分化型甲状腺癌骨转移预测模型的构建

Construction of predictive model of bone metastasis in differentiated thyroid carcinoma

目的:研究分化型甲状腺癌骨转移危险因素,并建立预测模型。方法:选择2014年1月至2019年1月来本院肿瘤科治疗的符合纳排标准的1 000例分化型甲状腺癌患者作为研究对象,根据是否发生骨转移将患者分成骨转移组(n=39)和非骨转移组(n=961),单因素分析两组患者一般资料,并对差异有统计学意义的单因素行非条件Logistic多因素分析,得出甲状腺癌骨转移的独立危险因素,建立并评估其预测价值。结果:两组患者年龄、病灶数、病理分型、颈部淋巴结是否清扫、肿瘤是否甲状腺外浸润及术后首次清甲时间等差异有统计学意义(P <0.05),而性别、手术方式、肿瘤最大直径、颈部淋巴结是否转移及TNM分期等差异无统计学意义(P> 0.05)。Logistic回归分析表明,术后清甲时间(≥6个月,OR=6.679,95%CI为4.576~9.750)、年龄(18~45岁,OR=0.018,95%CI为0.001~0.507)、颈部淋巴结清扫(OR=3.074,95%CI为2.468~3.829)、肿瘤甲状腺外浸润(OR=2.140,95%CI为1.067~4.292)及多病灶癌(OR=2.296,95%CI为1.136~4.640)等5项指标差异具有统计学意义(P <0.05);Logistic回归模型logit(P)=-6.421+1.899X1-4.037X2+1.123X3+0.761X4+0.831X5,其中X1代表术后清甲时间(≥6个月),X2代表年龄(18~45岁),X3代表颈部淋巴结清扫,X4代表肿瘤甲状腺外浸润,X5代表多病灶癌。术后清甲时间(≥6个月)、年龄(18~45岁)、颈部淋巴结清扫、肿瘤甲状腺外浸润及多病灶癌的ROC曲线下面积分别为0.869(95%CI为0.831~0.907)、0.631 (95%CI为0.564~0.698)、0.801 (95%CI为0.750~0.852)、0.708 (95%CI为0.645~0.772)及0.816 (95%CI为0.768~0.864),均有一定的预测价值。结论:术后清甲时间、年龄、淋巴结清扫、肿瘤甲状腺外浸润及多病灶癌等是分化型甲状腺癌骨转移的危险因素,该模型有望用于指导预测骨转移,提高患者生存率。

Objective: To explore the construction of predictive model of bone metastasis in differentiated thyroid carcinoma. Methods: 1 000 differentiated thyroid carcinoma patients from January 2014 to January 2019 treated in our hospital were selected as study objectives,according to the occurrence of bone metastasis,the patients were divided into bone metastasis group( n = 39) and non-bone metastasis group( n = 961). The basic information of patients was compared for single factor analysis,and the significantly different single factors were analyzed by nonconditional logistic regression analysis,the dangerous factors for DTC bone metastasis and the forecast model were established to determine its predictive value. Results: There were significant differences in age,thyroid focus,number of thyroid lesions,pathological classification,neck lymph node dissection,tumor extra thyroid invasion and the time of thefirst thyroid clearance between the two groups( P < 0. 05). There was no significant difference in gender,mode of operation,maximum diameter of tumor,metastasis of cervical lymph nodes and TNM stage;Logistic regression analysis showed the postoperative nail clearance time( OR = 6. 679,95% CI 4. 576-9. 750),age( 18-45 years)( OR = 0. 018,95% CI 0. 001-0. 507),cervical lymph node dissection( OR = 3. 074,95%CI 2. 468-3. 829),tumor extra thyroid infiltration( OR = 2. 140,95% CI 1. 067-4. 292) and multifocal carcinoma( OR = 2. 296,95% CI 1. 136-4. 640). The ROC curve showed that the AUC of postoperative nail clearance time( ≥ 6 months),age( 18-45 years old),cervical lymph node dissection,tumor extrathyroid invasion and multiple focal carcinoma were 0. 869( 95% CI 0. 831-0. 907),0. 631( 95% CI 0. 564-0. 698),0. 801( 95% CI 0. 750-0. 852),0. 708( 95% CI 0. 645-0. 772) and 0. 816( 95% CI 0. 768-0. 864),respectively,all of which had a certain prediction value. Conclusion: Postoperative nail clearance time,age,lymph node dissection,extrathyroid infiltration and multiple focal carcinoma are risk factors for bone metastasis of differentiated thyroid carcinoma. The model is expected to guide the prediction of bone metastasis and improve the survival rate of patients with differentiated thyroid cancer.