Risk factors and prediction model for elevated ki-67 index in pituitary neuroendocrine tumors

Abstract

Objective To investigate the risk factors of elevated ki-67 index in patients with newly diagnosed pituitary neuroendocrine tumors (PitNETs) so as to construct a nomogram model for preoperative prediction of ki-67 index. Methods Preoperative clinical data and pathological results of tumor tissue samples of 449 patients with PitNETs who were initially diagnosed at our hospital from January 2015 to December 2019 were collected and retrospectively analyzed. All the patients were randomly enrolled at a ratio of 7:3 to either the training cohort (n=314) or the validation cohort (n=135). According to the level of ki-67 index, they were divided into low level (ki-67 < 3%) and high level (ki-67 ≥3%) groups, and the clinical data were compared between them. Multivariate LASSO regression and binary logistic regression analyses were adopted in the training cohort to identify independent risk factors of ki-67 index, and a nomogram prediction model was accordingly established. The predictive performance of the model was evaluated using receiver operating characteristic-area under the curve (ROC-AUC), calibration curve (obtained by Bootstrap method in 1000 internal validation), and decision curve analysis (DCA), followed by external validation in the validation cohort. Results Regression analyses showed that age (OR=0.97, 95%CI : 0.95~0.99), maximum tumor diameter (OR=1.56, 95%CI : 1.21~2.03), and free thyroxine (FT4; OR=0.93, 95%CI : 0.87~0.99) were the independent risk factors for the increase of ki-67 index. Then based on the above 3 variables, a nomogram model was constructed, whose ROC-AUC values in the training and validation cohorts were 0.692 (95%CI: 0.629 4~0.755 1) and 0.691 (95%CI : 0.591 3~0.790 1), respectively. The calibration curves indicated that the predicted value was in good conformity with the measured value, and the DCA showed a net return within a threshold range of 0.1~0.35, suggesting certain clinical significance of the model. Conclusion Age, maximum tumor diameter and FT4 are the independent risk factors for elevated ki-67 index in newly diagnosed PitNETs patients. Our constructed nomogram model possesses high differentiation and good calibration in predicting ki-67 index preoperatively, which is helpful for surgical protocol selection and postoperative management.

Keywords

• pituitary neuroendocrine tumors
• ki-67 index
• prediction model
• nomogram