Tumor mutational burden (TMB) has emerged as a clinically relevant biomarker that may be associated with immune checkpoint inhibitor efficacy. Standardization of TMB measurement is essential for implementing diagnostic tools to guide treatment. Here we describe the in-depth evaluation of bioinformatic TMB analysis by whole exome sequencing (WES) in formalin-fixed, paraffin-embedded samples from a phase III clinical trial. In the CheckMate 026 clinical trial, TMB was retrospectively assessed in 312 patients with non-small-cell lung cancer (58% of the intent-to-treat population) who received first-line nivolumab treatment or standard-of-care chemotherapy. We examined the sensitivity of TMB assessment to bioinformatic filtering methods and assessed concordance between TMB data derived by WES and the FoundationOne® CDx assay. TMB scores comprising synonymous, indel, frameshift, and nonsense mutations (all mutations) were 3.1-fold higher than data including missense mutations only, but values were highly correlated (Spearman’s r = 0.99). Scores from CheckMate 026 samples including missense mutations only were similar to those generated from data in The Cancer Genome Atlas, but those including all mutations were generally higher. Using databases for germline subtraction (instead of matched controls) showed a trend for race-dependent increases in TMB scores. WES and FoundationOne CDx outputs were highly correlated (Spearman’s r = 0.90). Parameter variation can impact TMB calculations, highlighting the need for standardization. Encouragingly, differences between assays could be accounted for by empirical calibration, suggesting that reliable TMB assessment across assays, platforms, and centers is achievable.

447 Background: Programmed death-1/programmed death ligand 1 (PD-1/PD-L1) inhibitors are FDA-approved for urothelial carcinoma (UC) among other tumor types. PD-L1 immunohistochemistry (IHC) testing for UC has evolved with several diagnostic approvals. As of August 2018, the Dako PD-L1 IHC 22C3 pharmDx is a companion diagnostic, while the Dako PD-L1 IHC 28-8 pharmDx and the Ventana PD-L1 (SP263) assays are complementary diagnostics. The Ventana PD-L1 (SP142) assay has complementary status for platinum-treated patients with UC but gained companion status in July 2018 for cisplatin-ineligible patients with PD-L1-expressing UC. Here, we assess PD-L1 testing with the 22C3, 28-8, and SP142 assays on real-world UC samples. Methods: Analyses on 55,652 tumor samples from Symphony Health Solutions were performed between October 2015 and March 2018. PD-L1 results were linked to clinical characteristics using unique identifiers. PD-L1 testing on UC samples was performed at NeoGenomics Laboratories, Inc. based on the manufacturer’s protocols at the time. Results from the Dako assays were reported as % of tumor cells with PD-L1 expression. Test failure (TF) was defined as the absence of adequate sample with measurable PD-L1. Turnaround time (TAT) was defined as the time from sample receipt by the laboratory to test-report delivery. Paired 28-8 and 22C3 testing was performed on 13 samples. Results: 251 confirmed PD-L1 IHC tests (0.4% of tests conducted for the entire dataset) were performed on 223 UC samples, 98.2% of which had quantifiable PD-L1 expression. Mean TAT was 3.0 (1.0–4.2) days and the TF rate was 2.0% for all tests. The 22C3, 28-8, and SP142 assays were used in 52.2%, 9.6%, and 38.3% of tests, respectively. Paired testing with the 22C3 and 28-8 assays was highly correlated (Spearman’s r = 0.94; n = 13). Further analyses of UC PD-L1 prevalence will be presented. Conclusions: Based on the total tests conducted by this US national reference laboratory, PD-L1 testing was not commonly requested for UC during the analysis period. However, PD-L1 tests that were conducted displayed high success rates and reasonable TAT. This is the first comparison of the 22C3 and 28-8 assays on real-world UC samples, adding to the literature on PD-L1 testing concordance.

Anti-tumor treatment outcomes in mouse experiments can be challenging to interpret and communicate accurately. In reporting these experiments, rigorous statistical considerations are commonly absent, although statistical applications have been proposed. We investigated the practicality and utility of different statistical strategies for the analysis of anti-tumor responses in a longitudinal mouse case study. Each analysis that we performed had different endpoints, investigated different questions, and was based on different assumptions. We found rudimentary visual and risk analysis insufficient without additional considerations, and upon further investigation we found improvements in key anti-tumor parameter estimates associated with a drug combination in our case study. We offer practical statistical considerations for investigating anti-cancer treatments in mice, applying a multi-tier statistical approach.

Targeting the programmed death-1/programmed death ligand 1 (PD-1/PD-L1) pathway has improved clinical outcomes, expediting the US FDA approval of 5 agents as treatment for several tumor types. PD-L1 immunohistochemistry (IHC) diagnostic assays have been developed to guide treatment with anti–PD-1/PD-L1 agents. The Dako PD-L1 IHC 22C3 pharmDx and Ventana PD-L1 (SP142) assays are FDA-approved diagnostics for non-small cell lung cancer (NSCLC) as a companion to pembrolizumab and complementary to atezolizumab, respectively. The Dako PD-L1 IHC 28-8 pharmDx is approved for nonsquamous NSCLC, complementary to nivolumab. We sought to overcome barriers to PD-L1 testing by characterizing the use of 22C3, 28-8, and SP142 PD-L1 IHC assays on real-world lung cancer samples.

PURPOSE CheckMate 568 is an open-label phase II trial that evaluated the efficacy and safety of nivolumab plus low-dose ipilimumab as first-line treatment of advanced/metastatic non-small-cell lung cancer (NSCLC). We assessed the association of efficacy with programmed death ligand 1 (PD-L1) expression and tumor mutational burden (TMB). PATIENTS AND METHODS Two hundred eighty-eight patients with previously untreated, recurrent stage IIIB/IV NSCLC received nivolumab 3 mg/kg every 2 weeks plus ipilimumab 1 mg/kg every 6 weeks. The primary end point was objective response rate (ORR) in patients with 1% or more and less than 1% tumor PD-L1 expression. Efficacy on the basis of TMB (FoundationOne CDx assay) was a secondary end point. RESULTS Of treated patients with tumor available for testing, 252 patients (88%) of 288 were evaluable for PD-L1 expression and 98 patients (82%) of 120 for TMB. ORR was 30% overall and 41% and 15% in patients with 1% or greater and less than 1% tumor PD-L1 expression, respectively. ORR increased with higher TMB, plateauing at 10 or more mutations/megabase (mut/Mb). Regardless of PD-L1 expression, ORRs were higher in patients with TMB of 10 or more mut/Mb (n = 48: PD-L1, >= 1%, 48%; PD-L1, < 1%, 47%) versus TMB of fewer than 10 mut/Mb (n = 50: PD-L1, >= 1%, 18%; PD-L1, < 1%, 5%), and progression-free survival was longer in patients with TMB of 10 or more mut/Mb versus TMB of fewer than 10 mut/Mb (median, 7.1 v 2.6 months). Grade 3 to 4 treatment-related adverse events occurred in 29% of patients. CONCLUSION Nivolumab plus low-dose ipilimumab was effective and tolerable as a first-line treatment of advanced/metastatic NSCLC. TMB of 10 or more mut/Mb was associated with improved response and prolonged progression-free survival in both tumor PD-L1 expression 1% or greater and less than 1% subgroups and was thus identified as a potentially relevant cutoff in the assessment of TMB as a biomarker for first-line nivolumab plus ipilimumab. (C) 2019 by American Society of Clinical Oncology

Background: A number of programmed death ligand 1 (PD-L1) immunohistochemistry (IHC) diagnostic (Dx) tests have been approved by the FDA to guide treatment (Tx) with programmed death-1/PD-L1 inhibitors. Here, we evaluate the utilization of the Dako PD-L1 IHC 28-8 and 22C3 pharmDx and Ventana PD-L1 (SP142) assays on real-world samples across multiple tumor types, characterize PD-L1 testing practices among physicians who order a PD-L1 test, and investigate how PD-L1 test results impact physicians’ Tx decision-making and use of immuno-oncology (I-O) Tx.Methods: 55,652 samples with clinical characteristics were provided by Symphony Health Solutions. NeoGenomics Laboratories, Inc assessed PD-L1 expression between October 2015–April 2018, according to manufacturers’ protocols at time of study. Clinical characteristics were matched to PD-L1 test results using unique identifiers for the 4714 patients (pts) whose diagnoses and treatment could be determined.Results: Across tumor types, 56% of pts received a test prior to first-line (1L) Tx. Most lung cancer (57%) or melanoma (MEL; 65%) pts had a PD-L1 test prior to 1L Tx, while most squamous cell carcinoma of the head and neck (70%) or urothelial carcinoma (73%) pts had a PD-L1 test after Tx initiation, in line with drug and complementary Dx approvals for these tumor types during the testing period. The percentage of lung cancer pts whose PD-L1 expression was tested prior to 1L Tx rose from 32% during Q4 2015–Q3 2016 to 63% during Q4 2016–Q1 2018, in line with FDA approval of a companion PD-L1 IHC Dx for non-small cell lung cancer. Regardless of PD-L1 test used, most lung cancer (73%) or MEL (93%) pts received I-O Tx, defined as nivolumab (NIVO; + ipilimumab [IPI] for MEL only), pembrolizumab (pembro) ± chemotherapy, atezolizumab, or IPI alone. Moreover, the majority of lung cancer pts who had ≥1% PD-L1 expression or MEL regardless of PD-L1 expression received 1L I-O Tx irrespective of the PD-L1 test used. Analysis of PD-L1 expression, Dx test used, and Tx received showed that, for MEL pts tested with 22C3, 28%, 17%, and 0% of pts received NIVO or IPI, while 28%, 43%, and 43% received NIVO+IPI, at the expression cutoffs of 0%, 1–49%, and ≥50%, respectively. Lung cancer pts with 1–49% PD-L1 expression, as determined by a 22C3 test, received NIVO or pembro monotherapy at similar rates (12–18%). Of those who had ≥50% PD-L1 expression, as determined by the 28-8 test only, 52% received pembro monotherapy vs other Tx.Conclusions: Since FDA approval of the first PD-L1 IHC Dx in 2015, analyses of PD-L1 testing practices and physicians’ Tx decision behavior at a US national reference laboratory have shown that physicians’ adoption of PD-L1 testing is responsive to FDA approvals and that most testing for lung cancer and MEL occurs before 1L Tx initiation. In lung cancer or MEL pts who had a PD-L1 test result available, Tx decisions do not appear to be tied to the specific intended use of the PD-L1 assay that was ordered.Citation Format: Gabriel S. Krigsfeld, Emily Prince, Kim Zerba, Vladislav Chizhevsky, Josette William Ragheb, James White. Real-world utilization of PD-L1 IHC testing and results across multiple tumor types [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 3993A.

151 Background: A number of programmed death ligand 1 (PD-L1) IHC diagnostic tests have been approved by the FDA to guide treatment with programmed death-1/PD-L1 inhibitors. We evaluated PD-L1 assay utilization and concordance across all tumor types, including melanoma (MEL) and squamous cell carcinoma of the head and neck (SCCHN), using real-world cancer samples tested at a single reference laboratory in the United States. Methods: 55,652 samples with clinical characteristics were provided by Symphony Health Solutions. NeoGenomics Laboratories, Inc assessed PD-L1 expression using the Dako PD-L1 IHC 28-8 or 22C3 pharmDx, or Ventana PD-L1 SP142 assay tests between October 2015 and April 2018, according to the manufacturers’ protocols at the time of the study. Clinical characteristics were matched to PD-L1 test results using unique identifiers. Data were analyzed by BioStat Solutions, Inc. Results: 61,568 tests from 55,652 patients were included. Across all 61,568 tests, 88.8% were carried out with 22C3, 9.8% with 28-8, and 1.4% with SP142. The 28-8 and 22C3 pharmDx assays showed high analytical concordance in all 3113 matched samples with PD-L1 expression data for both assays (OPA = 96.2%, Spearman’s r = 0.96; Table). The failure rate across all tests was 3.5%, and 96.6% of patients had a quantifiable test result. Average test turnaround time was 3.1 days, with a 9.1-fold increase in total number of tests performed between Q4 2015 and Q1 2018 (from 1163 to 10,544). Subgroup analyses for 678 patients (1.2%) with MEL and 270 patients (0.5%) with SCCHN will be presented. Conclusions: Despite an increased demand for PD-L1 testing in the US, most patients received a result in 2–4 days. These real-world analytical data support the potential interchangeability of the Dako PD-L1 IHC 28-8 and 22C3 pharmDx assays in clinical practice for assessing tumor cell membrane PD-L1 expression across tumor types, and provide context on the evolution and adoption of PD-L1 testing for patients with cancer. [Table: see text]

Immune checkpoint inhibitors targeting the programmed death-1/programmed death ligand 1 (PD-1/PD-L1) pathway have improved clinical outcomes in patients with cancer. Testing for PD-L1 with immunohistochemistry (IHC) is performed to inform therapy decisions. Among the FDA-approved PD-L1 IHC diagnostic assays, the Dako PD-L1 IHC 22C3 pharmDx assay is approved as a companion diagnostic to pembrolizumab in non-small cell lung cancer (NSCLC) and the Dako PD-L1 IHC 28-8 pharmDx assay as a complementary diagnostic to nivolumab for nonsquamous NSCLC. Previous studies have compared the 2 assays in NSCLC and urothelial carcinoma. Here, we present a comparison of the 28-8 and 22C3 assays on the largest real-world dataset of lung cancer samples reported to date.

2593 Background: A multiparameter tumor inflammation assay based on gene expression profiling (TIA-GEP) can extend the utility of IHC to interrogate the tumor microenvironment (TME). Using CD8 expression assessed by IHC (CD8-IHC) as a surrogate for inflammation, statistical modelling was used to develop a specific gene signature on the TIA-GEP panel to predict CD8-IHC. The correlation between TIA-GEP and CD8-IHC and the prevalence of inflammation were explored across multiple tumor types. Methods: Levels of inflammation were measured by CD8-IHC and TIA-GEP on 1778 procured samples across 12 tumor types. Quality control metrics involved sample input quality, technical errors, and inter-run variability. Generalized linear models were used to identify an inflammation score that predicts the CD8-IHC score in melanoma and SCCHN tissue. The predictive accuracy of this signature was also examined in 10 additional tumor types. Results: Assessment of TME inflammation by CD8-IHC was consistent with that observed by TIA-GEP in multiple tumor types. The range of inflammation varied across different tumor types, with relatively lower inflammation range and scores in SCLC, ovarian, and prostate cancers, and higher values in NSCLC, melanoma, SCCHN, and gastric cancers. R2 x 100 values reflecting percent variation in CD8-IHC associated with TIA-GEP ranged from 62.4% to 79.2% ( P < 0.0001) for all tumor types except prostate cancer (32.5%). Low correlation in prostate cancer may be a result of low prevalence of inflammation by CD8-IHC. Estimated linear regression slopes between CD8-IHC and TIA-GEP ranged from 0.74 in SCLC to 1.27 in gastric cancer. Conclusions: The results suggest that the inflammation signature is a robust potential diagnostic tool predicting inflammation in the TME. The inflammation signature not only correlates with CD8-IHC for multiple tumor types, but also leverages the alternative benefits associated with TIA-GEP, which include information related to tumor inflammation-associated biomarkers and flexibility in exploring the value of other genomic signatures.

Abstract Background: Several studies have demonstrated the potential clinical utility of TMB as a biomarker of immune checkpoint blockade in multiple tumor types, including lung cancer. Prior analyses have been retrospective and exploratory, with varying cutoffs and methodologies used to assess TMB. The establishment of a clinically validated cutoff for nivo + ipi is required to confirm the clinical utility of TMB as a predictive biomarker in future studies. CheckMate 568 (NCT02659059) is a large, single-arm, phase 2 study of nivo + ipi in 1L NSCLC. TMB was assessed using Foundation One CDx™ (Foundation Medicine, Inc.) to identify an appropriate TMB cutoff to select patients for 1L nivo + ipi therapy. This cutoff was subsequently validated through a preplanned analysis in CheckMate 227, which evaluated progression-free survival in patients with high TMB as a co-primary endpoint (NCT02477826). Methods: In CheckMate 568, 288 patients with chemotherapy-naive stage IV NSCLC received nivo 3 mg/kg Q2W + ipi 1 mg/kg Q6W for up to 2 years. EGFR- and ALK-targetable NSCLC were excluded. The primary endpoint was objective response rate (ORR) as per independent review. ORR by TMB was a secondary endpoint. TMB was assessed using the validated FoundationOne® CDx assay (Foundation Medicine, Inc.). TMB classification performance with receiver operating characteristic (ROC) curves was used to determine an appropriate TMB cutoff associated with enhanced efficacy of nivo + ipi in 1L NSCLC. Prevalence analyses of TMB and PD-L1 were also conducted. Results: Baseline characteristics and efficacy were similar in both the all treated and TMB-evaluable populations. With a minimum follow-up of 3 months, ORR was 27% in all treated patients. ORR increased in patients with higher TMB, and plateaued with the threshold of ≥10 mutations (mut)/Mb (ORR: 4%, 10%, 44%, and 39% in patients with TMB <5, <10, ≥10, and ≥15 mut/Mb, respectively). ROC analysis of TMB vs ORR for nivo + ipi demonstrated optimal classification performance at 10 mut/Mb. The safety profile for nivo + ipi was consistent with previous reports and no new safety signals were observed. Conclusions: TMB ≥10 mut/Mb was associated with enhanced response to nivo + ipi regardless of PD-L1 expression, with ORRs >40%. This response rate, coupled with the known durability of responses to immuno-oncology agents, compares favorably to historical data with platinum-doublet chemotherapy. Therefore, a cutoff of ≥10 mut/Mb was chosen to define the TMB patient population for the co-primary efficacy endpoint in CheckMate 227. Citation Format: Suresh S. Ramalingam, Matthew D. Hellmann, Mark M. Awad, Hossein Borghaei, Justin Gainor, Julie Brahmer, David R. Spigel, Martin Reck, Kenneth J. O'Byrne, Luis Paz-Ares, Kim Zerba, Xuemei Li, William J. Geese, George Green, Brian Lestini, Joseph D. Szustakowski, Han Chang, Neal Ready. Tumor mutational burden (TMB) as a biomarker for clinical benefit from dual immune checkpoint blockade with nivolumab (nivo) + ipilimumab (ipi) in first-line (1L) non-small cell lung cancer (NSCLC): identification of TMB cutoff from CheckMate 568 [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr CT078.

Aims At the time of analysis, two widely used, drug-specific, tumour-cell programmed death ligand 1 (PD-L1) assays were approved by the US Food and Drug Administration for anti-PD-1 therapies: the Dako PD-L1 immunohistochemistry (IHC) 28-8 pharmDx assay and the Dako PD-L1 IHC 22C3 pharmDx assay. Given that the majority of current PD-L1 testing in US clinical practice is performed at commercial reference laboratories, we aimed to evaluate the concordance of the 28-8 and 22C3 assays in a real-world setting. Methods Matched PD-L1 IHC 28-8 and 22C3 results from routine assessment were obtained from 1930 patients, including 412 confirmed to have lung cancer, submitted from hospitals in over 38 US states/territories. Biopsies were stained, reviewed and scored by trained/certified pathologists at a single cancer reference laboratory between 2015 and 2017. Rate of concordance between assay findings was assessed by Bland-Altman analysis; overall per cent agreement (OPA), positive per cent agreement and negative per cent agreement; and Cohen's kappa. Results PD-L1 IHC 28-8 and 22C3 displayed strong correlation across all samples and in samples with a confirmed lung cancer diagnosis irrespective of biopsy site. The OPA was 97%-98% for all samples, depending on the expression level defining PD-L1 positivity. In the Bland-Altman analysis, the mean difference in percentage of tumour cells positively stained for PD-L1 between the paired assay findings was -0.80% for all samples and -0.93% in samples with a confirmed lung cancer diagnosis. Conclusions These data, in conjunction with recent findings, support the analytical concordance of the PD-L1 IHC 28-8 and 22C3 assays for assessing per cent tumour-cell membrane PD-L1 expression.

Background: Tumor mutation burden (TMB) is emerging as a potential predictive biomarker for the efficacy of immune checkpoint inhibitors. TMB has traditionally been evaluated using whole exome sequencing (WES), an established method for analysis of genomic alterations that requires both tumor and germline DNA. In the phase 3 CheckMate 026 study (NCT02041533) of nivolumab vs chemotherapy in first-line non-small cell lung cancer (NSCLC), exploratory analyses were performed to determine the value of TMB as a predictive biomarker in 312 tumor samples using WES. Patients with baseline TMB in the upper tertile were associated with an increased objective response rate to nivolumab and longer median PFS, relative to patients with TMB in the lowest 2 tertiles and compared with chemotherapy-treated patients in the upper tertile (Carbone et al. N Engl J Med. 2017). Increasingly, targeted cancer gene panels, including the FoundationOne ® (F1) assay, are being assessed as an alternative to WES for clinical use. To determine the feasibility of transitioning from WES to a potential in vitro diagnostic next-generation sequencing (NGS) platform for TMB analysis, we performed a study comparing values generated by WES with those generated by the F1 assay. Methods: Using the 2 hybridization-capture/NGS methods (WES and F1), TMB was assessed in formalin-fixed, paraffin-embedded NSCLC tumor samples. For WES, coding regions of 21,522 genes were analyzed, with TMB defined as the total number of missense mutations in the tumor exome. The F1 assay is a targeted gene panel of 315 cancer-related genes (Frampton et al. Nat Biotechnol. 2013), with TMB defined as the number of somatic mutations per megabase of sequenced tumor genome. Results: TMB data from 44 NSCLC samples were available on both F1 and WES platforms. We determined an empirical method to convert TMB values derived from WES into those derived from F1 to establish calibration between the 2 methodologies. TMB assessed by WES and by F1 was highly correlated (Spearman9s r=0.9). As an additional test of concordance between the platforms, we examined agreement between the assays around the median WES value of 148 mutations. This value projects to 7.64 mutations per megabase on F1. The overall agreement between F1 and WES around these values was 86% (95% Wilson CI, 73-94). Positive and negative agreements between the 2 hybridization-capture/NGS methods were both 86% (95% Wilson CI, 67-95). Conclusion: Bridging TMB analysis by WES to F1 facilitates the transition of WES-derived biomarker data to the clinical in vitro diagnostic F1. This study demonstrates the feasibility of harmonization of TMB results across testing platforms, which provides alternative TMB testing options. Citation Format: Joseph D. Szustakowski, George Green, William J. Geese, Kim Zerba, Han Chang. Evaluation of tumor mutation burden as a biomarker for immune checkpoint inhibitor efficacy: A calibration study of whole exome sequencing with FoundationOne ® [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 5528.

PURPOSE:This phase II study evaluated the efficacy and safety of ixabepilone as neoadjuvant therapy for invasive breast cancer not amenable to breast conservation surgery. Gene expression studies were undertaken using genes that were identified as potentially associated with sensitivity/resistance to ixabepilone in prior preclinical investigations.PATIENTS AND METHODS:Patients with invasive breast cancer >or= 3 cm were eligible. Ixabepilone 40 mg/m(2) was administered as a 3-hour intravenous infusion on day 1 of a 21-day cycle for four or fewer cycles.RESULTS:One hundred sixty-one patients were treated. The overall complete pathologic response (pCR) rate was 18% in breast and 29% in estrogen receptor (ER) -negative patients. Gene expression data were available for 134 patients. ER gene expression (ER1) was inversely related to pCR in breast and had a positive predictive value (PPV) of 37% and negative predictive value (NPV) of 92%. A 10-gene penalized logistic regression (PLR) model developed from 200 genes predictive of ixabepilone sensitivity in preclinical experiments included ER and tau and had higher PPV (45%) and comparable NPV (89%) to ER1. Grade 3 to 4 adverse events (AEs) were reported for 32% of patients. Except for neutropenia and leukopenia, all grade 3 to 4 AEs occurred in <or= 3% of patients. Reversible peripheral neuropathy was experienced by 3% of patients. CONCLUSION ER, microtubule-associated protein tau, and a 10-gene PLR model that included ER were identified as predictors of ixabepilone-induced pCR.RESULTS:indicate an inverse relation between ER expression levels and ixabepilone sensitivity. Neoadjuvant ixabepilone demonstrated promising activity and a manageable safety profile in patients with invasive breast tumors.

2525 Background: Ixabepilone (BMS-247550) is a microtubule stabilizing agent with demonstrable therapeutic value in taxane- refractory breast cancer (BC) patients. Biomarkers to predict either ixabepilone or paclitaxel activity in BC patients have previously been reported. However, markers that differentiate response to the two agents have yet to be identified. This study sought to discover predictive markers that will enable patient selection to differentially enhance response to ixabepilone or paclitaxel in ER-negative (ER-) patients. Materials and Methodologies: Pre-treatment gene expression profiles were generated for 62 ER- patients treated with ixabepilone in clinical study CA163080, and 51 ER- patients treated with T/FAC (paclitaxel and fluorouracil-doxorubicin-cyclophosphamide) in clinical study MDA133. Biomarkers differentially predictive of complete pathological response in breast were identified through gene set enrichment analysis (GSEA) or classification by threshold gradient descent (TGD). Gene knockdown by siRNA was used to study some of these candidate markers. Results: Four candidate models that differentiate response to ixabepilone treatment and taxane-containing therapy were identified. Two of the models, found by GSEA, are based on expression levels for single microtubule-related genes: transforming, acidic coiled-coil containing protein 3 (TACC3) and chromosome condensation protein G (HCAP-G). The potential of HCAP-G as a differential marker was supported by siRNA studies. Two of the models, found by TGD, are based on expression levels for 26 and 20 genes. Areas under the ROC curves for the models applied to each study separately are given in the table . Conclusions: We have identified four predictive models that differentiate response in a clinical trial of ixabepilone from that in a trial of T/FAC. A clinical trial is under way to further evaluate their utility for differentiating response to ixabepilone- and taxane-containing regimens. [Table: see text] [Table: see text]

586 Background: Previous studies have confirmed the activity of ixabepilone (BMS-247550, a semi-synthetic analog of epothilone) in metastatic BC. The present study was designed to determine the pathological response obtained with ixabepilone as neoadjuvant therapy for BC and to obtain tumor samples for analysis of gene expression and identification of potential predictors of response to ixabepilone. Methods: Women with invasive stage IIA-IIIB breast cancer with tumors ≥3 cm diameter received 40 mg/m2 ixabepilone as a 3-hour infusion on Day 1 for up to four 21-day cycles, followed by surgery within 3–4 weeks of completion of chemotherapy. Biopsies for analysis of mRNA expression were obtained both pre- and post-therapy. Adjuvant chemotherapy with an anthracycline combination regimen followed by radiotherapy and tamoxifen were administered as indicated. Pathological response was assessed using the Sataloff criteria. Results: A total of 164 patients were enrolled. Data were available for 47 patients for this preliminary report. Median age was 57 years (range 27–77) and 100% were ECOG 0. Median tumor diameter was 4 cm (range 3–10), Grade 1/2/3/unk was 6%/47%/23%/23% and ER status of +/-/unk was 49%/47%/4%. A complete pathological response (pCR) in the breast was achieved in 10 patients (21%), of whom 8 (17%) also had pCR in the axillary lymph nodes. Grade 3/4 neutropenia occurred in 21%/13%. Common grade 2/3 toxicities on-treatment included arthralgia/myalgia (28%/6%), neuropathy (11%/2%), and mucositis (6%/2%). Conclusions: Based on this early analysis, ixabepilone has manageable toxicity and a pCR rate which compares favorably to that reported for studies of single agent taxanes (ranging from 3–20%). This analysis will be updated at the ASCO annual meeting. Author Disclosure Employment or Leadership Consultant or Advisory Role Stock Ownership Honoraria Research Funding Expert Testimony Other Remuneration Bristol-Myers Squibb Bristol-Myers Squibb Bristol-Myers Squibb Bristol-Myers Squibb Bristol-Myers Squibb