¹⁸F-FDG PET radiomics for predicting progression-free survival in locally advanced cervical squamous cell carcinoma

To develop a fluorine-18-fluorodeoxyglucose (¹⁸F-FDG) positron emission tomography (PET) radiomics-based nomogram model for predicting progression-free survival (PFS) in locally advanced cervical squamous cell carcinoma (LACSC) patients undergoing concurrent chemoradiotherapy (CCRT).

A retrospective study included 241 LACSC patients treated with CCRT, divided into training (n=168) and validation (n=73) sets. Lesion segmentation, radiomics feature extraction and screening were performed on ¹⁸F-FDG PET images of each patient, and radiomics scores (Rad-scores) were calculated.Univariate and multivariate Cox regression analyses were used to identify independent prognostic factors and create a combined model and nomogram. Predictive performance was assessed using time-dependent receiver operating characteristic (ROC) curves, area under the curve (AUC), and consistency index (C-index). Calibration curves evaluated nomogram accuracy, and decision curve analysis (DCA) assessed nomogram clinical applicability.

The Rad-score calculated from five optimal radiomics features and FIGO stage were independent predictors of PFS in LACSC patients.The C-index values for the FIGO stage, Rad-score, and combined model were 0.586, 0.692, and 0.727 in the training set, and 0.612, 0.668, and 0.698 in the validation set, respectively. The combined model showed excellent predictive ability for PFS at 12, 18, and 24 months, with training set AUCs of 0.805, 0.738, and 0.719, and validation set AUCs of 0.670, 0.744, and 0.741, respectively. The calibration curves confirmed a good agreement between predicted and actual progression probabilities, with DCA revealing significant clinical net benefits.

The ¹⁸F-FDG PET radiomics-based nomogram effectively predicted PFS in LACSC patients and could support individualized treatment decisions and accurate prognostic evaluations.