Computed Tomography-Based Radiomics to Predict FOXM1 Expression and Overall Survival in Patients with Clear Cell Renal Cell Carcinoma.
Acad Radiol
; 2024 Mar 12.
Article
in En
| MEDLINE
| ID: mdl-38480074
ABSTRACT
RATIONALE AND OBJECTIVES:
To establish a computed tomography (CT)-based radiomics model to predict Fork head box M1(FOXM1) expression levels and develop a combined model for prognostic prediction in patients with clear cell renal cell carcinoma (ccRCC). MATERIALS ANDMETHODS:
A total of 529 patients were utilized to assess the prognostic significance of FOXM1 expression and were subsequently categorized into low and high FOXM1 expression groups. 184 patients with CT images were randomly divided into training and validation cohorts. Radiomics signature (Rad-score) for predicting FOXM1 expression level was developed in the training cohort. The predictive performance was evaluated using receiver operating characteristic (ROC) curves. A clinical model based on clinical factors and a combined model incorporating clinical factors and Rad-score were developed to predict ccRCC prognosis using Cox regression analyses. The concordance index(C-index) was employed to assess and compare the predictive capabilities of the Rad-score, TNM stage, clinical model, and combined model. The likelihood ratio test was used to compare the models' performance.RESULTS:
The Rad-score demonstrated high predictive accuracy for high FOXM1 expression with areas under the ROC curves of 0.713 and 0.711 in the training and validation cohorts. In the training cohort, the C-indexes for the Rad-score, TNM Stage, clinical model, and combined model were 0.657, 0.711, 0.737, and 0.741, respectively. Correspondingly, in the validation cohort, the C-indexes were 0.670, 0.712, 0.736, and 0.745. The combined model had the highest C-index, significantly outperforming the other models.CONCLUSION:
The Rad-score accurately predicts FOXM1 expression levels and is an independent prognostic factor for ccRCC.
Full text:
1
Collection:
01-internacional
Database:
MEDLINE
Language:
En
Journal:
Acad Radiol
Journal subject:
RADIOLOGIA
Year:
2024
Document type:
Article
Affiliation country:
China
Country of publication:
Estados Unidos