Analysis of prognostic factors in chemo-naïve patients with advanced NSCLC and poor performance status (PS): Cox regression analysis of two phase III trials

7113

Background: Two phase III trials in chemo-naïve PS2 patients with advanced NSCLC compared paclitaxel poliglumex (PPX) to either gemcitabine or vinorelbine (STELLAR 4), or PPX/carboplatin to paclitaxel/carboplatin (STELLAR 3). While overall survival (OS) was similar between treatment arms in both studies, individual patient characteristics may be predictive of benefit. Methods: STELLAR 3 and STELLAR 4 enrolled 400 and 381 chemo-naïve PS2 patients, respectively. The impact of pre-defined baseline characteristics on OS was evaluated by univariate and step-wise multivariate Cox regression analysis. Treatment differences between subgroups were also estimated by Cox analysis. Results: Univariate Cox analysis of potential risk factors showed pre-baseline weight loss, extra-thoracic metastasis, and a low lung cancer symptom (LCS) score to be highly significant (p < 0.001). In STELLAR 4, tobacco use was also a highly significant risk factor. Important primary baseline factors predicting survival as determined by multivariate analysis are summarized in the table . Other baseline factors evaluated but not predictive of survival included gender, number of comorbidities, age, and history of tobacco use. Treatment differences between subgroups were not statistically significant; however, a strong trend towards improved survival was observed for women receiving PPX in STELLAR 4 compared to those in the control-arm (HR = 0.65; p = 0.069). In contrast, men had similar survival between treatment arms (HR = 1.08; p = 0.579). Conclusion: In this large PS2 patient population, weight loss, presence of extra-thoracic metastasis, low LCS scores, and high LDH were found to be important clinical determinants of survival. In addition, significant differences in survival based on geographic region in STELLAR 3 highlight the importance of stratification by region.

[Table: see text]

[Table: see text]

Analysis of prognostic factors in patients with advanced relapsed/refractory NSCLC: Cox regression analysis of a randomized phase III trial comparing docetaxel and paclitaxel poliglumex (PPX)

7040

Background: A phase III trial compared PPX to docetaxel as 2nd-line treatment in pts with relapsed/refractory advanced NSCLC (STELLAR 2). While overall survival was similar between arms, the need for supportive measures to manage the effects of myelosuppression was significantly reduced in the PPX arm. The current analysis was performed to evaluate determinants of survival in the 2nd-line treatment of NSCLC. Methods: STELLAR 2 enrolled 849 pts, 427 on PPX and 422 on docetaxel; all patients were included in the analysis. Randomization between the study arms was stratified by tumor stage, performance status (PS), start of frontline chemotherapy (< 4 mo vs more than 4 mo), gender, and prior taxane therapy. Univariate and multivariate Cox regression analyses were performed to evaluate the impact of baseline characteristics on overall survival (OS). Results: At randomization, 29% of pts had received prior taxane, 14% were PS2, 80% had stage IV disease, and 31% had started frontline therapy within the prior 4 months. Risk factors significantly affecting survival as determined by multivariate analysis are listed in the table . These factors were consistent when treatment was added to the model. Prior exposure to taxane was not predictive of survival; tumor stage was a significant univariate predictor (p=0.0349), but had relatively less impact in the multivariate model. Conclusion: These analyses identified several factors associated with reduced survival benefit from standard second line therapy. Consequently, alternative treatment strategies may be necessary in patients with poor prognosis. For example, more tolerable agents may enhance the benefit/toxicity ratio in these patients.

[Table: see text]

[Table: see text]

EGFR mutations (muts), IHC and FISH status, and chromosome 7 gene copy number combined with pAkt expression as potential predictors of survival in non-small cell lung cancer (NSCLC) patients (pts) treated with gefitinib (GEF)

7182

Background: EGFR and pAkt expression by immunohistochemistry (IHC), muts, and FISH status have been identified as possible molecular predictors for GEF efficacy in NSCLC (Cappuzzo, et. al, JNCI, 2005). The goal of this study was to independently evaluate these findings regarding survival (surv), and to assess the predictive value of mean chromosome 7 copy number/cell (C7). Methods: 150 consecutive Expanded Access Trial pts with >1 week GEF therapy were included for analysis. IHC (present vs not detected) was performed for 87 pts, and 58 pts were analyzed for muts by SSCP, mut-specific PCR, and sequencing. Tissue from 81 pts was evaluated for EGFR and C7 gene copy numbers by fluorescence in situ hybridization (FISH). Results: 150 pts (77 female, 73 male; median (md) age 67; 85 adenocarcinoma) received GEF; md follow-up was 5.8 months (mo). Overall response was 8% (2 CR, 10 PR); 56 pts had stable disease. Md Kaplan-Meier surv was 5.9 mo. IHC revealed that 47/87 pts (54%) had EGFR+, and 36/75 pts (48%) had pAkt + tumors. pAkt+ pts had significantly (sig) longer surv than pAkt− pts (11.4 vs 5.8 mo, p < .05). High polysomy was seen in 36/81 pts (44%) who were designated FISH+; 45 pts were FISH−. EGFR IHC and FISH positivity were not sig associated with surv. C7 was defined as low (<3.6, 63 pts) or high (≥3.6, 18 pts); md surv was 6.6 and 17.1 mo, respectively, p < .01. Muts were found in 17/58 tumors (29%). Md surv for pts with and without muts was 23.8 and 7.9 mo, respectively, p < .07. EGFR IHC− pAkt− pts (18 pts) had sig shorter surv than 57 pts with any pos value (4.7 vs 8.8 mo, p < .02). Double-positive pts had sig longer surv than pts with any neg value. Conclusions: These findings resemble but do not duplicate those reported by Cappuzzo, et al. Additionally, high C7, alone or combined with pAKT, may be an important predictor for GEF efficacy in NSCLC. Further studies of C7, a technically simple and reproducible FISH assay, are warranted.

[Table: see text]

[Table: see text]

Gastrointestinal hemorrhage in advanced non-small cell lung cancer (NSCLC) patients treated with erlotinib and celecoxib

7172

Background: Erlotinib (E) was associated with superior survival in a phase III trial of previously treated advanced NSCLC patients (pts). Celecoxib (C) has been shown to potentiate the apoptotic and growth inhibitory effects of E in pre-clinical models. Methods: This was a phase II trial of E plus C in advanced NSCLC pts that failed one prior chemotherapy regimen. Primary endpoint: efficacy; secondary endpoint: toxicity. Pts received C (400mg b.i.d.) and E (150mg daily) until disease progression. Planned accrual: 40 pts. Results: 26 pts with stage IIIB/IV NSCLC were enrolled. Patient (pt) characteristics: male 65%; median age 66; ECOG performance status 0/1- 96%. Eighteen pts had tissue available for FISH and EGFR mutation analysis: 50% had chromosome 7 polysomy (> 4 copies per cell); none had EGFR gains (>2 EGFR/chromosome 7). Two pts had an EGFR gene mutation (1 exon 19, 1 exon 21). Response results: partial response- 2 pts (1 with exon 19 mutation), stable disease- 8 pts, and progressive disease- 16 pts (1 with exon 21 mutation). Median progression free survival (PFS) and overall survival (OS): 1.9 and 10.2 months, respectively. Grade 3/4 upper gastrointestinal bleeding (GIB) occurred in 4 pts prompting study closure. One pt was on therapeutic dalteparin and two pts receiving warfarin developed marked INR prolongation (INR >10). The fourth pt had a history of peptic ulcer disease. Platelet counts at time of GIB: 142 - 559. Three pts had endoscopy and gastric or duodenal ulcers were found in all three cases. No pts were taking anti-acid medication at the time of GIB. No other pts were on therapeutic anticoagulation. Three pts without upper GIB were taking low-dose aspirin. Other toxicities: 85% grade 1/2 rash; 65% grade 1/2 diarrhea, 30% grade 1/2 nausea, 30% grade 1/2 fatigue (one grade 3 fatigue); one grade 3 pneumonitis, one grade 3 esophageal stricture. Conclusions: These observations suggest that C plus E may be associated with increased incidence of gastrointestinal ulceration and GIB and that the regimen should not be given to pts with a previous history of peptic ulcer disease or to pts requiring therapeutic anticoagulation. Based on response rate, PFS, and OS in this group of pts, it appears that results with E and C are similar to those reported for E alone.

[Table: see text]

Pemetrexed and carboplatin plus bevacizumab as first-line therapy for advanced non-squamous non-small cell lung cancer: preliminary safety results of a phase II trial

17111

Background: Bevacizumab is a novel antiangiogenic agent that has been shown to improve response rates and survival of patients with advanced non-squamous NSCLC when added to paclitaxel and carboplatin. Pemetrexed is a multitargeted antimetabolite that has shown activity in NSCLC as a single agent and when combined with carboplatin. Because the combination of pemetrexed and carboplatin has activity comparable to that of other standard platinum doublets and promising toxicity profile (Zinner, 2005), the addition of bevacizumab to this regimen is investigated. Methods: This single cohort, phase 2 study evaluates the safety and efficacy of the combination of pemetrexed and carboplatin plus bevacizumab in patients with untreated non-squamous NSCLC. Eligibility requires ECOG performance status 0–1, Stage IIIB (malignant effusion) or Stage IV non- squamous NSCLC, no evidence of CNS metastases, no anticoagulation. Treatment consists of pemetrexed 500 mg/m2 over 10 minutes, carboplatin AUC 6 over 30 minutes, and bevacizumab 15 mg/kg over 30–90 minutes. Treatment is repeated every 21 days for 6 cycles. For patients who have either stable disease or partial response, pemetrexed 500 mg/m2 and bevacizumab 15 mg/kg are continued every 21 days until progression of disease or toxicity. All patients receive folic acid, vitamin B12 and steroid prophylaxis. Tumor response is assessed using RECIST every 2 cycles during treatment with carboplatin and then every 3 cycles during treatment with pemetrexed and bevacizumab alone. Results: From 8/2005 to 12/2005, 10 (of planned 50) patients with Stage IIIB and IV non-squamous NSCLC have been enrolled and treated. Patient characteristics are: median age: 65 (48–71), 20% female, 80% male, 30% stage IIIB, 70% stage IV. Median number of cycles delivered is 5 (range 1–9). No patient has discontinued therapy secondary to progressive disease or toxicity to date. 6 patients are evaluable for response: 1 PR, 1 minor response (24% reduction), 4 SD. No grade 3/4 toxicities have been experienced. Conclusions: This is a highly tolerable and active regimen with little toxicity to date. Updated response and toxicity data will be forthcoming. Supported by Genentech Inc and Lilly Pharmaceuticals.

[Table: see text]

Effect of gender on outcome in two randomized phase III trials of paclitaxel poliglumex (PPX) in chemonaïve pts with advanced NSCLC and poor performance status (PS2)

7039

Background: Women with NSCLC are younger and are more likely to be non-smokers than the overall NSCLC population. Estrogen seems to contribute to these differences in lung cancer biology, but its effect on treatment efficacy has not been well described. Methods: Two phase III trials in chemo-naïve PS2 patients with advanced NSCLC compared PPX to either gemcitabine or vinorelbine (STELLAR 4), or PPX/carboplatin to paclitaxel/carboplatin (STELLAR 3). Eligibility criteria were identical and pts were stratified for gender and age. Analysis of overall survival (OS), the primary study endpoint, showed similar survival between treatment arms. A trend to improvement with PPX for females but not males in both studies prompted an exploratory combined analysis of the 198 women in STELLAR 3 (93/400 pts) and STELLAR 4 (105/381 pts) using univariate and multivariate Cox regression analysis of OS. Results: Combined log-rank analysis of STELLAR 3 and 4 shows improved OS for females receiving PPX vs control (9.5 mo vs 7.8 mo; HR=0.70; p=0.03) and no difference in males (7.3 mo vs 6.9 mo; HR=1.06; p=0.53). Cox multivariate analysis identified treatment with PPX (HR=0.64; p=0.019), smoking (HR=1.50; p=0.037), and presence of extra-thoracic metastases (HR=1.76; p=0.003) as independent prognostic factors in these women. In the combined analysis, median survival advantage for PPX-treated pts was greater in women <55 years old (10.0 vs 5.2 mo, p=0.038) compared to women 55 or older (8.9 vs 8.6 mo, p=0.134). Pretreatment estrogen levels were available for 86 pts in STELLAR 3; pts with estrogen >30 pg/dl had a significant survival benefit when receiving PPX compared to paclitaxel (10.2 vs 5.5 mo; p=0.039). Conclusions: While the efficacy of PPX is comparable to other treatment options in NSCLC, PPX may be more effective female pts, particularly premenopausal women, compared to females in the control arms. Preclinical data suggest that estrogen enhances PPX distribution to lung tissue and upregulates the rate-limiting metabolic enzyme cathepsin B in NSCLC. To prospectively test the effect of estrogen on OS, a phase III trial (PIONEER) has been initiated to compare PPX to paclitaxel in chemo-naïve PS2 females with NSCLC.

[Table: see text]

A phase II, multicenter, randomized clinical trial to evaluate the efficacy and safety of bevacizumab in combination with either chemotherapy (docetaxel or pemetrexed) or erlotinib hydrochloride compared with chemotherapy alone for treatment of recurrent or refractory non-small cell lung cancer

7062

Background: Bevacizumab is a recombinant, humanized anti-VEGF MAb. Erlotinib is a potent, reversible, highly selective and orally available EGFR tyrosine-kinase inhibitor. Both compounds have demonstrated a survival benefit in the treatment of NSCLC: bevacizumab when added to chemotherapy in the first line setting, and erlotinib when given alone in the 2nd/3rd line. In addition, a single arm phase I/II study of the combination of bevacizumab and erlotinib has shown encouraging survival and response rate data, with a favorable safety profile (Sandler et al, PASCO 2004). Methods: A multicenter, randomized phase II trial was conducted to evaluate the safety of combining bevacizumab with chemotherapy (docetaxel or pemetrexed), or with erlotinib; and to make a preliminary assessment of the efficacy of combining bevacizumab with chemotherapy or erlotinib relative to chemotherapy alone, as measured by progression-free survival. All patients had histologically confirmed non-squamous NSCLC and had experienced disease progression (clinical or radiological) during or following one platinum-based regimen for advanced stage disease. Randomization was on a 1:1:1 basis to docetaxel or pemetrexed plus placebo (arm 1) v docetaxel or pemetrexed plus bevacizumab (arm 2) v bevacizumab plus erlotinib (arm 3). Patients remained in the treatment phase of the study until documented radiographic or clinical disease progression or through 52 weeks of study treatment. Results: Between August 2004 and November 2005, 120 patients were randomized and treated. To date, there have been 68/85 required PFS events; arms 1 and 2 therefore remain blinded. Demographics from the first 85 patients for arms 1 & 2 v 3 are as follows: median age 65 v 68;% male/female 64/36 v 44/56; ethnicity % white/black/asian/other 83/12/2/3 v 74/15/4/7; % ECOG PS 0/1/2 46/52/2 v 50/50/0; % adenocarcinoma/BAC&BAC-like/Other 67/7/26 v 69/8/23; % current/previous/never-smoker 15/71/14 v 11/74/15. Conclusions: Efficacy and safety results will be presented at the meeting.

[Table: see text]

Associations between chromosome 7 aneusomy and EGFR copy number with survival and response to gefitinib in NSCLC

7171

Background: EGFR, the presumed target of gefitinib and erlotinib, has been studied (expression, gene copy number, & mutations) for predicting response to these tyrosine kinase inhibitors (TKI), in non-small cell lung cancer (NSCLC). ’High polysomy’ and ’amplification’ of the EGFR gene, as defined by Cappuzzo et al. JNCI 97:643, using fluorescence in situ hybridization (FISH), showed significant association with objective response (Resp) and survival. We also measured EGFR and chromosome 7 (C7) copy numbers by FISH in 81 gefitinib-treated NSCLC patients (12 Resp). Using Cappuzzo’s FISH± parameter alone we saw similar trends but no statistical significance in the 81 patient group. Therefore we sought to optimize FISH parameters for these patients. Methods: FISH was performed (paraffin sections) using a 2-color probe set (Vysis LSI EGFR/CEP 7), median 80 cells per specimen. >50 parameters were derived from the data (e.g. mean EGFR/cell, C7/cell etc) and compared, using different threshold values, to Resp and survival. Results: The best single parameter associated with survival was the average C7/cell. Applying upper and lower thresholds of 3.6 and 2.0 C7/cell to delineate moderate from extreme ratios yielded median survival of 177 and 465 d, respectively (Kaplan-Meier, p < .002). A single threshold of 3.6, separating low from high, yielded survival of 201 and 522 d, respectively (p < .01). For these thresholds C7/cell was not associated with Resp, however, thresholds could be found for which both survival and Resp were significant. The best single parameter associated with Resp was average EGFR/cell. Of the 70 patients with EGFR/cell≤6.0, 63 (90%) did not respond while 5 of 11 patients (45%) with EGFR/cell >6.0 responded (p < .01). No EGFR/cell threshold could be found for which both survival and Resp were significant. Many 2-parameter combinations provided significant associations with both survival and Resp. Conclusions: Several FISH-derived parameters were significantly associated with either survival or Resp to gefitinib and a subset were associated with both. These parameters must be tested on independent data sets to determine their value in directing TKI therapy.

[Table: see text]

MAX2--a convenient index to estimate the average per patient risk for chemotherapy toxicity; validation in ECOG trials

Cancer patients, especially the elderly, present with a highly variable susceptibility to toxicity from chemotherapy. To estimate correctly a patient's risk for toxicity, both the average toxicity of a chemotherapy regimen and patient-related variables need to be assessed. However, treatment toxicities are typically reported item by item, not summarised per patient. We tested an index derived from a pilot study, the MAX2, on the ECOG database. Studies including 20 or more patients aged 70 years and older per arm were selected. Four studies were identified, representing 2526 patients, 410 (16%) being elderly. The association of the MAX2 index with the per patient incidence of grade 4 haematological and/or grade 3 or 4 non-haematological toxicity was highly significant, both for the overall group and for the elderly subgroup. The MAX2 index is a convenient and reproducible way of comparing the average per patient risk for toxicity from chemotherapy across several regimens.

Docetaxel and cisplatin in patients with advanced non small-cell lung cancer (NSCLC): a multicenter phase II trial

We examined the safety and efficacy of the docetaxel/cisplatin combination in patients with advanced, previously untreated NSCLC and evaluated changes in quality of life over time. Docetaxel was administered before cisplatin (both 75 mg/m2, 1-hour infusions) every 3 weeks to 47 patients with stage IIIB or stage IV NSCLC. Patients also received premedication of oral dexamethasone. The median age (range) of patients was 62 (45-78) years and 26 patients (55.3%) had adenocarcinoma. Of the 40 patients evaluable for response, one achieved a complete response and 14 had partial responses; the response rate was 37.5% (95% confidence intervals; 22.5, 52.5). In the intent-to-treat population the overall response rate was 31.9%. Time to response ranged from 3 to 20 weeks, and the median duration of response was 34.6 weeks. Median survival and median time to progression were 11.3 months and 18.9 weeks, respectively. One-year survival was 40%. Grade 3 or 4 neutropenia and febrile neutropenia were observed in 74.4% and 12.8% of patients, respectively. Severe asthenia was seen in 14.9% of patients. Other grade 3 or 4 toxicities included nausea (eight patients), vomiting (five), neurosensory effects (six), neuromotor effects (five), diarrhea (four), and infection (three). There was an improvement in emotional well-being; however, the overall quality of life score did not change with treatment. Docetaxel administered in combination with cisplatin is an active regimen in patients with NSCLC. This regimen of docetaxel (75 mg/m2) and cisplatin (75 mg/m2) repeated at 3-week intervals is being evaluated in an ongoing Eastern Cooperative Oncology Group (ECOG) randomized study in patients with advanced and metastatic NSCLC.