Related Experiment Video
Updated: Aug 22, 2025
Author Spotlight: Genetically Engineered Mouse Models and Pathological Characterization of Neurofibromatosis Type 1 Associated Tumors
Published on: May 17, 2024
Nomogram for preoperative estimation of histologic grade in gastrointestinal neuroendocrine tumors
Zhi-Qi Wu1,2, Yan Li3, Na-Na Sun4
1Department of Laboratory Medicine, The First Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu, China.
Background:
The treatment strategies and prognosis for gastroenteropancreatic neuroendocrine tumors were associated with tumor grade. Preoperative predictive grading could be of great benefit in the selection of treatment options for patients. However, there is still a lack of effective non-invasive strategies to detect gastrointestinal neuroendocrine tumors (GI-NETs) grading preoperatively.
Methods:
The data on 147 consecutive GI-NETs patients was retrospectively collected from January 1, 2012, to December 31, 2019. Logistic regression was used to construct a predictive model of gastrointestinal neuroendocrine tumor grading using preoperative laboratory and imaging parameters.The validity of the model was assessed by area under the receiver operating characteristic curve (AUC), calibration curve, and decision curve analysis (DCA).
Results:
The factors associated with GI-NETs grading were age, tumor size, lymph nodes, neuron-specific enolase (NSE), hemoglobin (HGB) and sex, and two models were constructed by logistic regression for prediction. Combining these 6 factors, the nomogram was constructed for model 1 to distinguish between G3 and G1/2, achieving a good AUC of 0.921 (95% CI: 0.884-0.965), and the sensitivity, specificity, accuracy were 0.9167, 0.8256, 0.8630, respectively. The model 2 was to distinguish between G1 and G2/3, and the variables were age, tumor size, lymph nodes, NSE, with an AUC of 0.847 (95% CI: 0.799-0.915), and the sensitivity, specificity, accuracy were 0.7882, 0.8710, 0.8231, respectively. Two online web servers were established on the basis of the proposed nomogram to facilitate clinical use. Both models showed an excellent calibration curve through 1000 times bootstrapped dataset and the clinical usefulness were confirmed using decision curve analysis.
Conclusion:
The model served as a valuable non-invasive tool for differentiating between different grades of GI-NETs, personalizing the calculation which can lead to a rational treatment choice.
