10.1245/ASO.2006.01.015
Annals of Surgical Oncology 13:291-301 (2006)
© 2006 Society of Surgical Oncology
Chemotherapy and Radiotherapy in the Treatment of Resectable Non–Small-Cell Lung Cancer
Eric D. Bernstein, MD, MPH,
Scott M. Herbert, MD and
Nasser H. Hanna, MD
Department of Medical Oncology, Indiana University, 535 Barnhill Drive, RT 473, Indianapolis, Indiana 46202
Correspondence: Address correspondence and reprint requests to: Nasser H. Hanna, MD; E-mail: nhanna{at}iupui.edu.
 |
ABSTRACT
|
|---|
Background: Surgical resection remains the cornerstone of therapy for early-stage disease and offers the best chance for cure. Local and distant failure rates, however, remain unacceptably high with surgery alone. Radiation and systemic chemotherapy have been used to reduce recurrences in early-stage disease. Neoadjuvant and adjuvant strategies both offer sound theoretical benefit, but evidence supporting this has been lacking. The publication of a meta-analysis in 1995 triggered a reevaluation of adjuvant chemotherapy. Four randomized trials reported in the last 2 years support the use of adjuvant platinum-based chemotherapy.
Methods: This article reviews the history of clinical trials evaluating neoadjuvant and adjuvant therapy in non–small-cell lung cancer.
Results: Adjuvant chemotherapy improves 5-year survival rates by approximately 5%–15% compared with surgery alone.
Conclusions: Surgical resection followed by adjuvant chemotherapy is the standard of care treatment for patients with completely resected stage I, II, and IIIA non–small-cell lung cancer.
Key Words: Lung neoplasms • Adjuvant chemotherapy • Radiotherapy • Non–small-cell lung carcinoma
|
INTRODUCTION
|
|---|
Lung cancer is the leading cause of cancer-related death in the United States, with an estimated 173,770 new cases (93,110 in men and 80,660 in women) and with 160,440 deaths (91,930 men and 68,510 women) expected in 2004.2 Non–small-cell lung cancer (NSCLC) accounts for >80% of lung cancer cases.
Surgery continues to provide the best hope for cure for patients with stage I or II disease. Five-year survival rates for stage I and II disease vary from 57% to 67% and 39% to 55%, respectively. Patients with completely resected stage III disease have 5-year survival rates of approximately 25%.2 Overall survival for patients with unresectable stage III disease is considerably worse, varying from 3% to 13%. Although surgery remains the cornerstone of treatment for early-stage NSCLC, there is a growing body of literature demonstrating that surgery combined with chemotherapy results in prolonged survival and improved cure rates compared with those with surgery alone. In the last three decades, mixed results from clinical trials failed to provide convincing evidence for a benefit with adjuvant chemotherapy. A meta-analysis published in 1995 suggested that there was a modest survival benefit from cisplatin-based chemotherapy.3 Since that publication, several well-powered randomized trials have shown a significant benefit to adjuvant chemotherapy in NSCLC. The relative benefits of adjuvant chemotherapy in lung cancer seem similar to those seen with breast and colon cancer.
|
NEOADJUVANT THERAPY
|
|---|
Neoadjuvant therapy has potential advantages over adjuvant therapy, which include treating microscopic metastatic disease earlier, in vivo assessment of tumor chemosensitivity, a lower risk of developing drug resistance, the delivery of chemotherapy to an uninterrupted tumor vascular bed, and better compliance with chemotherapy. The potential disadvantages of neoadjuvant therapy include delaying potentially curative surgery and increasing morbidity and mortality associated with surgery after chemotherapy. Some trials report an increase in perioperative mortality after neoadjuvant therapy, especially after right pneumonectomy;4,5 however, a study of 380 patients who underwent resection for NSCLC at M. D. Anderson reported no increase in perioperative morbidity or mortality.6 Patients with breast cancer seem to achieve a similar benefit from neoadjuvant or adjuvant chemotherapy because, in part, of equally high compliance (90%–92%) with both approaches.7 In contrast, only 45% to 60% of patients with NSCLC are able to complete their chemotherapy without dose reductions or delays, as compared with >90% of patients who complete neoadjuvant chemotherapy.8
There are potential biological advantages to giving chemotherapy after surgery as compared to before surgery. Resection of primary lung tumor leads to accelerated growth of disease,9 which results in more cells in their growth phase after surgery and, therefore more susceptible to chemotherapy. Additionally, chemotherapy is generally more effective in treating minimal volume disease as compared to grossly apparent disease.
|
NEOADJUVANT CHEMOTHERAPY
|
|---|
Enthusiasm for the use of neoadjuvant chemotherapy against NSCLC was initially generated by two small studies in stage III disease. In a study from M. D. Anderson, 60 patients with clinical stage IIIA disease (as defined by staging criteria at that time) were randomized to receive cyclophosphamide, etoposide, and cisplatin for three cycles followed by surgery or surgery alone.10 The study planned to enroll 130 patients, but accrual was stopped early because of a statistically significant improvement in survival for patients who received chemotherapy (median survival, 64 vs. 11 months; P < .008) and an improved 3-year survival (56% vs. 15%). Accrual for a second study by Rosell et al.11 that compared neoadjuvant mitomycin, ifosfamide, and cisplatin versus surgery alone in stage IIIA NSCLC was also stopped after 60 patients because of a statistically significant improvement in survival that favored chemotherapy. The median survival was significantly longer in the chemotherapy arm than in the surgery-alone arm (22 vs. 10 months; P = .005). However, the results of these studies are difficult to interpret given the unexpectedly poor survival in the surgery-alone arm.
In the largest phase III study of neoadjuvant chemotherapy reported to date, Depierre et al.12 randomized patients with NSCLC to surgery alone (n = 187) or two cycles of mitomycin C, ifosfamide, and cisplatin followed by surgery (n = 186). Patients with a clinical response to induction chemotherapy received an additional two courses of adjuvant chemotherapy. This study has been criticized because of a significant excess of N2 disease in the chemotherapy arm (40% vs. 28%; P = .06). In addition, 24% of patients in the surgery-alone arm and 31% of those in the combined surgery/chemotherapy arm did not undergo mediastinal lymph node sampling or dissection. Additionally, more patients in the surgery alone arm received postoperative radiotherapy (48% vs. 28%). Seventy-three percent of patients with clinical stage N0/1 were found to have pathologic stage N0/1 in the surgery-alone arm, and only 56% of those with clinical stage N2 had pathologic stage N2; this highlights the need for accurate nodal staging in randomized trials. There were more pneumonectomies in the surgery-alone arm (55.7% vs. 48.6%; P = .3), but there was a non–statistically significant increase in perioperative mortality in the chemotherapy arm (9% vs. 5%; P = .16). Despite these limitations, after adjustment for stage of disease, lower rates of distant recurrence (P = .01) and improved 4-year survival rates in the chemotherapy plus surgery group were reported compared with the surgery-alone group (43.9 vs. 35.3%; P = .09). In contrast to the studies from Roth et al.10 and Rosell et al.11 improvement in survival was limited to patients with stage I and II disease.
Additional evidence for the potential role of neoadjuvant chemotherapy in NSCLC was provided by a multicenter phase II trial that evaluated docetaxel plus cisplatin.13 The study was designed to evaluate safety and toxicity while identifying factors that predict a lack of benefit from surgery. The study included 90 patients with mediastinoscopically proven stage IIIA (T1–3 pN2, M0) NSCLC and a performance status 
2. The therapy was reasonably well tolerated and had acceptable levels of perioperative morbidity (17%) and mortality (3%). This supports the findings of the M. D. Anderson analysis that modern neoadjuvant chemotherapeutic regimens do not result in a significant increase in perioperative morbidity or mortality.6 Mediastinal downstaging occurred in 60% of patients, and a complete pathologic response was achieved in 15% of patients. Mediastinal downstaging from N2 to N0 or N1 was strongly predictive of 3-year survival (61% vs. 11%; P < .0001). The authors concluded that in patients with mediastinal disease resistant to neoadjuvant chemotherapy, resection should not be performed.
Most clinical trials in patients with advanced NSCLC favor a two-drug regimen that includes cisplatin. Carboplatin has been touted as being easier to tolerate and is often substituted for cisplatin in metastatic disease. However, in metastatic disease, carboplatin has been inferior to cisplatin in some trials,14,15 whereas in others the two have been equivalent.16,17 The Bimodality Lung Oncology Trial evaluated neoadjuvant carboplatin plus paclitaxel in a phase II trial.8 After a negative mediastinoscopy, 94 patients with stage IB to IIIA (T3N1) NSCLC were treated with paclitaxel and carboplatin followed by surgery and adjuvant chemotherapy between 1996 and 1999. Despite a negative preoperative mediastinoscopy, 36% of patients were found to have N2 disease at the time of operation. Preoperative compliance rates were high with this regimen: 94% of patients received all of their preoperative chemotherapy, as compared with 45% who completed all of their postoperative chemotherapy. The perioperative mortality of 1% was lower than that with many of the trials using neoadjuvant cisplatin.10,11 The 5-year survival rate was 48%.
Pisters et al.18 presented the S9900 trial evaluating neoadjuvant chemotherapy in patients with stage IB to IIIA (T3N1 only) disease at the 41st meeting of the American Society of Clinical Oncology (ASCO) in 2005. Three hundred thirty-five patients were randomized to receive three cycles of neoadjuvant carboplatin and paclitaxel versus no preoperative treatment. Neither arm was to receive postoperative treatment. This study was closed early after the reporting of positive results in several trials using postoperative chemotherapy. There was a trend toward better survival for the neoadjuvant treatment that did not reach statistical significance. This was similar to the advantage seen in the Depierre trial,12 which also failed to reach significance. Only 77% of patients were able to complete neoadjuvant chemotherapy, as compared with the 94% seen in the phase II Bimodality Lung Oncology Trial. The compliance rate was similar to rates seen in recent postoperative trials.
In an effort to further define a subset of patients who benefit from adjuvant chemotherapy, Kim et al.19 presented data from an adjuvant trial conducted in Japan. Poor-prognostic variables identified were lymphovascular invasion, tumor size, N stage, and histology. The use of adjuvant chemotherapy showed a significant improvement in both disease-free (P = .049) and overall (P = .060) survival in the subset of patients with lymphovascular invasion. An ongoing phase III trial study in Europe is comparing the strategies of neoadjuvant versus adjuvant chemotherapy in patients with resected disease.
|
NEOADJUVANT RADIATION AND CHEMORADIATION
|
|---|
Neoadjuvant radiation or chemoradiation has been evaluated in several randomized trials. Warram 20 reported that some patients are downstaged with preoperative radiation. Although radiation was shown to improve resectability in superior sulcus tumors as far back as 1961,21 two randomized trials from the early 1970s did not show any survival advantage with preoperative radiation alone.20,22
Neoadjuvant chemoradiation has become the standard of care for treating superior sulcus tumors on the basis of two phase II trials reported in 2003.23,24 Phase II studies have also reported that, in a subset of patients, aggressive treatment with chemoradiation can downstage mediastinal disease and make resection feasible in cases for which this would otherwise be impossible.25–27 Two studies have reported that approximately 50% of patients deemed unresectable were able to have R0 resection after neoadjuvant chemoradiotherapy.25,26 In a Southwest Oncology Group phase II trial, patients with stage IIIA (n = 75) or IIIB (n = 51) NSCLC were treated with two cycles of etoposide plus cisplatin and concurrent radiotherapy to 45 Gy.25 More than 50% of patients with N2 disease were downstaged to N0 after chemoradiation. Despite this, almost 40% of patients had a local recurrence, and most patients developed distant metastases. The overall 3-year survival was 27%, but in the subset of patients whose mediastinal disease was eradicated by chemoradiation, the 3-year survival increased to 44%. In patients with persistent mediastinal disease, the 3- year survival was only 18%.
These encouraging results led to a phase III inter-group study designed to evaluate the role of surgery in patients with pathologic N2 disease.28 Between 1994 and 2001, 411 patients received chemoradiation with etoposide plus cisplatin and were randomized to surgery or no surgery. Both arms received two cycles of consolidation chemotherapy. The treatment-related mortality was higher in the surgery arm (7% vs. 1.6%), which was associated with right-sided pneumonectomy. Eighty-eight percent of patients in the surgery arm had R0 resections. Patients in the surgery arm had prolonged 3-year disease-free survival rates (29% vs. 19%; P = .02), but there was no difference in overall 3-year survival (38% vs. 33%; P = .51). Downstaging of mediastinal disease to N0, regardless of T status, by chemoradiation was again predictive of survival; this subset of patients had a 3-year overall survival of 50%.
Despite the negative survival results with the addition of surgery, enthusiasm remains for surgery in patients with resectable N2 disease. The high mortality rates in the surgery arm could have masked an overall survival benefit. There has been concern that preoperative radiation may have increased perioperative mortality on this trial. However, Thomas et al.29 reported results of a phase III trial in 2004 in which 558 patients were randomized to receive neoadjuvant chemotherapy with radiation delivered either before or after surgery. There was no difference in treatment-related mortality (5.6% vs. 5.3%), progression-free survival, or overall survival between groups.29 This question is being further addressed in the planned phase III intergroup trial comparing preoperative concurrent chemoradiotherapy versus preoperative chemotherapy alone in resectable N2 patients.
|
SURGICAL CONSIDERATIONS
|
|---|
There is evidence that performing a complete mediastinal lymph node dissection may also influence survival. Palpation and visual inspection of lymph nodes is only 71% sensitive in detecting pathologic disease.30 Analysis of cases from the Eastern Cooperative Oncology Group 3590 trial (a randomized prospective trial of adjuvant therapy in patients with completely resected stage II and IIIa NSCLC) showed improved survival for patients who underwent complete mediastinal lymph node dissection as compared with those who underwent systematic sampling of their mediastinal nodes.31 Both methods were equally effective in detecting the presence of N2 disease, but multilevel N2 involvement was detected in a much higher percentage of patients after complete dissection as compared with systematic sampling. The median survival was 57.5 months for patients who had undergone complete lymph node dissection and 29.2 months for patients who had systematic sampling (P = .004). In a subgroup analysis, the survival benefit was restricted to patients with right-sided disease, with survival of 66.4 vs. 24.5 months (P = .001), favoring complete dissection. This question has more recently been addressed in a phase III trial, but results have yet to be reported.
Although there are no randomized trials showing improved survival with complete lymph node dissection, the need for accurate staging of the mediastinum cannot be overemphasized. With newer data showing benefit from different surgical or adjuvant approaches in specific stages, accuracy in staging is increasingly important. Mediastinoscopy remains the gold standard for staging, whereas positron emission tomography imaging seems to be superior to computed tomography, with 91% sensitivity and 86% specificity for mediastinal disease.32 Positron emission tomography is also superior to computed tomography in assessing the response to chemoradiation.33
|
ADJUVANT RADIATION
|
|---|
Patients with T1/2N0 disease have 5-year survival rates ranging from 50% to 70% after resection, with few local recurrences (<10%). However, local failure rates and cures decrease with increasing lymph node number and number of stations. The use of adjuvant radiation began as an effort to improve local control and increase cure rates.
Multiple randomized phase III studies have evaluated the role of adjuvant radiation in resected lung cancer. Local control has been consistently improved, with no improvement in overall survival.34–38 The Lung Cancer Study Group trial 773 randomized 230 patients with stage II or III squamous cell carcinoma to observation or radiation after surgery.34 Radiotherapy reduced the locoregional failure rate from 21% to 3% for node-positive patients. There was no difference between arms in overall survival (P = .678). There were numerically more non–cancer-related deaths in the radiation arm (P = not significant). The Medical Research Council trial LU11, which randomized 308 patients (1986–1993) to adjuvant radiation versus observation, also failed to show a survival advantage for radiotherapy.36 This trial also included only patients with stage II (63%) and stage IIIA (37%) disease (68%squamous cell). The median survival time was 19 months in the observation arm compared with 17.5 months for the radiation group (P = not significant). Both trials showed a trend toward improved survival for patients with N2 disease that did not reach statistical significance.
In 1998, a postoperative radiation meta-analysis evaluated 9 randomized phase III trials that included a total of 2128 patients with a median follow-up of 3.9 years.39 There was a statistically significant adverse effect of radiotherapy on survival, with a hazard ratio of 1.21 (P = .001). Postoperative radiation reduced overall survival from 55% to 48% at 2 years. The composite end points of local recurrence-free survival and distant recurrence-free survival favored surgery alone (hazard ratio, 1.16). Most events in these groups were deaths rather than recurrences. Subgroup analysis of the information showed that the detrimental effect of postoperative radiation was greatest for those with stage I disease. Patients with stage II disease showed a smaller increase in mortality, whereas patients with stage III disease showed neither benefit nor detriment.
After the completion of these trials and the subsequent meta-analysis, routine use of adjuvant radiation fell from favor. The postoperative radiation meta-analysis has, however, been criticized. Performance status data were available in only three of the trials. The trials differed considerably in the amount of total radiation given (30–60 Gy) and the tumor volume treated. Seven of nine of the trials used cobalt-60 radiation, a technique now considered obsolete. Finally, most of these trials enrolled patients throughout the 1980s by using computed tomographic planning and radiation-delivery systems that have been markedly improved on. In fact, recent trials incorporating more modern techniques of radiation delivery do not show similar detrimental effects of radiation. Machtay et al.40 found that the adjuvant radiation did not increase the risk of intercurrent death but that age and radiation dose showed trends toward decreased survival.
Although patients with stage I and II disease should not routinely receive adjuvant radiation, patients with high-risk disease (N2) may benefit from this modality. An improvement in survival for patients with N2 disease was seen in the Lung Cancer Study Group trial, and a similar trend was observed in the Medical Research Council trial. The postoperative radiation meta-analysis, however, was unable to demonstrate this beneficial effect. A retrospective study attempted to categorize patients with N2 disease into subsets that would benefit from adjuvant radiation. "High-risk" patients were those with an increased number of involved nodes, higher T stage, and discordant disease (referring to upper or lower lobe tumors with opposite pole mediastinal involvement). These high-risk patients who received adjuvant radiation had a 4-year survival of 37%, compared with 4% for those who did not receive radiation,41 but prospective studies have not confirmed this finding. Phase III trials incorporating modern adjuvant radiation are warranted to further assess this modality, especially in advanced N2 disease.
|
ADJUVANT CHEMORADIATION
|
|---|
Initial attempts to use adjuvant chemoradiation were disappointing. Four trials examined in the 1995 meta-analysis failed to show a survival improvement with this approach. Radiotherapy can be given concurrently or sequentially with chemotherapy. The concurrent approach to chemoradiation results in more toxicity than the sequential approach,42–45 but the concurrent approach decreases the duration of therapy, takes advantage of the radiosensitizing effects of chemotherapy, and delivers control of both local and distant disease at the earliest possible time. In an effort to maximize compliance with therapy, two phase III studies examined sequential chemoradiation in the adjuvant setting. Dautzenberg et al.46 randomized 267 patients from 1982 to 1986 to receive either postoperative radiotherapy (60 Gy) or three courses of cyclophosphamide, doxorubicin, cisplatin, vincristine, and lomustine followed by similar radiation. With a minimum follow-up of 6 years, no significant difference was seen in disease-free survival (P = .47) or 5-year survival (19% vs. 18%; P = .68). Local relapse was similar in both groups (P = .27), and distant metastasis occurred more frequently in the radiation group (P = .09). Wolf et al.47 randomized 150 patients with completely resected N2 disease to receive 3 cycles of cisplatin and ifosfamide followed by 50 Gy of radiation or radiation alone. Enrollment occurred from 1996 to 2000 but was stopped early because of slow recruitment and the inability to achieve the power needed to evaluate 6-month prolongation of survival. There was a trend toward increased survival in the patients who received radiation alone (P = .7).
Adjuvant concurrent chemoradiation was tested in a phase III intergroup trial (E3590) led by the Eastern Cooperative Oncology Group.31 A total of 488 patients with completely resected stage II or IIIa NSCLC were enrolled from 1991 to 1997 and randomized to receive 50.4 Gy of adjuvant radiation or four cycles of cisplatin and etoposide with the same radiation given concurrently. The two groups were well balanced; approximately 55% of patients in each group had N2 disease, and 50% underwent complete lymph node dissection. One hundred three patients were deemed ineligible because of a lack of adherence to the required mediastinal lymph node sampling and dissection technique. The median survival was 39 months in the radiation group and 38 months in the combined chemoradiation group (P = .56). The time to recurrence and patterns of recurrence did not differ significantly between the two groups. The simultaneous administration of chemotherapy and radiation was associated with a higher incidence of serious side effects. A similar percentage of both groups completed the full course of radiotherapy, but only 69% of patients were able to complete all four cycles of chemotherapy. Although there was no difference in treatment-related mortality between groups, there was a significant increase in morbidity in the chemoradiation group.
|
ADJUVANT CHEMOTHERAPY
|
|---|
Adjuvant Chemotherapy With Standard Agents
Early trials with adjuvant chemotherapy in NSCLC failed to show a survival benefit. Most of these trials evaluated alkylating agents and biological response modifiers that have little efficacy in NSCLC. With the increased use of cisplatin during the 1970s and 1980s, results began to favor the use of adjuvant chemotherapy. A meta-analysis published in 1995 containing 4357 patients examined 14 trials using adjuvant chemotherapy.3 The authors found decreased survival with older regimens using alkylating agents and a trend toward improved survival with platinum-based chemotherapy. Eight of these trials (1394 patients) contained cisplatin-based regimens. Six of the eight trials favored the use of adjuvant cisplatin-based therapy, with an overall hazard ratio of .87 (P = .08). The 5-year overall survival was increased by 5% (P = .08).3 This was reaffirmed in a 2000 Cochrane Collaboration meta-analysis that showed a 5% absolute reduction in mortality with adjuvant chemotherapy as compared with surgery alone for NSCLC.48 Although these data were provocative, routine use of adjuvant platinum chemotherapy could not be recommended on the basis of a subgroup analysis of a meta-analysis. Furthermore, most of these trials examined a small number of patients in an era when preoperative staging was often inadequate and platinum chemotherapy was associated with substantial toxicity: i.e., before serotonin antagonists and growth factor support. Several clinical trials evaluating platinum-based chemotherapy have been reported since the publication of these meta-analyses.
In 2002, results from the Adjuvant Lung Project Italy (ALPI) trial were reported.49 Patients (n = 1209) with completely resected stage I, II, or IIIA disease were randomized to receive adjuvant treatment with three cycles of mitomycin C, vindesine, and cisplatin or to a control group. The median age of the patients was 61 years, 50% had squamous cell carcinoma, and 55% underwent complete mediastinal lymph node dissection. Patients were allowed to receive radiation beginning 3 to 5 weeks after chemotherapy at the discretion of the treating center (43% received radiation). There was no significant difference in overall survival between groups (hazard ratio, .96; P = .589). Initially, there was a trend toward decreased survival in the chemotherapy arm. In the first year, there were 90 deaths in the chemotherapy arm as compared with 69 deaths in the control arm. The curves then overlapped from 2 to 4 years with a separation favoring chemotherapy after 4 years, although this was not statistically significant. Of those randomized to chemotherapy, 69% were able to complete treatment, and 51% required dose adjustment.
The largest trial to date of adjuvant chemotherapy in NSCLC was published in 2004. The International Adjuvant Lung Trial (IALT)50 enrolled 1867 patients from 1995 to 1999. Patients with completed resected stage I, II, or III disease were eligible for randomization. The median age of the patients was 59 years, and 46% of tumors were squamous cell. The study design allowed flexibility in the selection of which stages of disease to include, the dose of cisplatin given per cycle, and the agent to be combined with cisplatin. Choices of doublet agent included etoposide (56%), vinorelbine (27%), vinblastine (11%), or vindesine (6%). Adjuvant radiation given sequentially to chemotherapy was planned for approximately 31% of the study population. Seventy-four percent of patients received a cumulative dose of at least 240 mg/m2 of cisplatin, and only seven patients (.8%) died from the toxic effects of chemotherapy. The hazard ratio for death was .86 (P < .03) favoring the chemotherapy group. Five-year survival rates for the chemotherapy group and the control group were 44.5% and 40.4%, respectively. No statistically significant effects were seen in analyses of prespecified subgroups, including doublet agent, dose of cisplatin, demographic characteristics, stage of disease, type of surgery, histological type, and performance status.
The ALPI trial failed to demonstrate the survival advantage for adjuvant chemotherapy observed in the 1995 meta-analyses and the IALT trial. Although the reasons for this are unclear, closer examination of the trials reveals some important differences (Table 1
). To maximize enrollment, both trials included a heterogeneous treatment group with varying stages of disease. The ALPI trial included more patients with early-stage disease. It may be that the smaller size of the ALPI trial (1201 vs. 1867) lacked the statistical power to demonstrate a survival advantage. More postoperative radiotherapy was given on the ALPI trial (43% vs. 21%–25%), and this may have had a detrimental effect on survival. The ALPI trial also incorporated a triplet combination using mitomycin C. In a subsequent meta-analysis presented in 2003, there was increased toxicity with three drug regimens, with no survival advantage.51 A similar regimen had also been associated with an excess of early deaths in a previous trial.12
Received for publication January 25, 2005.
Accepted for publication October 12, 2005.
TOP
ABSTRACT
INTRODUCTION
