ABSTRACT: This is a multi-center prospective case control study aiming to compare different methods of risk stratification models in predicting the risk of gastric cancer development.
Project description:Background Bladder cancer represents a heterogeneous disease with distinct clinical challenges. Non-muscle invasive bladder cancer (NMIBC) typically presents as indolent and slow-growing, yet a critical clinical challenge remains: identifying which patients will progress to muscle-invasive disease requiring radical interventions. Early detection of progression propensity is essential, as once muscle invasion occurs, the risk of distant metastasis increases substantially, and treatment shifts from conservative TURBT (Transurethral Resection of Bladder Tumor) to aggressive surgical interventions with significant morbidity. Current risk stratification methods fail to adequately predict this transition in approximately 30% of cases, highlighting the urgent need for more accurate prognostic tools. Objective This retrospective study aimed to develop and validate a transcriptomics-based mRNA score for predicting early NMIBC recurrence, comparing its performance against traditional risk stratification methods. Methods We analyzed mRNA expression profiles from primary retrospective NMIBC tumor specimens (n=25) collected between [2018-2022]. Traditional risk stratification tools, including EORTC scoring, were applied alongside our novel mRNA-based risk score to evaluate predictive accuracy for recurrence. Results The transcriptomics-based mRNA score demonstrated a median prediction accuracy of 90% across 10,000 resampling iterations for predicting early NMIBC recurrence, significantly outperforming traditional EORTC risk scores. Our comprehensive gene set identified 435 differentially expressed genes associated with recurrence. Kaplan-Meier analysis showed significantly different recurrence-free survival between high and low mRNA risk score groups (Bonferroni corrected p-value<0.0001). Conclusions This retrospective analysis confirms that mRNA expression-based risk stratification provides superior predictive accuracy compared to conventional clinicopathologic risk tools. Implementation of this gene signature could potentially reduce over-investigation and improve surveillance cost-effectiveness after TURBT in patients with primary high-risk NMIBC. These findings may transform the clinical management paradigm by enabling more personalized follow-up protocols based on molecular risk assessment.
Project description:The project will aim to identify and determine subgroups of patients with different risks of progression to gastric cancer and to assess appropriate follow-up intervals. Implementing risk stratification only high risk individuals will be offered and performed endoscopic surveillance.
Project description:Gastric cancer (GC) is associated with high mortality rates and an unfavorable prognosis at advanced stages. In addition, there are no effective methods for diagnosing gastric cancer at an early stage or for predicting the outcome for the purpose of selecting patient-specific treatment options. Therefore, it is important to investigate new methods for GC diagnosis. We designed a custom microarray of gastric cancer. The customized microarray contained 1042 canceration and prognosis related genes identical to the probes on the Agilent microarray. DNA microarray profilling analysis was performed on gastric cancer tissues and premalignant tissues (20 samples per group).
Project description:Background: Follicular lymphoma (FL) is an indolent malignancy of germinal center B cells with highly variable patient outcomes. Recently, a 23-gene predictor score was proposed for predicting progression-free survival. In the past, we had shown that the m7-FLIPI, a clinico-genetic risk model, allows to improve patient stratification compared to clinical risk models alone. The multitude of prognostic tools in FL raises the question whether they identify common biology. Methods: In this study, we applied a modified risk score (MRS) to an independent gene-expression dataset of FL patients treated with rituximab in combination with chemotherapy. Results: Using supervised and unsupervised approaches, we showed that the MRS identifies patient groups with diverging outcomes in our dataset. In addition, using gene set enrichment and network identification, we discovered associations between the MRS, the m7-FLIPI, EZH2 mutation status and FOXP1 expression. Conclusions: Our findings lend support to expression of dark-zone related genes as a key determinant of poor outcome following rituximab and chemotherapy.
Project description:Revised risk estimation and treatment stratification of low- and intermediate-risk neuroblastoma patients by integrating clinical and molecular prognostic markers. To optimize neuroblastoma treatment stratification, we aimed at developing a novel risk estimation system by integrating gene expression-based classification and established prognostic markers. Gene expression profiles were generated from 709 neuroblastoma specimens using customized 4x44K microarrays. Classification models were built using 75 tumors with contrasting courses of disease. Validation was performed in an independent test set (n=634) by Kaplan-Meier estimates and Cox regression analyses. Combination of gene expression-based classification and established prognostic markers improves risk estimation of LR/IR neuroblastoma patients. We propose to implement our revised treatment stratification system in a prospective clinical trial.
Project description:The risk of locoregional or distant failure in advanced HPV-negative head and neck squamous cell carcinoma (HNSCC) patients is high. However, no suitable markers for stratification are clinically available. Thus, we aimed to identify a microRNA(miRNA)-signature predicting disease recurrence. For this purpose the miRNA profiles from 162 HNSCC samples were analysed with regard to identification of a low-complex porgnostic signature. The data set consists of a discovery dataset (n=85) and a validation dataset (n=77). The study resulted in a prognostic 5-miRNA signature significantly predicting the relevant clinical endpoint freedom from recurrence.
Project description:The clinical course of Coronavirus disease 2019 (COVID-19) displays a wide variability, ranging from completely asymptomatic forms to diseases associated with severe clinical outcomes. To reduce the incidence COVID-19 severe outcomes, innovative molecular biomarkers are needed to improve the stratification of patients at the highest risk of mortality and to better customize therapeutic strategies. MicroRNAs associated with COVID-19 outcomes could allow quantifying the risk of severe outcomes and developing models for predicting outcomes, thus helping to customize the most aggressive therapeutic strategies for each patient. Here, we analyzed the circulating miRNA profiles in a set of 12 hospitalized patients with severe COVID-19, with the aim to identify miRNAs associated with in-hospital mortality.
Project description:The paper "Metabolomic Machine Learning Predictor for Diagnosis and Prognosis of Gastric Cancer" addresses the need for non-invasive diagnostic tools for gastric cancer (GC). Traditional methods like endoscopy are invasive and expensive. The authors conducted a targeted metabolomics analysis of 702 plasma samples to develop machine learning models for GC diagnosis and prognosis. The diagnostic model, using 10 metabolites, achieved a sensitivity of 0.905, outperforming conventional protein marker-based methods. The prognostic model effectively stratified patients into risk groups, surpassing traditional clinical models.
I have successfully reproduced the diagnosis model from the paper. This machine learning-based system differentiates GC patients from non-GC controls using metabolomics data from plasma samples analyzed by liquid chromatography-mass spectrometry (LC-MS). The model focuses on 10 metabolites, including succinate, uridine, lactate, and serotonin. Employing LASSO regression and a random forest classifier, the model achieved an AUROC of 0.967, with a sensitivity of 0.854 and specificity of 0.926. This model significantly outperforms traditional diagnostic methods and underscores the potential of integrating machine learning with metabolomics for early GC detection and treatment.