摘要目的:评估eCura评分系统在早期胃癌内镜黏膜下剥离术（endoscopic submucosal dissection，ESD）非治愈性切除后治疗策略选择上的辅助作用。方法:2015年1月—2019年6月福建省立医院110例早期胃癌ESD非治愈性切除病例纳入回顾性队列研究，根据eCura评分系统分成低危组79例、中危组22例和高危组9例。采用ROC曲线分析检验eCura评分系统预测淋巴结转移的诊断效能；使用Logistic回归分析探寻eCura评分系统风险分层对淋巴结转移的影响；使用Kaplan-Meier法评估癌症生存率，生存率比较使用log-rank检验。结果:35例患者于ESD术后追加了标准外科手术，包括低危组22例、中危组8例和高危组5例，其中5例发生淋巴结转移，包括低、中危组各1例和高危组3例。ROC曲线下面积为0.857（95% CI：0.697～0.952， P=0.001），eCura评分系统评分的最佳截断值设置为3分时约登指数为0.7，达最大值，对应的敏感度和特异度分别为80%和90%。Logistic回归分析结果显示：中危组发生淋巴结转移的概率约是低危组的3.00倍（95% CI：0.17～54.57， P=0.458）；高危组发生淋巴结转移的概率约是低危组的31.50倍（95% CI：2.14～463.14， P=0.012）。随访时间12～58个月，中位随访时间40个月，10例复发，包括低危组4例和中、高危组各3例，其中2例来源于ESD术后追加标准外科手术者中，且均来自低危组，其余8例均来源于ESD术后未追加标准外科手术者中。Kaplan-Meier生存率曲线分析结果显示：低危组中追加手术者的癌症生存率略低于未追加手术者（ P=0.319），中危组中追加手术者的癌症生存率高于未追加手术者（ P=0.296），高危组中追加手术者的癌症生存率明显高于未追加手术者（ P=0.013）。 结论:eCura评分系统能够较好地辅助早期胃癌ESD非治愈性切除后治疗策略的选择，能够较为准确地预测后续淋巴结转移和复发风险。对于eCura评分系统评估为低风险类别者，密切随访可能是一种可接受的选择；而对于eCura评分系统评估为高风险类别者，追加标准外科手术治疗可能更有利于改善预后。

abstractsObjective:To evaluate the adjuvant role of the eCura scoring system in selecting appropriate treatment strategies after non-curative endoscopic submucosal dissection (ESD) of early gastric cancer (EGC) patients.Methods:The clinicopathological data of 110 EGC patients who underwent non-curative ESD at Fujian Provincial Hospital from January 2015 to June 2019 were retrospectively analyzed. According to the eCura score, patients were divided into three lymph node metastasis (LNM) risk groups: low-risk group (79 cases), middle-risk group (22 cases), and high-risk group (9 cases). The receiver operator characteristic (ROC) curve analysis was used to test the diagnostic efficacy of eCura scoring system in predicting LNM. Logistic regression analysis was used to explore the influence of risk stratification of eCura scoring system on LNM. Kaplan-Meier method was used to evaluate cancer survival rate, which was then compared with log-rank test.Results:Thirty-five patients underwent additional standard surgery after ESD, including 22 in the low-risk group, 8 in the middle-risk group, and 5 in the high-risk group. Among them, 5 cases had LNM, including 1 case in the low-risk group and the middle-risk group respectively and 3 cases in the high-risk group. The area under the ROC curve was 0.857 (95% CI: 0.697-0.952, P=0.001), and when the cut-off value of the eCura score was set at 3, the Yuden index reached the maximum value of 0.7, with the corresponding sensitivity and specificity of 80% and 90%, respectively. Logistic regression analysis showed that the probability of LNM in the middle-risk group was about 3.00 times (95% CI: 0.17-54.57, P=0.458) as high as that in the low-risk group, and the probability of LNM in the high-risk group was about 31.50 times (95% CI: 2.14-463.14, P=0.012) of that in the low-risk group. The follow-up time was 12 to 58 months, and the median follow-up time was 40 months. There were 10 cases of recurrence, including 4 cases in the low-risk group, 3 cases in the middle-risk group and 3 cases in the high-risk group, of which 2 cases in the low-risk group were from those of additional standard surgery after ESD, and the remaining 8 cases were from those who did not receive additional standard surgery after ESD. Kaplan-Meier survival curve analysis showed that the survival rate of patients with additional surgery in the low-risk group was similar to that of patients without ( P=0.319), and the survival rate of patients with additional surgery in the middle-risk group was also similar to that of patients without ( P=0.296). The survival rate of patients with additional surgery in the high-risk group was significantly higher than that of those without ( P=0.013). Conclusion:The eCura scoring system can assist the selection of treatment strategies after non-curative resection of EGC, and can accurately predict the risk of subsequent LNM and recurrence. Close follow-up may be an acceptable option for patients with low risk of LNM, and additional standard surgical treatment may be more conducive to improving the prognosis in patients with high risk of LNM.