Adjuvant Chemotherapy for Resected Non–Small-Cell Lung Cancer
Adjuvant Chemotherapy for Resected Non–Small-Cell Lung Cancer
Because of the high rate of distant disease recurrence, the 5-year survival of patients who have undergone complete surgical resection of localized non–small-cell lung cancer (NSCLC) is approximately 50%. Initial results from early studies of adjuvant postoperative chemotherapy reported an adverse effect of alkylating agent and older chemotherapy regimens on survival. Cisplatin-based combinations were the first to show a survival advantage. A 1995 meta-analysis of these studies suggested a 13% reduction in the hazard ratio for death (HR = 0.87), leading to a 5% survival benefit at 5 years. Still, these trials involved limited numbers of patients (N = 1,394), and the results failed to reach statistical significance (P = .08). Of the five largest subsequent randomized trials of platinum-based adjuvant therapy, three showed a significant survival advantage. Although it is impossible to determine the reasons for the differing outcomes of these studies, several key features distinguish them, and the data suggest that medically fit patients with resected stage IB or II NSCLC should be offered chemotherapy with a platinum/ new drug combination.
Surgery is potentially curative and represents the most appropriate treatment modality for medically fit patients who present with localized non-small-cell lung cancer (NSCLC; stage I, II, and IIIA without N2 involvement). Despite complete surgical resection, however, recurrence rates remain unsatisfactorily high. The majority of patients who relapse do so with distant metastatic disease, highlighting the need for effective systemic adjuvant therapy. Until recently, convincing evidence for the benefit of postoperative adjuvant chemotherapy has been lacking. However, recent studies have demonstrated the value of modern platinumbased chemotherapy and also of uracil/tegafur (UFT) in defined populations of patients with resected NSCLC. The overall 5-year survival of patients following resection of NSCLC is only about 50%. The major determinant for survival is the stage of the tumor, reflecting the importance of tumor size and particularly lymph node involvement. In American Joint Committee on Cancer (AJCC) pathologic stage I cancer survival is approximately 64.6%, while in stage II disease survival falls to 41.2%. Survival is highest in pathologic stage IA disease (pIA), with 5-year survival rates of 69% to 83% depending on the surgical series. In contrast, the 5-year survival for pIIIA is only about 23%. Other factors that influence survival include histologic subtype (improved survival in squamous cell carcinoma vs adenocarcinoma), involvement of multiple lymph nodes, involvement of multiple levels of N2 disease, and the presence of extranodal extension. Molecular markers such as Ras, p53, and EGFR, although thought to be important, have not consistently been shown to be of prognostic significance. Rationale for Adjuvant Therapy The majority of deaths occur as a result of recurrent disease. Analysis of the pattern of failure in patients who relapse following primary resection of NSCLC reveals that local recurrence occurring in isolation accounts for only about one-quarter of relapses (Table 1).[3-10] Distant recurrence alone or in combination with local recurrence occurs considerably more frequently. The brain is the most frequent site of recurrent metastatic disease, followed by bone and the contralateral lung. These data draw attention to the importance of unrecognized micrometastatic disease present at the time of operation. It is of concern that sensitive immunohistochemical techniques (using antibody to cytokeratin 18) have demonstrated the presence of micrometastatic disease in the bone marrow of as many as 28% to 60% of patients undergoing surgical resection thought not to have evidence of extrathoracic disease after conventional staging.[11-13] The rationale for adjuvant systemic therapy is that postoperative chemotherapy can eliminate residual sites of micrometastatic disease following surgical resection of the primary tumor and prevent the subsequent emergence of incurable clinical disease. It is based on the observation in experimental models that there is an inverse relationship between tumor burden and potential curability by drugs.[14,15] This may be mediated by factors such as the increased proportion of actively dividing cells in micrometastatic deposits, reduced opportunities for the emergence of drug resistance, and improved access of drug to tumor cells. Support for this concept is provided by the demonstrated effectiveness of adjuvant chemotherapy in breast cancer, colorectal cancer, and a growing list of other tumor types including early-stage ovarian cancer and soft-tissue sarcomas of the extremities. In NSCLC, early trials of postoperative adjuvant chemotherapy revealed a detrimental effect of alkylating agent and older chemotherapy regimens on survival. Subsequent studies performed with cisplatin-based combinations were the first to suggest a survival advantage. The magnitude of the survival benefit in these early studies, when seen, was small and often counterbalanced by the toxicities of the agents used. However, recent studies using more effective and less toxic modern platinum- based combinations have provided convincing evidence for the value of adjuvant chemotherapy for NSCLC. These studies and those evaluating a parallel strategy of postoperative adjuvant therapy with UFT treatment are discussed below. Platinum-Based Adjuvant Chemotherapy Trials of adjuvant cisplatin-based chemotherapy for completely resected NSCLC commenced in the 1970s. Several of these early trials evaluated cisplatin in combination with cyclophosphamide and doxorubicin, a combination widely used in patients with advanced disease at that time. These included the Lung Cancer Study Group (LCSG) trial 772 in which 141 patients with resected stage II or III NSCLC were randomized to CAP chemotherapy (six cycles of monthly cyclophosphamide at 400 mg/m2, doxorubicin [Adriamycin] at 40 mg/m2, and cisplatin [Platinol] at 40 mg/m2 ) or intrapleural BCG (TheraCys, TICE BCG) and 6 months of oral levamisole (Ergamisol). Although the recurrence rate was significantly reduced in patients receiving CAP, the improvement in median survival failed to reach statistical significance. Another LCSG study (LCSG 791) randomized 172 patients to postoperative CAP plus thoracic radiotherapy or to thoracic radiotherapy alone. Again improvements were observed in median survival which did not reach statistical significance (20 vs 13 months, P = .113). A third LCSG study (LCSG 801) randomized 269 patients with resected T1, N1 or T2, N0 NSCLC to four cycles of CAP chemotherapy or no treatment. No difference in time to recurrence or median survival was found with CAP chemotherapy. A small Finnish study found a beneficial effect on 5- and 10-year survival of six cycles of postoperative CAP chemotherapy compared with no treatment in 110 patients. However, after adjustment for an imbalance in randomization (more patients underwent pneumonectomies in the observation arm and thus had potentially less favorable disease), the results lost significance. Other randomized trials evaluated postoperative cisplatin and vindesine (alone or followed by UFT) or cisplatin in combination with doxorubicin followed by UFT. Many of these studies suffered from difficulties in accruing patients, heterogeneity in patient populations, and difficulties in delivering full doses of chemotherapy. All failed to demonstrate an unequivocal survival advantage for adjuvant platinum-containing chemotherapy. Meta-analysis
Recognition of the difficulties in detecting small but potentially real survival benefits in individual trials with limited sample size led to a metaanalysis using updated individual patient data conducted by the Non-Small Cell Lung Cancer Collaborative Group reported in 1995. The meta-analysis incorporated data from 52 randomized trials encompassing 9,387 patients. Included were eight trials (1,394 patients, 614 deaths) that evaluated the role of cisplatin-based combination chemotherapy following surgery. These trials used cisplatin in a range of doses (50 to 240 mg/m2 total dose) and in various combinations with doxorubicin, cyclophosphamide, and vindesine. A 13 % reduction in the risk of death was observed in favor of chemotherapy plus surgery compared with surgery alone, which translated to an absolute benefit from chemotherapy of 5% at 5 years (Figure 1). However, the result fell short of statistical significance with 95% confidence intervals (CIs) ranging from a detriment of 1% to a benefit of 10% for adjuvant chemotherapy (hazard ratio [HR] = 0.87; P = .08). This contrasted with the clear benefit of chemotherapy for patients with advanced disease where a 27% reduction in the risk of death was demonstrated (HR = 0.73; P < .0001). ALPI and IALT Trials
In spite of the lack of statistical significance, the suggestion of activity of adjuvant cisplatin-based chemotherapy evident in meta-analysis encouraged the initiation of a large number of randomized studies of postoperative adjuvant chemotherapy, many of which have been recently reported (Table 2).[24-31] The two largest studies and arguably the only studies sufficiently powered to detect an improvement in survival of the size suggested by the meta-analysis were the Adjuvant Lung Project Italy (ALPI) and the International Adjuvant Lung Trial (IALT) study. These two randomized trials provided apparently contradictory results.
- ALPI Trial-The ALPI study randomized 1,209 patients with resected stage I (42%), II (31%), or IIIA (27%) NSCLC to receive chemotherapy with three postoperative cycles of the triplet combination MVP (mitomycin at 8 mg/m2on day 1, vindesine at 3 mg/m2 on days 1 and 8, and cisplatin at 100 mg/m2 on day 1 every 3 weeks) or no chemotherapy. Postoperative radiotherapy was permitted at the discretion of the participating centers and was given to about 65% of patients in the MVP arm and 82% of the patients in the control arm. Results were reported on 1,088 patients (13 patients were ineligible, and 108 patients from one center were excluded because of concerns about data integrity). After median follow-up of 64.5 months no significant difference in overall survival was observed (HR = 0.96, 95% CI = 0.81-1.13; P = .589). Progression- free survival favored chemotherapy, but was not statistically significant (HR = 0.89, 95% CI = 0.76-1.03; P = .128). Only 69% received the planned three cycles of MVP chemotherapy.
- IALT Trial-The IALT was a large international study that evaluated the effect of adjuvant chemotherapy with cisplatin plus a vinca alkaloid or etoposide compared with no adjuvant therapy in patients with resected NSCLC. A total of 1,867 patients with completely resected stage I-III NSCLC were recruited by 148 centers in 33 countries. The study utilized an open choice design to facilitate accrual. Each participating center could determine the pathologic stages of disease to include, the dose of cisplatin given per cycle, the drug that was combined with cisplatin, and the postoperative radiotherapy policy. The chemotherapy options included 3 or 4 cycles of cisplatin (with doses ranging from 80 to 120 mg/m2) in combination with etoposide (49.3%), vinorelbine (26.8%), vinblastine (11.0%), or vindesine (5.8%). A total of 73.8% of patients received at least 240 mg/m2 of cisplatin. In the chemotherapy group 7.8% did not receive any chemotherapy. Seven patients (0.8%) died from chemotherapy related toxicity. Postoperative radiotherapy was given to slightly more patients in the control group (27.7%) than in the chemotherapy group (22.9%). In contrast to the ALPI study, after a median follow-up of 56 months in the IALT study both overall and disease- free survival were found to be significantly improved in the group receiving chemotherapy. An absolute improvement in 2-year survival of 3.6% (70.3% vs 66.7%) and 5-year survival of 4.1% (44.5% vs 40.4%) was seen in the chemotherapy group (HR = 0.86, 95% CI = 0.76-0.98; P < .003). Disease-free survival was also improved in the chemotherapy arm (P < .003) with an absolute difference in deaths or recurrence of 5.1% at 5 years. Subset analyses showed the greatest improvement in overall survival in patients with resected stage III disease.
- Discrepancies Between ALPI and IALT Trials-There are several possible explanations for the discrepancy between the positive IALT study and the negative ALPI trial. A mitomycincontaining triplet combination (MVP) was used in the ALPI trial, while in the IALT trial several cisplatin-based doublet combinations were used. Recent randomized studies have shown that two-drug regimens are less toxic and possibly more efficacious than three-drug regimens such as MVP.[32-35] The negative impact of toxic chemotherapy is likely to be of particular importance in the population of NSCLC patients who frequently suffer from multiple comorbidities and is likely to negate any beneficial effect on survival. Consistent with this is the excess of early deaths (< 12 months from randomization) seen in the chemotherapy arm of the ALPI study (90 patients in the MVP arm and 69 patients in the control arm). Furthermore, improved delivery of chemotherapy was possible in IALT (74%) compared with the ALPI study (69%), although a subset analysis from ALPI did not suggest that this was sufficient to explain the negative result of the study. A further contributory factor is likely to be the use of postoperative thoracic radiotherapy (PORT). A recent update of the PORT meta-analysis[ 36] based on the results of 10 randomized controlled trials and 2,232 individuals continued to show, as did the 1998 meta-analysis, that there was a detrimental effect of surgery followed by postoperative radiotherapy, with an 18% increased relative risk of death (HR = 1.18; P = .02). Postoperative radiotherapy, which is likely to contribute additive toxicity to adjuvant chemotherapy, was used in more patients in the ALPI study than in the IALT study.