Construction and analysis of a predictive model for the efficacy of ¹³¹I therapy in locally advanced or metastatic differentiated thyroid carcinoma

Objective To identify factors influencing the efficacy of ¹³¹I therapy in patients with locally advanced or metastatic differentiated thyroid carcinoma (DTC) and develop a predictive model to inform clinical prognosis evaluation.

Methods A retrospective analysis was conducted on 437 patients with locally advanced or metastatic DTC who underwent ¹³¹I therapy at Anhui Provincial Cancer Hospital between January 2016 and September 2023. The cohort included 168 males and 269 females, with an age of (46.8±0.7) years. Patients were categorized into effective and ineffective groups on the basis of treatment response criteria. Collected variables included gender, age, pathological type, lesion diameter, and serological markers such as stimulated thyroid-stimulating hormone (TSH), stimulated thyroglobulin (sTg), and stimulated thyroglobulin antibody (TgAb). The sTg/TSH ratio and sTg/TgAb ratio were calculated. Measurement data were compared using Mann-Whitney U test, and count data were compared using χ² test. Variables with significant differences in univariate analysis were entered into multivariate Logistic regression to identify independent predictors of treatment efficacy. A nomogram prediction model was developed using R4.4.2 software on the basis of these factors. Model performance was assessed using the receiver operating characteristic (ROC) curve, with the area under the curve (AUC) calculated. Calibration curves and Hosmer-Lemeshow test were used to evaluate goodness of fit.

Results The effective group comprised 269 patients, and the ineffective group had 168 patients. Significant intergroup differences were observed in lesion diameter (0.76 (0.50, 1.09) cm vs. 1.89 (1.50, 2.10) cm), sTg (3.46 (0.72, 10.55) ng/ml vs. 736.25 (85.15, 4 672.58) ng/ml), sTg/TSH ratio (0.06 (0.01, 0.19) vs. 32.10 (2.94, 212.10)), and sTg/TgAb ratio (0.31 (0.02, 1.27) vs. 104.16 (57.37, 136.41)) (Z=from –17.36 to –11.63, all P<0.001). Multivariate Logistic regression identified lesion diameter (OR=4.229, 95%CI: 1.161–15.400, P=0.029), sTg level (OR=1.236, 95%CI: 1.125–1.358, P<0.001), sTg/TSH ratio (OR=1.074, 95%CI: 1.015–1.137, P=0.014), and sTg/TgAb ratio (OR=1.028, 95%CI: 1.005–1.051, P=0.029) as independent predictors of ¹³¹I treatment efficacy. The nomogram model demonstrated an AUC of 0.881 (95%CI: 0.820–0.960, P=0.017), with a sensitivity of 79.6% and a specificity of 85.9%. The calibration curves showed good agreement between the predicted and observed outcomes, and the Hosmer-Lemeshow test indicated excellent goodness of fit (χ²=0.914, P=0.996).

Conclusions A predictive model for the efficacy of ¹³¹I therapy in locally advanced or metastatic DTC was successfully established on the basis of lesion diameter, sTg level, sTg/TSH ratio, and sTg/TgAb ratio. The model demonstrates favorable predictive efficacy, and it may serve as a useful tool for prognostic evaluation in patients with locally advanced or metastatic DTC.