Epidemiological studies have shown an increasing trend in the incidence of papillary thyroid cancer (PTC). Therefore, the question of how to make clinical decisions when an investigation reveals atypical or suspicious cells arises. The aim of this study was to develop a nomogram for assessment of the individual risk of malignancy of thyroid nodules based on clinical, biochemical and ultrasonographic indicators. This retrospective study analyzed data from 2993 patients who had undergone thyroid surgery, with patients randomly allocated to a training cohort ( n = 2095) and a validation cohort ( n = 898) at a ratio of 7:3. Predictor selection was conducted through a two-step process: first, univariate analysis identified variables significantly associated with PTC ( P < 0.05). Subsequently, these significant variables were entered into a multivariate logistic regression analysis to determine independent risk factors. The final nomogram prediction model was constructed based on the coefficients of these independent factors. The model’s performance was rigorously evaluated via internal validation. Discrimination was assessed by the area under the receiver operating characteristic curve (AUC), calibration was evaluated using calibration curves accompanied by the Hosmer-Lemeshow goodness-of-fit test, and clinical utility was examined through decision curve analysis (DCA). Univariate analysis selected twenty-two variables that may be factors affecting the occurrence of PTC ( P < 0.05). Multivariate analysis determined younger age (odd ratio OR = 0.956, 95% confidence interval CI, 0.945–1.574, P < 0.001), drinking (OR = 1.862, 95%CI, 1.013–3.469, P = 0.047), enlarged cervical lymph nodes (ECLN) (OR = 1.586, 95%CI, 1.116–2.283, P = 0.012), hypoechoic (OR = 29.07, 95%CI, 14.572–66.880, P < 0.001), irregular shape (OR = 6.838, 95%CI, 4.945–9.534, P < 0.001), calcification (OR = 1.583, 95%CI, 1.124–2.233, P = 0.008), higher thyroid-stimulating hormone (TSH)(OR = 1.141, 95%CI, 1.053–1.237, P = 0.001) and lower serum potassium (K) (OR = 0.478, 95%CI, 0.327–0.695, P < 0.001) were finalized as risk factors for PTC ( P < 0.05). A nomogram of PTC was constructed based on influencing factors. The nomogram underwent internal validation, which showed good discrimination in the training and validation groups, with an area under the curve (AUC) of 0.869 (95% CI, 0.853–0.886) and 0.872 (95% CI, 0.847–0.897), respectively. The calibration of the prediction model was evaluated by the Hosmer-Lemeshow goodness-of-fit test; the Hosmer-Lemeshow goodness-of-fit test values were P = 0.923 and P = 0.608, respectively. This indicates that the model aligns well with the observed data. The clinical utility assessed by decision curve analysis (DCA) demonstrates that the nomogram provides a superior net benefit across a wide range of threshold probabilities compared to alternative strategies. This indicates that the nomogram is a clinically useful tool for selecting patients who would benefit from intervention. A nomogram capable of accurately quantifying the risk of papillary thyroid carcinoma in thyroid nodules was developed by integrating clinical, biochemical, and ultrasonographic indicators. This predictive tool is designed to assist clinicians in optimizing clinical decision-making and avoiding unnecessary invasive procedures.
Li et al. (Thu,) studied this question.