SPECT/CT textural features applicability in differentiated thyroid cancer response assessment after radioiodine therapy

BACKGROUND : Despite the significant advancement of medical image analysis methods, including radiomics technologies, their routine integration into clinical practice remains limited. One promising direction is the use of textural features derived from hybrid imaging modalities, such as single-photon emission computed tomography combined with computed tomography (SPECT/CT). These features reflect the spatial characteristics of radiopharmaceutical distribution and can potentially be used to predict treatment response. OBJECTIVE: Assess the prognostic value of textural features extracted from post-therapeutic SPECT/CT images in evaluating the response to radioiodine therapy (RIT) in patients with differentiated thyroid cancer (DTC). MATERIALS AND METHODS : The study included 53 patients with DTC who underwent post-therapeutic SPECT/CT imaging 72 hours after administration of sodium iodine I-131. A total of 88 accumulation areas in the residual thyroid gland (thyroid gland) tissue and 61 metastatic foci of prostate cancer were analyzed. Disease status (remission or recurrence) was assessed six months after RIT based on clinical, laboratory, and imaging criteria. Logistic regression models and receiver operating characteristic (ROC) analysis were used to evaluate the predictive value of the extracted textural features. Feature selection was performed using mRmR, Lasso, and conventional statistical criteria. RESULTS : Diagnostic models based on textural features were developed and tested separately for residual thyroid tissue and metastatic DTC lesions. The model based on features from metastatic lesions demonstrated high predictive performance (AUC = 0.88), while the model based on residual thyroid tissue showed moderate prognostic value (AUC=0.61). CONCLUSION : This study demonstrates the feasibility of using radiomics based on SPECT/CT-derived I-131 uptake textural features to predict outcomes of radioiodine therapy in DTC. The application of these features may enhance the accuracy of recurrence risk stratification and contribute to more personalized treatment strategies