Abstract:

Objective To evaluate the clinical efficacy of endoscopic ultrasound combined with multi-band mucosectomy for the treatment of small gastric elevated lesions， to identify the risk factors for full-thickness resection （EFR） during the procedure， and to construct a risk prediction model. Methods Clinical data of 303 patients with a total of 336 small gastric elevated lesions who underwent endoscopic ultrasound examination and multi-band mucosectomy at our hospital from August 2015 to August 2025 were retrospectively analyzed. The clinical characteristics， endoscopic findings， endoscopic ultrasound features， intraoperative and postoperative complications， postoperative pathology， and follow-up outcomes were reviewed. Patients were divided into a non-full-thickness resection group （n = 155） and a full-thickness resection group （n = 181） based on whether perforation occurred during the procedure. Clinical data， endoscopic and endoscopic ultrasound features， postoperative pathology， and complications were compared between the two groups. Univariate and multivariate analyses were performed to identify risk factors for intraoperative full-thickness resection. A nomogram prediction model was constructed using R software， and its predictive performance was evaluated. Results Postoperative pathology revealed leiomyoma in 163 lesions （48.5%） and gastrointestinal stromal tumor （GIST） in 158 lesions （47.0%）. The majority of lesions were located in the gastric fundus （64.9%）. The median operative time was 14.00 （10.00， 20.00） minutes， with a mean of 16.05 minutes. The median number of hemoclips used for wound closure was 4.0 （4.0， 5.0）. Postoperative bleeding occurred in 5 cases， all of which were successfully managed by repeat endoscopy or conservative medical treatment. Postoperative abdominal pain occurred in 25 cases and fever in 24 cases， all of which resolved after medical treatment. No recurrence or metastasis was observed during follow-up. Multivariate logistic regression analysis showed that gastric fundus lesions （OR = 7.464， 95%CI： 4.219 - 13.204）， age （OR = 1.042， 95%CI： 1.012 - 1.072）， number of ligation bands （OR = 12.257， 95%CI： 3.406 - 44.110）， origin from the muscularis propria （OR = 9.763， 95%CI：1.112 - 85.716）， endoscopic ultrasound calcification （OR = 3.260， 95%CI： 1.257 - 8.455）， and extraluminal growth pattern （OR = 4.524， 95%CI： 1.179 - 17.365） were independent risk factors for intraoperative full-thickness resection. The nomogram model demonstrated good discriminative ability with an AUC of 0.805 （95%CI： 0.759 - 0.851）， a specificity of 74.8%， a sensitivity of 71.3%， a optimal cut-off value of 0.603， and a Youden index of 0.461， indicating good predictive accuracy. Conclusions Endoscopic ultrasound combined with multi-band mucosectomy is safe and effective for the resection of small gastric elevated lesions. Advanced age， lesions located in the gastric fundus， an increased number of ligation bands， origin from the muscularis propria， extraluminal growth pattern， and EUS revealed calcification are risk factors for full-thickness resection during the procedure. Preoperative assessment should focus on these risk factors， and adequate preparation of hemostatic devices such as hemoclips should be made for intraoperative precaution.

Key words: endoscopic ultrasound, multi-band mucosectomy, small gastric elevated lesions, gastrointestinal stromal tumor, myoma of smooth muscle