Lung cancer is the most common tumor in the Western world, and its incidence is growing, especially among women. Nonsmall-cell lung cancer (NSCLC) accounts for 75% to 80% of all bronchogenic carcinomas. At the time of diagnosis, 60% to 70% of patients will have either unresectable locally advanced or metastatic disease. The median survival for metastatic disease with palliative radiotherapy or other best supportive care measures ranges from 4 to 5 months, with only 10% to 15% of patients alive at 1 year. Cisplatin(Drug information on cisplatin) (Platinol)-based chemotherapy has improved symptoms, quality of life, survival (by 2 to 3 months), and the 1-year survival (by 10%) of patients.[3,4] However, because the long-term prognosis of advanced NSCLC is still disappointing, more effective chemotherapy is desperately needed.
Carboplatin(Drug information on carboplatin) (Paraplatin) has a more favorable toxicity profile, apart from myelosuppression, than its parent analog cisplatin and has been shown to prolong the survival of patients with advanced NSCLC. Carboplatin as single-agent therapy produced the best 1-year survival rate with the least toxicity in a five-arm Eastern Cooperative Oncology Group (ECOG) study of cisplatin combinations and analogues.
Paclitaxel(Drug information on paclitaxel) (Taxol) has also been tested as a single agent in metastatic NSCLC, and has demonstrated a response rate of 20% to 24%. Of note, the 1-year survival appears favorable.[6,7] When these two drugs are combined, response rates range from 15% to 40%, median survival from 8.3 to 12 months, and 1-year survival from 30% to 50%. Toxicity, including thrombocytopenia, is acceptable.[8-11]
Our group investigated the dose-response relationship of paclitaxel in combination with carboplatin. It appears that full doses of both drugs can be administered without additional toxicity and that a dose-response effect exists with paclitaxel regarding time to progression.
Gemcitabine(Drug information on gemcitabine) (Gemzar), a new pyrimidine antimetabolite, has demonstrated a response rate of 20% in platinum-refractory, advanced NSCLC, with a median duration of response of 29 weeks. When this drug was combined with cisplatin, response rates ranged from 35% to 50%, median survival from 8 to 10 months, and 1-year survival from 35% to 50%. However, the combinations of gemcitabine and cisplatin or gemcitabine and carboplatin are associated with severe myelosuppression.[15,16] Since gemcitabine and paclitaxel have different mechanisms of action and are devoid of overlapping toxicities, the combination of these drugs seems a reasonable approach to the treatment of advanced NSCLC.
There are very few studies of the paclitaxel/gemcitabine combination. Initial reports showed that response rates ranged from 29% to 58% and that toxicity was limited.[17,18] Based on this information, the Hellenic Cooperative Oncology Group (HeCOG) initiated a phase III randomized trial to compare the combination of paclitaxel and carboplatin with the combination of paclitaxel and gemcitabine. The objectives of this study were response rates, time to progression, median survival, 1-year survival, and toxicity. An additional end point was a comparison of quality of life.
Standard eligibility criteria were used in selecting patients for this trial. To enroll, patients had to:
(1) be more than 18 years of age with histologically documented inoperable, recurrent, or metastatic NSCLC (inoperable IIIA, IIIB, or IV stage according to the American Joint Committee on Cancer staging criteria);
(2) have an ECOG performance status of 0 to 2;
(3) have measurable or evaluable disease in nonirradiated fields unless subsequent disease was documented; and
(4) have a life expectancy of at least 12 weeks and adequate hematologic (white blood cell count ³ 4,000/µL, platelet count ³ 100,000/µL), renal (creatinine £ 1.4 mg/dL and creatinine clearance ³ 60 mL/min), and liver (bilirubin £ 1.2 mg/dL, normal alkaline and gamma-glutamyl transferase) function tests.
Patients with active infections and a history of other neoplasms, except for basal cell carcinoma of the skin or in situ carcinoma of the cervix, were excluded from the study. Moreover, patients previously treated with cytotoxic chemotherapy or those with active cardiac disease or pre-existing motor or sensory neuropathy higher than grade 2, according to the World Health
Organization (WHO) scale, were not candidates for the protocol. Women of childbearing age were required to have a pregnancy test within 48 hours of enrollment in the study.
All study patients signed an informed consent, according to Helsinki declarations and to our institutional policies.
Patients eligible for the protocol were randomized to receive either (1) a 3-hour inravenous (IV) infusion of paclitaxel 200 mg/m² followed by a 1-hour IV infusion of carboplatin at an area-under-the-plasma-concentration curve (AUC) of 6, according to the Calvert formula (group A) or (2) the same dose of paclitaxel followed by a 30-minute IV infusion of gemcitabine 1,000 mg/m² on days 1 and 8 of each cycle (group B). Paclitaxel was always administered prior to carboplatin or gemcitabine (Table 1). Treatment was repeated every 3 weeks. All patients received at least two cycles of combination chemotherapy.
If the patients disease did not progress, treatment was continued for a maximum of six cycles. If there was evidence of improvement after the fourth and sixth cycles had been completed, patients could continue treatment at the physicians discretion. All patients received antihypersensitivity prophylaxis before paclitaxel infusion as well as antiemetic pretreatment with 16 mg ondansentron (Zofran) IV infusion 30 minutes prior to treatment, followed by 8 mg ondansentron orally twice a day for 3 consecutive days. No growth factors were used on a routine basis.
If the absolute neutrophil count (ANC) was more than 2,000/µL and the platelet count more than 100,000/µL on the day of treatment, then full doses of the drugs were administered. If hematologic recovery was not achieved on day 35, the patient was withdrawn from the protocol.
Four levels of dose modifications were used for hematologic and nonhematologic toxicity (Table 1). If at any time during treatment, WHO Grade 3 neutropenia and/or thrombocytopenia occurred, drug doses were decreased to level 1. For grade 4 neutropenia and/or thrombocytopenia, drug doses at level 2 were readministered. If neutropenic fever was observed, doses were decreased to level 3. Granulocyte-hematopoietic growth factor (G-CSF) was used for 7 days if the ANC was < 500/µL or in the case of neutropenic fever.
Dose adjustments were also made if severe nonhematologic toxicity developed. Thus, for grade 2 neurotoxicity and hepatotoxicity or grade 3 diarrhea, dose level 1 was readministered. For grade 3 hepatotoxicity, the dose was decreased to level 2. Once the doses were reduced, no dose re-escalation was permitted.
Finally, patients had to discontinue treatment if they had symptomatic arrhythmias or atrioventricular block (except first degree), grade 3 neurotoxicity, or grade 4 hepatotoxicity not reversible within 15 days.
Patients were staged and response was assessed by clinical examination, chest x-ray, and computed tomography (CT). Abdominal CT scans, liver or adrenal ultrasound, and bone scans were performed as indicated.
Disease parameters were measured at least every 8 weeks, and chest x-rays were repeated monthly. All histopathology slides were reviewed by the same group of pathologists.
The usual criteria were followed for definition of response. Time to progression was measured from the start of treatment to disease progression, and survival from initiation of chemotherapy until death. Toxicity was evaluated according to the WHO grading system. The European Organization for Research and Treatment of Cancer (EORTC) Quality of Life instrument for lung cancer was used to assess quality of life.
Beginning in February 1998, 127 eligible patients were randomized, including 63 to group A and 64 to group B. The characteristics of the patients, stage of disease, tumor characteristics, and sites of metastases are listed in Tables 2, 3, and 4, respectively. The majority of the patients were male smokers with good performance status and multiple sites of metastatic disease.
Patients could receive more than 90% of their planned dose (Table 5). There was a trend toward higher response rates in group B (Table 6), although this was not statistically significant (37.5% vs 21.8%, P = .12). With a median follow-up of approximately 4.6 months, time-to-disease progression did not differ in either group (A vs B, 7.1 vs 7.25 months, respectively) (Table 7). Median survival, although not yet reached, did not differ significantly (Table 8).
As expected, hematologic toxicity was generally mild with 9.6% grade 3 and 4 neutropenia in group A and 3.1% in group B (Table 9). Again, no statistically significant difference was found. Grade 3 and 4 thrombocytopenia was rare and occurred only in group A.
Among the nonhematologic toxicities (Table 10), alopecia occurred in both treatment groups, whereas mild arthralgias and myalgias were more common in group B.
Of note, no severe neurotoxicity (grade 3 or 4) was reported for group B, while 5% was reported for group A. Finally, severe diarrhea and severe arrhythmias were recorded for only one patient in group A and one in group B, respectively.
The combination of paclitaxel/carboplatin, which is effective and well tolerated, has become the standard treatment for advanced NSCLC.
Our experience with the published phase II study was very encouraging not only in regard to response but also in regard to lack of major toxicity. Furthermore, in a randomized trial in which we used a standard dose of carboplatin in both arms and two different doses of paclitaxel (225 mg/m² vs 175 mg/m²), the higher dose was more effective in terms of prolonged time-to-disease progression.
However, with this dose, toxicity was significantly more pronounced, mainly neutropenia and neurotoxicity, and usually occurred in elderly patients. For this reason, we decided to use a paclitaxel dose of 200 mg/m² in the next protocols of our group.
The paclitaxel/gemcitabine combination, which is also active and devoid of serious toxicity, seems very attractive for the treatment of advanced NSCLC. Gemcitabine, both as a single agent and in combination with paclitaxel (although in a very limited number of trials), has shown significant activity and has a low toxicity profile.[13-18]
The preliminary results of this randomized study have demonstrated a trend toward higher response rates favoring the paclitaxel/gemcitabine combination, although time-to-disease progression and median survival were similar.
Of note, patients in both groups were able to receive their planned dose of cytotoxic drugs without significant toxicities. However, it is too early to draw a definitive conclusion.
The study is ongoing. Its maturation will permit us to provide more precise and perhaps valuable information.