[Risk factor analysis of lymph node metastasis in endometrial carcinoma combined with molecular types].

Objective: To investigate the relationships between molecular types of the cancer genome atlas (TCGA) of patients with endometrial carcinoma (EC) and lymph node metastasis and other clinicopathological features. Methods: The clinical pathological information of 295 patients with EC who underwent initial inpatient surgical treatment and accepted the detection of the molecular types of TCGA with next-generation sequencing technology at Peking University People's Hospital were collected during April 2016 and May 2022. The TCGA molecular typing of EC was divided into four types: POLE-ultramutated (15 cases), high microsatellite instability (MSI-H; 50 cases), copy-number low (CNL; 175 cases), and copy-number high (CNH; 55 cases). The differences of clinical pathological features among different molecular types and the risk factors of lymph node metastasis were analyzed retrospectively. Results: Among 295 patients with EC, the average age was (56.9±0.6) years. (1) There was a statistically significant difference in lymph node metastasis (0, 8.0%, 10.3% and 25.5%) among the four molecular types (χ2=12.524, P=0.006). There were significant differences in age, stage, pathological type, grade (only endometrioid carcinoma), myometrium invasion, lymphatic vascular space infiltration, and estrogen receptor among the EC patients of four molecular types (all P<0.05). Among them, while in the patients with CNH type, the pathological grade was G3, the pathological type was non-endometrioid carcinoma, and the proportion of myographic infiltration depth ≥1/2 were higher (all P<0.05). (2) Univariate analysis suggested that pathological type, grade, myometrium infiltration depth, cervical interstitial infiltration, lymphatic vascular space infiltration, and progesterone receptor were all factors which significantly influence lymph node metastasis (all P<0.01); multivariate analysis suggested that the lymphatic vascular space infiltration was an independent risk factor for lymph node metastasis (OR=5.884, 95%CI: 1.633-21.211; P=0.007). (3) The factors related to lymph node metastasis were different in patients with different molecular types. In the patients with MSI-H, the non-endometrioid carcinoma of pathological type was independent risk factor for lymph node metastasis (OR=29.010, 95%CI: 2.067-407.173; P=0.012). In the patients with CNL, myometrium infiltration depth≥1/2 (OR=4.995, 95%CI: 1.225-20.376; P=0.025), lymphatic vascular space infiltration (OR=14.577, 95%CI: 3.603-58.968; P<0.001) were the independent risk factors for lymph node metastasis. While in the CNH type patients pathological type of non-endometrioid carcinoma (OR=7.451, 95%CI: 1.127-49.281; P=0.037), cervical interstitial infiltration (OR=22.938, 95%CI: 1.207-436.012; P=0.037), lymphatic vascular space infiltration (OR=9.404, 95%CI: 1.609-54.969; P=0.013), were the independent risk factors for lymph node metastasis. Conclusions: POLE-ultramutated EC patients have the lowest risk of lymph node metastasis, and CNH patients have the highest risk of lymph node metastasis. The risk factors of lymph node metastasis of different molecular types are different. According to preoperative pathological and imaging data, lymph node metastasis is more likely to occur in patients with non-endometrioid carcinoma in MSI-H and CNH type patients, and lymph node metastasis is more likely to occur in patients with myometrium infiltration depth ≥1/2 in CNL type patients.