable the development of prognostic models based on dynamic changes in the tumor immune microenvironment in patients with hepatocellular carcinoma and portal vein tumor thrombus

BackgroundHepatocellular carcinoma (HCC) patients exhibiting portal vein tumor thrombosis (PVTT) face a high risk of rapid malignant progression and poor outcomes, with this issue being compounded by a lack of effective treatment options. The integration of bulk RNA-sequencing (RNA-seq) and single-cell RNA-seq (scRNA-seq) datasets focused on samples from HCC patients with PVTT has the potential to yield unprecedented insight into the dynamic changes in the tumor microenvironment (TME) and associated immunological characteristics in these patients, providing an invaluable tool for the reliable prediction of disease progression and treatment responses.MethodsscRNA-seq data from both primary tumor (PT) and PVTT cells were downloaded from the Gene Expression Omnibus (GEO) database, while the International Cancer Genome Consortium (ICGC) and Cancer Genome Atlas (TCGA) databases were used to access bulk RNA-seq datasets. scRNA-seq, clustering, GSVA enrichment, mutational profiling, and predictive immunotherapeutic treatment analyses were conducted using these data with the goal of systematically assessing the heterogeneity of PT and PVTT cells and establishing a model capable of predicting immunotherapeutic and prognostic outcomes in patients with HCC.ResultsThese analyses revealed that PVTT cells exhibited patterns of tumor proliferation, stromal activation, and low levels of immune cell infiltration, presenting with immune desert and immune rejection-like phenotypes. PT cells, in contrast, were found to exhibit a pattern of immunoinflammatory activity. Core PVTT-associated genes were clustered into three patterns consistent with the tumor immune rejection and immune desert phenotypes. An established clustering model was capable of predicting tumor inflammatory stage, subtype, TME stromal activity, and patient outcomes. PVTT signature genes were further used to establish a risk model, with the risk scores derived from this model providing a tool to evaluate patient clinicopathological features including clinical stage, tumor differentiation, histological subtype, microsatellite instability status, and tumor mutational burden. These risk scores were also able to serve as an independent predictor of patient survival outcomes, responses to adjuvant chemotherapy, and responses to immunotherapy. In vitro experiments were used to partially validate the biological prediction results.ConclusionThese results offer new insight into the biological and immunological landscape of PVTT in HCC patients, By utilizing individual patient risk scores, providing an opportunity to guide more effective immunotherapeutic interventional efforts.

Coronary inflammation stratifies risk in patients undergoing coronary CT angiography: the Oxford Risk Factors and Non-Invasive Imaging (ORFAN) study

Abstract Background Patients with chest pain symptoms often undergo coronary CT angiography (CCTA) for the diagnosis of obstructive coronary artery disease (CAD) to guide revascularisation. However, acute cardiac events could occur in the absence of obstructive CAD. Further risk stratification of patients could potentially be useful by quantifying coronary inflammation using the Fat Attenuation Index (FAI) Score, and an artificial intelligence (AI)-assisted prognostic model that captures the inflammatory risk by integrating CCTA-derived metrics (including FAI Score and plaque burden) and the patient’s clinical risk factors. Purpose To evaluate the risk of cardiac events among patients without obstructive CAD, and the prognostic value of integrating coronary inflammation in risk stratifying this patient group. Methods From the Oxford Risk Factors and Non-invasive imaging (ORFAN) study, patients undergoing CCTA in 7 UK Hospitals (n=40,091) were followed up for the incidence of cardiac mortality over a median(IQR) of 2.7(1.4-5.3) years. In a nested cohort of 3,393 patients with long-term follow up, we evaluated the long-term prognostic value of FAI Score and the performance of AI-Risk algorithm over a median(IQR) follow up of 7.7(6.4-9.1) years. Reclassification of risk category from AI-Risk was compared with contemporary clinical risk score (QRISK3). Results Patients without obstructive CAD (n=32,533, 81.1%) accounted for 63.7% of all cardiac deaths (n=1,754) during follow up, highlighting the substantial proportion of cardiac events in absolute terms given a large proportion of patients without obstructive CAD. In the nested cohort with long-term follow up, FAI Score in any of the 3 epicardial coronary arteries was significantly associated with cardiac mortality with or without obstructive CAD. In the whole cohort, the highest quartile of FAI Score in LAD showed 20-times higher risk of cardiac death compared with the lowest quartile (Panel A). Integrating coronary inflammation with clinical risk factors and plaque burden, patients with very-high AI-Risk classification showed 6-fold higher risk of cardiac mortality compared to low/medium risk, with good alignment of the incident predicted events from AI-Risk and observed events. (Panel B) The predicted cardiac mortality from AI-Risk also demonstrated good calibration with the observed events in the whole population. (Panel C) Finally, AI-Risk significantly reclassify the risk above clinical risk factor-based prediction (QRISK3) (Panel D). Conclusion Patients with high coronary inflammation measured by FAI Score showed significantly elevated risk for cardiac mortality. The AI-enabled absolute risk prediction algorithm leads to significant reclassification of patients undergoing routine CCTA, and can be used as a precision medicine tool.

Prospective and External Validation of Machine Learning Models for Short- and Long-Term Mortality in Acutely Admitted Patients Using Blood Tests

Background: Predicting mortality in emergency departments (EDs) using machine learning models presents challenges, particularly in balancing simplicity with performance. This study aims to develop models that are both simple and effective for predicting short- and long-term mortality in ED patients. Our approach uses a minimal set of variables derived from one single blood sample obtained at admission. Methods: Data from three cohorts at two large Danish university hospitals were analyzed, including one retrospective and two prospective cohorts where prognostic models were applied to predict individual mortality risk, spanning the years 2013–2022. Routine biochemistry analyzed in blood samples collected at admission was the primary data source for the prediction models. The outcomes were mortality at 10, 30, 90, and 365 days after admission to the ED. The models were developed using Light Gradient Boosting Machines. The evaluation of mortality predictions involved metrics such as Area Under the Receiver Operating Characteristic Curve (AUC), sensitivity, specificity, negative predictive values, positive predictive values, and Matthews correlation coefficient (MCC). Results: A total of 43,648 unique patients with 65,484 admissions were analyzed. The models showed high accuracy, with very good to excellent AUC values between 0.87 and 0.93 across different time intervals. Conclusions: This study demonstrates that a single assessment of routine clinical biochemistry upon admission can serve as a powerful predictor for both short-term and long-term mortality in ED admissions.

Synthetic data for privacy-preserving clinical risk prediction

Synthetic data promise privacy-preserving data sharing for healthcare research and development. Compared with other privacy-enhancing approaches—such as federated learning—analyses performed on synthetic data can be applied downstream without modification, such that synthetic data can act in place of real data for a wide range of use cases. However, the role that synthetic data might play in all aspects of clinical model development remains unknown. In this work, we used state-of-the-art generators explicitly designed for privacy preservation to create a synthetic version of ever-smokers in the UK Biobank before building prognostic models for lung cancer under several data release assumptions. We demonstrate that synthetic data can be effectively used throughout the medical prognostic modeling pipeline even without eventual access to the real data. Furthermore, we show the implications of different data release approaches on how synthetic biobank data could be deployed within the healthcare system.

The pro-tumor activity of INTS7 on lung adenocarcinoma via inhibiting immune infiltration and activating p38MAPK pathway

Lung adenocarcinoma (LUAD) is the most common lung cancer, accounting for 19.4% of all cancer deaths. Our previous study discovered that INTS7 expression was upregulated in LUAD, while the precise mechanism by which INTS7 exerts pro-cancer effects remains unknown. In our study, shRNA was used to knockdown the expression of INTS7 in A549 cells. Cancer behaviors in vitro were determined by CCK8 and transwell assays. Xenograft mice models were constructed to detect the tumorigenesis in vivo. Immunofluorescence and toluidine blue staining were used to test the immune infiltration. Bioinformatics analysis was adopted to predict the potential signaling pathways and construct INTS7-derived genomic prognostic model. Western blot was utilized to confirm the molecular pathways. In total, downregulation of INTS7 suppressed proliferation, invasion and migration of A549 cells, as well as tumor growth. Bioinformatics and western blot analysis indicated that p38MAPK pathway participated in the regulatory mechanism of INTS7. Moreover, INTS7 expression was negatively correlated with infiltration of memory B cells and mast cells, while positively correlated with infiltration of macrophages M2. A nomogram, including INTS7-derived risk score, was used to estimate individual’s survival probability. Generally, our findings provided comprehensive understanding of the molecular mechanisms about INTS7, and targeting INTS7 may represent a potential therapy for LUAD.

Identification and validation of a prognostic model based on three TLS-Related genes in oral squamous cell carcinoma

BackgroundThe tertiary lymphoid structures (TLSs) have an immunomodulatory function and have a positive impact on the survival outcomes of patients with oral squamous cell carcinoma (OSCC). However, there is a lack of standard approaches for quantifying TLSs and prognostic models using TLS-related genes (TLSRGs). These limitations limit the widespread use of TLSs in clinical practice.MethodsA convolutional neural network was used to automatically detect and quantify TLSs in HE-stained whole slide images. By employing bioinformatics and diverse statistical methods, this research created a prognostic model using TCGA cohorts and explored the connection between this model and immune infiltration. The expression levels of three TLSRGs in clinical specimens were detected by immunohistochemistry. To facilitate the assessment of individual prognostic outcomes, we further constructed a nomogram based on the risk score and other clinical factors.ResultsTLSs were found to be an independent predictor of both overall survival (OS) and disease-free survival in OSCC patients. A larger proportion of the TLS area represented a better prognosis. After analysis, we identified 69 differentially expressed TLSRGs and selected three pivotal TLSRGs to construct the risk score model. This model emerged as a standalone predictor for OS and exhibited close associations with CD4 + T cells, CD8 + T cells, and macrophages. Immunohistochemistry revealed high expression levels of CCR7 and CXCR5 in TLS + OSCC samples, while CD86 was highly expressed in TLS- OSCC samples. The nomogram demonstrates excellent predictive ability for overall survival in OSCC patients.ConclusionsThis is the first prognostic nomogram based on TLSRGs, that can effectively predict survival outcomes and contribute to individual treatment strategies for OSCC patients.