Alexander Ploner

Gene expression profiling spares early breast cancer patients from adjuvant therapy: derived and validated in two population-based cohorts

IntroductionAdjuvant breast cancer therapy significantly improves survival, but overtreatment and undertreatment are major problems. Breast cancer expression profiling has so far mainly been used to identify women with a poor prognosis as candidates for adjuvant therapy but without demonstrated value for therapy prediction.MethodsWe obtained the gene expression profiles of 159 population-derived breast cancer patients, and used hierarchical clustering to identify the signature associated with prognosis and impact of adjuvant therapies, defined as distant metastasis or death within 5 years. Independent datasets of 76 treated population-derived Swedish patients, 135 untreated population-derived Swedish patients and 78 Dutch patients were used for validation. The inclusion and exclusion criteria for the studies of population-derived Swedish patients were defined.ResultsAmong the 159 patients, a subset of 64 genes was found to give an optimal separation of patients with good and poor outcomes. Hierarchical clustering revealed three subgroups: patients who did well with therapy, patients who did well without therapy, and patients that failed to benefit from given therapy. The expression profile gave significantly better prognostication (odds ratio, 4.19; P = 0.007) (breast cancer end-points odds ratio, 10.64) compared with the Elston–Ellis histological grading (odds ratio of grade 2 vs 1 and grade 3 vs 1, 2.81 and 3.32 respectively; P = 0.24 and 0.16), tumor stage (odds ratio of stage 2 vs 1 and stage 3 vs 1, 1.11 and 1.28; P = 0.83 and 0.68) and age (odds ratio, 0.11; P = 0.55). The risk groups were consistent and validated in the independent Swedish and Dutch data sets used with 211 and 78 patients, respectively.ConclusionWe have identified discriminatory gene expression signatures working both on untreated and systematically treated primary breast cancer patients with the potential to spare them from adjuvant therapy.

False discovery rate, sensitivity and sample size for microarray studies

MOTIVATION In microarray data studies most researchers are keenly aware of the potentially high rate of false positives and the need to control it. One key statistical shift is the move away from the well-known P-value to false discovery rate (FDR). Less discussion perhaps has been spent on the sensitivity or the associated false negative rate (FNR). The purpose of this paper is to explain in simple ways why the shift from P-value to FDR for statistical assessment of microarray data is necessary, to elucidate the determining factors of FDR and, for a two-sample comparative study, to discuss its control via sample size at the design stage. RESULTS We use a mixture model, involving differentially expressed (DE) and non-DE genes, that captures the most common problem of finding DE genes. Factors determining FDR are (1) the proportion of truly differentially expressed genes, (2) the distribution of the true differences, (3) measurement variability and (4) sample size. Many current small microarray studies are plagued with large FDR, but controlling FDR alone can lead to unacceptably large FNR. In evaluating a design of a microarray study, sensitivity or FNR curves should be computed routinely together with FDR curves. Under certain assumptions, the FDR and FNR curves coincide, thus simplifying the choice of sample size for controlling the FDR and FNR jointly.

Intrinsic molecular signature of breast cancer in a population-based cohort of 412 patients

BackgroundMolecular markers and the rich biological information they contain have great potential for cancer diagnosis, prognostication and therapy prediction. So far, however, they have not superseded routine histopathology and staging criteria, partly because the few studies performed on molecular subtyping have had little validation and limited clinical characterization.MethodsWe obtained gene expression and clinical data for 412 breast cancers obtained from population-based cohorts of patients from Stockholm and Uppsala, Sweden. Using the intrinsic set of approximately 500 genes derived in the Norway/Stanford breast cancer data, we validated the existence of five molecular subtypes – basal-like, ERBB2, luminal A/B and normal-like – and characterized these subtypes extensively with the use of conventional clinical variables.ResultsWe found an overall 77.5% concordance between the centroid prediction of the Swedish cohort by using the Norway/Stanford signature and the k-means clustering performed internally within the Swedish cohort. The highest rate of discordant assignments occurred between the luminal A and luminal B subtypes and between the luminal B and ERBB2 subtypes. The subtypes varied significantly in terms of grade (p < 0.001), p53 mutation (p < 0.001) and genomic instability (p = 0.01), but surprisingly there was little difference in lymph-node metastasis (p = 0.31). Furthermore, current users of hormone-replacement therapy were strikingly over-represented in the normal-like subgroup (p < 0.001). Separate analyses of the patients who received endocrine therapy and those who did not receive any adjuvant therapy supported the previous hypothesis that the basal-like subtype responded to adjuvant treatment, whereas the ERBB2 and luminal B subtypes were poor responders.ConclusionWe found that the intrinsic molecular subtypes of breast cancer are broadly present in a diverse collection of patients from a population-based cohort in Sweden. The intrinsic gene set, originally selected to reveal stable tumor characteristics, was shown to have a strong correlation with progression-related properties such as grade, p53 mutation and genomic instability.

Quadrivalent Human Papillomavirus Vaccine Effectiveness: A Swedish National Cohort Study

Background Incidence of condyloma, or genital warts (GW), is the earliest possible disease outcome to measure when assessing the effectiveness of human papillomavirus (HPV) vaccination strategies. Efficacy trials that follow prespecified inclusion and exclusion criteria may not be fully generalizable to real-life HPV vaccination programs, which target a broader segment of the population. We assessed GW incidence after on-demand vaccination with quadrivalent HPV vaccine using individual-level data from the entire Swedish population. Methods An open cohort of girls and women aged 10 to 44 years living in Sweden between 2006 and 2010 (N > 2.2 million) was linked to multiple population registers to identify incident GW in relation to HPV vaccination. For vaccine effectiveness, incidence rate ratios of GW were estimated using time-to-event analyses with adjustment for attained age and parental education level, stratifying on age at first vaccination. Results A total of 124 000 girls and women were vaccinated between 2006 and 2010. Girls and women with at least one university-educated parent were 15 times more likely to be vaccinated before age 20 years than girls and women whose parents did not complete high school (relative risk ratio = 15.45, 95% confidence interval [CI] = 14.65 to 16.30). Among those aged older than 20 years, GW rates declined among the unvaccinated, suggesting that HPV vaccines were preferentially used by women at high risk of GW. Vaccination effectiveness was 76% (95% CI = 73% to 79%) among those who received three doses of the vaccine with their first dose before age 20 years. Vaccine effectiveness was highest in girls vaccinated before age 14 years (effectiveness = 93%, 95% CI = 73% to 98%). Conclusions Young age at first vaccination is imperative for maximizing quadrivalent HPV vaccine effectiveness.

Bias in the estimation of false discovery rate in microarray studies

MOTIVATION The false discovery rate (FDR) provides a key statistical assessment for microarray studies. Its value depends on the proportion pi(0) of non-differentially expressed (non-DE) genes. In most microarray studies, many genes have small effects not easily separable from non-DE genes. As a result, current methods often overestimate pi(0) and FDR, leading to unnecessary loss of power in the overall analysis. METHODS For the common two-sample comparison we derive a natural mixture model of the test statistic and an explicit bias formula in the standard estimation of pi(0). We suggest an improved estimation of pi(0) based on the mixture model and describe a practical likelihood-based procedure for this purpose. RESULTS The analysis shows that a large bias occurs when pi(0) is far from 1 and when the non-centrality parameters of the distribution of the test statistic are near zero. The theoretical result also explains substantial discrepancies between non-parametric and model-based estimates of pi(0). Simulation studies indicate mixture-model estimates are less biased than standard estimates. The method is applied to breast cancer and lymphoma data examples. AVAILABILITY An R-package OCplus containing functions to compute pi(0) based on the mixture model, the resulting FDR and other operating characteristics of microarray data, is freely available at http://www.meb.ki.se/~yudpaw CONTACT yudi.pawitan@meb.ki.se and alexander.ploner@meb.ki.se.

Association of Varying Number of Doses of Quadrivalent Human Papillomavirus Vaccine With Incidence of Condyloma

IMPORTANCE Determining vaccine dose-level protection is essential to minimize program costs and increase mass vaccination program feasibility. Currently, a 3-dose vaccination schedule is recommended for both the quadrivalent and bivalent human papillomavirus (HPV) vaccines. Although the primary goal of HPV vaccination programs is to prevent cervical cancer, condyloma related to HPV types 6 and 11 is also prevented with the quadrivalent vaccine and represents the earliest measurable preventable disease outcome for the HPV vaccine. OBJECTIVE To examine the association between quadrivalent HPV vaccination and first occurrence of condyloma in relation to vaccine dose in a population-based setting. DESIGN, SETTING, AND PARTICIPANTS An open cohort of all females aged 10 to 24 years living in Sweden (n = 1,045,165) was followed up between 2006 and 2010 for HPV vaccination and first occurrence of condyloma using the Swedish nationwide population-based health data registers. MAIN OUTCOMES AND MEASURES Incidence rate ratios (IRRs) and incidence rate differences (IRDs) of condyloma were estimated using Poisson regression with vaccine dose as a time-dependent exposure, adjusting for attained age and parental education, and stratified on age at first vaccination. To account for prevalent infections, models included a buffer period of delayed case counting. RESULTS A total of 20,383 incident cases of condyloma were identified during follow-up, including 322 cases after receipt of at least 1 dose of the vaccine. For individuals aged 10 to 16 years at first vaccination, receipt of 3 doses was associated with an IRR of 0.18 (95% CI, 0.15-0.22) for condyloma, whereas receipt of 2 doses was associated with an IRR of 0.29 (95% CI, 0.21-0.40). One dose was associated with an IRR of 0.31 (95% CI, 0.20-0.49), which corresponds to an IRD of 384 cases (95% CI, 305-464) per 100,000 person-years, compared with no vaccination. The corresponding IRDs for 2 doses were 400 cases (95% CI, 346-454) and for 3 doses, 459 cases (95% CI, 437-482). The number of prevented cases between 3 and 2 doses was 59 (95% CI, 2-117) per 100,000 person-years. CONCLUSIONS AND RELEVANCE Although maximum reduction in condyloma risk was seen after receipt of 3 doses of quadrivalent HPV vaccine, receipt of 2 vaccine doses was also associated with a considerable reduction in condyloma risk. The implications of these findings for the relationship between number of vaccine doses and cervical cancer risk require further investigation.

Multidimensional local false discovery rate for microarray studies

MOTIVATION The false discovery rate (fdr) is a key tool for statistical assessment of differential expression (DE) in microarray studies. Overall control of the fdr alone, however, is not sufficient to address the problem of genes with small variance, which generally suffer from a disproportionally high rate of false positives. It is desirable to have an fdr-controlling procedure that automatically accounts for gene variability. METHODS We generalize the local fdr as a function of multiple statistics, combining a common test statistic for assessing DE with its standard error information. We use a non-parametric mixture model for DE and non-DE genes to describe the observed multi-dimensional statistics, and estimate the distribution for non-DE genes via the permutation method. We demonstrate this fdr2d approach for simulated and real microarray data. RESULTS The fdr2d allows objective assessment of DE as a function of gene variability. We also show that the fdr2d performs better than commonly used modified test statistics. AVAILABILITY An R-package OCplus containing functions for computing fdr2d() and other operating characteristics of microarray data is available at http://www.meb.ki.se/~yudpaw.

Correlation test to assess low-level processing of high-density oligonucleotide microarray data

BackgroundThere are currently a number of competing techniques for low-level processing of oligonucleotide array data. The choice of technique has a profound effect on subsequent statistical analyses, but there is no method to assess whether a particular technique is appropriate for a specific data set, without reference to external data.ResultsWe analyzed coregulation between genes in order to detect insufficient normalization between arrays, where coregulation is measured in terms of statistical correlation. In a large collection of genes, a random pair of genes should have on average zero correlation, hence allowing a correlation test. For all data sets that we evaluated, and the three most commonly used low-level processing procedures including MAS5, RMA and MBEI, the housekeeping-gene normalization failed the test. For a real clinical data set, RMA and MBEI showed significant correlation for absent genes. We also found that a second round of normalization on the probe set level improved normalization significantly throughout.ConclusionPrevious evaluation of low-level processing in the literature has been limited to artificial spike-in and mixture data sets. In the absence of a known gold-standard, the correlation criterion allows us to assess the appropriateness of low-level processing of a specific data set and the success of normalization for subsets of genes.

Blood biomarker levels to aid discovery of cancer-related single-nucleotide polymorphisms: kallikreins and prostate cancer.

Polymorphisms associated with prostate cancer include those in three genes encoding major secretory products of the prostate: KLK2 (encoding kallikrein-related peptidase 2; hK2), KLK3 (encoding prostate-specific antigen; PSA), and MSMB (encoding β-microseminoprotein). PSA and hK2, members of the kallikrein family, are elevated in sera of men with prostate cancer. In a comprehensive analysis that included sequencing of all coding, flanking, and 2 kb of putative promoter regions of all 15 kallikrein (KLK) genes spanning ≈280 kb on chromosome 19q, we identified novel single-nucleotide polymorphisms (SNP) and genotyped 104 SNPs in 1,419 cancer cases and 736 controls in Cancer Prostate in Sweden 1, with independent replication in 1,267 cases and 901 controls in Cancer Prostate in Sweden 2. This verified prior associations of SNPs in KLK2 and in MSMB (but not in KLK3) with prostate cancer. Twelve SNPs in KLK2 and KLK3 were associated with levels of PSA forms or hK2 in plasma of control subjects. Based on our comprehensive approach, this is likely to represent all common KLK variants associated with these phenotypes. A T allele at rs198977 in KLK2 was associated with increased cancer risk and a striking decrease of hK2 levels in blood. We also found a strong interaction between rs198977 genotype and hK2 levels in blood in predicting cancer risk. Based on this strong association, we developed a model for predicting prostate cancer risk from standard biomarkers, rs198977 genotype, and rs198977 × hK2 interaction; this model had greater accuracy than did biomarkers alone (area under the receiver operating characteristic curve, 0.874 versus 0.866), providing proof in principle to clinical application for our findings. Cancer Prev Res; 3(5); 611–9. ©2010 AACR.

Adiposity as a cause of cardiovascular disease: a Mendelian randomization study

BACKGROUND Adiposity, as indicated by body mass index (BMI), has been associated with risk of cardiovascular diseases in epidemiological studies. We aimed to investigate if these associations are causal, using Mendelian randomization (MR) methods. METHODS The associations of BMI with cardiovascular outcomes [coronary heart disease (CHD), heart failure and ischaemic stroke], and associations of a genetic score (32 BMI single nucleotide polymorphisms) with BMI and cardiovascular outcomes were examined in up to 22,193 individuals with 3062 incident cardiovascular events from nine prospective follow-up studies within the ENGAGE consortium. We used random-effects meta-analysis in an MR framework to provide causal estimates of the effect of adiposity on cardiovascular outcomes. RESULTS There was a strong association between BMI and incident CHD (HR = 1.20 per SD-increase of BMI, 95% CI, 1.12-1.28, P = 1.9.10(-7)), heart failure (HR = 1.47, 95% CI, 1.35-1.60, P = 9.10(-19)) and ischaemic stroke (HR = 1.15, 95% CI, 1.06-1.24, P = 0.0008) in observational analyses. The genetic score was robustly associated with BMI (β = 0.030 SD-increase of BMI per additional allele, 95% CI, 0.028-0.033, P = 3.10(-107)). Analyses indicated a causal effect of adiposity on development of heart failure (HR = 1.93 per SD-increase of BMI, 95% CI, 1.12-3.30, P = 0.017) and ischaemic stroke (HR = 1.83, 95% CI, 1.05-3.20, P = 0.034). Additional cross-sectional analyses using both ENGAGE and CARDIoGRAMplusC4D data showed a causal effect of adiposity on CHD. CONCLUSIONS Using MR methods, we provide support for the hypothesis that adiposity causes CHD, heart failure and, previously not demonstrated, ischaemic stroke.