The novel and effective nonplatinum, nontaxane combination of gemcitabine and vinorelbine in advanced nonsmall cell lung carcinoma

Potential for decreased toxicity and combination with biologic therapy

Abstract

BACKGROUND

Gemcitabine and vinorelbine are two of the most active third-generation agents for the treatment of advanced nonsmall cell lung carcinoma (NSCLC). The authors conducted a formal Phase II trial to evaluate the efficacy of this combination in both untreated and previously treated patients with Stage IIIB (with pleural effusion) or Stage IV NSCLC.

METHODS

A total of 78 patients were treated on the current Phase II trial of front-line or second/third-line therapy with gemcitabine and vinorelbine in NSCLC. Eligible patients manifested either untreated disease (n = 42) or had received at least one but not more than two prior chemotherapy regimens (n = 36). The median age was 59 years (range, 33–79) with 57 men (73%) and 21 women (27%). The median performance status was one (range, one to two). The initial eight patients (four untreated and four previously treated) were treated at a previously established maximum tolerated dose of vinorelbine (30 mg/m²) and gemcitabine (1000 mg/m²) on Days 1, 8, and 15, with significant myelosuppression seen in five out of eight patients requiring dose omission in the first cycle. The next 70 patients received a reduced dose of vinorelbine (25 mg/m²) followed by gemcitabine (900 mg/m²) on Days 1, 8, and 15.

RESULTS

Seventy eight patients were treated. Fifteen (36%) of the 42 evaluable patients who received front-line therapy had objective responses and 14 (33%) had stable disease. In the patients with prior treatment, 6 (17%) of 36 patients had partial response and 18 patients (50%) had stable disease. Median survival time for the previously untreated patient group was 10.1 months, with a one year survival of 43% and a two year survival rate of 32%. For the previously treated patients, the median survival time was 8.5 months, with a one year survival rate of 30%. Toxic effects were notable for significant myelosuppression, with ≥Grade 3 granulocytopenia seen in 55% of the patients on the untreated arm and 67% of the patients on the previously untreated arm. Additionally, 9.5% and 13.9% (untreated and previously treated), respectively, of these patients experienced Grades 3 and 4 thrombocytopenia at some point in their treatment. A full dose delivery analysis showed that this myelosuppression resulted in Course 1, Day 15 skipped doses (even at the reduced dose level) in 42% of previously untreated patients and 47% of pretreated patients. Other side effects seen at Grades 3 and 4 in previously untreated and treated patients included anemia (9.5% and 2.8%), asthenia (4.8% and 5.5%), infection (14.3% and 5.6%), pain (9.5% and 19.4%), and pulmonary complications (4.8% and 13.8%).

CONCLUSIONS

Gemcitabine/vinorelbine is an active, well-tolerated combination in both front-line and second/third-line therapy for Stage IIIB/IV NSCLC. The response rate, median survival rate, and one year survival rate compare favorably with platinum-based regimens. The toxicity profile of the gemcitabine/vinorelbine combination was quite favorable, with minimal Grade 3 and 4 toxic effects aside from granulocytopenia, which resulted in numerous Day 15 skipped doses but no significant febrile neutropenia or infection. The combination of gemcitabine and vinorelbine could be a useful regimen in standard clinical practice and has the potential for efficient combination with biologic/targeted therapy. Multiple randomized trials of this combination versus platinum combinations are now ongoing. Cancer 2002;95:340–353. © 2002 American Cancer Society.

DOI 10.1002/cncr.10629

Lung carcinoma is the leading cause worldwide of cancer-related death, with 156,900 deaths recorded in the United States alone in the year 2000.1–3 Over 80% of newly diagnosed lung carcinomas are nonsmall cell lung carcinomas (NSCLC), with the predominant histologic subtypes in the U.S. being adenocarcinoma and squamous cell carcinoma.1, 2 Incremental improvements in therapy for nonsmall cell lung carcinoma has lead to improvement in the five year survival rate from 9% in 1960 to 15% by 2001.1 The vast majority of patients with this disease present with either metastatic or locally advanced lung carcinoma, which is rarely amenable to ultimate cure. A modest survival benefit has been shown in multiple meta-analyses for patients who receive cisplatin-based chemotherapy compared with best supportive care. More effective therapeutic drug combinations are therefore needed.4–6

The last two decades have seen modest progress in therapy for Stage IV NSCLC. Several different meta-analyses have indicated that platinum-based chemotherapy is superior to best supportive care;4–6 however, the toxicity of cisplatin-containing regimens can be substantial, with potential nausea, vomiting, neurotoxicity, nephrotoxicity, and volume depletion. Recently, a number of new agents, including paclitaxel, docetaxel, topotecan, CPT-11, gemcitabine, and vinorelbine have been tested in Stage IV NSCLC.7 Randomized trials incorporating some of these agents with platinum have shown benefit over older regimens but demonstrate therapeutic equivalence, among the newer combinations with some regimens having slightly less toxicity. In some cases, these agents have been used without platinum with the goal of avoiding the nephrologic and neurologic toxicity of this agent. In addition, it has been suggested that these nonplatinum combinations might be ideal for elderly or poor performance status patients. The nonplatinum, nontaxane combinations are especially interesting, since they could potentially avoid both nephrotoxicity and the neurotoxicity of the taxanes.

Gemcitabine (difluorodeoxycytidine) is a pyrimidine analogue that was initially synthesized as a potential antiviral compound. Activation of deoxycytidine kinase eventually results in gemcitabine triphosphate, which is incorporated into DNA, resulting in chain termination.8 Gemcitabine diphosphate inhibits ribonucleotide reductase, thereby diminishing intracellular difluorodeoxycytidine triphosphate and enhancing the incorporation of gemcitabine triphosphate into DNA. Gemcitabine has been shown to be active against a variety of solid tumors both in vitro and in several human tumor xenografts.9 Multiple Phase II and several Phase III studies incorporating this agent, in combination with cisplatin, have been completed. Single-agent response rates in NSCLC are approximately 20%, with faily mild toxic effects (myelosuppression, transient transaminase elevations) in most cases.10 Phase III studies have compared gemcitabine/cisplatin to cisplatin alone or to other cisplatin-containing regimens in advanced metastatic disease and have shown that the gemcitabine/cisplatin combination is an active regimen with response rates ranging from 31% to 41%11–13 and that the gemcitabine and cisplatin combination was superior to cisplatin alone in terms of survival rate in Stage IIIB/IV NSCLC (median of 9.1 months compared to 7.6 months, P = .004).13

Vinorelbine is a second-generation semi-synthetic vinca alkaloid whose antitumor activity is related to its ability to depolymerize microtubules and disrupt the mitotic spindle apparatus.14 Vinorelbine differs structurally from the other vinca alkaloids due to the substitutions on the catharanthine ring rather than on the vindoline ring. Furthermore, its enhanced affinity for mitotic tubules instead of axonal microtubules is likely to account for its better efficacy and improved toxicity profile over other vinca alkaloids.15 A number of Phase II trials in Advanced NSCLC have shown single-agent response rates of 8% to 37%.16–19 A large, pivotal European Phase III trial compared single-agent vinorelbine with vinorelbine/cisplatin as well as with the European standard of vindesine/cisplatin.20 Cisplatin/vinorelbine significantly improved the response rate compared with cisplatin/vindesine and single-agent vinorelbine (30% versus 19% and 14%, respectively). Median length of survival was also significantly better (40 weeks versus 32 weeks and 31 weeks). As a result of this trial, cisplatin/vinorelbine became the new European standard cisplatin-containing regimen, and single-agent vinorelbine was considered a reasonable option for patients who could not tolerate cisplatin. Similar trials were reported in the U.S., which showed the superiority of cisplatin and vinorelbine over vinorelbine alone21 and equivalence of cisplatin/vinorelbine to carboplatin/taxol.22 A subsequent Phase III trial in elderly NSCLC patients compared single-agent vinorelbine therapy with best supportive care.23 In this study, vinorelbine therapy resulted in improvement in quality of life and in median length of survival.

As both agents had shown significant efficacy, both alone and in combination with cisplatin in the treatment of Stage IIIB/IV NSCLC, we sought to affirm the feasibility and efficacy of the combination in advanced NSCLC, especially since our prior preclinical study showed an additive activity for these agents.24 We employed a 28 day schedule to mimic the schedule used for vinorelbine and gemcitabine as single agents, hoping to maintain a dose intensity equivalent to that recommended for single-agent use. A Phase I trial showing the tolerability and efficacy of this combination led us to investigate this combination in both the front-line and second–or third-line treatment of NSCLC.24 In the current single-center study, we looked at the same regimen in two simultaneous Phase II cohorts.

METHODS

Patient Selection/Eligibility

The current study was stratified into two Phase II cohorts: those patients untreated and those with one or two prior regimens. Patients were considered eligible for the current study based on the following common criteria between the front-line and second/third-line arms: age of at least 18 years a performance status of ECOG 0-2 (Eastern Collaborative Oncology Group [ECOG] > 1 for patients more than 70 years old); a minimum life expectancy of 12 weeks; creatinine and bilirubin < 2.0 mg/dL; and adequate hematologic parameters: hemoglobin ≥ 10 g/dL, white blood cell count > 3,000/μL, absolute neutrophil count > 2000/μL, platelet count > 150,000/μL. Patients with brain metastasis were eligible if they were neurologically stable after whole brain radiation therapy. Brain CT/MRI scans were required prestudy only in symptomatic patients.

Patients were ineligible for the current study if they had evidence of acute intercurrent illness, significant cardiac conduction abnormalities, or known neurologic impairment greater than National Cancer Institute (NCI) toxicity criteria Grade 1 at study entry or documented untreated central nervous system metastasis. Pregnant women and nursing mothers were excluded. All patients signed a written informed consent document approved by the institutional review board. No prior chemotherapy or radiation therapy was allowed within three weeks of study entry.

Patients were eligible for the previously untreated arm of the study if they had received no prior chemotherapy for their NSCLC. Patients were eligible for the previously treated arm if they had been treated with no more than two prior regimens for their Stage IIIB/IV NSCLC and if neither of those regimens contained either gemcitabine or vinorelbine.

Pretreatment and Followup Evaluation

Pretreatment evaluation consisted of a complete history and physical examination, posteroanterior chest x-ray, complete blood count, and serum chemistry analysis. Computed tomographic (CT) scans of the chest to the level of the adrenal glands were obtained in all patients. In addition, CT scans of the abdomen and radionuclide bone scans were performed as indicated. All pretreatment scans were performed within 14 days of study entry. Patients had weekly complete blood counts and liver function tests every four weeks during the study. Computed tomographic scans for disease response were performed every four weeks. An off-study exam was performed within 30 days of the patient leaving the study, if possible. Dose modifications on Days 8 and 15 were as follow: if the granulocyte count was > 1,500/μL and the platelet count was > 75,000/μL, the full dose of each drug was given; however, if the granulocyte count was 1,000–1,500/μL and the platelet count was 50,000–75,000/μL, then doses were reduced by 25% for each drug. For patients with a granulocyte count of < 1,000/μL or platelet count of < 50,000/μL, the dose of each drug was held. Held doses were not made up. For subsequent courses, the drugs were given at full doses once the granulocyte count recovered to > 1,500/μL and the platelet count to > 75,000/μL. If fever and neutropenia occurred during a cycle, doses were reduced by 25% for the next course.

Dose modifications for nonhematologic toxic effects included: a bilirubin laboratory value of Grade 3 (50% dose reduction of vinorelbine, 25% dose reduction of gemcitabine) or Grade 4 (75% of vinorelbine, 50% of gemcitabine), Grade 3 myalgia (25% of vinorelbine), Grade 4 myalgia (50% of vinorelbine), and Grade 4 nausea and vomiting (25% of both drugs).

Treatment Plan/Drug Administration

Treatment was given in the outpatient setting. Patients from the study were treated at two dose levels: dose level 0 (vinorelbine, 30 mg/m² and gemcitabine, 1000 mg/m²) and dose level-1 (vinorelbine, 25 mg/m² followed by gemcitabine, 900 mg/m²; Fig. 1). All but the first eight patients were treated initially at dose level-1. Patients on this trial were premedicated with compazine 10 mg PO (oral)/intravenously (IV) and benadryl 25–50 mg PO. Ondansetron 8–16 mg IV was used only if patients experienced nausea during the first week of gemcitabine/vinorelbine treatment. Patients were treated on a four week schedule, receiving vinorelbine over 10 minutes, followed by gemcitabine over 30 minutes on Days 1, 8, and 15. All patients were required to have central venous access. Courses were repeated every 28 days for a recommended six courses, but further treatment could be given at the discretion of the treating physician in stable or responding patients. Chemotherapy was discontinued in patients with progressive disease or unacceptable toxic effects.

Figure 1. Schema for the current trial showing the 28 day treatment course. Drugs were administered on Days 1, 8, and 15 at the dose levels shown. The initial eight patients were treated with the 0 dose level, and then the remainder (Patients 9–78) with the −1 level as shown.

Dose Intensity

The percentage dose intensity of gemcitabine and vinorelbine was calculated as a ratio of the average dose of gemcitabine or vinorelbine delivered in milligrams per meter squared per week to the maximum possible average weekly dose of gemcitabine; the resulting fraction was multiplied by 100%. The number of held or reduced doses was determined by dividing the total doses theoretically possible by the actual doses delivered or reduced. These calculations were made using dose level 0 (gemcitabine, 1000 mg/m² and vinorelbine; 30 mg/m²) for the first eight patients and dose level −1 (gemcitabine, 900 mg/m² and vinorelbine, 25 mg/m²) for the subsequent 70 patients, as that was the intended dose at study entry for each group.

Toxicity Evaluation

All patients were considered assessable for toxicity and survival if they received at least one treatment of gemcitabine and vinorelbine. The NCI common toxicity criteria were used. Several patients were registered for the study but ultimately did not receive treatment at The University of Texas M.D. Anderson Cancer Center, having decided to return to their local physician. Data from these patients was excluded from the results.

Response and Survival Assessment

Response was assessed using World Health Organization response criteria, such that a complete response was validated when disease completely disappeared on radiographic images for a minimum of four weeks without any new disease. A partial response occurred if there was a greater than or equal to 50% reduction in the sum of the products of the diameters of all measurable lesions. Disease progression was documented when a greater than or equal to 25% increase in the sum of the products of all measurable lesions was seen. Stable disease was defined by the remaining patients.

Statistical Analysis

The optimal two stage design for Phase II clinical trials proposed by Simon was used for two concurrent Phase II studies.25 Using a 10% Type I error and 90% power (targeting a response rate of 40% for untreated patients and 20% for patients with prior therapy), the sample size was 37 for each group. The first step (14 patients) for the untreated group required three responses while the treated group (14 patients) required two patients to respond. A minimum of 11 responses was needed in the front-line group and 4 in the second-line group to reach this goal. All survival curves were generated using a Kaplan-Meier estimate. Survival time and disease free survival time data were collected for all patients on an intent-to-treat basis. Length of survival was defined as the time between the initiation of treatment and death. If death did not occur, survival time was considered censured at the last known date alive. Progression free survival was defined as the time between initiation of treatment and disease progression. All the analysis was performed using SAS and S-plus software (SAS: Cary, NC).

RESULTS

Patient Characteristics

Between January 1998 and May 1999, seventy eight patients were enrolled in the current study at The University of Texas at M. D. Anderson Cancer Center. The study was stratified according to whether or not patients had received prior chemotherapy. Table 1 details the characteristics of the patients on the trial. For the total group, median age was 60 years, and all patients had performance status ECOG 1 or 2, with 57 men and 21 women on the trial. Histologic subtypes of disease included adenocarcinoma (51%), squamous cell carcinoma (22%), large cell carcinoma (1%), bronchoalveolar cell carcinoma (3%) and unclassified NSCLC (22%). There were no appreciable demographic differences between the untreated and previously treated groups.

Table 1. Patient Characteristics

Characteristic	Full study	Untreated	Prior treatment
a One patient had three prior regimens (not known prior to study entry) and was included for analysis.
Total number of patients	78	42	36
Median age in years (range)	60 (33–79)	63.5 (44–79)	57.5 (33–70)
Performance status
1	73 (94%)	41 (98%)	32 (89%)
2	5 (6%)	1 (2%)	4 (11%)
Gender
Men	57 (73%)	33 (79%)	24 (67%)
Women	21 (27%)	9 (21%)	12 (33%)
Histologic disease type
Adenocarcinoma	40 (51%)	22 (52%)	18 (50%)
Squamous cell carcinoma	17 (22%)	7 (17%)	10 (28%)
Bronchoalveolar	2 (3%)	2 (5%)	0 (0%)
Sarcomatoid	1 (1%)	0 (0%)	1 (3%)
Large cell	1 (1%)	0 (0%)	1 (3%)
Unspecified	17 (22%)	11 (26%)	6 (16%)
Disease stage in patients
IIIB	9 (12%)	8 (19%)	1 (3%)
IV	69 (88%)	34 (81%)	35 (97%)
No. of prior chemotherapy regimens
0	42	42	0
1	31	0	31
2	4	0	4
3a	1	0	1s

In the untreated group, 34 (81%) of the patients had Stage IV disease and 8 (19%) had Stage IV. All but one of these IIIB patients manifested T4 staging with a malignant pleural effusion. The previously treated group had 35 patients (97%) with Stage IV disease.

Of the 42 front-line patients who were registered on the trial, four (10%) were unable to complete two courses of therapy on protocol. Three patients received their first course of treatment at The University of Texas M. D. Anderson Cancer Center and subsequent courses in their hometowns. Data from these latter treatments were unobtainable. One patient died from pneumocystis carinii pneumonia during his first course of therapy. All 36 second-line patients were evaluable.

Breakdown of Prior Therapy

Of the 36 patients with prior chemotherapy, 20 patients had received carboplatin and paclitaxel as prior therapy and 10 others had received a regimen containing platinum. Therefore, thirty of the thirty six patients who received prior therapy had progressed either during or after some form of platinum therapy. Four patients had received two prior chemotherapy regimens and one patient had received three. This latter patient was technically not eligible but since this was not realized until after auditing, this patient was included in the analysis.

Response/Time to Progression

All patients were included for survival and response assessment in an intent-to-treat analysis (Table 2). Of the 42 patients on the front-line chemotherapy arm, 15 (36%) experienced a partial response as their best response that was confirmed by radiographic assessment a minimum of four weeks later. Fourteen (33%) patients had stable disease, and 9 (21%) patients had disease progression at the first eight week interval. Four patients (10%) were unevaluable.

Table 2. Response Data for Evaluable Patients

Type of treatment	No. of patients	Percentage
First line treatment (42)
Partial response	15	36%
Stable disease	14	33%
Progressive disease	9	21%
Inevaluable	4	10%
Second/third line treatment (36)
Partial response	6	17%
Stable disease	18	50%
Progressive disease	12	33%

Of the 36 patients receiving gemcitabine/vinorelbine as second- or third-line chemotherapy, six (17%) achieved a partial response and eighteen (50%) had stable disease at the time of the first eight week radiographic assessment. Only 12 patients (33%) had progressive disease on this regimen despite prior platinum therapy in the majority. Figure 2 shows the serial radiographic assessment of one patient who was previously treated with 10 courses of carboplatin and paclitaxel, developed disease progression (including brain metastases) for which he received radiation, and then received second-line therapy with gemcitabine and vinorelbine for approximately two years. As shown, in December 1997, he had widespread pulmonary lesions with bilateral disease. Excellent interval responses were then seen in February and June of 1998 and in March of 1999. This gentleman went on to receive 26 courses of the drugs in the second-line setting. It was also notable that this man was able to keep working during therapy and had no allergic reactions or alopecia.

Figure 2. Near complete response in a 41-year-old man who received gemcitabine and vinorelbine in the second-line setting. A) Baseline computed tomographic (CT) scan of the lung (12/97) shows a large mass in the right lower lobe with multiple bilateral pulmonary metastases representing overt disease progression after treatment with carboplatin and taxol. B) A followup CT after Course 2 2/98) showed a decrease in size of the right lower lobe mass and decreased pulmonary metastases. C) A followup CT scan after Course 6 (6/98) showed a significant decrease in the size of the right lower lobe mass. The metastatic pulmonary lesions are almost completely resolved on this scan. D) A followup CT scan after Course 15 (3/99) again showed the near complete disappearance of the right lobe lower mass. The multiple metastatic pulmonary lesions have cleared.

Length of Survival

In an intent-to-treat analysis for all 42 front line patients, median overall survival time for the previously untreated patients (Fig. 3A) was 10.1 months, with a one year survival rate of 43% and a two year survival rate of 32%. For the previously treated patients (Fig. 4A), median survival time was 8.5 months with a one year survival rate of 30%. Time to tumor progression was also evaluated for each subgroup. For the untreated patients, the median time to tumor progression was 6.2 months (Fig. 3B). For the previously treated patients, median tumor progression free survival was 4.6 months (Fig. 4B).

Figure 3. Kaplan-Meier survival curves for patients treated with gemcitabine and vinorelbine with no prior chemotherapy treatment. A) The median overall survival time was 10.1 months (95% confidence interval [CI] of 6.3–27.6 months), one year survival was 42.9% (95% CI of 30.2–60.8%), and two year survival was 31.6% (95% CI of 19.2–52.1%). B) The median progression free survival was 6.2 months (95% CI of 4.5–11.7 months), one year progression free survival was 29.2% (95% CI of 17.8–47.9%), and two year survival was 20.4% (95% CI of 10.8–38.7%).

Figure 4. Kaplan-Meier survival curves for patients who had received prior therapy. A) The median overall survival was 8.5 months (95% CI of 6.7–12.3 months), one year survival was 30% (95% confidence interval [CI] of 18.2%–50.6%), and two year survival was 12.1% (95% CI of 4.9–30.3%). B) The median progression free survival was 4.6 months (95% CI of 2.8–6.6 months), the one year progression free rate was 13.9% (95% CI of 6.2–33.3%), and the two year rate was 8.3% (95% CI of 2.8–24.6%).

Hematologic Toxic Effects

Major hematologic toxic effects encountered on this trial are shown in Table 3: 55% of previously untreated patients (29% Grade 3, 26% Grade 4) and 67% of previously treated (42% Grade 3, 25% Grade 4) patients experienced Grades 3 and 4 granulocytopenia; 9.5% of previously untreated patients had Grade 3 and 13.9% of previously treated patients (8% Grade 3, 6% Grade 4) experienced Grade 3 or 4 thrombocytopenia. Despite this, only 14.3% of previously untreated patients and 5.6% of previously treated patients experienced Grades 3 and 4 infection. Only six (12%) total patients in both arms experienced fever with neutropenia, and only one of these patients (in the previously treated arm) had Grade 3 toxicity of this effect.

Table 3. Toxicity by Patients Broken Down by Prior Therapy (Worst by Patient)

Side effect	Untreated (42 patients) Grade/no. of patients					Prior Treatment (36 patients) Grade/no. of patients
	1	2	3	4	% 3/4	1	2	3	4	% 3/4
Allergic reaction	8	1	0	0	0	5	2	0	0	0
Alopecia	11	2	0	0	0	1	5	0	0	0
Anemia	20	14	4	0	9.5	12	15	6	1	2.8
Arthralia/myalgia	9	2	0	0	0	12	2	0	0	0
Asthenia	25	9	2	0	4.8	12	16	2	0	5.5
Cardiac	5	1	2	0	4.8	3	5	0	1	2.8
Constipation	9	2	1	0	2.4	9	3	1	0	2.8
Diarrhea	3	2	1	0	2.4	2	1	0	0	0
Edema	6	2	0	0	0	7	2	0	0	0
Fever	7	10	0	0	0	8	9	1	0	2.8
Fever/neutropenia	1	1	0	0	0	2	1	1	0	2.8
Gastrointestinal	9	0	0	0	0	7	2	1	0	2.8
Granulocytopenia*	3	7	12	11	55	3	7	15	9	67
Hyperglycemia	0	0	0	0	0	0	1	0	1	2.8
Infection	4	6	4	2	14.3	2	8	1	1	5.6
Mucositis	7	2	0	0	0	1	2	0	0	0
Nausea/vomiting	7	8	1	0	2.4	11	3	3	0	8.3
Neurologic toxicity	20	4	0	0	0	19	5	1	0	2.8
Pain	15	3	4	0	9.5	13	3	7	0	19.4
Phlebitis	0	6	1	0	2.4	2	2	0	0	0
Pulmonary	26	9	2	0	4.8	13	8	4	1	13.8
Skin reaction	7	9	1	0	2.4	5	4	0	0	0
Thrombocytopenia	26	8	4	0	9.5	23	7	3	2	13.9
Thrombosis	0	1	2	0	4.8	0	3	0	0	0
Transaminitis/Liver Function Tests	0	0	1	0	2.4	0	0	0	1	2.8

Nonhematologic Toxic Effects

Pain usually but not always at the injection site was the most significant nonhematologic toxicity (9.5% of previously untreated patients and 19.4% of previously treated patients). This number is high despite the fact that patients were required to have a central line placed. This is likely because M. D. Anderson Cancer Center favors long lines, which track through the anticubital fossa and, hence, are prone to infiltration at the insertion site. Grades 3 and 4 pulmonary toxicity were seen in 4.8% of the untreated and 13.8% of the previously treated patients. Of the 29 hospitalizations on protocol, nine patients (31%) were hospitalized for pneumonia and three (4%) had pneumocystis carinii pneumonia. Bactrim prophylaxis was added for patients on steroid therapy toward the end of the trial. Asthenia at Grade 3 or 4 was observed in 4.8% of untreated and 5.5% of treated patients. Notably, there was little, if any, alopecia or neuropathy seen.

Dose Modifications

As reported above, the primary side effects caused by the combination of gemcitabine and vinorelbine were hematologic and occurred in large part because the regimen was administered on Days 1, 8, and 15. Dose delivered by cycle is quantitated in Table 4. With 174 total courses, the median number of courses of therapy administered per patient was 4 in the untreated group (range, 1–14) and 3 in the treated group (range, 1–26), with 168 total four week courses.

Table 4. Median Number of Courses Broken Down by Treated and Untreated Subgroups

Subgroup	Total coursesa	Median no. of courses	Range
a One course is equal to 28 days of therapy.
Untreated	174	4	1–14
Treated	168	3	1–26
Total	342	4	1–26

Dose Intensity of Gemcitabine/Vinorelbine

Table 5 and Table 6 show the dose intensity of gemcitabine and vinorelbine delivered over the first three courses for the first 8 patients who received the full dose (Fig. 5) and the next 70 patients who received the reduced dose level (Fig. 6). At the reduced dose level of 900 mg/m² of gemcitabine and 25 mg/m² of vinorelbine, previously untreated patients who completed three courses of therapy (25 patients) received 84.3% of the targeted vinorelbine dose and 83.8% of gemcitabine dose at Course 3. Using the same analysis for the previously treated group at the third course (a total of 18 patients), 76.1% of the vinorelbine and 80.6% of the gemcitabine was delivered. The results for patients completing only one or two course of therapy are quite similar, as shown in Tables 5 and 6.

Table 5. Doses Delivered Over the First Three Coursesa (Patients 1–8)

	Course 1	Course 2	Course 3
a Based on the 0-dose level: 1000 mg/m^2 gemcitabine + 30 mg/m^2 vinorelbine.
	n = 4	n = 4	n = 3
Untreated
Vinorelbine	83.8%	77.6%	77.3%
Gemcitabine	83.3%	80.0%	80.0%
	n = 4	n = 3	n = 2
Previously treated
Vinorelbine	75.3%	71.2%	83.6%
Gemcitabine	75.3%	71.6%	87.3%

Table 6. Doses Delivered Over the First Three Coursesa (Patients 9–78)

	Course 1	Course 2	Course 3
a Based on the −1 dose level: 900 mg/m^2 gemcitabine + 25 mg/m^2 vinorelbine.
	n = 38	n = 30	n = 25
Untreated
Vinorelbine	80.0%	84.2%	84.3%
Gemcitabine	80.6%	81.8%	83.8%
	n = 32	n = 26	n = 18
Previously treated
Vinorelbine	75.7%	84.2%	76.1%
Gemcitabine	84.2%	77.5%	80.6%

Figure 5. Proposed schema for the combination of cytotoxic chemotherapy with biologic therapy in advanced nonsmall cell lung cancer: (NCSLC) The goal of early detection is to identify patients with local disease, who can be treated with either surgery or radiation. However, the majority of patients with NSCLC die of metastatic disease, and many have disease not amenable to local therapy (with metastasis to brain, liver & bone). For them, a background chemotherapy regimen (i.e., gemcitabine/vinorelbine) can be added to a biologic modality as shown. Following maximal cytotoxic therapy, the response can be maintained and metastasis hopefully prevented to distant organs using a biologic agent (or agents) as a maintenance therapy.

Skipped Drug Doses

Held doses of gemcitabine and vinorelbine were not made up. Tables 7 and 8 show reduced or skipped doses of gemcitabine and vinorelbine by course and week. Table 7 shows the number of patients who had reduced or skipped doses in Patients 1–8 who received the increased dose of both drugs. In the four untreated patients, two (50%) skipped their Week 3 (Day 15) dose in Course 1. Of the four previously treated patients, three (75%) skipped their Week 3 dose in Course 1. It is for this reason that drug doses were reduced for the subsequent 70 patients. Table 8 shows reduced or skipped doses of drug for Patients 9–78. The slightly reduced drug doses had little effect on the Week 3 (Day 15) missed drug frequency. For the first course in the previously untreated group, 42% of the patients skipped Week 3. Week 3 was skipped in 47% of the previously treated patients in Course 1. For the untreated and previously treated patients who received three courses, the Week 3 dose was missed 24% and 33% of the time, respectively. Many Day 15 doses were held, which was the main drawback of the 28 day schedule.

Table 7. Number of Patients Who Had Reduced or Skipped Doses of Drug (Patients 1–8)

	Previously untreated Vinorelbine/gemciatine			Previously treated Vinorelbine/gemciatine
	Not reduced	Reduced	Skipped	Not reduced	Reduced	Skipped
		Course 1 (n = 4)			Course 1 (n = 4)
Week 1 (Day 1)	4 (100%)	0	0	4 (100%)	0	0
Week 2 (Day 8)	4 (100%)	0	0	4 (100%)	0	0
Week 3 (Day 15)	2 (50%)	0	2 (50%)	1 (25%)	0	3 (75%)
		Course 2 (n = 4)			Course 2 (n = 3)
Week 1 (Day 1)	2 (50%)	2 (50%)	0	1 (33%)	2 (67%)	0
Week 2 (Day 8)	2 (50%)	2 (50%)	0	1 (33%)	2 (67%)	0
Week 3 (Day 15)	2 (50%)	0	2 (50%)	1 (33%)	1 (33%)	1 (33%)
		Course 3 (n = 3)			Course 3 (n = 2)
Week 1 (Day 1)	1 (33%)	2 (67%)	0	1 (50%)	1 (50%)	0
Week 2 (Day 8)	1 (33%)	2 (67%)	0	1 (50%)	1 (50%)	0
Week 3 (Day 15)	1 (33%)	1 (33%)	1 (33%)	1 (50%)	1 (50%)	0

Table 8. Number of Patients Who Had Reduced or Skipped Doses of Drug (Patients 9–78)

	Previously untreated Vinorelbine/gemcitabine			Previously treated Vinorelbine/gemcitabine
	Not reduced	Reduced	Skipped	Not reduced	Reduced	Skipped
		Course 1 (n = 38)			Course 1 (n = 32)
Week 1 (Day 1)	37 (97%)	1 (3%)	0	32 (100%)	0	0
Week 2 (Day 8)	29 (76%)	5 (13%)	4 (11%)	21 (66%)	6 (19%)	5 (16%)
Week 3 (Day 15)	19 (50%)	3 (8%)	16 (42%)	9 (28%)	8 (25%)	15 (47%)
		Course 2 (n = 30)			Course 2 (n = 26)
Week 1 (Day 1)	26 (87%)	3 (10%)	1 (3%)	22 (85%)	4 (15%)	0
Week 2 (Day 8)	23 (77%)	7 (23%)	0	20 (77%)	5 (19%)	1 (4%)
Week 3 (Day 15)	15 (50%)	6 (20%)	9 (30%)	15 (58%)	3 (12%)	8 (31%)
		Course 3 (n = 25)			Course 3 (n = 18)
Week 1 (Day 1)	21 (84%)	4 (16%)	0	13 (72%)	5 (28%)	0
Week 2 (Day 8)	18 (72%)	6 (24%)	1 (4%)	11 (61%)	5 (28%)	2 (11%)
Week 3 (Day 15)	12 (48%)	7 (28%)	6 (24%)	7 (39%)	5 (28%)	6 (33%)

DISCUSSION

The current study assessed the combination of gemcitabine and vinorelbine in both the front- and second/third-line settings and showed significant activity with acceptable toxicity. The results are quite encouraging. Median survival time on the front-line arm was 10.1 months, and on the prior therapy arm, 8.5 months. Both front-line and second/third-line survival figures are favorable when compared to others in the literature for platinum containing regimens. However, 42–47% of patients who received one course of therapy had omission of the Day 15 gemcitabine/vinorelbine dose secondary to myelosuppression. This rarely translated into significant febrile neutropenia, but the degree of myelosupression in a substantial proportion of patients suggests that administration of this combination on a three week instead of a four week schedule may be preferable.

Other investigators have explored the gemcitabine/vinorelbine combination (Table 9). Frasci et al.26 compared the combination of gemcitabine and vinorelbine versus vinorelbine alone in a randomized trial of over 120 elderly patients (age ≥70). The median age on their trial was 74 years, with 59% of patients having Stage IV NSCLC and 41% having Stage IIIB disease. The trial was halted after 120 patients were accrued as the gemcitabine/vinorelbine arm had a significant improvement in median (29 weeks versus 18 weeks; P < 0.01) and one year survival (30% versus 13%; P < 0.01). In addition, response was significantly higher with the combination arm (22% versus 15%; P < 0.01). Grades 3 and 4 neutropenia occurred more often with the gemcitabine/vinorelbine combination (38% versus 28%). Thirteen percent of patients who received gemcitabine/vinorelbine experienced Grades 3 or 4 thrombocytopenia versus 8% of patients who received only vinorelbine. The overall survival in this trial was lower than the value in the current study but only elderly patients were included. Recently, the multicenter Italian lung cancer in the elderly study was presented in abstract form (MILES study). This study compared single agent gemcitabine or vinorelbine with the combination of gemcitabine and vinorelbine in an elderly population.27 In this initial report, there was no significant benefit for the combination over the single agents alone, with one year survival of 26% with gemcitabine, 41% with vinorelbine, and 31% with gemcitabine and vinorelbine. Dose delivery data, however, was not available yet from the abstract report.

Table 9. Gemcitabine + Vinorelbine Combinations

Reference	Prior therapy	Gemcitabine (mg/m^2)	Vinorelbine (mg/m^2)	No. of patients	Response rate (%)	Median survival (mo)	One year survival rate (%)	Primary toxicities/comments
a Day 1, 8, 15 every 28 days, but if the patient could not receive Day 15, the cycle was switched to 21 days.
Gridelli et al.30	No	1000–1200 Days 1, 8 every 21 days	25–30 Days 1, 8 every 21 days	126	26%	7.6	33%	27% Grade 3/4 neutropenia
Esteban et al.47	No	1000–1500 Days 1, 8 every 21 days	25–30 Days 1, 8 every 21 days	32	43%	NR	NR	13% Grade 3/4 neutropenia
Lilenbaum et al.29	No	1000–1250 Days 1, 8 every 21 days	25 Days 1, 8 every 21 days	32	25%	8.3	38%	38% Grade 3/4 neutropenia
Gridelli et al.48	No	1000 Days 1, 8 every 21 days	25 Days 1, 8 every 21 days	49	18%	NR	NR	12% Grade 3/4 neutropenia Elderly (all subjects ≥ 70 years)
Beretta et al.49	No	1000 Days 1, 8 every 21 days	25 Days 1, 8 every 21 days	43	35%	8	31%	35% Grade 3/4 neutropenia, 12% reversible transaminase elevation Elderly (≥ 65 years) and/or cisplatin contraindicated
Lorusso et al.28	No	1200 Days 1, 8 every 21 days	30 Days 1, 8 every 21 days	52	37%	12.5	52%	37% Grade 3/4 leukopenia, 2% Alanine aminotransferase/Astartate aminotransferase toxicity
Frasci et al.26	No	1200 Days 1, 8 every 21 days	30 Days 1, 8 every 21 days	60	22%	6.7	30%	38% Grade 3/4 neutropenia Elderly (all subjects ≥ 70 years)
Bajetta et al.50	No	1250 Days 1, 8 every 21 days	25 Days 1, 8 every 21 days	54	30%	12	49%	11% Grade 3 neutropenia, 9% Grade 3 flu-like syndrome
Chen et al.51	No	800 Days 1, 8, 15 every 28 days	20 Days 1, 8, 15 every 28 days	40	73%	11	48%	53% Grade 3/4 neutropenia, 15% interstitial pneumonitis
Feliu et al.52	No	1000 Days 1, 8, 15 every 28 days	25 Days 1, 8, 15 every 28 days	49	26%	12	33%	12% Grade 3/4 neutropenia Elderly (≥ 70 years) and/or cisplatin contraindicated
Baron et al.53	No	1000 Days 1, 8, 15 every 28 days	25 days 1, 8, 15 every 28 days	40	29%	8.5	40%	5% Grade 3 anemia and neutropenia
Laack et al.54	No	1000 Days 1, 8, 15 every 28 days	30 days 1, 8, 15 every 28 days	70	41%	8.3	34%	40% Grade 3/4 neutropenia
Krajnik et al.55	No	1200 Days 1, 8, 15 every 28 days	25 days 1, 8, 15 every 28 days	78	19%	7	32%	35% Grade 3/4 neutropenia
Isokangas et al.56	No	1200 Days 1, 15 every 28 days	35 days 1, 15 every 28 days	33	46%	8	24%	24% Grade 3/4 neutropenia
Hirsh et al.57	Yes	1000 flexiblea	25 flexiblea	34	53%	11.1	50	5% Grade 3 anemia and neutropenia
Hainsworth et al.31	Yes	1000 Days 1, 8, 15 every 28 days	20 days 1, 8, 15 every 28 days	50	18%	6.5	20%	36% Grade 3/4 neutropenia 85% prior taxane and platinum
Barr et al.32	Yes	750 Days 1, 8, 15 every 28 days	30 Day 1 15 Day 8 30 Day 15 every 28 days	27	20%	NR	NR	41% bone marrow suppression, 39% requiring Granulocyte Colony Stimulating Factor Patients who Failed to respond to prior taxane therapy
Herrero et al.58	Yes	1200 Days 1, 8, 15 every 28 days	25 Days 1, 8 every 28 days	16	6%	5.8	NR	11% cycles Grade 3 neutropenia Prior platinum therapy

Lorusso et al.28 reported on the Phase II combination of gemcitabine and vinorelbine in the treatment of patients with advanced NSCLC. The investigators administered the treatment on Days 1 and 8 of a 21 day schedule, with 30 mg/m² of vinorelbine and gemcitabine at 1,200 mg/m². Fifty two evaluable patients with Stage IIIB (54%) and Stage IV (46%) NSCLC showed a response rate of 36% and a median survival time of 12.4 months (14.4 months for Stage IIIB and 9.6 months for Stage IV). A clinical benefit response (improvement in performance status, weight gain, and pain reduction) was observed in 56% of patients. The higher survival rate in this study likely reflects the inclusion of more Stage IIIB patients.

Lilenbaum et al.29 conducted another Phase II study evaluating the combination of gemcitabine and vinorelbine in patients with advanced NSCLC. Thirty two patients with Stage IIIB or IV NSCLC were treated with gemcitabine at 1250 mg/m² over 30 minutes (1000 mg/m² for the first six patients) and vinorelbine at 25 mg/m² over 6 minutes, both administered intravenously on Days 1 and 8 every 21 days. Thirty two eligible patients were treated and eight (25%) had partial responses. A median survival time of 8.3 months and a one year survival of 38% were seen. Grades 3 and 4 neutropenia were observed in 13% and 25% of the 148 treatment cycles, respectively, with one episode of neutropenic sepsis. In their study, dose omission was not an issue on Day 15, since this week was planned as an off therapy week, thereby allowing for a higher percentage of the planned dosage to be delivered. Therefore, it might be preferential to use this schedule to avoid expected dose reductions.

The benefits of the Day 1 and Day 8 schedule are confirmed by the study of Gridelli et al.30 who conducted a randomized Phase II study to optimize the dose. They established the dose of gemcitabine at 1000 mg/m² and vinorelbine at 25 mg/m² for further study achieving an aggregate response rate of 26% and a median survival of 33 weeks.

All of these trials, as well as others shown in Table 9 demonstrated that the combination of gemcitabine and vinorelbine is active in the front-line therapy of NSCLC. In addition, the current trial showed activity in the second-line setting, which was previously shown only for docetaxel in a randomized fashion and for gemcitabine in Phase II trials. In the current study, 18% of previously treated individuals had disease that responded to treatment. This result is similar to the findings of Hainsworth et al., who observed a response rate of 18% among 50 patients with progressive NSCLC after one previous chemotherapy regimen.31 Barr et al. examined a group of 27 patients with metastatic NSCLC who had not responded to chemotherapy containing taxane. Using a gemcitabine dose of 750 mg/m² and a vinorelbine dose of 30 mg/m² on Days 1, 8, and 15 every 28 days, these investigators obtained a response rate of 20%.32 In the current study, we saw striking responses in patients whose disease had overtly progressed on front-line chemotherapy.

The state of the-art choice of front-line therapy for NSCLC is currently open. Two randomized intergroup trials that sought a favorable regimen among five different platinum-based chemotherapy regimens failed to yield decisive results in favor of any particular combination.33 The ECOG 1594 trial randomized 1,200 patients to a combination of gemcitabine (1,000 mg/m², Days 1 and 8) and cisplatin (100 mg/m², Day 1), docetaxel (75 mg/m², Day 1) followed by cisplatin (75 mg/m², Day 1), or docetaxel (225 mg/m²) followed by carboplatin to an AUC (Area under the curve) of 6, with all three arms compared with a control arm of paclitaxel at 135 mg/m² infused over 24 hours followed by cisplatin at 75 mg/m². Median survival time and response rate were comparable among the four arms, with a median survival time of approximately eight months and a one year survival rate of approximately 33%. Responses ranged from 15.3% for carboplatin and paclitaxel to 21.3% for gemcitabine and cisplatin.33

A trial by Kelly et al. compared approximately 400 patients with Stage IIIB/IV NSCLC who were selected at random to receive either paclitaxel (225 mg/m²) with carboplatin (AUC = 6) or vinorelbine (25 mg/m², Day 1, 8, and 15) and cisplatin (100 mg/m², Day 1). This trial also failed to show a statistically significant survival advantage for either regimen. However, while the 27% response rate, eight months median survival time, and 33% one year survival rate were equivalent between the two arms, the authors recommended the paclitaxel and carboplatin combination for further study based on an increased percentage of patients on that arm who were able to complete therapy.22

Thus, at least five platinum-based chemotherapy regimens are considered acceptable front-line combinations for patients with Stage IIIB/IV NSCLC. In addition, docetaxel was active in the second-line treatment of NSCLC, as shown by two randomized trials by Shepherd et al.34 and Fossella et al.35 with docetaxel emerging as the drug of choice in this setting. As previously indicated, the rationale for a nonplatinum-based regimen is strong because cisplatin-based chemotherapy causes significant side effects including nausea, vomiting, nephrotoxicity, and neurotoxicity. The current trial combined two active third-generation NSCLC drugs, gemcitabine and vinorelbine, and yielded active response rates, comparable survival rates, and fewer and less severe toxic effects than were seen in previous Phase II or III studies of carboplatin/cisplatin-based chemotherapy. The arguments for a nonplatinum-based regimen are strengthened by a recent randomized trial by Georgoulias et al.36, 37 wherein docetaxel in combination with cisplatin was compared with docetaxel and gemcitabine. In that trial, although response and survival rates were equivalent between the two arms, the docetaxel and gemcitabine combination had a statistically significantly lower incidence of Grades 3 and 4 toxic effects. Similar data have been shown by Kosmidis et al. looking at paclitaxel with gemcitabine.38

In summary, the current data show encouraging front-line and second-line response rates (31 of the 36 prior therapy patients were second line) for the combination of gemcitabine and vinorelbine with acceptable side effects. In light of its tolerability and efficacy, this combination should be directly compared with platinum based regimens. A completed randomized Phase II study from the Minnie Pearl Cancer Research Network showed that the toxicity of the gemcitabine/vinorelbine combination compared favorably with two platinum-containing triplets and a taxane-containing doublet, with similar response and survival to what we show in the current report.39 In addition, several Phase III studies are underway, including the Italian/Canadian study (vinorelbine/gemcitabine versus cisplatin/vinorelbine versus cisplatin/gemcitabine) and in a United States multicenter trial (vinorelbine/gemcitabine versus carboplatin/paclitaxel).

Finally, studies of these two potent cytotoxic drugs in combination with novel biologic agents such as inhibitors of epidermal growth factor receptor, antibodies to vascular endothelial growth factor, or farnesyl transferase inhibitors merit investigation.40–45 Given that most patients with cancer die from metastatic disease, multiple targeted therapies have been tested in this area. One potential use for these agents (Fig. 5) is to combine them with chemotherapy in advanced disease in a setting of known synergy. Once the maximal cytotoxic effect has been attained, the targeted agents can then be used alone as maintenance therapy to prevent new or further metastatic spread. If this paradigm becomes a reality several Epidermal Growth Factor Receptor (EGFR) antagonists are now in Phase III study),46 then it will be critical to employ chemotherapeutic agents with the least possible side effects to allow for the best long term quality of life. Nonplatin, nontaxane regimens such as gemcitabine and vinorelbine could be perfectly suited for this next era of cancer care. The results of these and future trials are awaited with interest in both elderly and younger patient subgroups as we continue to search for the best possible treatment regimen we can provide for patients with NSCLC.

Acknowledgements

The authors thank Julia Starr and Bich Tran for their editorial assistance.