骨肉瘤预后个体化预测模型列线图的建立

 目的 通过对骨肉瘤患者的临床、病理学、影像学及随访资料进行预后因素分析，建立骨肉瘤预后预测模型列线图，并验证其准确度。方法 收集1998至2008年确诊且符合入组标准的235例骨肉瘤患者组成建模组，2009年的55例骨肉瘤患者组成验证组。单因素生存分析采用Kaplan-Meier法绘制生存曲线，Log-rank法进行统计学分析；应用Cox比例风险模型进行多因素分析，确定独立预后因子；然后应用R软件建立预测模型列线图，内部验证运用Bootstrap法，外部验证运用验证组，一致性指数（C-index）用来评价模型准确度，并绘制出列线图预测和实际观察的五年生存率校准曲线。结果 建模组和验证组五年总体生存率分别为46.1%±6.7%和61.8% ±12.9%。多因素分析结果显示，病理性骨折、入院时碱性磷酸酶水平、肿瘤大小、肿瘤分期和术后化疗次数是独立预后因素。校准曲线显示列线图预测与实际观察的五年生存率有很好的一致性。列线图预测五年生存率的C-index为0.74（95%CI，0.70~0.78），明显高于Enneking分期系统。应用验证组进行外部验证，列线图、Enneking分期和美国癌症联合委员会（AJCC）分期系统的C-index分别为0.71、0.54和0.56，提示列线图比Enneking分期及AJCC分期的预后预测准确性更高。结论 成功建立的预测骨肉瘤患者总体生存列线图能实现个体化预测，且与其他预后预测系统相比更直观、准确。

Establishment and assessment of a nomogram for predicting the individual prognosis of osteosarcoma

Objective To develop an accurate prognostic nomogram for osteosarcoma and to validate it in order to provide scientific basis for risk evaluation, though analyzing the clinicopathologic, radiologic and follow-up data of osteosarcoma patients. Methods All of 235 osteosarcoma patients who diagnosed in the department of Musculoskeletal Oncology in the first affiliated hospital of Sun Yat-Sen University during 1998 to 2008 and met the inclusion criteria, were included in a primary cohort, and 55 patients diagnosed in 2009 were included in a validation cohort. In univariate survival analysis of the primary cohort, Kaplan-Meier method was used to plot Survival curves, and the log-rank test to compare differences. In the multivariate analysis, the Cox proportional hazard model had been used to identify the independent prognostic factors. Then a nomogram created by R software was validated by Bootstrap resampling as internal validation and by the validation cohort as external validation. Concordance index (C-index) was applied to the predictive evaluation of the nomogram and calibration curves were drawn to test the nomogram-prediction and actual observation of the five-year overall survival rate. Results The five-year overall survival rates of the primary cohort and validation cohort were 46.1%±6.7% and 61.8%±12.9%, respectively. In multivariate analysis, pathologic fracture, pre-ALP, tumor maximum diameter, Enneking stage and postoperative chemotherapy had been identified as independent prognostic factors. Good concordance of survival probability between nomogram-prediction and actual observation was shown by the calibration curves. The C-index of the nomogram for predicting survival, 0.74（95% CI，0.70-0.78）, was statistically higher than that of the Enneking system（C-index: 0.54）. Of the validation cohort, C-index of the nomogram, Enneking and American Joint Committee on Cancer (AJCC) stage system was 0.71, 0.54 and 0.56, respectively, indicating the nomogram had a higher accuracy in prognostic prediction than the others. Conclusion The nomogram as proposed is an effective tool for predicting the overall survival of osteosarcoma patients. Compared with Enneking and AJCC stage system, the nomogram was more accurate and visible.