Cox PH Model Research Articles

Abstract PD10-04: Predictive genomic markers to chemotherapy and adjuvant trastuzumab via whole genome expression DASL profiling in the N9831 adjuvant study

Abstract Background: Adding adjuvant trastuzumab to chemotherapy for patients (pts) with resected HER2+ breast cancer (BC) improves disease free and overall survival (Perez EA, et al. J Clin Oncol 2011). Understanding which pts are most likely to benefit from these therapies is a goal of our team. Methods: Baseline 1639 annotated tumor specimens from pts enrolled in the phase III N9831 adjuvant trastuzumab trial (NCT00005970) were processed via the >29,000 gene probe DASL technology (Illumina) to identify genomic predictors of pt outcome to Arm A chemotherapy or Arm C chemotherapy plus concurrent trastuzumab. Samples were spotted in 96 well plates, with appropriate replicates. Extensive quality control and internal validation were performed. The association of each gene with recurrence-free survival (RFS) was determined for each treatment arm separately using CoxPH models adjusted for clinical and tumor risk factors. A p < 0.0001 was used to define significance. A geneset analysis (using 408 Kegg/BioCarta genesets) was performed to determine pathways associated with RFS. Results: PRPF4B, EWSR1, BMI1, CNNM1, FOXO4, CLIP2, MCCC1, CHRM2 were found to be significantly associated with RFS for Arm A and AAK1, DAZAP2, TP53INP1, MACF1 and ADHFE1 were associated with RFS for Arm C. The most significant pathway associated with RFS in Arm A was the KEGG hypertrophic cardiomyopathy pathway and for Arm C, the BioCarta pertussis toxin-insensitive CCR5 signaling in macrophage pathway. Conclusions: We have identified 8 genes associated with RFS for pts treated with adjuvant chemotherapy, and 5 genes associated with RFS for pts treated with concurrent trastuzumab and chemotherapy in pts with HER2 early stage BC. In addition, geneset analysis identified BioCarta and Kegg pathways associated with RFS for each arm. Validation in an independent cohort is expected to be completed by the end of 2012. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr PD10-04.

Abstract 18244: The Association Between South Asian Ethnicity and Long-term Survival Among Patients Undergoing Coronary Artery Bypass Grafting

Introduction: South Asians (SA) have a high burden of coronary artery disease (CAD) often requiring coronary artery bypass grafting (CABG). However, CABG outcomes in this population have not been explored. The objective of this study was to assess the survival of patients of SA descent undergoing CABG in a large, well described Canadian cohort. Methods: All patients who underwent CABG between 1996 and 2010 were identified from the Alberta Provincial Project for Outcomes Assessment in Coronary Heart Disease registry. Ethnicity was determined using the Nam Pehchan surname software. Differences in long-term survival were compared between SA and European (EC) patients using adjusted and then unadjusted cox proportional hazards models. SA subjects were then matched 1:1 with EC counterparts using a non-parsimonious propensity model to balance groups; models were then re-run with the matched cohort. Results: Of the 20608 patients undergoing CABG, 608 were SA. SA patients were younger (63.8 vs 65.8 years) and had a higher prevalence of diabetes (39.5 vs 27.7%), but a lower prevalence of smoking (13.3 vs 25.3%) than EC patients (all p<0.001). There were 5382 deaths long-term with a median follow-up of 7.2 years. Ninety-one deaths occurred in SA patients, 5291 in the EC patients. Kaplan-Meier survival analysis demonstrated a lower incidence of death in the SA patients (log rank, p<0.0001) (Figure). In the unadjusted Cox PH model, SA ethnicity was associated with reduced mortality (Hazard Ratio 0.53, 95% CI 0.43, 0.66), an association that persisted in the adjusted model (HR 0.63, 95%CI 0.51, 0.78). All SA subjects were matched to EC counterparts. In the matched cohort SA ethnicity remained associated with reduced mortality (HR 0.68, 95%CI 0.52, 0.90). Conclusion: In a large prospective Canadian registry, SA undergoing CABG appear to have improved long-term survival compared to EC patients. The reasons behind this striking difference in mortality deserve further exploration.

Delayed Graft Function Is Associated with an Increased Risk of Early Death Post-Transplant in Elderly Kidney Transplant Recipients

The proportion of elderly kidney transplant recipients worldwide continues to grow over time. Even in patients aged over 70, transplantation confers a long-term survival benefit compared to dialysis. However, there remains a higher risk of death in the first 125 days post-transplant in these patients. Therefore, understanding risk factors for death early post-transplant is important in the elderly. In most regions, elderly patients are more likely to receive a deceased donor transplant kidney from older donors, which may confer a DGF. We sought to examine the impact of DGF on early post-transplant survival in elderly transplant recipients. Methods: Using data from the USRDS, we utilized Cox PH models to examine the association of DGF with death within the first 6 months post-transplant in deceased donor kidney transplant recipients aged 18-39 (n=17,366), 40-49 (n=18,047), 50-59 (n=21,333), 60-69 (n=15,333), and 70+ years (n=4308) during 1997-2007. Results: The incidence of DGF was similar in all age groups, ranging from 20-23%. The table outlines the hazards for death in the first 6 months post-transplant in those with DGF compared to those without DGF in each age group, unadjusted and adjusted for donor, recipient, and transplant factors. In each model, DGF was associated with a higher hazard of death within the first 6 months in recipients aged 70 and older, whereas this was not seen in other age groups.Table: [Hazard for Death (including death after graft loss]Conclusions: DGF is associated with a higher risk of death in the early post-transplant period in elderly recipients. These results highlight the need to tailor allocation of organs and peri-operative management in the elderly to minimize the risk DGF and its associated impact on early death post-transplant.

1113PD - The Melanoma Risk Loci as Determinants of Melanoma Prognosis

ABSTRACT Background Genetic risk factors of human cancer emerge as promising markers of clinical outcome. Recent melanoma genome-wide scans (GWAS) have identified loci associated with disease risk, nevi or UV/pigmentation, but prognostic potential of these variants is yet to be determined. In this study, we performed the first-to-date systematic evaluation of the association between established melanoma risk loci and progression. Methods 891 melanoma patients prospectively accrued and followed up at NYU Medical Center were studied. We examined the association of 108 melanoma susceptibility single nucleotide polymorphisms (SNPs), selected or imputed from recent GWAS on melanoma, nevi or pigmentation, with recurrence-free survival (RFS) and overall survival (OS). The genotyping was performed using Sequenom I-plex. Univariate and multivariate Cox PH model were used to test association between each SNP and RFS and OS. Multivariate adjustments included age at diagnosis, gender, tumor stage, histological subtype, thickness, ulceration status, and mitotic index. ROC curves were used to measure utility of SNPs in predicting 3-year recurrence. Results Strong associations were observed from the multivariate analysis for rs7538876 (RCC2) with RFS (HR = 1.48, 95% CI 1.20 – 1.83, p = 0.0002) and rs9960018 (DLGAP1) with both RFS and OS (HR = 1.43, 95% CI = 1.07-1.91, p = 0.0159, HR = 1.52, 95% CI = 1.09-2.22, p = 0.0138, respectively). Subgroup analyses for ulceration and tumor thickness have identified several SNPs which show associations for RFS and OS among these different groups. In addition, we identified the classifiers rs7538876 and rs9960018, that when combined with stage and histological type, achieve a higher area under the ROC curve (AUC = 84%), compared to the baseline (AUC = 78%), in predicting 3-year recurrence. Conclusions Our data revealed associations between specific melanoma susceptibility variants and worse disease outcome. The strength of associations observed for rs7538876 and rs9960018 suggest biological implication of these loci in melanoma recurrence. The observed predictive patterns of associated variants with clinical outcome provide the evidence for the potential utilization of genetic markers in melanoma prognostication. Disclosure All authors have declared no conflicts of interest.

The melanoma risk loci as determinants of melanoma prognosis.

8557 Background: Genetic risk factors of human cancer emerge as promising markers of clinical outcome. The recent melanoma genome-wide scans (GWAS) have identified loci associated with the disease risk, nevi or UV/pigmentation, but the prognostic potential of these variants is yet to be determined. In this study, we performed the first-to-date systematic evaluation of the association between established melanoma risk loci and melanoma progression. Methods: 891 melanoma patients prospectively accrued and followed up at NYU Medical Center were studied. We examined the association of 108 melanoma susceptibility single nucleotide polymorphisms (SNPs), selected or imputed from recent GWASs on melanoma, nevi or pigmentation, with recurrence-free survival (RFS) and overall survival (OS). The genotyping was performed using Sequenom I-plex. Cox PH model was used to test the association between each SNP and RFS and OS adjusted by age at diagnosis, gender, tumor stage and histological subtype. ROC curves were used to measure predictive utility of SNPs in predicting 3-year recurrence. Results: The strong association was observed for rs7538876 (RCC2) with RFS (HR=2.445, 95% CI 1.57 – 3.8, p=0.0006) and rs9960018 (DLGAP1) with both RFS and OS (HR=4.7, 95% CI=2.11-10.43, p=0.0061, HR=1.55, 95% CI=1.11-2.17, p=0.0094, respectively) using adjusted multivariate analysis. In addition, we identified the classifier with rs7538876 and rs9960018, stage and histological type at primary tumor diagnosis, achieving a higher area under the ROC curve (AUC=84%) compared to the baseline (AUC=78%) in predicting 3-year recurrence. Univariate survival analyses have identified associations of several SNPs with ulceration and/or tumor thickness. Conclusions: Our data revealed an association between specific melanoma susceptibility variants and worse clinicopathological variables at the time of diagnosis as well as worse disease outcome. The strength of associations observed for rs7538876 and rs9960018 suggest biological implication of these loci in melanoma recurrence. The observed predictive patterns of associated variants with clinical outcome provide for the first time evidence for the potential utilization of genetic markers in melanoma prognostication.

Association of inherited genetic variation with clinical outcome in patients with advanced renal cell carcinoma treated with mTOR inhibition.

4543 Background: Mammalian target of rapamycin (mTOR)-targeted therapy is standard in patients with metastatic renal cell carcinoma (mRCC). Predictive biomarkers for response are lacking. We sought to characterize outcome after mTOR inhibition based on germline variation in 2 critical genes in the mTOR pathway: phosphoinositide-3-kinase catalytic alpha (PIK3CA) and mTOR. Methods: 76 Americans of European ancestry treated with temsirolimus or everolimus for mRCC were included. Germline DNA was genotyped for all common genetic variation (minor allele frequency [MAF] >0.05) across PIK3CA and mTOR, tagging with single nucleotide polymorphisms (SNPs) at r2>0.8. Associations between genotype and progression-free survival (PFS) and overall survival (OS) from the start of mTOR-targeted therapy were measured using univariate Cox model. Results: Among 76 patients, 66% were poor or intermediate risk and 89% received prior systemic therapies. Median follow up was 23.7 months. Median PFS and OS were 4.3 and 12.7 months, respectively. Two intronic PI3KCA SNPs were significantly associated with PFS and OS, and a SNP in the 5’ UTR of mTOR was associated with OS. Associations were maintained when adjusted for age, gender and MSKCC RCC risk categories using Cox PH model (Table). The PI3KCA SNPs are in modest linkage disequilibrium (r2 ~0.36). Conclusions: We suggest that inherited variation at PIK3CA is associated with PFS and OS after mTOR inhibition. If results are validated, prospective studies could explore the role of genotyping in treatment selection for mRCC. Further work will be needed to identify causal alleles and define the mechanism underlying the associations. [Table: see text]

A gene expression signature of MEK pathway activation to predict survival in triple-negative breast cancer.

1024 Background: Tumor cell proliferation measured by Ki67 in the surgically removed tumor after neoadjuvant chemotherapy (NAC) has been shown to predict patient outcome in breast cancer. It is unclear from these studies if breast cancer subtype may account in part for the predictive ability of Ki67. Thus, we tested whether Ki67 score in the surgically-resected residual tumor (RT) after NAC predicted outcome in a cohort of triple negative breast cancer (TNBC). Gene expression profiling was performed to identify molecular subtype and test the association with gene modules indicative of signaling pathway activation. Methods: Ki67 was scored in the RT of 89 patients with stage II-III TNBC (ER/PR/HER2 negative by IHC at diagnosis) that had been treated with NAC. Expression levels for 450 genes were quantified by Nanostring. Ki67, node and menopause status, age, therapy type (± taxanes), molecular subtype, and gene expression scores were tested for association to RFS and OS using univariate and multivariate CoxPH models. Results: Ki67 score in the post-NAC RT demonstrated a wide range (1.5-77.7%; median: 36.2%). Twenty seven % of RTs were 3+ HER2 by IHC. Molecular subtype in the RT was as follows: 64% Basal-like; 20% HER2-enriched; 6% LumA; 6% LumB; 4% Normal-like. In univariate analyses to respectively predict RFS and OS, node status (p=0.005 and p=0.02), number of nodes (p=0.003 and p<0.001), and the score of a gene expression module of MEK pathway activation (p=0.04 and p=0.01,) were significant. Basal-like subtype approached significance for RFS (p=0.1) and OS (p=0.05). In multivariate analyses, node status (p=0.002 for RFS and p<0.001 for OS) and MEK pathway activation score (p=0.04 and p=0.01) were significant predictors. Conclusions: Ki67 in the RT after NAC was not predictive of outcome in a cohort of TNBC. However, a gene expression signature of activated MEK was negatively associated with outcome. These data are consistent with the reported preclinical activity of MEK inhibitors against basal-like breast cancer cells and a possible role of this signaling pathway in chemotherapy resistance. They also support deep sequencing studies to identify genetic alterations in the RAS/MEK/MAPK pathway in TNBC.

Combined survival analysis of cardiac patients by a Cox PH model and a Markov chain

The control and treatment of dyslipidemia is a major public health challenge, particularly for patients with coronary heart diseases. In this paper we propose a framework for survival analysis of patients who had a major cardiac event, focusing on assessment of the effect of changing LDL-cholesterol level and statins consumption on survival. This framework includes a Cox PH model and a Markov chain, and combines their results into reinforced conclusions regarding the factors that affect survival time. We prospectively studied 2,277 cardiac patients, and the results show high congruence between the Markov model and the PH model; both evidence that diabetes, history of stroke, peripheral vascular disease and smoking significantly increase hazard rate and reduce survival time. On the other hand, statin consumption is correlated with a lower hazard rate and longer survival time in both models. The role of such a framework in understanding the therapeutic behavior of patients and implementing effective secondary and primary prevention of heart diseases is discussed here.

Long-term results from EORTC24971/TAX323: Comparing TPF to PF in patients with unresectable squamous cell carcinoma of the head and neck (SCCHN).

5530 Background: At a median follow-up (FU) of 32.5 months, the original EORTC24971/TAX323 study showed that compared with standard cisplatin/infusional fluorouracil (PF), induction chemotherapy (ICT) with the addition of docetaxel (TPF) prior to radiotherapy significantly improved progression-free survival (PFS) and overall survival (OS) with less toxicity and an improved quality of life in patients with unresectable SCCHN (NEJM 2007; 357: 1695; BJC 2010; 103: 1173). The present report describes survival data after a median FU of 8.6 years and long-term side effects (feeding tube dependency, tracheostomy, gastrostomy and second malignancies). Methods: Data were collected in a long-term FU form and analyzed in the intent-to-treatment population, using a Cox PH model adjusted for treatment, tumor site, T&N stage and performance status. The two treatment arms were also compared with an unstratified logrank test at a 2-sided 5% significance level. The clinical database was locked on January 10, 2011. Results: FU forms were obtained from 308 (86%) of the 358 randomized patients (156 TPF, 152 PF). PFS remained significantly better with TPF compared with PF (hazard ratio [HR] unadjusted 0.71 (95% CI, 0.57-0.89), p=0.003, medians of 12.7 vs 8.6 months, and 5-years PFS 22.9% vs 13.5%, mainly due to less locoregional progression). A similar picture was observed for OS (HR unadjusted 0.74 (95% CI, 0.59-0.94), p=0.011, medians of 18.8 vs 14.5 months, 5-years OS 27.5 vs 18.6%). These data were confirmed in the Cox PH model. Long-term side effects in TPF/ PF arms were: tracheostomy 13/9 ; feeding tube dependency 6/10; gastrostomy 19/19; second malignancy 14/6 (of which lung 42.9%/50%; H&N 42.9%/33.3%, and GI 14.3%/16.7%). Conclusions: This long-term update supports the conclusions of the final analysis that TPF is superior to PF as ICT for patients with unresectable SCCHN.

Association of baseline CEA, VEGF, and soluble VEGF receptor-2 with treatment outcomes in two randomized phase III trials of cediranib in metastatic colorectal cancer (mCRC).

3590^ Background: Cediranib is an oral VEGFR tyrosine kinase inhibitor. In the HORIZON (HZ) II and III studies in mCRC, cediranib plus standard chemotherapy (CT) showed improvements in PFS vs CT alone (HR 0.84, P=0.012; HZII), but not vs bevacizumab plus CT (HR 1.1, P=0.119; HZIII). To explore potential prognostic/predictive factors, we conducted a retrospective analysis of baseline (BL) levels of CEA, VEGF and sVEGFR2. Methods: In HZII, 860 pts received FOLFOX or XELOX with cediranib 20 mg (n=502) or placebo (n=358). In HZIII, 1,422 pts received mFOLFOX6 with cediranib 20 mg (n=709) or bevacizumab (n=713). Plasma VEGF and sVEGFR2 were measured centrally by ELISA; serum CEA was measured locally. Cut-off values were 98 pg/mL for VEGF, 11,586.6 pg/mL for sVEGFR2 and 50 ng/mL for CEA, based on the median BL values across HZII and III. HRs were derived from a Cox PH model with covariates for performance status, CT received, BL liver function and study phase. The results presented relate to the primary endpoints (HZII: PFS, OS; HZIII: PFS). Results: For each factor, BL data were available for >85% patients in each study. Prognostic factors: High BL VEGF was associated with a worse overall outcome for PFS in HZII (HR 1.41 [1.21, 1.65]) and HZIII (HR 1.20 [1.04, 1.38]), and OS in HZII (HR 1.35 [1.12, 1.63]). BL sVEGFR2 was not prognostic for PFS in HZII (HR 0.99 [0.85, 1.15]) or HZIII (HR 0.98 [0.86, 1.13]), or OS in HZII (HR 0.92 [0.77, 1.10]). High BL CEA was associated with a worse overall outcome for PFS in HZII (HR 1.34 [1.14, 1.57]) and HZIII (HR 1.20 [1.04, 1.37]), and OS in HZII (HR 1.63 [1.36, 1.96]). Predictive factors: There was no strong evidence that BL VEGF, sVEGFR2 or CEA predicted for PFS or OS outcome to cediranib treatment in HZII or HZIII; however, the groups with low BL sVEGFR2 levels were associated with a trend to an improved cediranib PFS effect in HZII and HZIII (HZII low HR 0.81, high HR 0.96; HZIII low HR 1.04, high HR 1.21). Conclusions: In this retrospective subset analysis of two large randomized Phase III trials, BL VEGF and CEA emerged as prognostic biomarkers for both PFS and OS. The prognostic value of CEA and VEGF should be considered when designing studies in mCRC.

Optimal variability sensitive condition-based maintenance with a Cox PH model

Condition based maintenance (CBM) is an important maintenance strategy in practice. In this paper, we propose a CBM method to effectively incorporate system health observations into maintenance decision making to minimise the total maintenance cost and cost variability. In this method, the system degradation process is described by a Cox PH model and the proposed framework includes simulation of failure process and maintenance policy optimisation using adaptive nested partition with sequential selection (ANP-SS) method, which can adaptively select or create the most promising region of candidates to improve the efficiency. Different from existing CBM strategies, the proposed method relaxes some restrictions on the system degradation model and taking the cost variation as one of the optimisation objectives. A real industry case study is used to demonstrate the effectiveness of our framework.

Abstract S5-1: Prospective Outcomes for Patients with Micro-Metastases and Macro-Metastases In Sentinel Nodes: NSABP B-32 Sentinel Node Trial

Abstract Background: Sentinel node biopsy (SNB) allows for a more detailed analysis of axillary nodes, increasing the detection of nodal metastatic disease. Prospective outcome results related to patients with micro-and macro-metastases in sentinel nodes from the NSABP B-32 trial are presented here. Methods: After stratification, women with operable invasive breast cancer and clinically negative nodes were randomly assigned to sentinel node resection (SNR) with immediate conventional axillary dissection (AD) [Group 1] or to SNR without AD [Group 2]. Group 2 patients with positive SNs underwent AD. Sentinel nodes underwent analysis by hematoxylin and eosin. Systemic therapy was at the discretion of the physician. Regional nodal irradiation for node positive lumpectomy patients and chest wall/regional node irradiation for node positive mastectomy patients was permitted. Univariate and multivariate analyses were carried out via Cox PH models to identify predictors of overall survival (OS) and disease-free survival (DFS). Results: From May 1999 through February 2004, 5,611 patients were entered into NSABP B-32. A total 1,390 patients with positive SN were identified; 1,389 of these patients had follow-up data. There were 422 patients with macro-metastases (>2mm), 312 with micro-metastases (>0.2, ≥2.0mm), and 626 with unknown status. Median time in study was 94 months. 97% of SN-positive patients received systemic adjuvant therapy. Significant univariate predictors of OS and DFS were age, receptor status, clinical tumor size, histologic grade, number of positive SNs, SN metastasis size, number of positive nodes, lymphovascular invasion, and systemic therapy. Multivariate analysis identified histologic grade (HR: 2.30, P<0.0001), SN macrometastasis (HR: 2.44, p=0.0003), systemic therapy (HR: 0.22, P<0.0001), age, and clinical tumor size (CTS) as significant predictors of OS (Table 1) with similar predictors for DFS. For DFS, a significant quadratic effect due to age (p=0.015) was observed where younger and older women had higher risk of having an event than did middle aged women. Conclusion: The most striking factors associated with poor OS and DFS were poor histologic tumor grade and macro-versus micro-SN metastases. Patients with micro-or macro-SN metastases who received systemic therapy had a 78% reduction in mortality and a 76% improvement of DFS. Supported by: NCI U10-CA-69651, -12027, -37377, -69974. Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr S5-1.

A Phase III Randomized Trial of Thalidomide (THAL) Plus Zoledronic Acid (ZLD) Versus Zoledronic Acid Alone In Patients with Early Stage Multiple Myeloma (MC0289)

AbstractAbstract 3053 Background:Patients with smoldering multiple myeloma (SMM) have a higher chance of progression to active MM than patients with monoclonal gammopathy of undetermined significance (MGUS). Since active MM remains incurable and since patients with SMM can have a long time to requiring treatment, observation remains an option for these patients. Bisphosphonates can prevent the bone complications of myeloma. The immunomodulatory (IMiD) drugs are well-tolerated and have documented anti-tumor activity in active MM. We hypothesized that treatment with the IMiD THAL and a bisphosphonate (ZLD) would prolong the time to progression (TTP) over the control arm of ZLD alone. This is the first report of this phase III trial of THAL/ZLD vs ZLD alone for patients with untreated SMM. Goals:The primary goal of this trial was to compare the TTP between patients treated with THAL/ZLD versus ZLD alone in asymptomatic MM. Secondary goals were progression rate at one year, response rate, duration of response, time to next therapy, and toxicity. Methods:Patients were required to have measurable disease as defined by either a serum monoclonal protein >1.0 g by protein electrophoresis or nephelometry or >200 mg of monoclonal protein in the urine on 24 hour electrophoresis or a measurable soft tissue plasmacytoma; >10% plasma cells as measured on the bone marrow aspirate, bone marrow biopsy, or labeling index; absolute neutrophil count >1500/μL; platelet count >100,000/μL; creatinine <2.0 mg/dL, and a performance status of 0, 1, or 2. Patients could not have symptomatic MM that required chemotherapy. The statistical plan was to accrue 120 eligible patients (60 per arm) over 4 years and after a minimum follow-up of 12 months the study would provide >90% power at a type I error rate of 0.05 to detect an increase in the median TTP from 12 months (ZLD) to 24 months (THAL/ZLD), and >80% power to detect an increase in median TTP from 12 to 21 months). Results:The study was activated in July 2003 and closed March 2009 due to slow accrual. Sixty-eight patients (35 Thal/ZLD; 33 ZLD) with a median age 63 years (range, 47–84) were randomized. The median TTP for Thal/ZLD versus ZLD was 2.4 years versus 1.2 years, respectively, P=0.02 one-sided log-rank. After adjusting for pre-specified stratification factors, the hazard ratio for TTP was 2.2 (one-sided p value = 0.01, univariate stratified Cox PH model) for ZLD compared to Thal/ZLD. 89% of patients on the Thal/ZLD arm were progression-free survival (PFS) at one year compared to 55% on ZLD alone (one-sided p<0.001, chi-square). Similar results were obtained (Table and Figures) when the CRAB (calcium, renal, anemia, bone) criteria were applied. Confirmed response was evaluated using the first 12 months of treatment. In the Thal/ZLD arm the response rate was 31% with a median duration of response of 5.1 years (95% CI: 1.9 - NA). There were no confirmed responses in the ZLD alone arm. The median overall survival has not been reached for either arm. In regards to toxicity, no grade 5 adverse events (AEs) have been reported. Thirty patients have reported grade 3+ AEs (17 Thal/ZLD; 13 ZLD, Fisher's exact p-value 0.47). Eight patients have reported grade 4 adverse events (5 Thal/ZLD; 3 ZLD, Fisher's exact p-value 0.71). Overall, only one grade 4 event was felt to be at least possibly related to study treatment – grade 4 neutropenia in Thal/ZLD. Conclusions:While the trial did not meet its planned accrual goals, there was a significant improvement in improved in outcomes in the Thal/ZLD arm compared to control (ZLD). Thal/ZLD produces anti-tumor responses that are quite durable with the median response duration of over 5 years. ZLD alone did not produce any anti-tumor responses. This study indicates that a non-chemotherapy approach can be effective in SMM and may be a useful strategy to test in future studies.Table:TTP & PFS Estimates for Protocol and CRAB Progression CriteriaCriteriaThal/ZLD (n=35)ZLD (n=33)p-value (unadjusted, log-rank, one-sided)p-value (adjusted for stratification factors using a univariate Cox model, one-sided)TTP Protocol2.4 years1.2 years0.020.01# of Events2124TTP CRAB4.1 years3.1 years0.030.02# of Events1520PFS Protocol2.4 years1.2 years0.020.005# of Events2325PFS CRAB3.8 years2.3 years0.030.008# of Events1822 [Display omitted] [Display omitted] Disclosures:Off Label Use: Thalidomide for smoldering myeloma. Lacy:Celgene: Research Funding. Dispenzieri:Celgene: Honoraria, Research Funding; Binding Site: Honoraria. Kumar:Celgene: Consultancy, Research Funding; Millennium: Research Funding; Merck: Consultancy, Research Funding; Novartis: Research Funding; Genzyme: Consultancy, Research Funding; Cephalon: Research Funding. Gertz:Celgene: Honoraria; Millennium: Honoraria.

Abstract 1186: PTEN expression predicts biochemical recurrence in prostate cancer

Abstract Prostate cancer is the most commonly diagnosed malignancy in men, but many patients have indolent disease that is unlikely to cause significant morbidity or mortality. Biomarkers that accurately predict disease severity would enable physicians to more appropriately manage patient care. Data from previous studies indicated that the molecular status of PTEN (phosphatase and tensin homolog) might be prognostic in prostate cancer patients after radical prostatectomy (RP). PTEN expression was evaluated on surgical specimens by IHC using anti-PTEN antibody 138G6 (Cell Signaling Technology) in 132 men treated with RP. Approximately half the patients also had neoadjuvant hormonal blockade therapy (BAT). PTEN IHC scores were derived from the weighted average of percent positive cells x intensity (0, 1 or 2). PTEN status (positive or negative) was determined using a predefined IHC score of ≤ 20 for negatives. Using this scoring system, 24% of the tumors were PTEN negative. Association with biochemical recurrence was evaluated using Cox PH models and likelihood ratio tests. Since we were testing a prior hypothesis about the prognostic utility of PTEN, the corresponding p-values are one-sided. In a univariate analysis on the overall series, PTEN status was associated with biochemical recurrence (p-value = 0.0046, HR = 2.3; 95% CI: 1.27, 4.33). Most of the clinical parameters typically associated with recurrence were affected by neoadjuvant therapy and rendered uninterpretable. Nevertheless, pathological tumor stage (p-value = 4.3 × 10−6, HR = 6.0; 95% CI: 2.35, 15.22) was still associated with outcome. In multivariate analyses, PTEN status was a significant predictor of outcome after adjusting for pathological stage (p-value = 0.009), neoadjuvant therapy (p-value = 0.014), or both (p-value = 0.04). These data support the hypothesis that PTEN status predicts outcome after RP, and suggest further, that PTEN may be a clinically important molecular marker for defining aggressive prostate cancer. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 1186.

Abstract A201: Mutation profiling and prognostic value of BRAF, HRAS, KRAS, MET, NRAS, PIK3CA, KRASG12V in colon cancer: Results from NSABP C-07

Abstract More than 20% of patients with resected stage II colon cancer and 50% of patients with resected stage III colon cancer will develop recurrence. To identify markers for recurrence and to find potentially new targets for treatment, mutations were profiled in tumors from NSABP clinical trial C-07. TypePLEX chemistry and the Mass Array system from Sequenom (SanDiego CA) were used to profile 238 common, hot-spot, cancer mutations in 19 genes in 239 colon cancer samples utilizing the OncoCarta panel. Mutations were detected in 7 different genes (ABL1, AKT1, BRAF, KRAS, MET, NRAS and PIK3CA) at 26 different nucleotide positions. Twenty-four assays which detected mutations in more than 1% of the samples were reconfigured into a new multiplexed panel, termed here as ColoCarta, and used to repeat the profiling of 24 mutant samples. These profiles were identical to those found with OncoCarta, demonstrating reproducibility. The method was also sensitive, requiring as little as 1ng of input DNA. In some assays it was possible to detect mutations that represented only 5% of the alleles, an important consideration for heterogeneous tumor tissue. An additional 869 tumors from C-07 were profiled with ColoCarta. Data from a 1035 eligible C07 patients with follow-up were analyzed with Cox PH models, controlling for treatment, number of positive nodes, age and gender. Each gene (BRAF, HRAS, KRAS, MET, NRAS and PIK3CA), a specific mutation KRASG12V and mutations in any gene were examined as dichotomous variables. Only NRAS mutations were significant predictors of recurrence (HR= 1.79, p=0.049) (Table 1). In conclusion, this study presents a rapid, sensitive, reproducible and cost effective method for mutation profiling of both infrequent (ABL1, AKT1, NRAS) and frequent (BRAF, KRAS and PIK3CA) mutations and also identified mutations new to colon cancer in the MET oncogene. MET has been shown to promote many metastatic activities and may provide a new target for colon cancer treatment in a subset of patients. Table 1 C07 Gene Mutation Subset, First Recurrence Variable (range Without Mutation Recur/N (% Recur) With Mutation Recur/N (% Recur) Dichotomous (Any Mutation) P-value Hazard Ratio* 95% CI BRAF (0,1) 237/879 (27.0) 46/153 (30.1) 0.5994 1.091 0.79–1.51 HRAS (0,1) 283/1029 (27.5) 1/6 (16.7) 0.8214 0.797 0.11-5.71 KRAS (0,1,2) 167/632 (26.4) 115/400 (28.8) 0.3482 1.121 0.88–1.42 KRAS G12V (0,1) 253/947 (26.7) 31/88 (35.2) 0.0714 1.410 0.97–2.05 MET (0,1,2,) 273/1002 (27.2) 11/33 (33.3) 0.4017 1.295 0.71–2.37 NRAS (0,1,2) 270/1001 (27.0) 12/29 (41.4) 0.0488 1.792 1.00–3.20 PIK3CA (0, 1, 2) 230/817 (28.2) 54/213 (25.4) 0.9390 1.012 0.75–1.36 Total Number of Mutations (0–3) 92/370 (24.9) 192/665 (28.9) 0.0874 1.246 0.97–1.60 * The hazard ratio compares the hazard in patients with the mutation to patients without the mutation. Citation Information: Mol Cancer Ther 2009;8(12 Suppl):A201.

Assessment of the relative success of sporozoite inoculations in individuals exposed to moderate seasonal transmission

BackgroundThe time necessary for malaria parasite to re-appear in the blood following treatment (re-infection time) is an indirect method for evaluating the immune defences operating against pre-erythrocytic and early erythrocytic malaria stages. Few longitudinal data are available in populations in whom malaria transmission level had also been measured.MethodsOne hundred and ten individuals from the village of Ndiop (Senegal), aged between one and 72 years, were cured of malaria by quinine (25 mg/day oral Quinimax™ in three equal daily doses, for seven days). Thereafter, thick blood films were examined to detect the reappearance of Plasmodium falciparum every week, for 11 weeks after treatment. Malaria transmission was simultaneously measured weekly by night collection of biting mosquitoes.ResultsMalaria transmission was on average 15.3 infective bites per person during the 77 days follow up. The median reappearance time for the whole study population was 46.8 days, whereas individuals would have received an average one infective bite every 5 days. At the end of the follow-up, after 77 days, 103 of the 110 individuals (93.6%; CI 95% [89.0–98.2]) had been re-infected with P. falciparum. The median reappearance time ('re-positivation') was longer in subjects with patent parasitaemia at enrolment than in parasitologically-negative individuals (58 days vs. 45.9; p = 0.03) and in adults > 30 years than in younger subjects (58.6 days vs. 42.7; p = 0.0002). In a multivariate Cox PH model controlling for the sickle cell trait, G6PD deficiency and the type of habitat, the presence of parasitaemia at enrolment and age ≥ 30 years were independently predictive of a reduced risk of re-infection (PH = 0.5 [95% CI: 0.3–0.9] and 0.4; [95% CI: 0.2–0.6] respectively).ConclusionResults indicate the existence of a substantial resistance to sporozoites inoculations, but which was ultimately overcome in almost every individual after 2 1/2 months of natural challenges. Such a study design and the results obtained suggest that, despite a small sample size, this approach can contribute to assess the impact of intervention methods, such as the efficacy vector-control measures or of malaria pre-erythrocytic stages vaccines.

Effect of Venous Thrombotic Events on Overall Survival in Multiple Myeloma: Analysis of Thrombotic Events Occurring in E4A03A Randomized Trial of Lenalidomide Plus High-Dose Dexamethasone (RD) Versus Lenalidomide Plus Low-Dose Dexamethasone (Rd) in Newly Diagnosed Multiple Myeloma, a Trial Coordinated by the Eastern Cooperative Oncology Group (ECOG).

Background: Venous thrombotic events (VTE) are a common complication of therapy with the lenalidomide plus dexamethasone regimen. The incidence of VTE with RD is approximately 20%, and can be lowered with the use of effective thromboprophylaxis, avoidance of erythropoietin, and the use of lower doses of dexamethasone. The goal of this study was to determine the impact of VTE on overall survival of patients with newly diagnosed myeloma by studying events occurring in ECOG E4A03 phase III trial of lenalidomide plus high (standard) dose dex (RD) versus lenalidomide plus low dose dex (Rd) in newly diagnosed myeloma (MM).Methods: Pts with untreated, symptomatic MM were eligible. Pts in the RD arm (Arm A) received lenalidomide 25 mg/day PO days 1–21 every 28 days plus dex 40 mg days 1–4, 9–12, and 17–20 PO every 28 days; pts in the Rd arm (Arm B) received lenalidomide at the same dose plus dex 40 mg days 1, 8, 15, and 22 PO every 28 days. The trial initially did not mandate routine thromboprophylaxis, but recommended that such treatment be considered. After the first 264 patients were enrolled the trial was amended to require mandatory thromboprophylaxis of aspirin for all patients, with a recommendation to use stronger thromboprophylaxis with either warfarin (target INR 2–3) or low molecular weight heparin among patients in the RD arm.Results: 445 pts (median age, 65 yrs) were accrued; 223 randomized to RD, 222 to Rd. Median follow-up time is 30 months. Overall VTE including deep vein thrombosis (DVT) and pulmonary embolism (PE) occurred in 18.5% of patients; 25.6% in Arm A and 11.4% in Arm B. Rates for the first 4 cycles of treatment were 20.2% in Arm A and 8.2% in Arm B, P<0.01. Rates did not change substantially before and after the prophylaxis amendment. A partial response (PR) or higher was seen in 82.1% of pts who experienced VTE compared with 74.6% of pts who did not experience VTE, P=0.19. Overall VGPR rates also were not inferior. Pts who had VTE, however, had significantly higher other grade 3–5 toxicities such as hyperglycemia (14.6% vs 7.5%, P=0.051), cardiac ischemia (4.9% vs 0.8%, P=0.002), non-neuropathic weakness (13.4% vs 6.4%, P=0.039), infection/pneumonia (17.1% vs 11.1%, P=0.138) and fatigue (18.3% vs 10.5%, P=0.060). In a Cox PH model, VTE status as a time-varying covariate was marginally significant: HR 1.54 95%CI (0.96–2.47), P=0.074, suggesting patients that develop VTE have a higher hazard of death.Conclusions: The occurrence of VTE may adversely affect the survival of patients with newly diagnosed myeloma receiving Rev-Dex. VTE was associated with a higher frequency of other serious adverse events. Prevention of VTE events is a priority. Besides lowering the dose of dexamethasone, studies investigating optimum thromboprophylaxis are needed.

Predictive value of serum MALDI-TOF proteomic profiling from patients treated with erlotinib and bevacizumab for survival in patients with non-small cell lung cancer (NSCLC) treated with erlotinib alone

11014 Background: A multicenter phase I/II study found that inhibition of VEGF and EGFR with the combination of bevacizumab and erlotinib had significant antitumor efficacy in NSCLC (Herbst, et al JCO 2005). We undertook a pilot project to determine if Matrix-Assisted Laser Desorption/Ionization- Time of Flight Mass Spectrometry (MALDI-TOF MS) proteomic analysis of pretreatment patient serum could be used to develop a profile predictive of improved time to progression and overall survival after treatment with the erlotinib and bevacizumab. Methods: Pretreatment serum was available for 35 of 40 patients with advanced NSCLC treated with bevacizumab and erlotinib. MALDI-TOF mass spectra were obtained with a Voyager-DE STR instrument (PerSeptive Biosystems) and the raw spectra were calibrated, baseline-corrected, and normalized prior to analysis. A Compound Covariate based Cox PH model (Tukey, CCT 1993) was fit to the normalized intensities at each m/z value to evaluate correlation with OS and PFS while accou...

Rethinking the role of nodal status in TNM staging of node-positive breast cancer

10584 Background: Among TNM staging criteria, the number of positive lymph nodes [PNOD] is the strongest predictor of patient outcome. Based on SEER data, Vinh Hung (2004) asserted that the ratio of the number of positive nodes to the number of nodes removed [NODR] may better predict outcome. We evaluated this assertion in patients from 13 node-positive NSABP phase III trials. Methods: The impact of PNOD, its logarithm (LPNOD) and NODR on survival (S) and recurrence free survival (RFS) were assessed via Cox PH models in 19,768 patients with node-positive breast cancer in 13 NSABP clinical trials. The cohort was split into 2 groups: 9 older trials (Protocols B-05, B-07, B-08, B-09, B-10, B-11, B-12, B-15 and B-16) consisting of 9,915 patients and 4 newer trials (B-22, B-25, B-28 and B-31) consisting of 9,853 patients. Only randomized eligible patients with follow-up were considered. Models were fit to the outcomes for the older trials and validated in the newer trials. Results were adjusted for age, protocol and tumor size. We used Wald z-scores, Nagelkerke R2, and residual analyses to assess the strength of the relationship between nodal status and outcomes. Results: Median time on study was 19.8 years in the older trials and 10.6 years in the newer trials. Ten-year models across the protocols revealed that the relative hazard ratio for each of the outcomes studies was logarithmic according to the number of positive nodes for = 30 positive nodes. Using LPNOD and NODR compared to PNOD improved the strength of association but did not greatly improve goodness of fit ( Table 1 1). For all endpoints, the strength of association and goodness of fit associated with nodal status and other TNM factors attenuated in newer protocols. Conclusions: Nodal status and other TNM factors currently used for breast cancer staging continue to be strong indicators of patient prognosis. However, these factors should be augmented to account for new genetic and biological markers. [Table: see text] No significant financial relationships to disclose.

Randomized phase III trial comparing vincristine, actinomycin, cyclophosphamide (VAC) with VAC/V topotecan/cyclophosphamide (TC) for intermediate risk rhabdomyosarcoma (IRRMS). D9803, COG study

9509 Background: IRRMS was defined as nonmetastatic alveolar (A) RMS or undifferentiated sarcoma (UDS); stage (stg) 2 and 3, group (gr) III embryonal (E) RMS; stg 4 ERMS &lt;10y. IRRMS had a 5 yr failure free survival (FFS) of 64% on IRS IV. Since T alone and with C has significant activity in RMS, we tested whether alternating VAC/VTC improved FFS for IRRMS compared with VAC alone. Methods: Patients (pts) &lt; 50 y with IRRMS with adequate organ function were randomized to 39 wks of VAC (V=1.5 mg/m2, A=0.045 mg/kg, C=2.2 g/m2 q 3 wk) vs VAC alternating with VTC (T=0.75 mg/m2/d ×5, C=250 mg/m2/d ×5 q 3 wk), with additional wkly V; doses adjusted for age &lt;3y; local therapy was after wk 12. Pts with parameningeal RMS/meningeal extension (PME) were nonrandomly treated with VAC and immediate XRT. Primary study endpoint was FFS. The study was designed with 80% power (5% 2-sided alpha level) to detect an increase in 5 yr FFS from 64% to 75%, a relative risk (RR) of 0.64, with VAC/VTC, requiring 518 randomized pts. Results reflect the 3rd of 4 planned analyses (75% information); O'Brien Fleming adjusted alpha level is 0.0166; and Cox PH model was used. Results: 620 eligible pts were entered from 9/99 to 8/05: 266 randomized to VAC, 253 to VAC/VTC, 101 PME patients were nonrandomly treated with VAC. Treatment strata were ERMS stg 2/3, gr III 37%; ERMS grp IV &lt;10 y 6%; ARMS/UDS stg 1 or gr I 16%; ARMS/UDS other 25%; PME 16%. 3% of pts were &lt;1 y, 69% 1–9 y, 28% ≥10 y; 13% of pts were stg 1; 26% stg 2; 54% stg 3; 7% stg 4. 77% pts had gr III tumors. Primary tumor sites were PM in 35% pts, extremity 15%, bladder/prostate 16%, retroperitoneal/perineal 15%, other 19%. Median follow up is 2.4 yrs. FFS at 2 yrs is 77% with VAC and 72% with VAC/VTC. Observed data are inconsistent with hypothesized improvement under VAC/VTC (RR=1.20, 98.34% CI: 0.78–1.85, p=0.3). Similar patients on IRS IV had 2 yr FFS of 77%. 2 yr survival (OS) with VAC: 91%; VAC/VTC: 87% (RR=0.94, CI: 0.55–1.59, p=0.8). There were 12 2nd malignancies and 10 deaths as first events. 4 of 10 deaths were from hepatopathy, which decreased in incidence following dose changes for A and C. Conclusions: In IRRMS, VAC/VTC does not significantly improve FFS nor OS versus VAC, and outcome is similar to IRS IV. No significant financial relationships to disclose.