AMG 925 br Because of the limited sample size of stage
Because of the limited sample size of stage I patients with SCC histology records from previous clinical trials, there was no evidence that would allow us to evaluate the efficacy of ACT in this AMG 925 particular population. In the pre-sent study among p-stage IA SCC patients ACT failed to provide any significant benefit in terms of either RFS or OS. These results were consistent with an earlier meta-analysis that evaluated ACT in stage I NSCLC and showed no improvement of OS among patients with stage IA NSCLC receiving platinum-based chemotherapy (HR,
Fig 4. Kaplan-Meier survival curves for stage IB patients: (A) recurrence-free survival and (B) overall survival. (CI ¼ confidence interval; HR ¼ hazard ratio; No. ¼ number.)
1.40; 95% CI, 0.95–2.06) . However, another meta-analysis demonstrated that postoperative tegafur-uracil could significantly improve survival in patients with stage IA NSCLC compared with surgical procedures alone . This discrepancy might be partly caused by the different chemotherapy agents used in these trials. Because tegafur-uracil has not been approved in China, most chemotherapy agents used in the present study were platinum-based.
The CALGB study was the only trial focusing on evaluating postoperative platinum-based chemotherapy in p-stage IB NSCLC patients. The results of this study demonstrated that ACT failed to provide RFS and OS benefits in p-stage IB NSCLC patients with a tumor that was 4 cm or more in diameter . However, in the cur-rent study our results suggested that postoperative chemotherapy could prolong OS for p-stage IB SCC pa-tients. This discrepancy may have been caused by the difference in study populations. First, the present study was a retrospective investigation; thus, treatment selec-tion bias could have been introduced, which may have affected outcomes. For example, patients in the ACT cohort were younger than those in the non-ACT cohort, XU ET AL 1687 ADJUVANT CHEMOTHERAPY FOR P-STAGE IB SCC
Table 2. Multivariate Cox Regression Analysis of Factors Associated With RFS and OS
Well or moderate
CI ¼ confidence interval; HR ¼ hazard ratio; LVI ¼ lymphovas- cular invasion; OS ¼ overall survival; PET ¼ positron emission tomography; RFS ¼ recurrence-free
survival; VPI ¼ visceral pleural invasion.
which may have favorably influenced their better OS. Therefore, PSM was carried out; the HRs of RFS and OS were 0.58 (95% CI, 0.35–0.96; p ¼ 0.033) and 0.49 (95% CI, 0.27–0.88; p ¼ 0.017), respectively, thus confirming the results before PSM. Furthermore, in the CALGB 9633 study most patients had been diagnosed with adenocar-cinoma. Previous evidence suggested that early-stage SCC might be associated with a higher probability of recurrence in comparison with adenocarcinoma patients [13–15]. Compared with adenocarcinoma SCC is more closely associated with cigarette smoking , which is one of the independent factors associated with a higher risk of recurrence. Unlike some of the stage I adenocar-cinoma patients, who present with a ground-glass opacity appearance on thin-section CT, almost all SCC patients present with a solid tumor appearance on thin-section
THORACIC 1688 XU ET AL Ann Thorac Surg
GENERAL CT. A solid tumor appearance on thin section reflects the proliferation of invasive tumor cells [20, 21]. In addition in another study stage I SCC patients showed higher pro-liferation (MIB-1) and Bcl-2 expression than patients with adenocarcinomas; thus, MIB-1 and Bcl-2 can be consid-ered as predictive factors for relapse . These results suggested that early-stage SCC patients are more likely to require clinical management to prevent recurrence. Be-sides the relatively higher risk of recurrence of SCC, clinical and preclinical data showing that SCC histology predicts a better treatment response to chemotherapy than adenocarcinoma histology may also partly explain the efficacy of ACT for stage IB SCC patients [23, 24].
It should be noted that a meta-analysis including six randomized controlled trials with a total of 2,007 patients showed benefits for both disease-free survival and OS; there was no significant difference in terms of benefit according to tumor histology . The pooled analysis performed by the LACE Collaborative Group had shown that the benefit of adjuvant therapies did not vary with classic clinical factors, except perhaps stage IA disease and a performance status of 2. Of the patients with stage IB disease in these previous analyses the ACT group trended toward benefit (HR, 0.93) but failed to reach statistical significance (95% CI, 0.78–1.10) . In our present study the difference in RFS between ACT and non-ACT cohorts failed to achieve a statistical signifi-cance, although a clear separation of Kaplan-Meier curves for RFS was apparent when these two cohorts were compared. The lack of a statistically significant difference may have arisen because of the limited sam-ple size. There were also other possibilities, such as improved general healthcare for patients seen more frequently during their ACT treatment. This might explain significantly improved OS but failure of statis-tical significance of PFS.