Randomized phase II study of erlotinib (E) or intercalated E with carboplatin/paclitaxel (CP) in chemotherapy-naive advanced NSCLC: Correlation of biomarker status and clinical benefit

8026

Background: E plus chemotherapy showed no additive effects in NSCLC but preclinical studies suggested that intercalation of E and chemotherapy could give synergy. Clinical studies suggested that EGFR mutations could aid in pt selection. KRAS mutation status of tumors was also evaluated. We conducted a randomized phase II study of E and E intercalated with CP in pts with chemonaive NSCLC. Methods: Stage IIIB/IV EGFR+ NSCLC pts were randomized to E 150 mg/d or CP d1 plus E days 2–15 q21 days (ECP). After 4 cycles, E continued until PD. Tumor tissue was evaluated by IHC (EGFR), FISH (EGFR gene copy number), and PCR amplification followed by DNA sequencing (EGFR and KRAS mutations). Results: Among 143 pts randomized 53% EGFR FISH+; 13% activating EGFR mutations and 8% non-activating EGFR mutations (evaluable samples=114); and 22% KRAS mutations (evaluable samples=130). No pt had both EGFR and KRAS mutations. EGFR-activating mutations were higher among females (19% vs 6% males), adenocarcinoma histology (17% vs 0% others), Asians (45% vs 7% non-Asian), and never smokers (29% vs 7% former and 0% current); KRAS mutations were higher in current smokers (41% vs 27% former and 0% never) and adenocarcinoma histology (22% vs 18% squamous). In the E arm, 6-mo PFS probability for the efficacy evaluable population (n=69) was significantly higher in pts with EGFR activating mutations vs no mutations (89% vs 25%, HR=0.17, P=0.001), numerically higher in pts with EGFR FISH+ vs FISH- (40% vs 22%, HR=0.61, P=0.07), and with KRAS wild type vs mutation+ (38% vs 12%, HR=0.56, P=0.10). Response rates, PFS and OS by type of EGFR/KRAS mutation will be presented. In the ECP arm, 6-mo PFS probability for the efficacy evaluable population (n=68) was numerically higher in pts with EGFR activating mutations (42% vs 29%, HR=0.72, P=0.53), numerically higher in pts with wild type KRAS (32% vs 9%, HR=0.57, P=0.08), and numerically lower in pts with EGFR FISH+ vs FISH- (23% vs 30%, HR=0.93, P=0.78). Conclusions: Activating EGFR mutations correlate with increased 6 mo PFS probability in 1st line therapy with E. EGFR FISH + and absence of KRAS mutation trend towards increased 6 mo PFS rate with E.

[Table: see text]

Analysis of IGF1R gene copy number by silver in situ hybridization (SISH) in non-small cell lung cancer (NSCLC)

7525

Background: IGF1R is a potential biomarker for predicting outcome of NSCLC patients (pts) treated with therapeutics targeting the IGF1R signaling pathway. SISH can be expeditiously analyzed on a bright field microscope given that the morphology of the tissue is observed concurrent to the scoring of discrete gene signals. Methods: An experimental SISH probe designed by Ventana Medical Systems (Tucson, AZ) was used to evaluate IGF1R gene copy number on a tissue microarray containing triplicate samples from 189 pts surgically treated for NSCLC (median follow-up 4 years and 5-year survival probability of 40%). Valid results, at least one core with sufficient tumor content for assessment, were obtained for 166 patients. There were 128 males; 93 squamous cell carcinomas (SCC), 47 adenocarcinomas, 5 large cell carcinomas and 21 other histologies. The pathological stages were I: 68, II: 37 and III/IV: 60. The mean number of IGF1R gene copies/nuclei/core was determined by a board certified pathologist counting 50 nuclei in each core. The mean number of IGF1R gene copies/nuclei/patient was determined by using, within the triplicates, the core with the highest gene copy number/nuclei. Results: The mean number of IGF1R gene copies/nuclei per patient was 2.26 (range: 1.12 - 7.56; standard deviation 0.81). The median copy number was 2.11. The pts were divided into two groups, those with 2.1 genes/nuclei or less and those with greater than 2.1 genes/nuclei. There was no statistical difference related to gender (p=0.422) or between pathological stages, (p=0.221). However, there was a highly significant difference between the two categories when considering histological pattern. Among patients with SCC, 66.3% had high copy number compared to 33.7 % in non-SCC histologies (p=0.008). Analysis of overall survival comparing pts with low vs. high IGF1R gene copy number revealed no statistical difference in the median survival: 1.5 yrs (95% CI 0–3.7 yrs) vs. 3.3 yrs (2.0–4.5 yrs) or the 3-year survival: 46% (35–57%) vs. 52% (41–63%). Conclusions: IGF1R gene copy number detected by SISH is higher in SCC than in other histologies of NSCLC, but does not associate with gender, pathological stage or survival. IGF1R SISH should be further explored as a predictive biomarker for IGF1R therapeutics.

[Table: see text]

ErbB-3 expression is associated with E-cadherin and their coexpression restores response to gefitinib in non-small-cell lung cancer (NSCLC)

BACKGROUND

Epidermal growth factor receptor (EGFR) inhibitors are effective in a subset of patients with non-small-cell lung cancer (NSCLC). We previously showed that E-cadherin expression associates with gefitinib activity. Here, we correlated the expressions of ErbB-3 and E-cadherin in NSCLC tumors and cell lines, their effect on response to gefitinib, and induction of both by the histone deacetylase (HDAC) inhibitors vorinostat and SNDX-275.

METHODS

Real-time RT-PCR was carried out on RNA isolated from 91 fresh-frozen NSCLC samples and from 21 NSCLC lines. Protein expression was evaluated with western blot and flow cytometry. Apoptosis was assessed using vibrant apoptosis assay.

RESULTS

Expressions of E-cadherin and ErbB-3 correlated significantly in primary tumors (r = 0.38, P < 0.001) and in cell lines (r = 0.88, P < 0.001). Cotransfection of ErbB-3 and E-cadherin in a gefitinib-resistant cell line showed enhanced apoptotic response to gefitinib. vorinostat and SNDX-275 induced ErbB-3 and E-cadherin in gefitinib-resistant cell lines. When gefitinib-resistant lines were treated with vorinostat and gefitinib, synergistic effects were detected in four of the five lines tested.

CONCLUSION

ErbB-3 and E-cadherin are coexpressed and induced by HDAC inhibitors. For tumors with low ErbB-3 and E-cadherin expressions, the combination of HDAC and EGFR-tyrosine kinase inhibitors increased expression of both genes and produced more than additive apoptotic effect.

EGFR regulation by microRNA in lung cancer: correlation with clinical response and survival to gefitinib and EGFR expression in cell lines

BACKGROUND

Allelic loss in chromosome 3p is one of the most frequent and earliest genetic events in lung carcinogenesis. We investigated if the loss of microRNA-128b, a microRNA located on chromosome 3p and a putative regulator of epidermal growth factor receptor (EGFR), correlated with response to targeted EGFR inhibition. Loss of microRNA-128b would be equivalent to losing a tumor suppressor gene because it would allow increased expression of EGFR.

PATIENTS AND METHODS

We initially showed that microRNA-128b is a regulator of EGFR in non-small-cell lung cancer (NSCLC) cell lines. We tested microRNA-128b expression levels by quantitative RT-PCR, genomic copy number by quantitative PCR, and mutations in the mature microRNA-128b by sequencing. We determined whether microRNA-128b loss of heterozygosity (LOH) in 58 NSCLC patient samples correlated with response to gefitinib and evaluated EGFR expression and mutation status.

RESULTS

We determined that microRNA-128b directly regulates EGFR. MicroRNA-128b LOH was frequent in tumor samples and correlated significantly with clinical response and survival following gefitinib. EGFR expression and mutation status did not correlate with survival outcome.

CONCLUSION

Identifying microRNA regulators of oncogenes could have far-reaching implications for lung cancer patients including improving patient selection for targeted agents, development of novel therapeutics, or development as early biomarkers of disease.

Survival benefit of gefitinib treated non-small-cell lung cancer patients with exon 19 deletion mutations in epidermal growth factor receptor (EGFR), detected by a new sensitive PCR-based method

10614

Background: As for two common types of EGFR mutations, patients with exon 19 deletion mutations have longer survival than those with the L858R point mutations in exon 21 after treatment with EGFR inhibitors. We have developed a simple, polymerase chain reaction (PCR)-based method to detect exon 19 deletion mutations, and evaluated survival benefit of gefitinib treated patients with exon 19 deletion mutations versus patients without these mutations. Patients and methods: Tumor tissue was microdissected under stereoscopic microscopy from formalin-fixed paraffin-embedded sections, and DNA was extracted from tumor cells with DNeasy (Qiagen). Our method consisted of two different semi-nested PCRs with the deletion screening PCR and the common deletion specific PCR. All of the known deletions present in cell lines were detected by this method without direct sequencing. The result was validated by sequencing of exon 19. 73 non-small-cell lung cancer (NSCLC) Japanese patients treated with gefitinib were analyzed with this method. Study group consisted of 28 females (38%), 29 never smokers (40%) and 57 patients with adenocarcinoma (78%). Results: The PCR-based method detected mutations at mutant to wild type DNA copy ratio of 1/600, and in samples as small as 30 ng of purified DNA. Exon 19 deletion mutations were found in 25 (34%) patients. This method was more sensitive than conventional sequencing. The sequencing was performed in 19 patients with mutations and could not detect 3 deletions. Among 60 assessable patients 14 had overall response (23%). Objective response rates to gefitinib were observed in 7/21 patients with exon 19 deletion mutations (33%), and 7/39 patients without exon 19 deletion mutations (17%) (P = .211). Patients with exon 19 deletion mutations survived significantly longer than those without exon 19 deletion mutations (P = .017). Conclusions: The PCR-based method to detect exon 19 deletion mutations is cost effective and very sensitive, compared to previously described methods. We demonstrated survival benefit in NSCLC patients with exon 19 deletion mutations treated with gefitinib, and our PCR-based method is easily applicable for clinical use.

No significant financial relationships to disclose.

Role of EGFR but not HER2 or HER3 gene copy number in predicting sensitivity of head and neck squamous cell carcinoma (SCCHN) cell lines to EGFR tyrosine kinase inhibitors

6063

Background: EGFR gene copy number has previously been reported to predict for improved overall survival in NSCLC patients treated with gefitinib (IRESSA) or erlotinib compared with placebo [JCO 2006;24:5034–42 & N Engl J Med 2005;353:133–44]. The utility of EGFR gene copy number as a predictive biomarker in other tumour types such as squamous cell carcinoma of the head and neck (SCCHN) is currently under clinical investigation. The present study examined a panel of 20 SCCHN cell lines to identify potential biomarkers predicting in vitro sensitivity to EGFR tyrosine kinase inhibitors (TKIs). Methods: A panel of 20 SCCHN cell lines was screened for sensitivity to gefitinib, vandetanib or erlotinib using a viable cell number endpoint, with G150 values determined for each cell line (inhibitor concentration required to give 50% growth inhibition). Cell lines were blinded and assessed for EGFR, HER2 and HER3 protein expression by ELISA, mutation status by dye-terminator sequencing, and gene copy number by fluorescence in situ hybridisation (FISH). Results: A broad range in sensitivity was observed for all compounds across the panel of 20 SCCHN cell lines (G150 ranging from 0.001uM to =10uM). 12 cell lines were positive for EGFR genomic gain. Sensitivity (GI50 <1uM) to all EGFR TKIs was seen in 11 lines and resistance (GI50 >8uM) in 5 lines. Of the sensitive cell lines, 9 were positive for EGFR genomic gain compared with only 1 of the resistant lines. Furthermore, EGFR protein expression also had a direct association with EGFR TKI sensitivity. In contrast, only 4 cell lines were positive for HER2 or HER3 genomic gain and there was no correlation with sensitivity. The most sensitive cell line was positive for EGFR genomic gain and was the only line to have an EGFR TK mutation (S768I in exon 20). Conclusions: EGFR gene copy number and protein expression appeared to have predictive value in identifying SCCHN cell lines sensitive to EGFR TKIs.

No significant financial relationships to disclose.

Combined EGFR targeted therapy in a novel in vivo pancreas cancer model

14063

Background: EGFR inhibitors have a definite but limited activity in pancreatic cancer (PaCa). We have reported enhanced activity of dual EGFR therapy with a small molecule inhibitor (erlotinib) plus a monoclonal antibody (cetuximab) in head and neck cancer models. Human xenografted tumors recapitulate better the pathobiology of cancer than existing cell lines. Here we explored a dual EGFR targeting approach using a direct PaCa xenograft model, and sought after markers predicting efficacy. Methods: PaCa specimens obtained at the time surgery were implanted in nude mice and expanded to develop cohorts of tumor bearing mice suitable for drug evaluation. Ten cases were expanded, and within each case 4 groups of 6–8 mice each were treated with vehicle, erlotinib, cetuximab, and the combination of both for 28 days. Gene mutation analysis, gene amplification by fluorescence-in-situ hybridization, and immunohistochemistry (IHC) were done with untreated samples. Results: Two cases were sensitive to both single agents, but the combination did not induce higher efficacy in those. Two additional cases that were resistant to both single agents became sensitive with the combination. No egfr mutations were detected. Three and four cases carried low and high egfr polysomy (defined as [[Unsupported Character - ]] 4 copies in 10–40% and [[Unsupported Character - ]] 40% of the cells, respectively). No correlation was found between egfr copy number and efficacy. By IHC sensitive cases had a lower Ki67 proliferation index, and higher EGFR and nuclear pMAPK staining than resistant cases. The degree of Ki67 decrease after therapy correlated with efficacy. In cases resistant to the single agents but sensitive to the combination nuclear pMAPK only decreased with the dual targeting. Cases with high egfr polysomy were more labile to pharmacodynamic effects after treatment (such as EGFR or pMAPK decreases). Conclusions: EGFR inhibitors showed modest single agent antitumor effect that was enhanced with dual EGFR therapy in PaCa. No genomic markers predicted efficacy, although high egfr polysomy was associated with deeper pharmacodynamic inhibition, conceivably suggesting a phenomenon of pathway addiction. Higher pathway activation by IHC was linked with higher activity.

No significant financial relationships to disclose.

Erlotinib as initial therapy in patients with advanced non-small cell lung cancer (NSCLC) and a performance status (PS) of 2: A SWOG phase II trial (S0341)

7536

Background: Patients (pts) with advanced NSCLC and PS 2 have an inferior survival compared with good PS pts. Single agent and combination chemotherapy have been used with modest success with toxicity often limiting treatment. Targeted agents such as the EGFR tyrosine kinase inhibitor erlotinib (E) offer an alternative which may confer comparable benefit with better tolerance. This phase II trial of E in unselected chemotherapy-naive pts with advanced NSCLC and PS 2 was performed to obtain preliminary data regarding efficacy and EGFR biology in this pt population, and to set the stage for a subsequent randomized trial of E vs.chemotherapy, in pts selected for EGFR expression. Methods: Eligibility: stage IIIB (pleural effusion)/IV NSCLC; measurable disease; PS 2; no prior chemotherapy/biologic treatment for NSCLC. Treatment: E 150 mg orally daily. Molecular correlative studies:EGFR protein expression (IHC), gene copy (FISH), mutation analysis. Results: Pts: 82; 73 eligible; 72 fully evaluable; age (median) 74.4; M/F 47%/53%; current/former smoker 91%; stage IIIB/V 12%/88%; adenoca 54%. Treatment was well tolerated. Five pts (7%) had a grade 4 toxicity (fatigue 3 pts; dyspnea 2 pts). Most common grade 3 toxicities: fatigue 9 pts (13%); rash 7 pts (10%); diarrhea 5 pts (7%); anorexia 5 pts (7%). There was 1 possible treatment related death due to pneumonitis. One complete (1%) and 5 (7%) partial responses were noted. Stable disease was seen in 25 pts (35%) for an overall disease control rate (DCR) of 43% (31 pts). Progression free survival: 2.1 months (95% CI 1.5 –3.1); Median survival: 5.0 months (95 % CI 3.5 –7.3). One year survival: 22% (95% CI 12 –32%). Analysis of molecular correlates is ongoing. Conclusions: Single agent erlotinib is a well tolerated treatment for chemotherapy- naive patients with advanced NSCLC and PS 2 with an overall DCR of 43% and median survival of 5 months. These efficacy results are comparable to the outcome seen in SWOG trial S0027 in PS 2 pts employing sequential vinorelbine and docetaxel. We hypothesize that pt selection by an EGFR biomarker strategy will improve results with E, and that E will be superior to chemotherapy in this selected population.This trial design is under development within SWOG at present.

[Table: see text]

Fluorescence in situ hybridization (FISH) subgroup analysis of TRIBUTE, a phase III trial of erlotinib plus carboplatin and paclitaxel in NSCLC

7570

Background: TRIBUTE was a phase III, placebo-controlled study of patients with previously untreated advanced NSCLC. Patients received erlotinib (E) (150 mg/d) or placebo, plus a course of 6 cycles of carboplatin and paclitaxel (CP), followed by maintenance E monotherapy (for those in the CP+E arm who were responding to treatment). 1,059 patients were randomized and treated (526 E; 533 placebo). There were no significant differences in OS, RR, or TTP between the two arms. In subgroup analyses of the pivotal 2nd line trial (BR21) of E in patients with relapsed NSCLC in which E significantly increased median survival (HR=0.71, p <0.0001), patients who scored positively for increased EGFR copy number by fluorescence in situ hybridization (FISH) exhibited prolonged OS (HR=0.44, p=0.008). Here we report on a similar subgroup analysis for TRIBUTE. Methods: FISH analysis was performed on all available tissue samples. FISH+ samples had a high level of polysomy (=four copies of the gene in =40% of cells), or gene amplification (presence of tight gene clusters, gene/chromosome per cell ratio =2, or =15 copies of the genes per cell in =10% of analyzed cells). Results: FISH analysis was successfully performed on 245 patients (121 E, 124 placebo). Outcome in the placebo patients from this subgroup was better than the overall population, suggesting that this subgroup may not be representative. Of the 100 patients (41%) that were FISH+, 33 had amplification and 67 had high polysomy. In FISH+ patients, OS was similar between those treated with CP+E and those treated with CP alone. However, FISH+ patients in the CP+E arm experienced a decrease in RR compared to those in the CP alone arm, and had a marginally significant longer TTP (HR=0.59, 0.35–0.99). The benefit in TTP appeared after approximately 6 months, during the maintenance portion of the trial. Conclusions: In this retrospective analysis, FISH+ did not predict survival benefit in TRIBUTE. A longer TTP, but a lower RR, was observed in the FISH+ patients. The lower RR in the CP+E arm in this group, taken together with the improved TTP during maintenance therapy, suggests that a non-concurrent combination approach (CP followed by E) warrants further investigation.

No significant financial relationships to disclose.

Validation of proteomic classifier for clinical benefit from erlotinib as first line treatment for advanced non-small cell lung cancer (ECOG 3503)

7508

Background: In this study we tested the ability of our serum mass spectrometry classifier of clinical benefit from gefitinib to classify pre-treatment sera and plasma samples from non-small cell lung cancer (NSCLC) patients treated first line with erlotinib in ECOG E3503. Methods: Pretreatment serum and plasma samples were available from 73 of the 96 previously untreated advanced NSCLC patients treated with single agent erlotinib on ECOG protocol 3503, 13 subjects had only serum samples, and 10 had only plasma samples. All of these samples analyzed in replicate by MALDI mass spectrometry (MS). A prediction algorithm we established based on a training cohort of 139 patients treated second or third line with gefitinib was used to classify these patients for survival and time to progression, and results obtained from serum and plasma were compared when both were available Results: We found that the signals for the 8 distinct mass-to-charge (m/z) features used in our classifier were highly concordant between serum and plasma samples from the same patient, and that there was no difference in the classification of the patients between serum and plasma when both were available. Therefore we classified all 96 patients using serum if available, and plasma if not. The classification algorithm very successfully classified patients into groups with good and poor survival (median survival of 306 days vs 107 days, p = 0.0007). With the available follow-up, the time to progression was also statistically significant in this group (p = 0.007, data not shown). In a Cox multivariate analysis including the most significant univariate parameters PS (0 vs. 1 vs 2), number of involved sites (=3 vs >3) and prior weight loss (<5% vs = 5%), the MALDI MS classification algorithm retained independent significance at p = 0.03, with a hazard ratio of 0.53. Conclusions: This multi-institutional ECOG study demonstrates that MALDI MS can assist in the pre-treatment selection of a subgroup of NSCLC patients who are likely to show improved survival after treatment with first line erlotinib.

No significant financial relationships to disclose.

Epidermal growth factor receptor (EGFR) immunohistochemistry: Comparison of antibodies (Abs) and cut points to predict benefit from gefitinib in a phase III placebo-controlled study in advanced non-small cell lung cancer (NSCLC)

7576

Background: Tumor tissues obtained from the ISEL phase III trial assessing the efficacy of gefitinib vs placebo in chemotherapy-pretreated NSCLC were used to evaluate two Abs (DAKO and Zymed) and assess whether different cut points of EGFR protein expression improved prediction of response and survival benefit from gefitinib. Methods: EGFR protein expression in tumor samples was assessed by immunohistochemistry using DAKO EGFR pharmDx kit (scoring percent of tumor cells with positive staining, predefined cut point of =10%) and Zymed monoclonal Ab clone 31G7 (scoring proportion of positive cells times staining intensity [scale 0–400], predefined cut point of =200). Results: Clinical characteristics of the patients (pts) assessed with DAKO (n=379) and Zymed (n=357) Abs reflected the overall study population (N=1692) with the exception of fewer never-smokers and Asians. Of the pts evaluated with DAKO/Zymed Abs, females represented 32%/31%; never-smokers, 13%/14%; Asians, 6%/4%; adenocarcinomas, 44%/42%; and 88%/88% of pts were refractory to most recent chemotherapy. With the above criteria, 70% of tumor samples were scored as positive using DAKO Ab and 68% using the Zymed Ab (agreement between assessments 76%). The objective response rates in gefitinib treated EGFR-positive pts defined with various cut points with DAKO Ab (=1% to =90%) varied between 8% and 12%, and with Zymed Ab (score =50 to =350), between 10% and 13%. Lower cut points with the DAKO Ab provided the best discrimination between EGFR positive and EGFR negative patients in terms of survival hazard ratios (HRs) comparing gefitinib to placebo, with a significant treatment/cut point interaction for the 10% cut point (p=0.049). A similar trend was noted for Zymed Ab, although the discrimination between HRs was less apparent and not significant for any cut point analyzed. Conclusion: Assessment with DAKO PharmDx kit, according to percentage of positive staining, may provide more accurate prediction of survival benefit for gefitinib-treated pts than assessment with Zymed Ab and staining index. Use of higher cut points to define positivity does not improve discrimination of the test.

No significant financial relationships to disclose.

High insulin-like growth factor 1 receptor (IGF1R) expression is associated with poor survival in surgically treated non-small cell lung cancer (NSCLC) patients (pts)

7550

Background: IGF1R represents a novel molecular target in NSCLC. We aimed to evaluate the association between IGF1R expression, clinical characteristics and survival in operable NSCLC pts. Methods: 184 consecutive surgically treated pts (median follow up of 32 months [range: 13–51 months]) were included in the study. There were 23% females, 55% pts with squamous cell carcinoma, 28% with adenocarcinoma, 3% with large cell carcinoma, 13% with mixed carcinoma and 1% with other types, 5% were never-smokers. Median age was 64 years (range: 37–85 years). Pathological stage I was diagnosed in 41%, stage II –22%, stage III –33% and stage IV –4% pts. Induction or adjuvant treatment was not routinely used. IGF1R was evaluated in tissue microarray by immunohistochemistry and scored by two observers according to H-score, taking into account proportion of cells with positive staining times staining intensity (scale 0–400). Results: IGF1R IHC expression was low in analyzed samples (median H-score 20, range: 0–223). Higher IGF1R score was observed in adenocarcinomas vs. squamous-cell carcinomas (p=0.02) and never-smokers vs. smokers (p=0.02). IGF1R score did not depend on stage (p=0.98) and sex (p=0.83). When analyzed as a continuous variable, IGF1R score tended to associate with worse survival in the entire cohort (p=0.10), and was significantly associated with inferior survival in patients with complete (R0) resections (p=0.04). When categorized with cut-off point of 50 (30% of patients positive), IGF1R expression was again associated with poor survival (p=0.05). This association was confined to pathological stage I (hazard ratio, HR=3.07; 95% confidence interval, CI=1.41–6.70, p=0.005) vs. all other stages (HR=1.03; 95% CI=0.63–1.70, p=0.90), and to patients with adenocarcinoma histology (HR=2.67; 95%CI=1.17–6.70, p=0.02 vs. all other histologies (HR=1.28; 95%CI=0.79–2.11, p=0.32). Conclusion: In this cohort of operable NSCLC pts, high IGF1R expression was associated with never-smoking history, adenocarcinoma histology, and poor survival, particularly in stage I disease and in patients with adenocarcinoma. These results support involvement of IGF pathway in the progression of early-stage NSCLC.

No significant financial relationships to disclose.

FISHing for lung cancer: Promising findings using multi-targeted interphase fluorescence in situ hybridization (FISH) for lung cancer detection in sputum

7567

Background: Lung CT imaging shows promise for early lung cancer detection, but may be insensitive for central lesions and will create a need for complementary biomarkers to guide clinical decisions. Methods: We have been conducting a prospective study to assess the performance of various biomarkers in sputum to detect lung cancer. Within a cohort of 3,269 people at high risk for lung cancer (over 30 pack-years of cigarette use and chronic obstructive lung disease) we have conducted a nested case-control study to assay stored samples from 85 incident lung cancer cases and 73 controls for cytologic morphology, DNA methylation, and multi-targeted FISH assays (LAVysion, Vysis/Abbott Molecular). The FISH assay targeted DNA sequences from centromere 6, 5p15.2, 7p12 (EGFR), and 8q24 (cMYC). Cytology was classified as abnormal if moderate or greater atypia was observed, methylation was abnormal if three or more of 8 selected genes were methylated, and FISH was abnormal if two or more scored cells were observed to contain chromosomal aneuploidy indicated by signal gain for at least two targets or signal gain for one and signal loss for two or more targets. Results: Among all subjects, regardless of the time between sputum collection and diagnosis, FISH was abnormal in 49/87 cases (56%) and 5/73 controls (7%) (odds ratio (OR) = 17.5, 95% CI = 6.4 to 47.8). Considering only the cases for which sputum had been collected within 18 months before the lung cancer diagnosis, FISH was abnormal in 37/49 cases (75%) and in 3/38 controls (8%) (OR = 36.0, 95% CI = 9.4 to 138). For this same time period, cytological atypia was abnormal in 37% of cases and 22% of controls (OR = 2.1, 95% CI = 0.8 to 5.3) and methylation was abnormal in 64% of cases and 36% of controls (OR = 4.5, 95% CI = 1.0 to 19.8). Conclusions: The LAVysion FISH assay is more sensitive than other currently available sputum tests for the prediction of invasive carcinoma in subjects and is highly specific with a low frequency of false positive and false negative results. The results suggest that chromosomal aneuploidy and by extrapolation, missegregation, are features of advanced premalignancy and should identify patients who are proximate to clinically overt lung cancer.

[Table: see text]

Epidermal growth factor receptor gene copy number and protein level are not associated with outcome of non-small-cell lung cancer patients treated with chemotherapy

BACKGROUND

Survival benefit of non-small-cell lung cancer (NSCLC) patients treated with epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors is predicted by high EGFR gene copy number and by strong EGFR protein expression. Clinical relevance of these features in patients treated with chemotherapy has not been reported.

PATIENTS AND METHODS

This study included 82 NSCLC patients treated with chemotherapy. There were 45% of females, 6% of never smokers and 45% of patients diagnosed with adenocarcinoma. EGFR gene copy number was evaluated by fluorescence in situ hybridization and EGFR protein level by immunohistochemistry.

RESULTS

High EGFR gene copy number and protein level were found in 33% and 71% of patients, respectively. Both markers were significantly associated (P = 0.01). For objective response and disease control, there was no difference between patients defined as negative or positive for both EGFR gene copy number (P = 0.39 and P = 1.00, respectively) and for EGFR protein (P = 1.00 and P = 0.80, respectively). There were no differences in progression-free and overall survival according to EGFR gene copy number (P = 0.76 and P = 0.82, respectively) and protein level (P = 0.67 and P = 0.62, respectively).

CONCLUSION

In chemotherapy-treated NSCLC patients, EGFR gene copy number was positively associated with protein level but none of the features were predictive for either treatment response or survival.

Combination of EGFR gene copy number and protein expression predicts outcome for advanced non-small-cell lung cancer patients treated with gefitinib

BACKGROUND

Biological markers for optimal selection of patient to epidermal growth factor receptor (EGFR)-targeted therapies are not established in advanced non-small-cell lung cancer (NSCLC).

PATIENTS AND METHODS

EGFR/HER2 gene copy number by FISH, EGFR protein and pAKT expression by immunohistochemistry (IHC) and EGFR and KRAS mutations were tested in 204 gefitinib-treated NSCLC patients.

RESULTS

Increased EGFR and HER2 gene copy number (FISH+), EGFR protein overexpression (IHC+), EGFR mutations and pAKT overexpression were all associated with significantly higher response rates (33%, 29%, 22%, 39% and 20% respectively). EGFR FISH+ (32%) and IHC+ (61%) correlated with improved survival, while EGFR mutations (27%), KRAS mutations (26%) and pAKT expression (69%) did not. In multivariate survival analysis EGFR FISH and IHC were independent predictive markers. EGFR FISH+/IHC+ patients (23%) had a median survival of 21 months versus 6 months for double-negative patients (30%).

CONCLUSION

Combination of EGFR FISH and IHC is effective predictor for benefit from gefitinib. Patients with double-negative results are unlikely to benefit in western NSCLC populations.

Chromosomal aneusomy in sputum predicts for lung cancer in nested case-control study

7201

Background: Lung cancer usually is disseminated at diagnosis and prognosis is poor for most patients. Heavy smokers are known to be at high risk factor for lung cancer and are target population for lung cancer prevention. Sputum has long been considered a potential source of material for biomarker based early detection but no tests have been validated. Methods: Chromosome aneusomy was sought in sputum of 114 cases and 110 controls from the Colorado Sputum Screening Cohort Study, which includes heavy smokers with airflow obstruction matched on age, gender, and date of sample collection. Subject characteristics were as follows: mean age 67 (± 8 years), 75% males, 36% current smokers, mean pack-year 71.2 (± 33.5). The FISH probe LAVysion (Vysis/Abbott) was used, including the DNA targets EGFR, MYCC, 5p15 and CEP6. Results: The multi-target set showed the highest sensitivity (0.78) and specificity (0.95) rates in specimens collected within 12 months of lung cancer diagnosis. The individual probes EGFR, MYCC, 5p15 and CEP6 showed, respectively, decreasing sensitivity rates (0.78, 0.67, 0.62, and 0.29) and increasing specificity rates (0.84, 0.91. 0.86, and 0.95). Combinations of two specific probes (AND) or of any of two probes (OR) have not favorably impacted these coefficients. Proportion of abnormal sputum specimens was higher in squamous cell carcinoma than in adenocarcinoma or small cell carcinoma, both considering the set of specimens collected 12 months prior to disease diagnosis (92%, 75%, 60%) and all specimens (80%, 58%, 44%). Conclusions: Chromosomal aneusomy in sputum was demonstrated in a nested case-control cohort to be a promising marker for prediction of lung cancer risk in heavy smokers with airflow obstruction. Evaluation of four DNA targets was more effective than any single marker or combination of markers, and the test had high sensitivity in patients with squamous cell carcinoma.

[Table: see text]

Prediction of clinical outcome in non-small cell lung cancer (NSCLC) patients treated with gefitinib using Matrix-Assisted Laser Desorption/Ionization-Time of Flight Mass Spectrometry (MALDI-TOF MS) of serum

7004

Background: Assessment of epidermal growth factor receptor (EGFR) gene copy number, mutational status, and protein levels may predict response and possibly survival following treatment with EGFR tyrosine kinase inhibitors, but all these methods require tumor tissue, highlighting the need for a non-invasive predictive test. We evaluated the ability of MALDI-TOF MS profiling of serum to predict which patients (Pts) with NSCLC were likely to benefit from gefitinib treatment. Methods: Serum from Pts with NSCLC, collected prior to treatment with gefitinib, was subjected to MALDI-TOF MS using Voyager DE-STR or DE-PRO instruments. Replicate mass spectra obtained at two institutions were submitted to a third party for processing (background subtraction, noise estimation, normalization, spectral alignment and peak identification) and selection of discriminating mass/charge (m/z) values from a training cohort of 70 Caucasian Pts. The predictive capability of the profile was then assessed in independent cohorts of NSCLC Pts. Results: Intra- and inter- laboratory reproducibility of MALDI spectra were excellent. A set of 11 m/z values (mass range 5 - 12.5 kDa) predictive of clinical benefit were identified in the training cohort and confirmed by leave-one-out cross validation. Spectra from 19/70 Pts in the training cohort were unclassifiable. For the remaining 51 Pts the algorithm discriminated groups more or less likely to benefit with respect to time to progression (median 3.0 vs. 1.5 mo., p = 0.0325) and overall survival (median 14.6 vs. 2.3 mo., p = 0.0128). Validation was performed in an independent cohort consisting of serum from 69 Japanese Pts. Spectra from 13/69 Pts were unclassifiable. For the remaining 56 Pts it was possible to identify a group with superior time to progression (median 14.8 vs. 2.1 mo., p = 0.0012) and overall survival (median 19.1 vs. 7.9 mo., p = 0.0102). Conclusions: MALDI-TOF MS of pretreatment serum may aid with the identification of subsets of NSCLC Pts that will benefit from treatment with gefitinib. This algorithm is currently being evaluated in an expanded cohort of Pts receiving gefitinib treatment and in Pts treated with erlotinib.

[Table: see text]

Prognostic significance of EGFR FISH and IHC in non small-cell lung cancer patients treated with chemotherapy alone

7193

Background: High EGFR gene copy number by fluorescence in situ hybridization (FISH) predicts response and survival benefit in non small-cell lung cancer (NSCLC) patients (pts) treated with EGFR tyrosine-kinase inhibitors, but its prognostic value remains debated. We aimed to evaluate the association of EGFR FISH, EGFR immunohistochemistry (IHC) and prognosis in NSCLC pts treated with chemotherapy alone. Methods: 85 pts treated with platinum-containing chemotherapy (median follow up of 15 months [range: 2–29 months]) were included in the study. There were 47% females, 35% of pts with performance status (PS) 0, 53% PS = 1 and 12% PS = 2, 6% of never-smokers. Median age was 62 years (range: 41–80 years). Stage I-IIIA was diagnosed in 7%, stage IIIB - 44% and stage IV - 48% pts. Adenocarcinoma was the most common histology (51% pts). EGFR FISH was performed using LSI EGFR SpectrumOrange/CEP 7 SpectrumGreen probe and IHC using DAKO PharmDx kit. Results: FISH results were available in 79 pts (93%), and EGFR FISH-positive tumors (high polysomy or gene amplification) were found in 28 pts (35%). IHC results were available in 81 pts (95%) and 25 pts (31%) were scored as positive (staining index ≥200). Distribution of clinical characteristics did not differ according to either FISH or IHC result. FISH-positive pts had higher EGFR IHC staining indices as compared with FISH-negative pts (median 160 vs. 60, p = 0.005, Mann-Whitney U test). Median survival of FISH-positive pts was 12.6 months vs. 8.1 months in FISH-negative pts (log-rank p = 0.68; HR = 0.88 [95% CI: 0.49–1.59]) and the respective figures for progression-free survival (PFS) were 7.5 vs. 4.9 months, log-rank p = 0.72; HR = 0.91 [95% CI: 0.55–1.51]. Median survival in IHC-positive vs. IHC-negative pts was 6.6 months vs. 9.2 months (log-rank p = 0.44; HR = 1.27 [95% CI: 0.69–2.36]). There was no significant difference in PFS (median of 4.8 vs. 5.3 months, respectively; log-rank p = 0.71; HR = 1.11 [95% CI: 0.64–1.92]). FISH and IHC remained insignificant in a Cox regression survival analysis. Conclusion: In this cohort of NSCLC patients treated with chemotherapy alone, EGFR FISH was associated with EGFR IHC and both features had no statistically significant influence on prognosis.

[Table: see text]

Studies of pemetrexed and gemcitabine, alone and in combinations, in human lung cancer models

17114

Background: Gemcitabine (GEM) is a deoxycytidine analog that inhibits DNA synthesis. Pemetrexed (ALIMTA, PEM) is a novel antifolate inhibiting multiple enzymes targets, including thymidylate synthase (TS). This study aimed at evaluating the antitumor effects of these antimetabolites against NSCLC and SCLC tumor models. Methods: In vitro growth inhibition (IC50) studies were done by 6-days MTT assays against a panel of 20 NSCLC and 17 SCLC cell lines. In vivo studies used only NSCLC H2122 tumor line, implanted either subcutaneously in athymic nude mice or orthotopically in athymic nude rats. Drugs were given via the ip route at the designated schedules. Results: Against NSCLC and SCLC cell lines, the averaged IC50s of GEM were 0.015 ± 0.008 μM and 0.055 ± 0.04 μM respectively. The corresponding averaged IC50s for PEM were 0.65 ± 0.2 μM and 0.091±0.018 μM respectively. When H2122 tumors reached 50–100mg, mice were treated with 10 daily doses of PEM at 100, 200 and 300 mg/kg, or three doses of GEM every 4 days at 30, 60 and 120 mg/kg. PEM delayed tumor growth by 12 to 18 days, and GEM delayed by 10 to 14 days, relative to vehicle control. Results of three combination regimens with GEM (30 mg/kg) and PEM (100 mg/kg) were: (1) GEM → PEM gave intermediate activities between the two single agents, but was toxic to animals; (2) PEM and GEM given concurrently were more active than single agents alone and delayed tumor growth by 12 days with some toxic side effects; (3) PEM → GEM was better than the single agents alone, and delayed tumor growth by ∼14 days without toxicity. Athymic nude rats bearing orthotopic H2122 tumors given PEM daily at 50, 100 and 200 mg/kg for 21 days had significantly prolonged survival, but not in a dose-dependent manner. PEM at 50 mg/kg was more effective than doses at 100 or 200 mg/kg. GEM was toxic to nude rats due to poor plasma deamination of GEM. Conclusions: In vitro, PEM was more potent against SCLC than NSCLC cell lines, but GEM had similar activities against all lung lines tested. Studies of H2122 xenografts in rodent supported PEM → GEM as the preferred sequence for the combined administration of these two drugs.

[Table: see text]

High gene copy number of epidermal growth factor receptor by fluorescence in situ hybridization is frequent in head and neck squamous cell carcinomas and associates with worse recurrence-free survival

5502

Background: Epidermal growth factor receptor (EGFR) is expressed in 90–95% of head and neck squamous cell carcinoma (HNSCC). High EGFR gene copy number and activating mutations correlate with response and survival in lung cancer upon the treatment with EGFR tyrosine kinase inhibitors, but such findings have not been reported for HNSCC. A better understanding of the EGFR pathway may improve the use of EGFR inhibitors and provide a rationale for combination therapy. Methods: Tumor samples were obtained from 47 patients and analyzed for increased EGFR copy numbers by fluorescence in situ hybridization (FISH) (Vysis/Abbot) (N = 47), by immunohistochemistry (IHC) using antibodies against EGFR, phospho-EGFR (Tyr1173) and phospho-AKT (Ser473) (N = 31), for activating mutations in EGFR exons 18, 19 and 21 (N = 29), and for gene expression using Affymetrix Human 133U Plus 2.0 GeneChip (N = 18). Results from these assessments were tested for associations with patient characteristics and clinical endpoints. Univariate analysis was performed to test for differences in recurrence-free survival (RFS) and overall survival. Results: Twenty-six (63%) of 41 samples with FISH result demonstrated EGFR high polysomy and/or gene amplification (FISH+). The FISH(+) group did not differ from the FISH(−) group with regard to age, sex, race, tumor grade and site, disease stage, total EGFR, phospho-EGFR, or phospho-AKT, although the total EGFR protein score was close to significance (p = 0.10, Chi2-test for trend). The FISH(+) were associated with worse RFS, although this association did not reach significance (p = 0.057, Wilcoxon test). No activating EGFR point mutations were found. Using the FISH results as a supervising parameter, microarray data was interrogated and expression of genes involved in wnt signalling, cell motility, G-protein regulation and epithelial-mesenchymal transition pathways were found to be enhanced in the FISH(+) group (FDR <25%). Conclusions: High gene copy number of EGFR by FISH is frequent in HNSCC and may be a poor prognosticator. This data suggests that FISH(+) may be a marker of clinical benefit upon EGFR inhibition in HNSCC and further investigation is indicated.

[Table: see text]

EGFR genomic gain in Japanese non-small cell lung cancer patients treated with gefitinib

7164

Background: Increased EGFR gene copy numbers evaluated by fluorescence in situ hybridization (FISH) is excellent predictor of clinical benefit from EGFR tyrosine kinase inhibitors (TKIs) in Caucasian non small-cell lung cancer (NSCLC) patients. This study was performed to verify how this marker performs in Oriental patients. Methods: A cohort of 59 Japanese patients with recurrent NSCLC after surgery was treated with gefitinib 250 mg daily. The cohort included 46% of females, 47% of never-smokers, 68% of patients had prior chemotherapy and 59% had stage III-IV at the time of surgery. Adenocarcinoma was the most common histology (85%). FISH was performed using EGFR /CEP 7 and PathVysion probes (Vysis/Abbott). Specimens were classified as FISH+ when showing gene amplification or high polysomy (≥4 copies of the gene in ≥40% of tumor cells). Effectiveness of gefitinib was concluded when there was 30% tumor shrinkage in imaging studies and/or elevated CEA level decreased to <50% of baseline. Results: FISH results were available in 51 tumors: 35 (69%) were EGFR+ and 29 (57%) were HER2+. There was a significant association between EGFR+ and HER2+ (p = 0.012) and between EGFR FISH+ and of EGFR mutations (p = 0.003). All 9 tumors exhibiting clustered EGFR gene amplification harbored mutations. Response rate assessed in 43 patients was 59% in EGFR+ and 33% in EGFR- (p = 0.11). Gefitinib treatment produced response in 5 (83%) cases with gene amplification and 12 (52%) cases with high polysomy. Response rate was 77% in 22 cases positive for FISH and mutation and 100% in 5 cases with EGFR mutation/FISH- cases. No responses were observed in EGFR WT patients, neither in 9 FISH+ nor in 10 FISH- cases. Survival was not significantly impacted by EGFR (p = 0.42) or HER2 (p = 0.40) FISH status, while EGFR mutation was significantly associated with longer survival (p = 0.005). Conclusion: Patients in this Japanese cohort had more frequently FISH+ status than in the Caucasian populations, which may contribute to higher sensitivity to EGFR TKIs. However, the predictive power of EGFR FISH status for TKI sensitivity seems to be lower supporting significant differences in the mechanisms of EGFR pathway activation in Oriental and Caucasian populations.

[Table: see text]