Objective: To establish and validate a predicting nomogram for cervical adenocarcinoma based on surveillance, epidemiology and end results (SEER) database and Chinese single-center data, and to explore the optimal treatment for cervical adenocarcinoma. Methods: This study selected 2 478 cervical adenocarcinoma patients from the SEER database as the training cohort, and 195 cervical adenocarcinoma patients from Cancer Hospital of Dalian University of Technology, Liaouing Cancer Hospital and Institute as an external validation cohort. Clinicopathological information and follow-up data of the two cohorts were collected. The radiotherapy group was defined as receiving comprehensive treatment based on concurrent chemoradiotherapy after initial diagnosis, while the surgery group was defined as receiving comprehensive treatment based on radical surgery. Log-rank test and cox regression were used to evaluate factors affecting the prognosis of cervical adenocarcinoma patients. A nomogram was drawn to predict the 3-year and 5-year overall survival rates of cervical adenocarcinoma patients, and then internal validation of the training cohort from SEER database and external validation of the hospital cohort were conducted. Results: (1) In the SEER database training cohort, there were 385 patients (15.54%, 385/2 478) in the radiotherapy group and 2 093 patients (84.46%, 2 093/2 478) in the surgery group. Overall survival time of the radiotherapy group was (55.8±51.3) months, while that of the surgery roup was (94.4±61.7) months, the difference between the two groups was statistically significant (χ2=256.44, P<0.001). Log-rank test showed that age, marital status, maximum of tumor diameters, pathological grade, International Federation of Gynecology and Obstetrics (FIGO) stage, and treatments were all significant factors affecting the overall survival time of cervical adenocarcinoma patients (all P<0.001). Multivariate Cox regression analysis showed that elder (>50 years old), single status, huge tumors (>4 cm), high pathological grades (G2, G3), and advanced FIGO stages (≥Ⅱa2 stage) were independent risk factors for the overall survival time of cervical adenocarcinoma patients (all P<0.05); compared with radiotherapy, surgery was a protective factor for the prognosis of cervical adenocarcinoma patients (HR=0.619, 95%CI: 0.494-0.777; P<0.001). Further analysis of locally advanced stage and Ⅲc stage of patients showed that surgery was a protective factor for the prognosis of cervical adenocarcinoma patients with a maximum tumor diameter >4 to <6 cm (HR=0.414, 95%CI: 0.182-0.942; P=0.036) in locally advanced stage and Ⅲc T1 to T2 stage (HR=0.473, 95%CI: 0.307-0.728; P=0.001). (2) The external validation cohort consisted of 39 patients (20.00%, 39/195) in the radiotherapy group and 156 patients (80.00%, 156/195) in the surgery group. The overall survival time of patients in the radiotherapy group was (51.7±34.3) months, while that of the surgery group was (63.1±26.6) months (χ2=28.41, P<0.001). Further analysis was conducted on locally advanced stage and Ⅲc stage patients, and multivariate Cox regression analysis was performed after propensity score matching, which showed that surgery was a protective factor for the prognosis of cervical adenocarcinoma patients with a maximum tumor diameter >4 to <6 cm in locally advanced stage (HR=0.141, 95%CI: 0.023-0.843; P=0.032) and Ⅲc T1 to T2 stage (HR=0.184, 95%CI: 0.036-0.947; P=0.043). (3) Establishment and internal and external validation of nomogram: based on the six factors screened out by the multivariate Cox regression model, the nomogram was developed to predict the prognosis of cervical adenocarcinoma patients. The consistency index of the internal and external validation were 0.801 and 0.766, respectively, and the calibration curves matched well with the ideal fitting line. Conclusions: The key to the treatment of cervical adenocarcinoma is to prioritize radical surgery for patients with conditions for radical tumor resection. Compared with concurrent chemoradiotherapy, patients with locally advanced stages (Ⅰb3, Ⅱa2), and Ⅲc (T1, T2) stages cervical adenocarcinoma could benefit from comprehensive treatment based on radical surgery. The nomogram of this study has been validated internally and externally, and show good survival prediction efficacy for cervical adenocarcinoma patients.
目的： 基于美国国立癌症研究所的监测、流行病学及最终结果（SEER）数据库及中国单中心数据建立并验证预测子宫颈腺癌患者预后的列线图，探讨子宫颈腺癌的优选治疗方式。 方法： 本研究从SEER数据库筛选出2 478例子宫颈腺癌患者作为训练队列，并选择辽宁省肿瘤医院收治的195例子宫颈腺癌患者作为外部验证队列，收集两个队列患者的临床病理资料及随访数据；放疗组为同步放化疗为主的综合治疗，手术组为根治性手术为主的综合治疗。采用log-rank检验和Cox回归方法分析影响子宫颈腺癌患者预后的相关因素，建立预测子宫颈腺癌患者3年、5年总生存率的列线图，并进行SEER数据库训练队列的内部验证及本院数据验证队列的外部验证。 结果： （1）SEER数据库训练队列中，放疗组385例（15.54%，385/2 478），手术组2 093例（84.46%，2 093/2 478）；放疗组患者的总生存时间为（55.8±51.3）个月，手术组为（94.4±61.7）个月，两组比较，差异有统计学意义（χ2=256.44，P<0.001）。log-rank检验显示，年龄、婚姻状态、肿瘤最大径、病理分级、国际妇产科联盟（FIGO）分期、治疗方式均为显著影响子宫颈腺癌患者总生存时间的因素（P均<0.001）。多因素Cox回归分析显示，高龄（>50岁）、单身状态、肿瘤最大径（>4 cm）、病理分级高（G2、G3）及FIGO分期晚（≥Ⅱa2期）均为影响子宫颈腺癌患者总生存时间的独立危险因素（P均<0.05）；而与放疗比较，手术是子宫颈腺癌患者预后的保护因素（HR=0.619，95%CI为0.494~0.777；P<0.001）。对局部晚期、Ⅲc期患者的进一步分析显示，手术是肿瘤最大径>4~<6 cm的局部晚期（Ⅰb3、Ⅱa2期；HR=0.414，95%CI为0.182~0.942；P=0.036）、Ⅲc T1~T2期（HR=0.473，95%CI为0.307~0.728；P=0.001）子宫颈腺癌患者预后的保护因素。（2）本院数据的外部验证队列中，放疗组39例（20.00%，39/195），手术组156例（80.00%，156/195）；放疗组患者的总生存时间为（51.7±34.3）个月，显著低于手术组的（63.1±26.6）个月（χ2=28.41，P<0.001）。对局部晚期、Ⅲc期患者进行进一步分析，经倾向性评分匹配后行多因素Cox回归分析，结果显示，手术治疗是肿瘤最大径>4~<6 cm的局部晚期（HR=0.141，95%CI为0.023~0.843；P=0.032）、Ⅲc T1~T2期（HR=0.184，95%CI为0.036~0.947；P=0.043）子宫颈腺癌患者预后的保护因素。（3）列线图的建立及内部、外部验证：基于SEER数据库训练队列多因素Cox回归分析筛选出的6个阳性因素，建立预测子宫颈腺癌患者预后的列线图。SEER数据库训练队列的内部验证、本院数据验证队列的外部验证，列线图的一致性指数分别为0.801、0.766，的校准曲线与理想拟合线均吻合良好。 结论： 子宫颈腺癌的治疗，重点在于对具有手术条件的患者首选根治性手术可获益；与同步放化疗相比，肿瘤最大径>4~<6 cm的局部晚期（Ⅰb3、Ⅱa2期）、Ⅲc期（T1、T2期）子宫颈腺癌患者能从以根治性手术为主的综合治疗中获益更多。本研究建立的列线图经内部及外部验证，对子宫颈腺癌患者具有良好的预后预测效能。.