Radiomics features of diffusion kurtosis imaging and dynamic contrast enhanced-MRI for predicting aggressive pathological characteristics of oral and oropharyngeal squamous cell carcinoma.

Abstract

Objective: To investigate the predictive value of radiomics features derived from diffusion kurtosis imaging and dynamic contrast-enhanced MRI for aggressive pathological characteristics of oral and oropharyngeal squamous cell carcinoma (SCC).

Study design: This prospective single-center study enrolled 157 patients with pathologically confirmed oral and oropharyngeal SCC, randomly divided into a training set (n = 109) and an internal validation set (n = 48). Additionally, a temporal validation cohort comprising 18 consecutive patients enrolled from November 2022 to September 2023 was utilized to further evaluate the model's robustness. Radiomics features were extracted from 15 parameter maps. Least absolute shrinkage and selection operator (LASSO) regression was used for feature selection, and logistic regression models were constructed.

Results: For T stage prediction, the combined DKI and DCE model achieved AUC values of 0.930 in the training set and 0.818 in the internal validation set. For lymph node metastasis (LNM) prediction, the combined model yielded AUC values of 0.978 and 0.854, respectively. In the temporal validation cohort, the combined model maintained favorable predictive performance, with AUCs of 0.900 for T stage and 0.800 for LNM.

Conclusions: Radiomics models based on combined diffusion kurtosis imaging and dynamic contrast-enhanced MRI demonstrated favorable predictive performance for T stage and LNM in oral and oropharyngeal SCC. The successful validation in a temporal cohort suggests the potential clinical utility of these models as noninvasive decision-support tools.