Background
Lung cancer remains the leading cause of cancer-related deaths worldwide [
1]. Non-small-cell lung cancer (NSCLC) accounts for 80 % of all lung cancer cases and approximately 40 % of patients with NSCLC present with locally advanced non-small-cell lung cancer (LA-NSCLC) at diagnosis [
2]. The standard-of-care treatment for LA-NSCLC is concurrent platinum-based chemotherapy and thoracic radiotherapy [
3‐
5], which yields superior survival compared with either radiotherapy alone or sequential chemoradiotherapy. However, the outcome of LA-NSCLC treated with concurrent chemoradiotherapy (CRT) remains disappointing, with a median survival of 12–23.2 months [
6,
7].
To improve survival, numerous studies have focused on exploring the feasibility and efficacy of consolidation chemotherapy (CCT) following concurrent CRT with discordant results. A phase II study of the Southwest Oncology Group (SWOG) 9504 [
8] treated patients with concurrent CRT followed by consolidation docetaxel and achieved a promising median survival of 26 months suggesting a possible benefit of CCT. However, the Hoosier Oncology Group (HOG) [
6], who published the only full article on a randomized phase III trial thus far, failed to replicate the encouraging outcome of SWOG 9504 by randomly delivering either docetaxel or observation after CRT. A recent pooled analysis [
2] of 45 studies showed that CCT provided no survival benefit for LA-NSCLC patients. However, a subgroup analysis demonstrated that Asian populations (mostly from Japan and Korea) tended to benefit from CCT, although this benefit did not meet statistical significance (HR = 0.84; 95 % CI, 0.68-1.04;
p = 0.105). Given the lack of substantial evidence from randomized phase III clinical trials, the definitive role of CCT in LA-NSCLC is unknown, especially in the Asian population. Therefore, our study attempted to evaluate the efficacy and toxicity of CCT after concurrent CRT at our institution.
Methods
Ethics statement
This retrospective study was approved by the ethics committee of the Cancer Hospital and Institute of Chinese Academy of Medical Sciences & Peking Union Medical College. Informed consent was exempted by the board due to the retrospective nature of this research. Patient records were anonymized and de-identified prior to analysis.
Eligibility
We retrospectively reviewed the clinical records of LA-NSCLC patients treated with concurrent CRT as an initial treatment at out institution between January 2001 and December 2010. The criteria for inclusion were defined as follows: (1) histologically or cytologically proven NSCLC; (2) clinically diagnosed as stage III disease according to the American Joint Committee on Cancer (AJCC) 2009 staging system; (3) treated with curative thoracic radiotherapy of no less than 50 Gy using intensity modulated radiotherapy (IMRT) or three-dimensional conformal radiotherapy (3D-CRT) with concurrent platinum doublet chemotherapy; (4) treatment responses evaluated 1 month after the completion of concurrent CRT in accordance with the Response Evaluation Criteria for Solid Tumors (RECIST) version 1.1 as complete response (CR), partial response (PR), and stable disease (SD).
Evaluation and follow-up
Complete blood cell counts (CBCs) and blood chemistry examinations were repeated once per week during the treatment period. The follow-up evaluations consisted of a physical examination, CBC, serum biochemistry, tumor marker, thoracic computed tomography (CT) scans, abdomen B-ultrasound examination, and other necessary imaging examinations as clinically indicated at intervals of 3 months for the first year, then every 6 months for the following 2 years, and annually thereafter. Local recurrence was defined as primary tumor recurrence, and regional recurrence was defined as recurrence in the mediastinum, hilum and supraclavicular fossa. Other sites of recurrence, including contralateral lung and metastatic lymph nodes in the neck or axilla, were defined as distant metastasis. Disease progression was determined based on a radiologic examination, histologic examination, or both. Treatment toxicities were graded according to the Common Terminology Criteria for Adverse Events (CTCAE) version 3.0.
Data analysis
Overall survival (OS) was defined from the beginning of concurrent CRT to the time of death due to any cause or last follow-up. Cancer specific survival (CSS) was defined from the beginning of concurrent CRT to the time of death due to lung cancer or last follow-up. Progression-free survival (PFS) was defined from the beginning of concurrent CRT to the time of tumor progression or last follow-up. Local regional progression-free survival (LRPFS) was defined from the beginning of concurrent CRT to the time of local regional progression or last follow-up. Distant metastasis-free survival (DMFS) was defined from the beginning of concurrent CRT to the time of appearance of metastatic disease or last follow-up. Survival analysis was performed using the Kaplan-Meier method and log-rank test. Univariate and multivariate analyses by use of a Cox-proportional hazards model were performed to evaluate potential prognostic factors for OS and PFS. Variables with p < 0.3 in univariate analyses were entered into multivariate analyses. The Pearson χ2 test was used to compare the baseline characteristics and incidence of specific toxicities between treatment groups. Cox proportional hazards models, stratified by age, sex, histology, pretreatment Karnofsky performance score (KPS), stage, treatment response and radiotherapy dose were used to estimate HRs and 95 % confidence intervals (CIs) and test for significance for OS. A statistically significant difference was defined as p < 0.05. All data were processed with SPSS version 19.0.
Discussion
The outcomes of LA-NSCLC are relatively poor, with a high possibility of residual disease after definitive CRT. Thus, many clinical trials have investigated the role of additional CCT. To date, three randomized phase III studies [
6,
9,
10] have been carried out to explore the efficacy and toxicity of CCT, among which only one has been published as a full article. HOG [
6] reported that consolidation docetaxel yielded no survival benefit (median OS, 21.2 months vs. 23.2 months;
p = 0.883) with an increased risk for grade 3/4 pneumonitis (9.6 % vs. 1.4 %;
p < 0.001), infections (11 % vs. 0 %;
p = 0.003), hospitalization (28.8 % vs. 8.1 %) and treatment-related death (5.5 % vs. 0 %;
p = 0.058). In the GILT [
9] study, consolidation oral vinorelbine (NVBo) and cisplatin (P) after NVBo plus P failed to prolong the median PFS (6.4 months vs. 5.5 months;
p = 0.630) and 4-year OS (25.3 % vs. 21.4 %). The multinational CCheIN trial [
10] reported that consolidation DP (docetaxel plus cisplatin) after concurrent weekly DP resulted in a PFS (median PFS, 9.1 months vs. 8.1 months;
p = 0.390) and a OS (median OS, 21.8 months vs. 20.6 months;
p = 0.490) that were similar to those of the observation group. A recently reported pooled-analysis including forty-one phase II/III studies with 3479 patients also failed to provide significant survival benefit of CCT for LA-NSCLC. Unlike HOG, the GILT study and CCheIN trial observed that the addition of CCT did not increase the toxicities. Despite the negative results mentioned above, many oncologists still attempt to deliver CCT for LA-NSCLC patients with good performance status after CRT in routine clinical practice, at least partially due to a poor survival rate of less than 20 % at 5 years and a significant survival benefit achieved by CCT in stage IV disease.
The long-term results of this retrospective study suggest that CCT further prolongs survival compared with CRT alone for LA-NSCLC without increased toxicities. Although more patients in the CCT group had a positive selection factors (female, younger age and a lighter history of smoking), the multivariate analysis was able to account for those selection bias and showed that CCT was a positive prognostic factor for OS and PFS. For patients in the CCT group, the encouraging median OS and 5-year OS were 27 months and 30.4 %, respectively, which were superior to those reported in randomized clinical trials [
6,
9,
10] and comparable to the survival results in SWOG 9504. The median OS and 5-year OS were 16 months and 22.5 %, respectively, in the non-CCT group, which were similar to the historical controls [
4,
7]. Although there was no difference regarding LRPFS or DMFS between the CCT group and the non-CCT group, CCT prolonged survival compared with CRT alone, which may be attributed to several reasons as follows. First, the multivariate analysis for PFS showed that CCT was an independent favorable prognostic factor (HR = 0.643; 95 % CI, 0.441–0.937;
p = 0.022), though we found that the LRPFS (
p = 0.265) and DMFS (
p = 0.779) outcomes were similar between the CCT and non-CCT group. The improvement in disease control may translate into improved survival. The improvement in disease control may translate into improved survival. The multivariate analysis for PFS showed that CCT was an independent favorable prognostic factor (HR = 0.643; 95 % CI, 0.441–0.937;
p = 0.022). A second explanation is that ethnicity may affect the efficacy of CCT. Our result is consistent with a recent pooled analysis [
2] that suggested that survival was better in Asian patients when CCT was delivered, though this improvement was not statistically significant. Soo et al.[
11] reported that the survival and response rate to chemotherapy were better in Asian patients with lung cancer, while the treatment-related toxicities were more severe than in Caucasian patients. To date, the exact mechanisms with which ethnicity affects the efficacy of CCT are unknown. The interethnic difference may be attributable to differences in the genetic backgrounds or environment and culture. Third, it should be noted that the actually delivered cycles of CCT in most studies were relatively lower (0.7 to 3.1, average: 1.5) than those observed in our study (the median number was 3 and 89.4 % of patients completed ≥2 cycles of CCT). Last, bias may be involved in such a retrospective study. The choice of oncologists and patients may influence the administration of CCT. Treatment compliance was higher in patients in the CCT group than in those in the non-CCT group because some patients refused CCT despite the oncologists’ suggestion. Treatment compliance could impact patients’ routine follow up and motivation for salvage treatment after progression, which influences the outcome. The reason why CCT resulted in no significant increase in toxicities may be increased use of IMRT (85.8 % vs. 71.1 %;
p = 0.010) and timely management of toxicity, as IMRT may decrease esophageal and pulmonary toxicity compared with 3D-CRT by increasing target conformity [
12,
13].
Our study also suggested that CCT may lead to significant OS benefit for males, patients with age < 60 years, non-squamous histology, pretreatment KPS ≥ 80, stage IIIb, SD and radiotherapy dose ≥ 60 Gy. It seems plausible that fit patients with higher risk of distant metastasis would benefit from CCT. Interestingly, the fact that the HR for patients achieving SD is favoring CCT, which is contrary to Jeremic [
14] holding the view that patients with a CR or a PR rather than those with a SD were likely to benefit from CCT. However, the number of patients with SD in our study was too small to draw a conclusion.
Prognostic factors are essential to understand the disease process, select treatments and design clinical trials. Numerous studies have investigated the prognostic factors for LA-NSCLC with inconsistent results. The commonly recognized favorable prognostic factors include stage IIIa, good performance status, non-significant weight loss, and female gender [
15‐
17]. In our study, the multivariate analyses identified pretreatment CEA ≥ 5 ng/ml, no CCT, and SD after CRT as predictive of worse OS. Age < 60 years, pretreatment CEA ≥ 5 ng/ml, and no CCT were significantly associated with poor PFS. Our study did not show a significant association between OS or PFS and the widely recognized prognostic factors mentioned above, which may be the result of a relatively small sample size and under-representation of patients with pretreatment KPS < 80 (5.4 %) and weight loss ≥ 5 % (18 %).
Similar to our results, a retrospective study [
18] reported that the clinical tumor response was significantly associated with OS. Kim et al. [
19] found a five-fold likelihood of long term survival for responders (CR or PR) compared to non-responders (SD or PD) (
p = 0.067). Because the clinical tumor response can be assessed soon after CRT, this approach may aid in the following treatment decision according to clinical tumor response to initial CRT because non-responders may need more aggressive treatment.
The prognostic role of age for LA-NSCLC is contradictory. A Radiation Therapy Oncology Group (RTOG)-based analysis [
17] found that age ≤ 70 years was associated with improved survival. Nevertheless, the secondary analysis of RTOG 9410 [
20] demonstrated that in patients treated with CRT, the median OS was longer for patients aged ≥ 70 years (22.4 months vs. 15.5 months,
p-value not provided). Numerous recent trials [
21‐
23] suggested that CRT yielded similar treatment outcome for fit older patients compared with younger patients, which agreed with our results that the elderly (age ≥ 60 years) were non-inferior to the young (age < 60 years) with respect to OS. The reason why age < 60 years acted as a negative predictor for PFS is unknown. The difference in the biological behavior between younger and older patients warrants further investigation.
Although our study is based on a relatively large sample size with a long follow-up period, it has some limitations. Like all other retrospective studies, our study is inevitably subject to multiple biases. Moreover, the CCT regimens were largely heterogeneous, which hindered our study from further exploring the most effective CCT regimen.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
LL participated in acquisition of data, analysis and interpretation of data, and drafting of the manuscript. NB, ZJ, J.Li, JW, XW, ZH, J.Lv, J.Liang, ZZ, YW, WY carried out the study during clinical observation, follow up, collected the clinical data for analysis. LW conceived of the study, participated in its design and coordination and revised the final manuscript. All authors have read and approved the final manuscript.