James Huang1, Frank C Detterbeck2, Zuoheng Wang2, Patrick J Loehrer3. 1. Division of Thoracic Surgery, Department of Surgery, Yale University School of Medicine and Department of Biostatistics, School of Epidemiology and Public Health, Yale University, New Haven, Connecticut. 2. Thoracic Service, Memorial Sloan-Kettering Cancer Center, New York City, New York. 3. HH Gregg Professor of Oncology, Division of Medical Oncology, Department of Internal Medicine, Indiana Universitychool of Medicine, Indianapolis, Indiana.
目前,关于胸腺恶性肿瘤还没有国际抗癌联盟(International Union Against Cancer, UICC)和美国癌症联合委员会(American Joint Committeeon Cancer, AJCC)的官方分期,但已经存在一些分期系统[,包括Masaoka分期[、Masaoka-Koga分期[、GETT分期[和TNM分期[。大多数中心和发表的文章都采用Masaoka分期系统,但自1995年后更多学者开始使用Koga改良分期系统(Masaoka-Koga分期),见表 1。也是ITMIG目前建议使用的分期。Koga改良分期与以往的Masaoka分期的不同之处在于,对于包膜侵犯但未侵透,Masaoka分期归为IIb期,而Masaoka-Koga分期归为I期。这是因为大多数病理学家认为包膜部分侵犯没有意义,而且生存数据似乎也证实这一点[。此外,该分期系统关于包膜完整胸腺瘤和侵袭性胸腺瘤的定义与ITMIG是一致的。另外一个不同之处是,与周围结构粘连或肉眼侵犯但未侵透纵隔胸膜或心包,Masaoka-Koga归为Ⅱb期,而Masaoka未给出明确的定义。尽管Masaoka-Koga分期系统得到广泛应用,但是也有很多定义不明确之处,详见其它有关分期的介绍。ITMIG和IASLC将承担对该分期系统的评价和统计验证,并提出可能的替代分期方法,这就要求收集的数据不止Masaoka-Koga分期,ITMIG数据库对此会单独说明。
Overall cause (A) and stage-specific (B) cause of death after resection of patients with thymoma. Results are an average of studies from 1980 to 2009 of100 patients reporting this data3.
Specific outcomes for a stage Ⅲ resected thymoma at 10 years, estimated from data regarding overall survival, cause of death, incidence of recurrence, and incidence of other cancers[.
胸腺瘤患者术后所有死因(A)和分期相关死因(B)。结果来自1980年至2009年样本量大于等于100例的研究结果的平均值。Overall cause (A) and stage-specific (B) cause of death after resection of patients with thymoma. Results are an average of studies from 1980 to 2009 of100 patients reporting this data3.生存评估方法Survival measuresⅢ期胸腺瘤术后患者的10年特异性生存,从总生存、死因、复发和第二原发肿瘤发生的相关数据评估[。Specific outcomes for a stage Ⅲ resected thymoma at 10 years, estimated from data regarding overall survival, cause of death, incidence of recurrence, and incidence of other cancers[.描述特定结局事件的精算结局曲线(比如单独计算无局部复发或特定病因的生存曲线)通常具有误导性,产生过于乐观的结果[。原因在于精算方法在统计学上要求各个事件之间相互独立,事实上并非如此,比如局部复发时间和远处复发时间很可能是相互关联的[。根据结局事件的关联程度不同,采用精算法预测一种结局事件的发生率容易比实际发生率低估30%-50%[。因此,最好的办法是先大体上分析死亡或者其他事件,然后再分析死亡或者发生事件的原因风险比例,从而避免不同原因之间的竞争风险[。因此,建议报道复发类型和死亡原因的风险比例。ITMIG推荐的评估胸腺肿瘤疗效的方法是计算其发生在任何部位的复发率,控制重症肌无力等相关疾病的能力应单独考虑。此外,受其他因素的影响死亡原因仅是次优的评估指标,当然不混杂复发的死亡是最理想的评估指标。因此,建议以无复发生存作为根治性切除术或放疗后完全缓解患者预后的最佳评估指标。对于未根治的患者,疾病进展时间是最好的评估指标,疾病进展时间也可以用于R1切除的患者,因为其存在残余病灶,见表 3。不难看出,这两个评估指标的终点是一样的,都是疾病复发,但选择哪个术语取决于治疗后是否仍然存在肿瘤。姑息性治疗的患者选择疾病进展时间,是肿瘤内科医生容易接受的评估指标。根治患者选择无复发生存而非复发时间是因为前者强调阳性结局的可能,而后者给人印象是复发仅为一种时间形式。
Variance in actuarial survival estimates by size of cohort for (A) a 10-year study duration and (B) a 5-year study duration. The vertical bars are 95% confidence intervals for the survival estimate at 5 and 10 years, based on a standard model of exponentially decreasing survival, a constant rate of accrual of patients during the course of the study until study termination, and no loss to follow-up. MST, median survival time.
队列样本量不同,持续10年的研究(A)和持续5年的研究(B)在精确估算生存上存在差异。竖线代表根据标准的指数生存模型(研究过程中患者是固定的并且零失访)估计出5年和10年生存的95%置信区间。Variance in actuarial survival estimates by size of cohort for (A) a 10-year study duration and (B) a 5-year study duration. The vertical bars are 95% confidence intervals for the survival estimate at 5 and 10 years, based on a standard model of exponentially decreasing survival, a constant rate of accrual of patients during the course of the study until study termination, and no loss to follow-up. MST, median survival time.
Authors: P Therasse; S G Arbuck; E A Eisenhauer; J Wanders; R S Kaplan; L Rubinstein; J Verweij; M Van Glabbeke; A T van Oosterom; M C Christian; S G Gwyther Journal: J Natl Cancer Inst Date: 2000-02-02 Impact factor: 13.506
Authors: L H Schwartz; M S Ginsberg; D DeCorato; L N Rothenberg; S Einstein; P Kijewski; D M Panicek Journal: J Clin Oncol Date: 2000-05 Impact factor: 44.544
Authors: J F Regnard; P Magdeleinat; C Dromer; E Dulmet; V de Montpreville; J F Levi; P Levasseur Journal: J Thorac Cardiovasc Surg Date: 1996-08 Impact factor: 5.209
Authors: James Huang; Nabil P Rizk; William D Travis; Gregory J Riely; Bernard J Park; Manjit S Bains; Joseph Dycoco; Raja M Flores; Robert J Downey; Valerie W Rusch Journal: J Thorac Cardiovasc Surg Date: 2009-07 Impact factor: 5.209
Authors: Jeremy J Erasmus; Gregory W Gladish; Lyle Broemeling; Bradley S Sabloff; Mylene T Truong; Roy S Herbst; Reginald F Munden Journal: J Clin Oncol Date: 2003-07-01 Impact factor: 44.544