Introduction: Bortezomib is a small-molecule proteasome inhibitor with single-agent activity in patients with non-small cell lung carcinoma (NSCLC) and synergy with gemcitabine in preclinical studies. This phase II study of bortezomib in combination with gemcitabine/carboplatin was conducted in chemotherapy-naive advanced NSCLC patients to assess efficacy and safety. Methods: Patients with selected stage IIIB/IV NSCLC, performance status 0–1, and no history of brain metastasis received up to six 21-day cycles of gemcitabine 1000 mg/m2, days 1 and 8, carboplatin area under curve 5.0, day 1, and bortezomib 1.0 mg/m2, days 1, 4, 8, and 11. Results: One-hundred-fourteen patients (52% adenocarcinoma, 85% stage IV) received a median of 3.6 treatment cycles. Median follow-up was >3 years. Median overall survival was 11 months; 1-year and 2-year survival rates were 47% and 19%, respectively. Median progression-free survival was 5 months; 1-year progression-free survival rate was 7%. Response rate was 23%, and disease control rate (responses + stable disease) was 68%. The most common grade 3/4 toxicities were thrombocytopenia (63%) and neutropenia (52%). One patient experienced febrile neutropenia. Grade 3/4 neuropathy occurred in 4%, and a further 6% experienced grade 2 sensory neuropathy. Conclusions: Bortezomib plus gemcitabine/carboplatin resulted in a notable survival benefit in patients with advanced NSCLC, with the anticipated primary toxicity of myelosuppression. Further studies designed to investigate the role of bortezomib in advanced NSCLC are warranted.

Background: With the emergence of Src inhibitors in clinical trials, improved knowledge of the molecular responses of cancer cells to these agents is warranted. This will facilitate the development of tests to identify patients who may benefit from these agents, allow drug activity to be monitored and rationalize the combination of these agents with other treatment modalities.Methods: This study evaluated the molecular and functional effects of Src inhibitor AZD0530 in human lung cancer cells, by Western blotting and reverse transcription-polymerase chain reaction, and by assays for cell viability, migration, and invasion.Results: Src was activated in four of five cell lines tested and the level corresponded with the invasive potential and the histologic subtype. Clinically relevant, submicromolar concentrations of AZD0530 blocked Src and focal adhesion kinase, resulting in significant inhibition of cell migration and Matrigel invasion. Reactivation of STAT3 and up-regulation of JAK indicated a potential mechanism of resistance. AZD0530 gave a potent and sustained blockage of AKT and enhanced the sensitivity to irradiation.Conclusions: The results indicated that AZD0530, aside from being a potent inhibitor of tumor cell invasion which could translate to inhibition of disease progression in the clinic, may also lower resistance of lung cancer cells to pro-apoptotic signals.

p53R2 is a p53-inducible ribonucleotide reductase that contributes to DNA repair by supplying deoxynucleotide triphosphate pools in response to DNA damage. In this study, we found that p53R2 was overexpressed in prostate tumor cell lines compared with immortalized prostatic epithelial cells and that the protein was induced upon DNA damage. We investigated the effects of p53R2 silencing on DNA damage in LNCaP cells (wild-type p53). Silencing p53R2 potentiated the apoptotic effects of ionizing radiation and doxorubicin treatment as shown by increased sub-G1 content and decreased colony formation. This sensitizing effect was specific to DNA-damaging agents. Comet assay and γ-H2AX phosphorylation status showed that the decreased p53R2 levels inhibited DNA repair. Silencing p53R2 also reduced the levels of p21WAF1/CIP1 at the posttranscriptional level, suggesting links between the p53-dependent DNA repair and cell cycle arrest pathways. Using LNCaP sublines stably expressing dominant-negative mutant p53, we found that the sensitizing effect of p53R2 silencing is mediated by p53-dependent apoptosis pathways. In the LNCaP sublines (R273H, R248W, and G245S) that have defects in inducing p53-dependent apoptosis, p53R2 silencing did not potentiate DNA damage–induced apoptosis, whereas p53R2 silencing was effective in a LNCaP subline (P151S) which retains the ability to induce p53-dependent apoptosis. This study shows that p53R2 is a potential therapeutic target that could be used to enhance the effectiveness of ionizing radiation or DNA-damaging chemotherapy in a subset of patients with prostate cancer. (Mol Cancer Res 2008;6(5):808–18)

Lung cancer is the leading cause of cancer death in men and women and frequently presents as advanced disease. The majority of lung cancers are of the non-small cell type, for which chemotherapy has demonstrated modest survival benefits at all stages of disease. Clearly, more effective therapies are needed. Agents that alter critical molecular cell growth pathways, so-called targeted therapies, are a growing area of research and development. Targeted therapies including drugs directed at the epidermal growth factor receptor (EGFR) and vascular endothelial growth factor (VEGF) have recently established a role in the treatment of advanced stage non-small cell lung cancer (NSCLC). These drugs and many others are undergoing investigation, either as single agents or in combination with cytotoxics or other targeted therapies, with dual goals of improved efficacy and reduced toxicity. Although progress has been made in target identification for lung cancer treatment, the ability to select groups of NSCLC patients who benefit from these therapies based on predictive markers remains a challenge. Ongoing studies correlating potential predictive biomarkers with patient outcome are designed to refine the use of developing targeted therapies, that is, to provide a rational basis for “targeted use of targeted therapies.” Despite some recent breakthroughs in identifying molecular signatures predictive of benefit, much remains to be learned. Using erlotinib and bevacizumab as prime examples, this chapter will review a dilemma currently facing both basic scientists and clinical investigators engaged in the study of NSCLC, namely how to develop and test paradigms for individualizing patient therapy.

Introduction: Bortezomib is a small-molecule proteasome inhibitor with single-agent activity in patients with non-small cell lung carcinoma (NSCLC) and synergy with gemcitabine in preclinical studies. The combination of gemcitabine and carboplatin is an accepted first-line treatment for advanced NSCLC. We conducted a phase I study of gemcitabine and carboplatin in combination with bortezomib. Methods: Bortezomib was administered on days 1, 4, 8, and 11, after gemcitabine on days 1 and 8, and carboplatin on day 1 of a 21-day cycle. Three escalating dose levels were evaluated: bortezomib 1.0 mg/m2/gemcitabine 800 mg/m2, bortezomib 1.0 mg/m2/gemcitabine 1000 mg/m2, and bortezomib 1.3 mg/m2/gemcitabine 1000 mg/m2, in combination with carboplatin AUC 5.0. Results: Twenty-six patients with advanced NSCLC were treated; 21 were chemotherapy-naive. The median age was 59 years (range, 34–74), and 23 patients were stage IV. The Karnofsky performance score was ≤80% in 10 and >80% in 16 patients. Dose-limiting toxicities were grade 3 thrombocytopenia with bleeding and febrile neutropenia accompanied by grade 4 thrombocytopenia and grade 3 hyponatremia. The maximum-tolerated dose was defined as bortezomib 1.0 mg/m2, gemcitabine 1000 mg/m2, and carboplatin AUC 5.0. The most common grade 3/4 toxicities were thrombocytopenia (rarely associated with bleeding), and neutropenia. Nine of 26 patients (35%) achieved partial response, and eight patients had stable disease. Conclusions: The combination of bortezomib 1.0 mg/m2, gemcitabine 1000 mg/m2, and carboplatin AUC 5.0 demonstrated manageable toxicities and encouraging activity in NSCLC. This regimen was used in a phase II study.

Because of the current controversy on the origin and clinical value of circulating KRAS codon 12 mutations in lung cancer, we screened 180 patients using a combined restriction fragment-length polymorphism and polymerase chain reaction (RFLP–PCR) assay. We detected KRAS mutations in 9% plasma samples and 0% matched lymphocytes. Plasma KRAS mutations correlated significantly with poor prognosis. We validated the positive results in a second laboratory by DNA sequencing and found matching codon 12 sequences in blood and tumor in 78% evaluable cases. These results support the notion that circulating KRAS mutations originate from tumors and are prognostically relevant in lung cancer.

15562 Background: Previous reports have suggested benefit of GC in a modest number of patients with RCC but with concerns regarding toxicity. We subsequently evaluated an adjusted dose regimen in pts with advanced unresectable or metastatic RCC. Methods: Pts who had not received prior chemotherapy where treated with G at 900mg/m2 days 1,8,15 and C at 625mg/m2 BID day 1- 21 of a 28 day cycle. Pts must have had adequate performance status, organ function and may have received prior immunotherapy. The primary endpoint was probability of response, confirmed and unconfirmed complete (CR) and partial responses (PR). A one-stage design for accrual was used and regimen would not be considered of interest if the true probability of response was <=5%. Results: 43 patients were enrolled of which 1 was ineligible and 2 were not analyzable. Median follow-up was 11.2 months (range: 1.3 to 18.5). One confirmed complete response (CR) and three unconfirmed partial responses (PR) were observed, for an overall response probability of 10% (95% CI: 3% to 24%). Four of the eligible patients had inadequate disease assessment and were assumed to be non-responders. 19 pts (48%) had stable disease (SD). The 6-month probability of being free from treatment failure was 20% (95% CI: 8% to 32%). The probability of survival at 6 months was estimated to be 75% (95% CI: 62% to 88%). Median survival was estimated to be 13 months (lower limit of 95% CI is 7 months, upper limit not estimable). One patient each experienced Grade 4 neutropenia, fatigue, thrombocytopenia and hemolysis/renal failure. Most common Grade 3 toxicities were neutropenia (12 pts), fatigue (5 pts), and leucopenia (4 pts). Conclusions: GC at this dose and schedule benefits a small percentage of pts with RCC with an acceptable toxicity profile. Given the RR and SD, GC may serve as a base regimen for combination therapy with targeted agents. No significant financial relationships to disclose.

BACKGROUND:Epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs) given concurrently with chemotherapy in 4 large randomized clinical trials did not improve patient outcomes compared with chemotherapy alone in advanced non-small-cell lung cancer (NSCLC). We hypothesized that the lack of benefit resulted from a negative interaction between chemotherapy and EGFR TKIs. MATERIALS AND METHODS:Herein, we report the cell cycle and apoptotic effects of treatment with erlotinib and docetaxel in the NSCLC cell lines A549 and Calu-1, both of which are mutant for K-ras and wild-type for EGFR. RESULTS:Treatment with erlotinib resulted in accumulation of cells in G(1) phase in A549 cells, with no evidence of apoptosis. Docetaxel treatment led to apoptosis as assessed by increased sub-G1 DNA content and cleavage of caspase 3 and poly (ADP-ribose) polymerase. The sequence of docetaxel followed by erlotinib resulted in significantly enhanced apoptosis compared with single-agent docetaxel in both cell lines. However, in the reverse sequence of erlotinib followed by docetaxel, a reduction of apoptosis was observed. We hypothesize that cell cycle arrest induced by erlotinib accounts for these findings in the presence of wild-type EGFR and that pharmacodynamic separation of the 2 drug classes will ameliorate these effects. CONCLUSION:These studies provide a rationale for intermittent dosing of EGFR TKIs with chemotherapy in order to enhance cytotoxicity.

This chapter presents recent advances in the molecular biology of lung cancer. This knowledge contributes to our understanding of the mechanisms of carcinogenesis, cancer progression, and metastasis and is essential for the subsequent development of new therapies targeted against disease-relevant genetic pathways. Proto-oncogenes are genes that contribute to malignant transformation when mutationally activated or overexpressed. In 1917, Rous reported that sarcoma in chickens could be caused by a transmissible agent. Activated proto-oncogenes contribute to carcinogenesis by driving the cell toward neoplasia. Histopathologic subtyping is becoming increasingly important for lung cancer therapy, especially with the advent of molecular targeted therapies, which have been shown to convey different risks and benefits in histologic subtypes of non–small cell lung cancer (NSCLC). The chapter describes only selected topics regarding the therapeutic implications of lung cancer biology.

Background: This phase I study was performed to determine the dose-limiting toxicity and maximum tolerated dose (MTD) of docetaxel in combination with bortezomib in patients with advanced non-small cell lung cancer (NSCLC) or other solid tumors.Methods: Patients were enrolled in cohorts of three over six dose levels. Each treatment cycle was 3 weeks long and consisted of one docetaxel infusion (day 1) and four bortezomib injections (days 1, 4, 8, and 11). Dose escalation and MTD determination were based on the occurrence of dose-limiting toxicities in cycle 1 only.Results: A total of 36 patients were enrolled, 26 of whom had NSCLC. All patients received at least one dose of study drug at one of five dose levels. The MTD of the combined regimen was determined to be 1.0/75 mg/m(2) bortezomib/docetaxel. The combination was generally well tolerated. Toxicities were manageable, and no additive toxicities were observed. The most common adverse events were fatigue (67% of patients), nausea (50%), diarrhea (39%), and neutropenia (39%). Two patients with NSCLC achieved a partial response, and seven (19%) patients achieved stable disease (including six patients with NSCLC).Conclusion: The combination of bortezomib and docetaxel was feasible and well tolerated in patients with advanced NSCLC or other solid tumors. The recommended phase II dose is bortezomib 1.0 mg/m(2) on days 1, 4, 8, and 11 plus docetaxel 75 mg/m(2) on day 1, cycled every 21 days. Therapeutic doses of docetaxel and bortezomib are achievable for this combination.

PURPOSE:Advanced bronchioloalveolar carcinoma (BAC) is a distinct subtype of non-small-cell lung cancer (NSCLC) for which there is currently no optimal therapy. Based on preclinical and clinical data suggesting relevance of the epidermal growth factor receptor (EGFR) axis in BAC, the Southwest Oncology Group initiated a phase II trial (S0126) to evaluate the EGFR tyrosine kinase inhibitor gefitinib in chemotherapy-naïve and chemotherapy-pretreated patients with advanced BAC.METHODS:A total of 136 eligible and assessable patients (101 untreated, 35 previously treated) received gefitinib 500 mg daily until progression or prohibitive toxicity.RESULTS:The median age was 68.0 years (range, 34.3 to 88.6); 51% were female; 89% had a performance status (PS) of 0% or 1% and 11% had a PS of 2. The Response Evaluation Criteria in Solid Tumors response rate was 17%, with 6% complete responses (CRs) among 69 previously untreated patients with measurable disease, and 9% with no CRs among 22 pretreated patients. Median survival was 13 months for both chemo-naïve (95% CI, 8 to 18) and previously treated patients (95% CI, 6 to 17). Overall survival at 3 years was 23% (95% CI, 14% to 32%). Toxicity consisted mainly of rash and diarrhea, but 2% of patients died of presumed interstitial lung disease. Exploratory subset analyses revealed improved survival among women (P = .031), patients developing a rash (P = .003), never-smokers (P = .061), and patients with a PS of 0 or 1 (P = .015).CONCLUSION:Gefitinib is an active agent in advanced stage BAC. Several subsets demonstrate significantly improved clinical outcomes.

This work explores spontaneous immortalization in keratinocytes, derived from two skin samples, that display naturally elevated telomerase activity. Serially passaged with 3T3 feeder layer support, the keratinocytes were examined for colony-forming ability, telomerase activity, telomere length, and finally gene expression using Affymetrix DNA microarrays. The cells initially exhibited normal karyotypes and low colony-forming efficiencies typical of normal epidermal cells, but after 40 passages (approximately 400 generations) colony-forming ability increased markedly, yielding immortalized lines exhibiting a small number of chromosomal aberrations and functionally normal p53. An improved protocol for analysis of microarray data permitted detection of 707 transcriptional changes accompanying immortalization including reduced p16(INK4A) mRNA. Telomerase activity was clearly elevated in cells even at low passage from both samples, and telomerase catalytic subunit mRNA was greatly elevated in those with elevated colony-forming ability. The data raise the possibility of an unusual natural phenotype in which aberrant telomerase regulation extends keratinocyte lifespan until rare variants evade senescence. In addition to revealing a potential tumor-prone syndrome, the findings emphasize the desirability of carefully minimizing the degree or timing of elevated expression of telomerase used to immortalize cells for therapeutic purposes.

Epidermal gro wth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs) given concurrently with chemotherapy do not improve patient outcomes compared with chemotherapy alone in advancedstage non–small-cell lung cancer (NSCLC). One potential explanation for this lack of benefit is a negative interaction or antagonism between chemotherapy and EGFR TKIs when delivered concomitantly. Support for this line of reasoning is provided by preclinical data demonstrating that EGFR TKIs induce primarily a cytostatic effect resulting from a G1 cell cycle arrest in cell lines with wildtype EGFR, reducing cell cycle phasedependent activity of chemotherapy, whereas they induce apoptotic cell death in tumors with EGFR-activating mutations. Because the great majority of NSCLC tumors consist of wild-type EGFR, sequence-specific interactions of EGFR TKI/chemotherapy combinations might negatively influence the efficacy of these regimens in patients with NSCLC. Further evidence is provided by EGFR mutational analysis in patient tumor specimens from the TRIBUTE study. Herein we provide the preclinical and clinical rationale for studies examining the concept of pharmacodynamic separation as a means for overcoming hypothesized antagonism of EGFR TKIs and chemotherapy.

13150 Background: AI CaP is an invariably fatal disease. While treatment with Doc, a microtubule-stabilizing taxane, improves survival, patient outcomes remain suboptimal. PS-341 inhibits degradation of cell cycle and tumor suppressor proteins resulting in cycle arrest and apoptosis. We hypothesized that the combination of Doc with PS-341 would abrogate the abnormal survival response seen in AI CaP and lead to improved tumor cell kill, but that results would be dependent on administration schedule due to interactive cell cycle kinetics. Methods: The PC3 cell line model of AI CaP was evaluated in vitro and in vivo to determine response to Doc or PS-341 alone, and in combination in sequences of PS-341→Doc, Doc→PS-341, and simultaneous (PS-341 + Doc). Cell cycle and protein analyses were performed by flow cytometry and Western blotting, respectively. For nu/nu mouse xenografts, 5 × 106 cells were injected subcutaneously into each flank. The agents were administered either together or 24hr apart, with all regimens given weekly [IP doses: Doc: 10 mg/kg; PS-341: 0.5 mg/kg]. Results: in vitro: Each combination showed an increased apoptotic sub-G1 population versus untreated cells, in addition to altered cell cycling in a sequence-specific manner. Of the combinations, PS-341 + Doc showed the largest sub-G1 while Doc→PS-341 had the lowest sub-G1 but the largest S-phase content; in vivo: PS-341 + Doc showed a cytotoxic effect (reduction in tumor volume) while the combinations of Doc→PS-341 and PS-341→Doc both showed growth inhibition (stabilization of tumor growth) as best response. Conclusions: Combinations of PS-341 and Doc have sequence specific cell cycle effects leading to increases in apoptosis (PS-341 + Doc) or cell cycle arrest (Doc→PS-341). Clinical validation of these findings is warranted. (ACS: CRTG-0019701-CCE) No significant financial relationships to disclose.

13012 Background: Cetuximab, an IgG1 monoclonal antibody directed against the epidermal growth factor receptor (EGFR), is FDA-approved at a 400 mg/m2 loading dose, followed by 250 mg/m2 weekly maintenance. Clinical activity of cetuximab has been reported to correlate with grade of skin toxicity, and skin toxicity is reported to increase with increasing dose. We therefore examined the safety and feasibility of escalating weekly cetuximab doses. We hypothesized that increased dose would correlate with rash severity as a surrogate for tumor response. Methods: Four dose levels were tested: cetuximab loading 400 mg/m2 and 250,300,350,400 mg/m2 weekly maintenance. Dose limiting toxicity (DLT) was defined as: grade 4 platelets, grade 3 platelets with bleeding, febrile neutropenia, grade 4 cutaneous toxicity, grade 3 cutaneous toxicity necessitating holding cetuximab for > 4 weeks or any other grade ≥ 3 non-hematologic toxicity. Rash was evaluated using two additional validated dermatology methods: acne lesion counting and global acne grading scale (Int J Derml.1997;36:416–18, J Am Acad Derm.1996;35:559–65, Arch Derm.1982;118:23–25). Results: Twelve patients with advanced solid tumors were treated, including 3 H&N, 2 pancreas, 2 breast, 2 lung, 2 colorectal and 1 bladder. Patient characteristics: age range 44–84, median 62; gender: 10 M; KPS ≥80/<80=7/5; prior chemo ≤1:>1=5:7; median cycles 2 (1–8). Treatment was generally well tolerated. There were no DLTs. The most common grade 3/4 toxicities were acneiform rash (1) and lymphopenia (2). The majority of patients (6) had a Grade 2 rash. In 10 evaluable patients, there were no responses; 3 patients had stable disease. Correlative science studies are ongoing evaluating EGFR expression and polymorphisms, pEGFR, pMAPK, pAKT, Ki67, p27 levels and K-ras mutations. Conclusions: 1) Cetuximab 400 mg/m2 loading dose and 400 mg/m2 weekly maintenance is feasible and well tolerated. Doses up to 400 mg/m2 did not portend increased toxicity and a MTD was not reached on this schedule. 2) Grade of rash did not increase with increasing doses of weekly cetuximab in this limited population. 3) To date, cetuximab has not demonstrated RECIST response in this cohort of pretreated patients with solid tumors. 20 additional patients are being evaluated at the highest dose level. No significant financial relationships to disclose.

4106 Background: Correlative studies that incorporate biomarkers to rapidly analyze response to new agents are needed. Unique to pancreatic cancer is the high incidence of KRAS mutations (over 90%). This pilot study evaluated plasma KRAS mutations for disease monitoring in LAPC pts treated on a Phase I trial combining CT-RT with the EGFR inhibitor, gefitinib. METHODS DNA was extracted from plasma of 11 pts collected at 3 timepoints: pre-gefitinib, pre-CT-RT, and post-gefitinib+CT-RT. Matched tissue DNA was obtained from 4 pts with available paraffin blocks. KRAS codon 12 mutations were detected using a two-stage RFLP-PCR assay. Cell line controls: Calu-1 (mutant KRAS) and LNCaP (wild-type KRAS). Mutations were confirmed by direct DNA sequencing. Results were related to pt clinical data. RESULTS KRAS mutations were detected in the pre-gefitinib plasma of 5/11 pts, and in the matched tumor tissue of 3/4 pts. Of the 5 pts with plasma KRAS mutations, 2 pts with no detectable mutant KRAS in the plasma post-gefitinib+CT-RT had overall survival of 8 and 21 months, whereas 2 pts who retained mutant KRAS had overall survival of only 2 and 5 months, and one pt withdrew early. Of the 3 tumor tissues containing mutant KRAS, the mutations were also detectable in the matched plasma in 2 pts (67%). KRAS codon 12 mutations spectrum: 4 GGT→GAT, 2 GGT→GTT and 1 GGT→AGT. CONCLUSIONS Plasma KRAS mutations are readily detectable in LAPC pts, and the clearance or persistence of plasma KRAS mutations after treatments reflected the clinical course in some cases. The use of plasma KRAS mutation as a marker of survival and response will be further assessed in a recently approved phase I trial using a proteasome inhibitor with chemoradiation at the University of Colorado. No significant financial relationships to disclose.

Background: In preclinical models, the proteasome inhibitor bortezomib (PS-341) inhibits the growth of small cell lung cancer (SCLC) by inhibiting the antiapoptotic Bcl-2 signaling pathway. We conducted a phase II trial of PS-341 in previously treated patients with platinum-sensitive and -refractory extensive stage SCLC to determine response rate, toxicity, and survival.Methods: Patients with histologically confirmed SCLC, measurable disease, Zubrod performance status 0-1, and previous treatment with platinum-based therapy were enrolled. They were stratified by platinum-sensitivity status: sensitive (relapse > 90 days after platinum) or refractory (progression during or <= 90 days after platinum). PS-341 was administered at 1.3 mg/m(2) intravenously on days 1, 4, 8, and 11 every 21 days.Results: Of 56 eligible patients, 28 were platinum sensitive and 28 refractory. Twenty-nine patients (52%) had received two or more previous chemotherapy regimens. One platinum-refractory patient had a confirmed partial response. A majority of assessable patients (91%) progressed. Median progression-free survival and overall survival were I month and 3 months, respectively. Ten patients (18%) discontinued treatment due to adverse events or side effects.Conclusion: Although PS-341 induced a response in a patient with platinum-refractory disease, it has limited single-agent activity in this heavily pretreated cohort. As shown in preclinical models, testing of PS-341 in combination with an apoptotic trigger such as chemotherapy, is a rational clinical approach. A trial of topotecan plus PS-341 has been initiated to test this concept.

7198 Background: OPN is a secreted glycoprotein with a diverse array of functions, including induction of uPA & increased cell migration. OPN has been shown to be elevated in a number of tumor types, & its downregulation reduces tumorigenicity & metastasis in tumor models. High levels have also been associated with tumor hypoxia/angiogenesis, as are vascular endothelial growth factor (VEGF) & plasminogen activator inhibitor (PAI-1). We hypothesized that secreted levels of these biomarkers would correlate with clinical outcome after treatment. Methods: Plasma concentrations of OPN, VEGF & PAI-1 were measured by ELISA in 160 NSCLC patients enrolled on the Southwest Oncology Group (SWOG) trial S0003 (paclitaxel/carboplatin ± the hypoxic cytotoxin tirapazamine). Post-treatment plasma samples were available in 56 patients. Results: Baseline OPN plasma levels correlated significantly with patient overall survival (OS). High interpatient variability was observed, with levels ranging from undetectable to 2560 ng/ml, (median: 606.5 ng/ml). When dichotomized, median OS was 11 months for patients below median OPN levels & 7 months for those above (p = 0.004). Survival decreases with increasing OPN concentration. Furthermore, OPN levels correlated with response rate (RR) (median responders: 497; median non-responders: 698 ng/ml. Wilcoxon rank-sum p = 0.03). No association between baseline levels of either VEGF or PAI-1 with RR or OS was observed. However, plasma levels of both PAI-1 & VEGF were significantly inter-related & trended together (p < 0.0001), & both decreased significantly after treatment (p = 0.0004 & 0.04, respectively). Median decrease: OPN: 17%, PAI: 44%, VEGF: 42%. No significant differences were observed between study arms, suggesting that OPN is prognostic in NSCLC, but not predictive for response to tirapazamine. Conclusions: 1) There is a great need for development of tumor biomarkers which can be serially assessed pre- & post-therapy. 2) High OPN plasma levels were significantly associated with reduced RR & OS for patients on this trial. OPN is a strong candidate for inclusion in a panel of prognostic (& perhaps predictive) markers for NSCLC. Supported by the Hope Foundation & R01-CA107228. No significant financial relationships to disclose.