Concordance between an FDA-approved companion diagnostic and an alternative assay kit for assessing homologous recombination deficiency in ovarian cancer.

OBJECTIVE: Authors evaluated the performance of a commercially available next-generation sequencing assay kit; this was based on genomic content from Illumina's TruSight™ Oncology 500 research assay that identifies BRCA variants and proprietary algorithms licensed from Myriad and, with additional genomic content, measures the homologous recombination deficiency (HRD) genomic instability score (GIS) in tumor tissue (TSO 500 HRD assay). METHODS: Data from the TSO 500 HRD assay were compared with data from the Myriad MyChoice®CDx PLUS assay (Myriad assay). Prevalence rates for overall HRD status and BRCA mutations (a deleterious or suspected deleterious BRCA1 or BRCA2 mutation or both) and assay agreement rates for HRD GIS and BRCA analysis were assessed in ovarian tumor samples. Pearson correlations of the continuous HRD GIS and analytic sensitivity and specificity were evaluated. RESULTS: The prevalence of overall HRD positivity was 51.2% (TSO 500 HRD assay) versus 49.2% (Myriad assay) and the prevalence of BRCA mutations was 27.6% (TSO 500 HRD assay) versus 25.5% (Myriad assay). After post-processing optimization, concordance of the HRD GIS was 0.980 in all samples and 0.976 in the non-BRCA mutation cohort; the area under the receiver operating characteristic curve was 0.995 and 0.992, respectively. CONCLUSIONS: Comparison between the Illumina and Myriad assays showed that overall HRD status, the individual components of BRCA analysis, and HRD GIS detection results were highly concordant (>93%), suggesting the TSO 500 HRD assay will approach the analytical accuracy of the FDA-approved Myriad assay.

Concordance between single-nucleotide polymorphism-based genomic instability assays and a next-generation sequencing-based homologous recombination deficiency test.

BACKGROUND: We evaluated the performance of single-nucleotide polymorphism (SNP) genotyping arrays OncoScan (Thermo Fisher Scientific, San Diego, CA) and Infinium CytoSNP-850K (CytoSNP; Illumina, Waltham, MA) for assessing homologous recombination deficiency (HRD) genomic instability. METHODS: DNA (pretreatment samples) across 20 tumor types was evaluated with OncoScan, CytoSNP, and the clinically validated HRD test. Copy number variation (CNV) and loss of heterozygosity (LOH) analyses were performed with ASCATv2.5.1. Aggregate HRD genomic metrics included LOH, telomeric-allelic imbalance number (TAI), and large-scale state transition (LST). Associations between BRCA mutation (BRCAm) status and the clinically validated HRD test metric (dichotomized at a clinical cut-off) were evaluated using area under the receiver operating characteristic (AUROC); Spearman ρ was calculated for continuous metrics. CNV segmentation and HRD genomic metrics were calculated (n = 120, n = 106, and n = 126 for OncoScan, CytoSNP and clinically validated HRD test, respectively). RESULTS: When assessed by SNP arrays, the genomic metric demonstrated good association with BRCAm (AUROC of HRD: OncoScan, 0.87; CytoSNP, 0.75) and the clinically validated test (cut-off, 42; AUROC of HRD: OncoScan, 0.92; CytoSNP, 0.91). The genomic metrics demonstrated good correlation with the clinically validated aggregate HRD test metric (ρ: OncoScan, 0.82; CytoSNP, 0.81) and for each component (ρ: OncoScan, 0.68 [LOH], 0.76 [TAI], and 0.78 [LST]; CytoSNP, 0.59 [LOH], 0.77 [TAI], and 0.82 [LST]). HRD assessed by SNP genotyping arrays and the clinically validated test showed good correlation. CONCLUSION: OncoScan and CytoSNP may potentially identify most HRD-positive tumors with appropriate clinically relevant cut-offs.

Homologous Recombination Deficiency: Concepts, Definitions, and Assays.

BACKGROUND: Homologous recombination deficiency (HRD) is a phenotype that is characterized by the inability of a cell to effectively repair DNA double-strand breaks using the homologous recombination repair (HRR) pathway. Loss-of-function genes involved in this pathway can sensitize tumors to poly(adenosine diphosphate [ADP]-ribose) polymerase (PARP) inhibitors and platinum-based chemotherapy, which target the destruction of cancer cells by working in concert with HRD through synthetic lethality. However, to identify patients with these tumors, it is vital to understand how to best measure homologous repair (HR) status and to characterize the level of alignment in these measurements across different diagnostic platforms. A key current challenge is that there is no standardized method to define, measure, and report HR status using diagnostics in the clinical setting. METHODS: Friends of Cancer Research convened a consortium of project partners from key healthcare sectors to address concerns about the lack of consistency in the way HRD is defined and methods for measuring HR status. RESULTS: This publication provides findings from the group's discussions that identified opportunities to align the definition of HRD and the parameters that contribute to the determination of HR status. The consortium proposed recommendations and best practices to benefit the broader cancer community. CONCLUSION: Overall, this publication provides additional perspectives for scientist, physician, laboratory, and patient communities to contextualize the definition of HRD and various platforms that are used to measure HRD in tumors.

Prognostic significance of T-cell-inflamed gene expression profile and PD-L1 expression in patients with esophageal cancer.

PURPOSE: The ability of the T-cell-inflamed gene expression profile (GEP) to predict clinical outcome in esophageal cancer (EC) is unknown. This retrospective observational study assessed the prognostic value of GEP and programmed death ligand 1 (PD-L1) expression in patients with EC treated in routine clinical practice. METHODS: Tumor samples of 294 patients from three centers in Denmark, South Korea, and the United States, collected between 2005 and 2017, were included. T-cell-inflamed GEP score was defined as non-low or low using a cutoff of -1.54. A combined positive score (CPS) ≥10 was defined as PD-L1 expression positivity. Associations between overall survival (OS) and GEP status and PD-L1 expression were explored by Cox proportional hazards models adjusting for age, sex, histology, stage, and performance status. RESULTS: Median age was 65 years; 63% of patients had adenocarcinoma (AC) and 37% had squamous cell carcinoma (SCC). Thirty-six percent of tumors were GEP non-low, with higher prevalence in AC (46%) than SCC (18%). Twenty-one percent were PD-L1-positive: 32% in South Korean samples versus 16% in non-Asian samples and 26% in SCC versus 18% in AC. GEP scores and PD-L1 CPS were weakly correlated (Spearman's R = 0.363). OS was not significantly associated with GEP status (non-low vs low; adjusted hazard ratio, 0.91 [95% CI, 0.69-1.19]) or PD-L1 expression status. CONCLUSION: Neither GEP nor PD-L1 expression was a prognostic marker in Asian and non-Asian patients with EC.

Association of programmed death ligand 1 expression with prognosis among patients with ten uncommon advanced cancers.

AIM: PD-L1 expression and high levels of microsatellite instability (MSI-H) may predict response to checkpoint inhibitors, but their prevalence and prognostic value are unknown in many cancers. METHODS: We retrospectively evaluated PD-L1 combined positive score (CPS) and MSI-H and their association with clinical outcomes among patients with ten advanced uncommon cancers. RESULTS: 398 of 426 patients (93%) had a valid PD-L1 result; most (242; 61%) had CPS ≥1. Prevalence of MSI-H tumors was 8/360. Median overall survival was shorter among patients with PD-L1 CPS ≥1 tumors after first-line treatment (23.0 vs 39.7 months, p = 0.014). CONCLUSION: PD-L1 was commonly expressed in solid tumors, and CPS ≥1 was associated with shorter overall survival. Prevalence of MSI-H was low.

Measuring Tumor Mutational Burden (TMB) in Plasma from mCRPC Patients Using Two Commercial NGS Assays.

Tumor tissue mutational burden (TMB) has emerged as a promising predictive biomarker for immune checkpoint therapy. Measuring TMB from circulating tumor DNA (ctDNA) found in plasma is attractive in tissue-constrained indications. We compared the performance of two plasma-based commercial TMB assays including the effect of two different collection methods. Our findings suggest that the two plasma based TMB assays are highly correlated and they are also both correlated with a tissue-based TMB assay for relatively high TMB samples. The two collection methods are also found to be very comparable. Plasma-based TMB assays may be mature enough to be clinically useful in mCRPC and potentially other indications.

Pre-Analytical Considerations for Successful Next-Generation Sequencing (NGS): Challenges and Opportunities for Formalin-Fixed and Paraffin-Embedded Tumor Tissue (FFPE) Samples.

In cancer drug discovery, it is important to investigate the genetic determinants of response or resistance to cancer therapy as well as factors that contribute to adverse events in the course of clinical trials. Despite the emergence of new technologies and the ability to measure more diverse analytes (e.g., circulating tumor cell (CTC), circulating tumor DNA (ctDNA), etc.), tumor tissue is still the most common and reliable source for biomarker investigation. Because of its worldwide use and ability to preserve samples for many decades at ambient temperature, formalin-fixed, paraffin-embedded tumor tissue (FFPE) is likely to be the preferred choice for tissue preservation in clinical practice for the foreseeable future. Multiple analyses are routinely performed on the same FFPE samples (such as Immunohistochemistry (IHC), in situ hybridization, RNAseq, DNAseq, TILseq, Methyl-Seq, etc.). Thus, specimen prioritization and optimization of the isolation of analytes is critical to ensure successful completion of each assay. FFPE is notorious for producing suboptimal DNA quality and low DNA yield. However, commercial vendors tend to request higher DNA sample mass than what is actually required for downstream assays, which restricts the breadth of biomarker work that can be performed. We evaluated multiple genomics service laboratories to assess the current state of NGS pre-analytical processing of FFPE. Significant differences in pre-analytical capabilities were observed. Key aspects are highlighted and recommendations are made to improve the current practice in translational research.

Data Interoperability of Whole Exome Sequencing (WES) Based Mutational Burden Estimates from Different Laboratories.

Immune checkpoint inhibitors, which unleash a patient's own T cells to kill tumors, are revolutionizing cancer treatment. Several independent studies suggest that higher non-synonymous mutational burden assessed by whole exome sequencing (WES) in tumors is associated with improved objective response, durable clinical benefit, and progression-free survival in immune checkpoint inhibitors treatment. Next-generation sequencing (NGS) is a promising technology being used in the clinic to direct patient treatment. Cancer genome WES poses a unique challenge due to tumor heterogeneity and sequencing artifacts introduced by formalin-fixed, paraffin-embedded (FFPE) tissue. In order to evaluate the data interoperability of WES data from different sources to survey tumor mutational landscape, we compared WES data of several tumor/normal matched samples from five commercial vendors. A large data discrepancy was observed from vendors' self-reported data. Independent data analysis from vendors' raw NGS data shows that whole exome sequencing data from qualified vendors can be combined and analyzed uniformly to derive comparable quantitative estimates of tumor mutational burden.

Analytical Validation of AmpliChip p53 Research Test for Archival Human Ovarian FFPE Sections.

The p53 tumor suppressor gene (TP53) is reported to be mutated in nearly half of all tumors and plays a central role in genome integrity. Detection of mutations in p53 can be accomplished by many assays, including the AmpliChip p53 Research Test. The AmpliChip p53 Research Test has been successfully used to determine p53 status in hematologic malignancies and fresh frozen solid tissues but there are few reports of using the assay with formalin fixed, paraffin-embedded (FFPE) tissue. The objective of this study was to describe analytical performance characterization of the AmpliChip p53 Research Test to detect p53 mutations in genomic DNA isolated from archival FFPE human ovarian tumor tissues. Method correlation with sequencing showed 96% mutation-wise agreement and 99% chip-wise agreement. We furthermore observed 100% agreement (113/113) of the most prevalent TP53 mutations. Workflow reproducibility was 96.8% across 8 samples, with 2 operators, 2 reagent lots and 2 instruments. Section-to-section reproducibility was 100% for each sample across a 60 µm region of the FFPE block from ovarian tumors. These data indicate that the AmpliChip p53 Research Test is an accurate and reproducible method for detecting mutations in TP53 from archival FFPE human ovarian specimens.

HCV genotyping from NGS short reads and its application in genotype detection from HCV mixed infected plasma.

Genotyping of hepatitis C virus (HCV) plays an important role in the treatment of HCV. As new genotype-specific treatment options become available, it has become increasingly important to have accurate HCV genotype and subtype information to ensure that the most appropriate treatment regimen is selected. Most current genotyping methods are unable to detect mixed genotypes from two or more HCV infections. Next generation sequencing (NGS) allows for rapid and low cost mass sequencing of viral genomes and provides an opportunity to probe the viral population from a single host. In this paper, the possibility of using short NGS reads for direct HCV genotyping without genome assembly was evaluated. We surveyed the publicly-available genetic content of three HCV drug target regions (NS3, NS5A, NS5B) in terms of whether these genes contained genotype-specific regions that could predict genotype. Six genotypes and 38 subtypes were included in this study. An automated phylogenetic analysis based HCV genotyping method was implemented and used to assess different HCV target gene regions. Candidate regions of 250-bp each were found for all three genes that have enough genetic information to predict HCV genotypes/subtypes. Validation using public datasets shows 100% genotyping accuracy. To test whether these 250-bp regions were sufficient to identify mixed genotypes, we developed a random primer-based method to sequence HCV plasma samples containing mixtures of two HCV genotypes in different ratios. We were able to determine the genotypes without ambiguity and to quantify the ratio of the abundances of the mixed genotypes in the samples. These data provide a proof-of-concept that this random primed, NGS-based short-read genotyping approach does not need prior information about the viral population and is capable of detecting mixed viral infection.

Navigating the rapids: the development of regulated next-generation sequencing-based clinical trial assays and companion diagnostics.

Over the past decade, next-generation sequencing (NGS) technology has experienced meteoric growth in the aspects of platform, technology, and supporting bioinformatics development allowing its widespread and rapid uptake in research settings. More recently, NGS-based genomic data have been exploited to better understand disease development and patient characteristics that influence response to a given therapeutic intervention. Cancer, as a disease characterized by and driven by the tumor genetic landscape, is particularly amenable to NGS-based diagnostic (Dx) approaches. NGS-based technologies are particularly well suited to studying cancer disease development, progression and emergence of resistance, all key factors in the development of next-generation cancer Dxs. Yet, to achieve the promise of NGS-based patient treatment, drug developers will need to overcome a number of operational, technical, regulatory, and strategic challenges. Here, we provide a succinct overview of the state of the clinical NGS field in terms of the available clinically targeted platforms and sequencing technologies. We discuss the various operational and practical aspects of clinical NGS testing that will facilitate or limit the uptake of such assays in routine clinical care. We examine the current strategies for analytical validation and Food and Drug Administration (FDA)-approval of NGS-based assays and ongoing efforts to standardize clinical NGS and build quality control standards for the same. The rapidly evolving companion diagnostic (CDx) landscape for NGS-based assays will be reviewed, highlighting the key areas of concern and suggesting strategies to mitigate risk. The review will conclude with a series of strategic questions that face drug developers and a discussion of the likely future course of NGS-based CDx development efforts.

Past, current and future approaches to querying MAPK pathway activation: status and clinical implications.

MAPK pathway activation related to cancer development has drawn a great deal of attention in the field of personalized medicine in recent years. Many different approaches and assays have been developed to query the activation of this pathway and to develop life-saving treatments. The goal of this review article is threefold. First, to provide a brief overview of the many mutation assays that have been used to detect MAPK pathway activation, and to compare pros and cons of these assay platforms. Second, to focus on one custom-designed multiplexing mutation assay that is currently used to support an ongoing clinical trial and to show the novel features of this assay and its relevance in addressing unmet clinical needs. Third, to provide future perspectives of these MAPK pathway gene mutation detection efforts and to suggest how what we have learned from past and current approaches should guide future developments.

Practical guidance for implementing predictive biomarkers into early phase clinical studies.

The recent U.S. Food and Drug Administration (FDA) coapprovals of several therapeutic compounds and their companion diagnostic devices (FDA News Release, 2011, 2013) to identify patients who would benefit from treatment have led to considerable interest in incorporating predictive biomarkers in clinical studies. Yet, the translation of predictive biomarkers poses unique technical, logistic, and regulatory challenges that need to be addressed by a multidisciplinary team including discovery scientists, clinicians, biomarker experts, regulatory personnel, and assay developers. These issues can be placed into four broad categories: sample collection, assay validation, sample analysis, and regulatory requirements. In this paper, we provide a primer for drug development teams who are eager to implement a predictive patient segmentation marker into an early clinical trial in a way that facilitates subsequent development of a companion diagnostic. Using examples of nucleic acid-based assays, we briefly review common issues encountered when translating a biomarker to the clinic but focus primarily on key practical issues that should be considered by clinical teams when planning to use a biomarker to balance arms of a study or to determine eligibility for a clinical study.

Development and validation of a clinical trial patient stratification assay that interrogates 27 mutation sites in MAPK pathway genes.

Somatic mutations identified on genes related to the cancer-developing signaling pathways have drawn attention in the field of personalized medicine in recent years. Treatments developed to target a specific signaling pathway may not be effective when tumor activating mutations occur downstream of the target and bypass the targeted mechanism. For instance, mutations detected in KRAS/BRAF/NRAS genes can lead to EGFR-independent intracellular signaling pathway activation. Most patients with these mutations do not respond well to anti-EGFR treatment. In an effort to detect various mutations in FFPE tissue samples among multiple solid tumor types for patient stratification many mutation assays were evaluated. Since there were more than 30 specific mutations among three targeted RAS/RAF oncogenes that could activate MAPK pathway genes, a custom designed Single Nucleotide Primer Extension (SNPE) multiplexing mutation assay was developed and analytically validated as a clinical trial assay. Throughout the process of developing and validating the assay we overcame many technical challenges which include: the designing of PCR primers for FFPE tumor tissue samples versus normal blood samples, designing of probes for detecting consecutive nucleotide double mutations, the kinetics and thermodynamics aspects of probes competition among themselves and against target PCR templates, as well as validating an assay when positive control tumor tissue or cell lines with specific mutations are not available. We used Next Generation sequencing to resolve discordant calls between the SNPE mutation assay and Sanger sequencing. We also applied a triplicate rule to reduce potential false positives and false negatives, and proposed special considerations including pre-define a cut-off percentage for detecting very low mutant copies in the wild-type DNA background.

High-throughput simultaneous analysis of RNA, protein, and lipid biomarkers in heterogeneous tissue samples.

BACKGROUND: With expanding biomarker discovery efforts and increasing costs of drug development, it is critical to maximize the value of mass-limited clinical samples. The main limitation of available methods is the inability to isolate and analyze, from a single sample, molecules requiring incompatible extraction methods. Thus, we developed a novel semiautomated method for tissue processing and tissue milling and division (TMAD). METHODS: We used a SilverHawk atherectomy catheter to collect atherosclerotic plaques from patients requiring peripheral atherectomy. Tissue preservation by flash freezing was compared with immersion in RNAlater®, and tissue grinding by traditional mortar and pestle was compared with TMAD. Comparators were protein, RNA, and lipid yield and quality. Reproducibility of analyte yield from aliquots of the same tissue sample processed by TMAD was also measured. RESULTS: The quantity and quality of biomarkers extracted from tissue prepared by TMAD was at least as good as that extracted from tissue stored and prepared by traditional means. TMAD enabled parallel analysis of gene expression (quantitative reverse-transcription PCR, microarray), protein composition (ELISA), and lipid content (biochemical assay) from as little as 20 mg of tissue. The mean correlation was r = 0.97 in molecular composition (RNA, protein, or lipid) between aliquots of individual samples generated by TMAD. We also demonstrated that it is feasible to use TMAD in a large-scale clinical study setting. CONCLUSIONS: The TMAD methodology described here enables semiautomated, high-throughput sampling of small amounts of heterogeneous tissue specimens by multiple analytical techniques with generally improved quality of recovered biomolecules.

cDNA targets improve whole blood gene expression profiling and enhance detection of pharmocodynamic biomarkers: a quantitative platform analysis.

BACKGROUND: Genome-wide gene expression profiling of whole blood is an attractive method for discovery of biomarkers due to its non-invasiveness, simple clinical site processing and rich biological content. Except for a few successes, this technology has not yet matured enough to reach its full potential of identifying biomarkers useful for clinical prognostic and diagnostic applications or in monitoring patient response to therapeutic intervention. A variety of technical problems have hampered efforts to utilize this technology for identification of biomarkers. One significant hurdle has been the high and variable concentrations of globin transcripts in whole blood total RNA potentially resulting in non-specific probe binding and high background. In this study, we investigated and quantified the power of three whole blood profiling approaches to detect meaningful biological expression patterns. METHODS: To compare and quantify the impact of different mitigation technologies, we used a globin transcript spike-in strategy to synthetically generate a globin-induced signature and then mitigate it with the three different technologies. Biological differences, in globin transcript spiked samples, were modeled by supplementing with either 1% of liver or 1% brain total RNA. In order to demonstrate the biological utility of a robust globin artifact mitigation strategy in biomarker discovery, we treated whole blood ex vivo with suberoylanilide hydroxamic acid (SAHA) and compared the overlap between the obtained signatures and signatures of a known biomarker derived from SAHA-treated cell lines and PBMCs of SAHA-treated patients. RESULTS: We found cDNA hybridization targets detect at least 20 times more specific differentially expressed signatures (2597) between 1% liver and 1% brain in globin-supplemented samples than the PNA (117) or no treatment (97) method at FDR = 10% and p-value < 3x10-3. In addition, we found that the ex vivo derived gene expression profile was highly concordant with that of the previously identified SAHA pharmacodynamic biomarkers. CONCLUSIONS: We conclude that an amplification method for gene expression profiling employing cDNA targets effectively mitigates the negative impact on data of abundant globin transcripts and greatly improves the ability to identify relevant gene expression based pharmacodynamic biomarkers from whole blood.

Refined anatomical isolation of functional sleep circuits exhibits distinctive regional and circadian gene transcriptional profiles.

Powerful new approaches to study molecular variation in distinct neuronal populations have recently been developed enabling a more precise investigation of the control of neural circuits involved in complex behaviors such as wake and sleep. We applied laser capture microdissection (LCM) to isolate precise brain nuclei from rat CNS at opposing circadian time points associated with wake and sleep. Discrete anatomical and temporal analysis was performed to examine the extent of variation in the transcriptional control associated with both identifiable anatomical nuclei and with light/dark cycle. Precise isolation of specific brain nuclei regulating sleep and arousal, including the LC, SCN, TMN, VTA, and VLPO, demonstrated robust changes in gene expression. Many of these differences were not observed in previous studies where whole brain lysates or gross dissections were used to probe for changes in gene expression. The robust and differential profiles of genomic data obtained from the approaches used herein underscore the requirement for careful anatomical refinement in CNS gene expression studies designed to understand genomic control within behaviorally-linked, but functionally isolated brain nuclei.

Effects of atmospheric ozone on microarray data quality.

A data anomaly was observed that affected the uniformity and reproducibility of fluorescent signal across DNA microarrays. Results from experimental sets designed to identify potential causes (from microarray production to array scanning) indicated that the anomaly was linked to a batch process; further work allowed us to localize the effect to the posthybridization array stringency washes. Ozone levels were monitored and highly correlated with the batch effect. Controlled exposures of microarrays to ozone confirmed this factor as the root cause, and we present data that show susceptibility of a class of cyanine dyes (e.g., Cy5, Alexa 647) to ozone levels as low as 5-10 ppb for periods as short as 10-30 s. Other cyanine dyes (e.g., Cy3, Alexa 555) were not significantly affected until higher ozone levels (> 100 ppb). To address this environmental effect, laboratory ozone levels should be kept below 2 ppb (e.g., with filters in HVAC) to achieve high quality microarray data.

Functional interactions between yeast translation eukaryotic elongation factor (eEF) 1A and eEF3.

The translation elongation machinery in fungi differs from other eukaryotes in its dependence upon eukaryotic elongation factor 3 (eEF3). eEF3 is essential in vivo and required for each cycle of the translation elongation process in vitro. Models predict eEF3 affects the delivery of cognate aminoacyl-tRNA, a function performed by eEF1A, by removing deacylated tRNA from the ribosomal Exit site. To dissect eEF3 function and its link to the A-site activities of eEF1A, we have identified a temperature-sensitive allele of the YEF3 gene. The F650S substitution, located between the two ATP binding cassettes, reduces both ribosome-dependent and intrinsic ATPase activities. In vivo this mutation increases sensitivity to aminoglycosidic drugs, causes a 50% reduction of total protein synthesis at permissive temperatures, slows run-off of polyribosomes, and reduces binding to eEF1A. Reciprocally, excess eEF3 confers synthetic slow growth, increased drug sensitivity, and reduced translation in an allele specific fashion with an E122K mutation in the GTP binding domain of eEF1A. In addition, this mutant form of eEF1A shows reduced binding of eEF3. Thus, optimal in vivo interactions between eEF3 and eEF1A are critical for protein synthesis.

A gene-expression signature as a predictor of survival in breast cancer.

BACKGROUND: A more accurate means of prognostication in breast cancer will improve the selection of patients for adjuvant systemic therapy. METHODS: Using microarray analysis to evaluate our previously established 70-gene prognosis profile, we classified a series of 295 consecutive patients with primary breast carcinomas as having a gene-expression signature associated with either a poor prognosis or a good prognosis. All patients had stage I or II breast cancer and were younger than 53 years old; 151 had lymph-node-negative disease, and 144 had lymph-node-positive disease. We evaluated the predictive power of the prognosis profile using univariable and multivariable statistical analyses. RESULTS: Among the 295 patients, 180 had a poor-prognosis signature and 115 had a good-prognosis signature, and the mean (+/-SE) overall 10-year survival rates were 54.6+/-4.4 percent and 94.5+/-2.6 percent, respectively. At 10 years, the probability of remaining free of distant metastases was 50.6+/-4.5 percent in the group with a poor-prognosis signature and 85.2+/-4.3 percent in the group with a good-prognosis signature. The estimated hazard ratio for distant metastases in the group with a poor-prognosis signature, as compared with the group with the good-prognosis signature, was 5.1 (95 percent confidence interval, 2.9 to 9.0; P<0.001). This ratio remained significant when the groups were analyzed according to lymph-node status. Multivariable Cox regression analysis showed that the prognosis profile was a strong independent factor in predicting disease outcome. CONCLUSIONS: The gene-expression profile we studied is a more powerful predictor of the outcome of disease in young patients with breast cancer than standard systems based on clinical and histologic criteria.