Development of a Multilayered Prognostic Model for Wilms' Tumor Based on Characteristic Lymphocyte Genes

To develop a prognostic nomogram for Wilms' tumor (WT) integrating genetic and clinical factors to improve evaluation accuracy and clinical utility.

RNA sequencing (RNA-seq) data from 125 WT patients and single-cell RNA (scRNA-seq) data from 2437 samples were analyzed using bioinformatics tools for data processing, including normalization and scaling with SCTransform, and cell clustering with Seurat. Principal component analysis (PCA) and Uniform Manifold Approximation and Projection (UMAP) were utilized for data visualization. Differential gene expression analysis identified pivotal genes for the Genetic Feature Prognostic Model for WT (GPM-WT). Univariate Cox regression analysis refined this model by incorporating clinical prognostic indicators. Survival analysis, Cox regression, and ROC curve assessments evaluated these models' prognostic capabilities. Immune cell infiltration and drug sensitivity were quantified, linking these to patient risk categories.

Six prognostic lymphocyte genes (KLRC1, APOC2, GBP2, SLA, MLLT3, and SIGLEC5) were identified for GPM-WT. Clinical factors, age and sex, were integrated to refine the model. The Lymphocyte Gene and Clinical Features Prognostic Nomogram (LGCPN-WT) effectively distinguished high from low-risk groups, predicting 2-5-year survival rates with area under the curve (AUC) values of 0.771, 0.774, 0.751, and 0.785. Elevated immune cell infiltration and enhanced drug sensitivity characterized the high-risk group, exhibiting significant responsiveness to chemotherapy, targeted, and immunotherapy treatments (p < 0.05).

The study developed an integrated LGCPN-WT model, significantly enhancing survival prediction accuracy and clinical utility for WT, thus supporting personalized treatment approaches.