Development and validation of an EUS-based nomogram for prediction of the malignant potential in gastrointestinal stromal tumors

Abstract

Background

Gastrointestinal stromal tumors (GISTs) are the most common mesenchymal tumors in the gastrointestinal (GI) tract that require different therapeutic interventions according to the malignancy. We aim to develop and validate a EUS (endoscopic ultrasonography)-based nomogram to predict malignant potential in patients with GIST.

Methods

258 patients with pathological diagnosis of gastric GISTs were enrolled retrospectively in our hospital from June 2015 to October 2020. Patients were randomly divided into the development cohort (DC, n = 179) and the validation cohort (VC, n = 79). We established a nomogram using lasso regression based on DC data. The predictive effectiveness of the nomogram was evaluated by the area under the receiver operating characteristic curve (AUC). Through bootstrapping, a consistency index (C-index) and calibration chart are developed to evaluate the reliability and accuracy of the nomogram.

Results

A total of 192 patients with low-malignant potential (very low and low-risk) GISTs and 66 patients with high-malignant potential (intermediate and high-risk) GISTs were included in this study. The nomogram was constructed with the following 6 EUS indicators: ulceration, hemorrhage, tumor shape, irregular border, transverse diameter, and necrosis. Internal and external validation showed that the nomogram had a good ability to predict the malignant potential of GISTs (AUC = 0.881 and 0.908, respectively). The calibration curve shows that the nomogram has a good agreement between predicted and actual probabilities for differentiating GISTs malignancy (C-index = 0.868 and 0.907, respectively).

Conclusions

This study developed and verified a EUS-based nomogram, which can effectively predict the malignant potential of patients with gastric GISTs.

Keywords:

• Gastrointestinal stromal tumors
• endoscopic ultrasonography
• nomogram
• prediction
• risk factor

Author contributions

(I) Conception and design: Xiaobin Sun; (II) Administrative support: Xiaobin Sun; (III) Provision of study materials or patients: Li Liu and Jing Chen; (IV) Collection and assembly of data: Jing Shan; (V) Data analysis and interpretation: Jing Shan and Li Liu; (VI) Manuscript writing: All authors; (VII) Final approval of manuscript: All authors.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the Foundation of Science and Technology Department of Sichuan province (2020YJ0485).