Prognosis of lasso-like penalized Cox models with tumor profiling improves prediction over clinical data alone and benefits from bi-dimensional pre-screening

Jardillier, Rémy; Koca, Dzenis; Chatelain, Florent; Guyon, Laurent

doi:10.1186/s12885-022-10117-1

Cited by 11 publications

(10 citation statements)

References 72 publications

(100 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…LASSO-penalized Cox regression is a method that combines LASSO regularization with the Cox proportional hazards model. It aims to identify a subset of features that are most important for survival analysis while minimizing overfitting [ 54 ].…”

Section: Resultsmentioning

confidence: 99%

Assessing Metabolic Markers in Glioblastoma Using Machine Learning: A Systematic Review

et al. 2023

View full text Add to dashboard Cite

Glioblastoma (GBM) is a common and deadly brain tumor with late diagnoses and poor prognoses. Machine learning (ML) is an emerging tool that can create highly accurate diagnostic and prognostic prediction models. This paper aimed to systematically search the literature on ML for GBM metabolism and assess recent advancements. A literature search was performed using predetermined search terms. Articles describing the use of an ML algorithm for GBM metabolism were included. Ten studies met the inclusion criteria for analysis: diagnostic (n = 3, 30%), prognostic (n = 6, 60%), or both (n = 1, 10%). Most studies analyzed data from multiple databases, while 50% (n = 5) included additional original samples. At least 2536 data samples were run through an ML algorithm. Twenty-seven ML algorithms were recorded with a mean of 2.8 algorithms per study. Algorithms were supervised (n = 24, 89%), unsupervised (n = 3, 11%), continuous (n = 19, 70%), or categorical (n = 8, 30%). The mean reported accuracy and AUC of ROC were 95.63% and 0.779, respectively. One hundred six metabolic markers were identified, but only EMP3 was reported in multiple studies. Many studies have identified potential biomarkers for GBM diagnosis and prognostication. These algorithms show promise; however, a consensus on even a handful of biomarkers has not yet been made

show abstract

Section: Resultsmentioning

confidence: 99%

Assessing Metabolic Markers in Glioblastoma Using Machine Learning: A Systematic Review

et al. 2023

View full text Add to dashboard Cite

show abstract

“…In a previous work, we advised equally Lasso, Elastic Net and Ridge as penalization methods, as they provide equal model performance [ 8 ]. The Lasso selects fewer genes, leading to a more parsimonious model.…”

Section: Methodsmentioning

confidence: 99%

“…Classically, in the case of high-dimensional datasets, a penalty term is used to constrain the coefficients of the model, and to select only a subset of genes. Different forms of penalties exist [ 6 ], but we will focus on the elastic net penalty [ 7 ] in this paper, as we have recently shown that they provide similar performances [ 8 ]. More recently, non-parametric machine learning algorithms have been proposed and adapted to deal with survival data, including random survival forest [ 9 ].…”

Section: Introductionmentioning

confidence: 99%

Optimal microRNA Sequencing Depth to Predict Cancer Patient Survival with Random Forest and Cox Models

Jardillier

Koca

Chatelain

et al. 2022

Genes

Self Cite

View full text Add to dashboard Cite

(1) Background: tumor profiling enables patient survival prediction. The two essential parameters to be calibrated when designing a study based on tumor profiles from a cohort are the sequencing depth of RNA-seq technology and the number of patients. This calibration is carried out under cost constraints, and a compromise has to be found. In the context of survival data, the goal of this work is to benchmark the impact of the number of patients and of the sequencing depth of miRNA-seq and mRNA-seq on the predictive capabilities for both the Cox model with elastic net penalty and random survival forest. (2) Results: we first show that the Cox model and random survival forest provide comparable prediction capabilities, with significant differences for some cancers. Second, we demonstrate that miRNA and/or mRNA data improve prediction over clinical data alone. mRNA-seq data leads to slightly better prediction than miRNA-seq, with the notable exception of lung adenocarcinoma for which the tumor miRNA profile shows higher predictive power. Third, we demonstrate that the sequencing depth of RNA-seq data can be reduced for most of the investigated cancers without degrading the prediction abilities, allowing the creation of independent validation sets at a lower cost. Finally, we show that the number of patients in the training dataset can be reduced for the Cox model and random survival forest, allowing the use of different models on different patient subgroups.

show abstract

“…Clinical and biochemical markers, such as stage and grade of tumor, are often used in the assessment of ccRCC prognosis [19]. Recently, we have shown that, for most cancers, survival prediction can be improved by incorporating clinical and transcriptomic data [20]. Additionally, there is an increasing interest in how ECM-related genes could be used to assess the survival of ccRCC patients [21][22][23][24].…”

Section: Introductionmentioning

confidence: 99%

COL7A1 Expression Improves Prognosis Prediction for Patients with Clear Cell Renal Cell Carcinoma Atop of Stage

Koca

Séraudie

Jardillier

et al. 2023

Cancers

Self Cite

View full text Add to dashboard Cite

Clear-cell renal cell carcinoma (ccRCC) accounts for 75% of kidney cancers. Due to the high recurrence rate and treatment options that come with high costs and potential side effects, a correct prognosis of patient survival is essential for the successful and effective treatment of patients. Novel biomarkers could play an important role in the assessment of the overall survival of patients. COL7A1 encodes for collagen type VII, a constituent of the basal membrane. COL7A1 is associated with survival in many cancers; however, the prognostic value of COL7A1 expression as a standalone biomarker in ccRCC has not been investigated. With five publicly available independent cohorts, we used Kaplan–Meier curves and the Cox proportional hazards model to investigate the prognostic value of COL7A1, as well as gene set enrichment analysis to investigate genes co-expressed with COL7A1. COL7A1 expression stratifies patients in terms of aggressiveness, where the 5-year survival probability of each of the four groups was 72.4%, 59.1%, 34.15%, and 8.6% in order of increasing expression. Additionally, COL7A1 expression was successfully used to further divide patients of each stage and histological grade into groups of high and low risk. Similar results were obtained in independent cohorts. In vitro knockdown of COL7A1 expression significantly affected ccRCC cells’ ability to migrate, leading to the hypothesis that COL7A1 may have a role in cancer aggressiveness. To conclude, we identified COL7A1 as a new prognosis marker that can stratify ccRCC patients.

show abstract

Prognosis of lasso-like penalized Cox models with tumor profiling improves prediction over clinical data alone and benefits from bi-dimensional pre-screening

Cited by 11 publications

References 72 publications

Assessing Metabolic Markers in Glioblastoma Using Machine Learning: A Systematic Review

Assessing Metabolic Markers in Glioblastoma Using Machine Learning: A Systematic Review

Optimal microRNA Sequencing Depth to Predict Cancer Patient Survival with Random Forest and Cox Models

COL7A1 Expression Improves Prognosis Prediction for Patients with Clear Cell Renal Cell Carcinoma Atop of Stage

Contact Info

Product

Resources

About