各位好！今天与大家分享一篇近期发表在JTCVS上的一篇文献，该研究回顾了NCDB数据库中仅基于肿瘤大小进行T分期临界人群的预后（如T1c/T2a;T2a/T2b;T2b/T3），作者尝试阐明肿瘤直径在临界T分期上下2mm的人群接受辅助治疗的情况与总生存的关系。来看看病理测量上的微小差异，在临床决策的视角上是怎样一番情况？
Living on the Edge: Role of Adjuvant Therapy After Resection of Primary Lung Cancer Within 2 Millimeters of a T-Stage Cutoff
Brooks V. Udelsman, MD MHS, Christina K. Bedrosian, BS, Eric S. Kawaguchi, PhD, Li Ding, MD, Williams D. Wallace, MD, Graeme Rosenberg, MD, Takashi Harano, MD, Sean Wightman, MD, Scott Atay, MD, Anthony W. Kim, MD, Gavitt Woodard, MD
26 October 2024
Purpose: We evaluated the use of systemic therapy and overall survival in patients with resected non-small cell lung cancer (NSCLC) whose pathologic tumor size was within 2mm of a T-stage cutoff.
目的：评估了病理肿瘤大小在T期临界值2 mm以内的切除非小细胞肺癌（非小细胞肺癌）患者综合治疗的使用情况和总体生存率。
Methods: Retrospective cohort study using the National Cancer Database of patients who underwent resection of tumors within 2mm of the T1c/T2a, T2a/T2b, and T2b/T3 T-stage cutoffs. Patients with nodal involvement or whose T-stage was based on pathologic features other than tumor size were excluded. A multistate model compared the primary outcomes of systemic therapy and overall survival.
方法：基于国家癌症数据库对T1 c/T2 a、T2 a/T2 b和T2 b/T3 T期截止点2 mm内接受肿瘤切除的患者进行了回顾性队列研究。排除有淋巴结受累或T分期基于肿瘤大小以外的病理特征的患者。多状态模型比较了综合治疗的主要结果和总生存期。
Results: From the NCDB, 18,490 patients were identified: 9,966 at the T1c/T2a cutoff, 5,593 at the T2a/T2b cutoff, and 2,931 at the T2b/T3 cutoff. Peaks in tumor size distribution occurred at 5mm intervals. Based on an expected normalized curve, 2,050 patients (11.1%) may have been under-staged. Use of systemic therapy was higher among patients with larger tumors at the T1c/T2a cutoff (7.1% vs. 8.9%; p<0.001), the T2a/T2b cutoff (20.0% vs. 25.5%; p<0.001), and the T2b/T3 cutoff (31.2% vs. 41.8%; p<0.001). In a multistate model, mortality was higher above the T1c/T2a cutoff (Hazard Ratio [HR] 1.10; p=0.01), T2a/T2b cutoff (HR 1.17; p<0.01), and T2b/T3 cutoff (HR 1.13; p=0.03). In patients who received systemic therapy, this trend was eliminated (HR 1.24; p=0.14, HR 0.79; p=0.07, and HR 1.23; p=0.09, respectively).
结果：从NCDB中，确定了18490名患者：T1c/T2a临界值为9，966人，T2a/T2b临界值为5593人，T2b/T3临界值为2，931人。肿瘤大小分布的峰值以5 mm的间隔出现。根据预期的标准化曲线，2，050名患者（11.1%）可能未进行充分分期。在T1c/T2a临界值（7.1% vs. 8.9%; p<0.001）、T2a/T2b临界值（20.0% vs. 25.5%; p<0.001）和T2 b/T3临界值（31.2% vs. 41.8%; p<0.001）时，肿瘤较大的患者中系统治疗的使用率较高。在多状态模型中，超过T1c/T2a临界点（危险比[HR] 1.10; p=0.01）、T2a/T2b临界点（HR 1.17; p<0.01）和T2b/T3临界点（HR 1.13; p=0.03）的死亡率较高。在接受系统治疗的患者中，这种趋势被消除（分别为HR 1.24; p=0.14、HR 0.79; p=0.07和HR 1.23; p=0.09）。
Conclusions: Rounding of tumor size for pathologic staging is common. While seemingly trivial, rounding may downstage patients and is associated with decreased rates of adjuvant therapy use and potentially worse overall survival.
结论：对肿瘤大小进行病理分期的四舍五入很常见。虽然看似微不足道，但四舍五入可能会降低患者的分期，并与辅助治疗使用率下降和总体生存率可能更差有关。
1. 随着第九版TNM分期的公布，pT还是沿用之前第八版的划分标准。理想的病理报告通常包含以下内容。
10.1111/his.15313
2.上细节：
首先本文是回顾了NCDB 2010-2019年期间的仅由肿瘤大小决定T分期且N0的患者。2.9cm的肿瘤和3.1cm的肿瘤虽然属于不同的T分期，但可能验证其生存差异和相似需要极大样本量。而且这种微小的差异可能会影响后续患者的诊疗决策。虽然有NCDB大样本量，但毫米级预后分析上还是略显单薄。数据库中固有的信息也使得分析维度上的临床意义较为有限。
其次，文章方法算是一个亮点部分使用了multistate (three-state) model和the Markov assumption, transition-specific Cox regression models。有兴趣的可以学习模仿。
第三，初看题目眼前一亮，细看结论四平八稳。值得思考的是在辅助治疗维度上，千级样本的大数据分析往往因为数据维度有限而无法发挥出理想中的价值。造成大而散乱的问题，即使是在资源丰富的美国也只能完成基本的流调级数据。真正临床驱动的数据目前还是以rct，中小型队列研究为主。
3.当然凡事总有例外。Rarely, when pathologic measurement is problematic and deemed inaccurate, using the clinical measurement as the pT size is recommended.
目录
1. INTRODUCTION
2. Methods
2.1 Data Source (Supplementary Data 1)
2.2 Inclusion and Exclusion Criteria
2.3 Comparison Groups, Covariates, and Outcomes
2.4 Statistical analysis (Supplementary Figure 1)
3. Results
3.1 Distribution of Tumor Size (Figure 1-2)
3.2 Characteristics of Patients at T-Stage Cutoffs (Supplemental Tables 1-3)
3.3 Receipt of Systemic Therapy (Table 1-2)
3.4 Overall Survival and Adjuvant Therapy in Multistate Model (Figure 3a-c)(Table 3)(Supplementary Tables 4-6)
4. Discussion (Figure 4)(Supplementary Table 7-11)
— 图表汇总—
2.1 Data Source
Supplementary Data 1
2.2 Inclusion and Exclusion Criteria
2.3 Comparison Groups, Covariates, and Outcomes
2.4 Statistical analysis
Supplementary Figure 1. Conceptual framework for a multistate model. The rate of each transition is calculated and expressed as a hazard rate.
3. Results
3.1 Distribution of Tumor Size
Figure 1. Inclusion and exclusion criteria using the National Cancer Database (NCDB) participant user file (PUF) non-small cell lung cancer (NSCLC).
Figure 2.Histogram of pathologic tumor size for patients with surgically resected tumors between 2.6cm and 5.4cm. Tumors below and above a T-stage cutoff colored red and blue.
3.2 Characteristics of Patients at T-Stage Cutoffs
Supplemental Tables 1. Patient and tumor characteristics at the T1c/T2a transition
Supplemental Tables 2. Patient and tumor characteristics at the T2a/T2b transition
Supplemental Tables 3. Patient and tumor characteristics at the T2a/T2b transition
3.3 Receipt of Systemic Therapy
Table 1. Adjuvant systemic therapy with 2mm of a T-stage cutoff
Table 2. Explanation for not receiving adjuvant systemic therapy among the included study patients with T2b and T3 tumors
3.4 Overall Survival and Adjuvant Therapy in Multistate Model
Figure 3. Mirror plots displaying overall survival after surgery among patients with tumors 2mm above and below each T-stage cutoff. In these plots, survival can be compared between patients just below and just above the T-stage cutoff with time increasing toward the center of the plot.
Table 3. Mortality and Systemic Therapy in Multistate Model. The likelihood of each state transition is compared between patients with tumors 2mm above and below a T-stage cutoff.
Supplementary Tables 4. Multi-state Model Including Covariates at the T1c/T2a Transition
Supplementary Tables 5. Multi-state Model Including Covariates at the T2a/T2b Transition
Supplementary Tables 6. Multi-state Model Including Covariates at the T2b/T3 Transition
4. Discussion
Figure 4. Graphical Abstract
Supplementary Table 7. Neoadjuvant systemic therapy at T-status transitions. Patients receiving adjuvant systemic therapy were excluded from analysis.
Supplementary Table 8. Sensitivity analysis with neoadjuvant chemotherapy group included in multi-state model at the T1c/T2a transition
Supplementary Table 9. Sensitivity analysis with neoadjuvant chemotherapy group included in multi-state model at the T2a/T2b transition
Supplementary Table 10. Sensitivity analysis with neoadjuvant chemotherapy group included in multi-state model at the T2b/T3 transition
Supplementary Table 11. Mortality and Systemic Therapy in Multistate Model. The likelihood of each state transition is compared between patients with tumors 1mm above and below a T-stage cutoff.
