内镜黏膜下剥离术后早期胃癌患者迟发性出血的危险因素及其预测价值研究

Risk factors for delayed hemorrhage in patients with early gastric carcinoma after endoscopic submucosal dissection and its predictive value

摘要:

目的 分析内镜黏膜下剥离术（ESD）后早期胃癌（EGC）患者迟发性出血的危险因素及其预测模型的应用价值。方法 回顾性分析2019年7月－2024年3月该院收治的386例行ESD治疗的EGC患者的临床资料，术后均接受短期（术后24 h～30 d）随访，根据是否发生迟发性出血，将患者分为发生组（50例）和未发生组（336例）。采用多因素Logistic回归模型，分析影响EGC患者ESD后发生迟发性出血的独立危险因素，并绘制受试者操作特征曲线（ROC curve），分析其预测价值。结果 发生组贲门胃底、病灶直径 ≥ 3 cm、黏膜下纤维化、黏膜下粗大血管和有溃疡或瘢痕的占比明显高于未发生组，差异均有统计学意义（P < 0.05）。多因素Logistic回归分析（引入水准为0.05）结果显示：病灶位置为贲门胃底（OR＾ = 1.088，95%CI：1.040～1.138）、病灶直径 ≥ 3 cm（OR＾ = 1.095，95%CI：1.057～1.135）、黏膜下纤维化（OR＾ = 1.130，95%CI：1.061～1.203）、黏膜下粗大血管（OR＾ = 1.177，95%CI：1.116～1.241）和有溃疡或瘢痕（OR＾ = 1.082，95%CI：1.057～1.108）是影响EGC患者ESD后发生迟发性出血的独立危险因素（P < 0.05）。建立预测模型：Logit（P） = -9.238 + 病灶位置为贲门胃底×0.084+病灶直径≥3 cm×0.091+黏膜下纤维化×0.122 + 黏膜下粗大血管×0.163 + 有溃疡或瘢痕×0.079（P < 0.05），似然比χ² = 131.09，DF = 8，P < 0.05，说明构建的预测模型有效。Hosmer-Lemeshow拟合优度检验显示，预测模型的拟合效果较好（P > 0.05）。ROC curve显示：截断值为12.00时，曲线下面积（AUC）为0.929，敏感度为88.00%（95%CI：0.757～0.955），特异度为83.04%（95%CI：0.786～0.869）。结论 接受ESD治疗的EGC患者，术后发生迟发性出血与病灶为贲门胃底、病灶直径 ≥ 3 cm、黏膜下纤维化、黏膜下粗大血管和有溃疡或瘢痕有关，据此构建的预测模型，预测价值较高。

Abstract:

Objective To analyze the risk factors for delayed hemorrhage in patients with early gastric carcinoma (EGC) after endoscopic submucosal dissection (ESD), as well as to explore the predictive value of predictive models.Methods The clinical data of 386 EGC treated with ESD patients from July 2019 to March 2024 were retrospectively analyzed. They were followed up for a short period of time (24 h~30 d postoperatively) after ESD, and were divided into the group of occurrence (50 cases) and the group of non-occurrence (336 cases) according to the occurrence or non-occurrence of delayed hemorrhage. To understand the clinical data of the patients, multivariate Logistic regression was used to analyze the risk factors for delayed hemorrhage in patients with EGC after ESD, and plot receiver operating characteristic curve (ROC curve) to analyze the predicted value.Results The proportions of cardia gastric fundus, lesion diameter ≥ 3 cm, submucosa with fibrosis, coarse blood vessels, and having ulcer or scar in the occurrence group were significantly higher than those in the non-occurrence group, the differences were statistically significant (P < 0.05). The results of multivariate Logistic regression analysis (with an introduction level of 0.05) showed that the lesion location was cardia-gastric fundus (OR＾ = 1.088, 95%CI: 1.040 ~ 1.138), the diameter of the lesion was ≥ 3 cm (OR＾ = 1.095, 95%CI: 1.057 ~ 1.135), and fibrosis under the mucosa (OR＾ = 1.130, 95%CI: 1.061 ~ 1.203), thick blood vessels under the mucosa (OR＾ = 1.177, 95%CI: 1.116 ~ 1.241), and the presence of ulcer or scar (OR＾ = 1.082, 95%CI: 1.057 ~ 1.108) were the risk factors for the occurrence of delayed hemorrhage in patients with EGC after ESD (P < 0.05). A prediction model was established: Logit （P） = -9.238 + Location of lesion as cardia-gastric fundus × 0.084 + Diameter of lesion ≥ 3 cm×0.091 + Submucosal with fibrosis × 0.122 + Submucosal with thick blood vessels × 0.163 + Ulcer or scar×0.079 (P < 0.05), the likelihood ratio test yielded χ2 = 131.09, DF = 8, P < 0.05, which indicated that the constructed prediction model was valid. Hosmer- Lemeshow goodness-of-fit test showed a good model fit (P > 0.05). The ROC curve showed an area under the curve (AUC) value of 0.929 at a cutoff value of 12.00, with a sensitivity of 88.00% (95%CI: 0.757 ~ 0.955) and a specificity of 83.04% (95%CI: 0.780 ~ 0.869).Conclusion Postoperative delayed hemorrhage in patients with EGC treated with ESD is associated with cardia-gastric fundus, lesion diameter ≥3 cm, submucosal fibrosis, thick blood vessels, and the presence of ulcer or scar, and the prediction value of the prediction model constructed accordingly is good.