KM Research Hub
← 返回论文库
无 KM 图 入排:待复核 图源:可得 入排置信度 0.45

Variable Selection via Fused Sparse-Group Lasso Penalized Multi-state Models Incorporating Molecular Data.

Kaya Miah, Jelle J Goeman, Hein Putter, Annette Kopp-Schneider, Axel Benner

Biom J 67(6) e70087

PMID41146443
PMCIDPMC12559784
DOI10.1002/bimj.70087
发表日期2025-10-28
入排依据:While the trial registration and publication metadata suggest an oncology Phase III randomized trial, the abstract describes a methodological/statistical analysis (proposing a variable selection algorithm) applied to trial data, not a formal report of the trial's efficacy results. The presence of a drug intervention is not established in the abstract, and no primary time-to-event endpoint is reported.
已检查 14 张 PubMed Figure,均未识别为 KM 图:FIGURE 1:AI 判定为 flow diagram;FIGURE 2:AI 判定为 other;FIGURE 3:AI 判定为 other;FIGURE 4:AI 判定为 other;FIGURE 5:AI 判定为 other;另有 9 张未列出。
摘要

摘要

In multi-state models based on high-dimensional data, effective modeling strategies are required to determine an optimal, ideally parsimonious model. In particular, linking covariate effects across transitions is needed to conduct joint variable selection. A useful technique to reduce model complexity is to address homogeneous covariate effects for distinct transitions. We integrate this approach to data-driven variable selection by extended regularization methods within multi-state model building. We propose the fused sparse-group lasso (FSGL) penalized Cox-type regression in the framework of multi-state models combining the penalization concepts of pairwise differences of covariate effects along with transition-wise grouping. For optimization, we adapt the alternating direction method of multipliers (ADMM) algorithm to transition-specific hazards regression in the multi-state setting. In a simulation study and application to acute myeloid leukemia (AML) data, we evaluate the algorithm's ability to select a sparse model incorporating relevant transition-specific effects and similar cross-transition effects. We investigate settings in which the combined penalty is beneficial compared to global lasso regularization. Clinical Trial Registration: The AMLSG 09-09 trial is registered with ClinicalTrials.gov (NCT00893399) and has been completed.

KM 提取结果

0 个 Figure · 0 个 KM 面板

尚未提取到 KM 面板。
其他 Figure(14,默认折叠)

这些图片未产生 KM 曲线结果,不在主页面重复展示。

FIGURE 1FIGURE 1 State chart of the multi‐state model for acute myeloid leukemia (AML) with nine states and eight possible transitions represented by arrows. Numbers denote the corresponding transition.
非 KM Figure
FIGURE 2FIGURE 2 Tuning parameter selection results for FSGLmstate: Mean generalized cross‐validation (GCV) statistics across all preselected combinations of penalty parameters ( α , γ ) over n sim = 225 simulation runs. The pair ( α , γ ) = ( 1 , 1 ) corresponds to the global lasso penalty.
非 KM Figure
FIGURE 3FIGURE 3 Boxplots of estimated regression coefficients based on simulated data of the nine‐state AML model with eight transitions and two binary covariates. X 1.3 and X 1.7 as well as X 1.4 and X 1.8 refer to transitions with true equal effects of covariate X 1 . Covariate X 2 has no true effect on any transition. Dots depict estimated covariate effects based on λ ^ opt,L and λ ^ opt,FSGL of each simulated data set. True underlying covariate effects β true are denoted as crosses ( × ).
非 KM Figure
FIGURE 4FIGURE 4 Variable selection results in terms of true positive rates (TPRs) and false discovery rates (FDRs) for LASSOmstate and FSGLmstate. Dots illustrate TPR and FDR of each simulated data set.
非 KM Figure
FIGURE 5FIGURE 5 Mean bias and mean squared error (MSE) of estimating the non‐zero covariate effects along with 95% Monte Carlo confidence intervals (MC‐CI). Dots illustrate mean bias and mean MSE of a single simulated data set.
非 KM Figure
FIGURE 6FIGURE 6 Event counts of the multi‐state model for acute myeloid leukemia (AML) with nine states and eight transitions based on the AMLSG 09‐09 trial data. Gray numbers indicate the corresponding transition.
非 KM Figure
FIGURE 7FIGURE 7 Stacked transition probabilities to all states from randomization derived from the multi‐state model for acute myeloid leukemia (AML) based on the AMLSG 09‐09 trial data. The distance between two adjacent curves represents the probability of being in the corresponding state. CR: Complete remission.
非 KM Figure
FIGURE 8FIGURE 8 Estimated regression effects of clinical and mutation variables by FSGLmstate separately for each transition derived from the nine‐state model for acute myeloid leukemia (AML) based on the AMLSG 09‐09 trial data. Larger crosses ( × ) depict non‐zero effects.
非 KM Figure
FIGURE A1FIGURE A1 Scenario B: Boxplots of estimated regression coefficients based on simulated data of the nine‐state AML model with eight transitions and P = 10 binary covariates. Dots depict estimated covariate effects based on λ ^ opt,L and λ ^ opt,FSGL of each simulated data set. True underlying covariate effects β true are denoted as crosses ( × ).
非 KM Figure
FIGURE A2FIGURE A2 Scenario B: Variable selection results in terms of true positive rates (TPR) and false discovery rates (FDR) for LASSOmstate and FSGLmstate. Dots illustrate TPR and FDR of each simulated data set.
非 KM Figure
FIGURE A3FIGURE A3 Scenario B: Mean bias and mean squared error (MSE) of estimating the non‐zero covariate effects along with 95% Monte Carlo confidence intervals (MC‐CI). Dots illustrate mean bias and mean MSE of each simulated data set.
非 KM Figure
FIGURE A4FIGURE A4 Scenario C: Boxplots of estimated regression coefficients based on simulated data of the nine‐state AML model with eight transitions and P = 10 binary covariates. Dots depict estimated covariate effects based on λ ^ opt,L and λ ^ opt,FSGL of each simulated data set. True underlying covariate effects β true are denoted as crosses ( × ).
非 KM Figure
FIGURE A5FIGURE A5 Scenario C: Variable selection results in terms of true positive rates (TPR) and false discovery rates (FDR) for LASSOmstate and FSGLmstate. Dots illustrate TPR and FDR of each simulated data set.
非 KM Figure
FIGURE A6FIGURE A6 Scenario C: Mean bias and mean squared error (MSE) of estimating the non‐zero covariate effects along with 95% Monte Carlo confidence intervals (MC‐CI). Dots illustrate mean bias and mean MSE of each simulated data set.
非 KM Figure