Length Bias Research Articles

Composite Partial Likelihood Estimation Under Length-Biased Sampling, With Application to a Prevalent Cohort Study of Dementia

The Canadian Study of Health and Aging (CSHA) employed a prevalent cohort design to study survival after onset of dementia, where patients with dementia were sampled and the onset time of dementia was determined retrospectively. The prevalent cohort sampling scheme favors individuals who survive longer. Thus, the observed survival times are subject to length bias. In recent years, there has been a rising interest in developing estimation procedures for prevalent cohort survival data that not only account for length bias but also actually exploit the incidence distribution of the disease to improve efficiency. This article considers semiparametric estimation of the Cox model for the time from dementia onset to death under a stationarity assumption with respect to the disease incidence. Under the stationarity condition, the semiparametric maximum likelihood estimation is expected to be fully efficient yet difficult to perform for statistical practitioners, as the likelihood depends on the baseline hazard function in a complicated way. Moreover, the asymptotic properties of the semiparametric maximum likelihood estimator are not well-studied. Motivated by the composite likelihood method (Besag 1974), we develop a composite partial likelihood method that retains the simplicity of the popular partial likelihood estimator and can be easily performed using standard statistical software. When applied to the CSHA data, the proposed method estimates a significant difference in survival between the vascular dementia group and the possible Alzheimer's disease group, while the partial likelihood method for left-truncated and right-censored data yields a greater standard error and a 95% confidence interval covering 0, thus highlighting the practical value of employing a more efficient methodology. To check the assumption of stable disease for the CSHA data, we also present new graphical and numerical tests in the article. The R code used to obtain the maximum composite partial likelihood estimator for the CSHA data is available in the online Supplementary Material, posted on the journal web site.

Assessment of method of breast cancer detection and Oncotype DX recurrence score.

581 Background: Data indicate that screen-detected (SD) cancers show more favourable pathological characteristics and have a better prognosis compared to symptomatic or interval cancers (SY). These findings can be partly explained by length time and lead time biases. OncotypeDX recurrence score (RS) provides prognostic information independent of traditional clinicopathological variables and provides a robust measure of tumour proliferation. The objective of this study was to determine whether the RS would vary depending on the method of detection (MOD) in patients with lymph node-negative, ER-positive, HER-2 negative breast cancer (BC). Methods: We included patients with ER-positive, lymph node-negative and HER2-negative BC who underwent OncotypeDX testing as part of routine care or a clinical trial. We retrospectively reviewed patient charts and extracted information on MOD (SD versus SY), clinicopathological variables and RS. Chi-squared test was used to test for differences between categorical variables and MOD. Analysis of variance (ANOVA) was used to investigate the relationship between RS and MOD and clinicopathological variables. Results: We identified 596 patients from 4 academic hospitals (Mater University Hospital n= 98 (16%), St Vincent’s Hospital n= 161 (27%), Waterford Regional Hospital n= 27 (4%) and University of Texas, MD Anderson Cancer Center n= 310 (52%)). There were 359 (60 %) SD- detected and 237 (40%) SY-detected breast cancers. Clinicopathological variables were similar for age at diagnosis (56 versus 52 years), tumour size (16mm versus 18mm), grade (63% versus 61% grade 2; 15% versus 15% grade I; 21% versus 24% grade 3) and LVI (18% versus 18%) for SY compared to SD, respectively. The number of patients in each group with low (51% versus 50%), intermediate (41% each) or high (8% versus 9%) RS was similar in the SD compared to SY groups, respectively. There was a statistically significant association between RS and tumour grade (P=<0.001) and lymphovascular invasion (P=<0.001). We did not observe a statistically significant association between RS and MOD (P=0.086). Conclusions: In patients with clinically intermediate breast cancer sent for OncotypeDX testing we found no association between RS and MOD.

55P Prognostic Role of Palpability in Screen Detected Breast Cancer (BC)

ABSTRACT Introduction Screen detected BC has excellent outcome due to lead time and length bias, but approximately 30% of the mortality reduction through screening remains unexplained. On the other hand, not all patients with screen detected BC have excellent outcome as some will eventually relapse. Therefore, new prognostic variables in screen detected BC are needed and we here study palpability as a candidate variable. Previously, Gill et al (J Med Screen, 2006; 13:98-101), did not find an independent prognostic value for palpability within screen detected BC. Patients and methods Primary operable invasive BC treated at UZ-Leuven (01/2000 – 12/2009). Preoperative systemic therapy (n = 407), metastases at diagnosis (n = 228), male patients (n = 28) or external surgery (n = 530) were exclusion criteria. Details on treatment and pathology are described in detail elsewhere (Brouckaert et al, Breast Cancer Res Treat., 2009 May; 115 (2):349-58). We included 1610 patients with screen detected BC. Missing data for palpability were 2. Palpability was assessed preoperatively by the treating physician. Multivariable Cox proportional hazard models were used adjusting for age, size, grade, nodal status, ER, PR and HER2 receptor status and adjuvant therapy. The proportional hazard assumption for the Cox model was evaluated by including time-dependent covariates in the models, which were included when significant. Results Median follow-up is 6.1 years. Most screen detected BC are grade 1 or 2 (72%), ER (90%) positive, PR (79%) positive and 52% are palpable. Palpability (Hazard ratio (HR) 2,73 95% CI 1,26-5,92 (p = 0,0107)) is an independent prognostic variable for distant metastasis free survival together with tumor size (HR 1,01 95% CI 1,00-1,03 (p = 0,0411)) and total number of positive lymph nodes (HR 1,11 95% CI 1,06-1,16 (p Conclusion Palpability is an independent prognostic variable in screen detected BC. Disclosure All authors have declared no conflicts of interest.

44P Prognostic Value of Detection Method by Breast Cancer (BC) by Phenotype

ABSTRACT Introduction Mammographic screening has shown to reduce BC mortality, but approximately 30% of the mortality reduction remains unexplained. Method of detection has independent prognostic value and we here assess its value in function of BC phenotypes. Patients and methods Primary operable invasive BC patients treated at the Multidisciplinary BC Centre UZ-Leuven (01/2000 – 12/2009) were included. Preoperative systemic therapy (n = 407), metastases at diagnosis (n = 228), male patients (n = 28) or external surgery (n = 530) were exclusion criteria. Details on treatment and pathology are described in detail elsewhere (Brouckaert et al., Breast Cancer Res Treat., 2009 May; 115(2): 349-58). BCs were allocated to screen detected (n = 1610) when diagnosed through organized or opportunistic screening (missing n = 91) and to palpable when palpable by the treating physician (missing n = 57). We defined BC phenotypes as Luminal A (ER and/or PR positive HER2 negative, grade 1/2), Luminal B1 (idem, grade 3), Luminal B2 (ER and/or PR and HER-2 positive), HER-2 like (ER/PR negative, HER-2 positive) and Triple negative (2011 St. Gallen criteria). Multivariable Cox proportional hazard models were used adjusting for age, size, grade, nodal status, ER, PR and HER2 receptor status and adjuvant therapy (see Table 1 ). The proportional hazard assumption for the Cox model was evaluated by including time-dependent covariates in the models, which were included when significant. Results Median follow-up 6.1 years. Table 1 . Adjusted HR (Screening versus non-) Distant metastasis free survival (p-value) Breast cancer specific survival (p-value) All (n = 4227) 0.42 (0.03) 0.49 (0.13) Luminal A (n = 2338) 0.79 (0.29) 0.63 (0.2) Luminal B1 (n = 924) 0.58 (0.03) 0.69 (0.23) Luminal B2 (n = 289) 0.37 (0.07) 0.53 (0.39) HER-2 like (n = 173) 0.39 (0.09) 0.46 (0.24) Triple negative (n = 412) 0.44 (0.04) 0.23 (0.01) Conclusion BC screening may be criticized to reduce mortality by simply diagnosing indolent tumors (length bias). However, the prognostic value of method of detection seems to have increasing value in increasingly more aggressive BC phenotypes. Method of detection may also prove to be a new and independent prognostic tool to assess outcome in triple negative BC. References on demand. Disclosure All authors have declared no conflicts of interest.

Lung cancer screening: Examining the issues

The goal of screening is to detect disease at a stage when cure or control is possible, thereby decreasing disease-specific deaths in the population. Many studies have attempted to demonstrate that lung cancer screening using chest radiography or computed tomography (CT) identifies patients with lung cancer and reduces cancer-related mortality. Until recently, there was no evidence confirming a reduction in disease-specific mortality with screening. Early cancer screening should result in a gradual population-wide stage shift toward earlier cancer stages over time, but stage shifting was not reported in early lung cancer screening studies. Lead-time, length-time, and overdiagnosis biases may each have an impact on screening studies reporting survival as an outcome. In this past year, the National Lung Screening Trial reported a significant reduction in cancer-related mortality as a result of screening with chest CT imaging. This will shape the direction of future screening programs.

RNA sequencing data: biases and normalization

Motivations. In recent years, RNA sequencing (RNA-seq) has rapidly become the method of choice for measuring and comparing gene transcription levels. Despite its wide application, it is now clear that this methodology is not free from biases and that a careful normalization procedure is the basis for a correct data interpretation. The most common normalization techniques account for: library size, gene or transcript length and sequence-specific biases such as GC-content effects. The aim of the present work is to investigate biases affecting RNA seq data and their effect on differential expression analysis. In order to reduce biases due to over-simplification of gene transcription models, we consider exon-based counts. Methods. We two used publicly available RNA-seq data sets from two-group comparison studies which are characterized by multiple technical replicates. We summarized read counts at exon level and investigated their dependence on sequence-specific covariates: GC-content and exon length. In addition, we considered the effect of library size correction on between-groups comparison and the impact of the above mentioned biases on the detection of differentially expressed exons. The assessment is performed on raw data, as well as on data normalized with different approaches: RPKM [1], library size scaling, based on Trimmed Mean of M-values (TMM) [2] and on Poisson goodness-of-fit statistic applied to non differentially expressed genes [3], and within-lane normalization based on loess regression of log-counts on GC-content and exon length [4]. We selected differentially expressed exons using the GLM-based version of edgeR [5] as it can consider an offset matrix which codifies counts normalization, that can be computed with the desired approach, and library size scaling factors specified by the user. Results. In our study, read counts show a significant dependence on exon length and a moderate dependence on GC-content. Exon length bias also affects differential expression analysis: longer exons tend to have lower P-values and to be selected as differentially expressed more frequently than shorter exons. The tested normalization techniques do not completely remove biases and, in particular, RPKM approach over-corrects for exon length bias. Moreover, the choice of the strategy for library size adjustment has a great impact on the direction of the detected differential expression. The results obtained on these data sets demonstrate that RNA-seq data normalization is still an open issue. Further efforts should be directed towards the clarification of the relationship between read counts and sequence-specific biases, which are, in turn, correlated to each other, and the definition of new models for their correction.

1894 POSITRON EMISSION TOMOGRAPHY-COMPUTED TOMOGRAPHY (PET-CT) ROLE IN PREDICTING MORTALITY IN PATIENTS WITH METASTATIC BLADDER CANCER

You have accessJournal of UrologyBladder Cancer: Metastatic Disease/Staging1 Apr 20121894 POSITRON EMISSION TOMOGRAPHY-COMPUTED TOMOGRAPHY (PET-CT) ROLE IN PREDICTING MORTALITY IN PATIENTS WITH METASTATIC BLADDER CANCER Maria Carmen Mir, Nathan Papa, Andrew Scott, Rustom Manecksha, Damien Bolton, Ian Davis, and Nathan Lawrentschuk Maria Carmen MirMaria Carmen Mir Melbourne, Australia More articles by this author , Nathan PapaNathan Papa Melbourne, Australia More articles by this author , Andrew ScottAndrew Scott Melbourne, Australia More articles by this author , Rustom ManeckshaRustom Manecksha Melbourne, Australia More articles by this author , Damien BoltonDamien Bolton Melbourne, Australia More articles by this author , Ian DavisIan Davis Melbourne, Australia More articles by this author , and Nathan LawrentschukNathan Lawrentschuk Melbourne, Australia More articles by this author View All Author Informationhttps://doi.org/10.1016/j.juro.2012.02.2050AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookTwitterLinked InEmail INTRODUCTION AND OBJECTIVES 18F-fluorodeoxyglucose (18F-FDG)-PET has assisted in predicting therapeutic response and outcome in patients with various tumors. Our aim was to investigate the correlation between 18F-FDG PET-CT metabolic uptake within tumour deposits and clinical outcomes in patients with metastatic urothelial carcinoma (UC) arising in the bladder. METHODS A total of 94 patients with biopsy-proven UC at a single centre underwent FDG-PET scans at the time of staging for metastatic disease. Data on staging, mortality and surgery was de-identified and linked through the statewide prospective Victorian Cancer Registry database. Patients were divided into PET-nodal metastasis (PET-N) positive and negative groups according to the presence or absence of FDG uptake on this original study (some had repeat and later studies but we focused on the initial staging study). Kaplan-Meier analysis was used to determine survival outcomes according to 18F-FDG-PET metabolic response. RESULTS In all, 49% patients (n=48) had a PET-N positive study, the remainder (n=46) PET-N negative. Groups were almost identical for clinical stage (based on transurethral biopsy, CT scan and clinical examination) and grade at the time of PET-CT. PET status had no significant difference from date of diagnosis of UC to overall survival (OS). However, on further analysis of survival after the first staging PET study, using a Cox regression and adjusting for age and gender, having a PET-M positive study had a HR = 1.86 (CI 1.05-3.30) for OS (Figure 1). The time to death of patients with a PET-M positive study appeared 86% quicker than if PET-M negative study. Further, those with a PET-M negative study had a 46% reduction in the rate of progression to death. Finally, the timing of the PET from original diagnosis was not significantly different between groups indicating no lead or length bias with PET study results. CONCLUSIONS Uptake of 18F-FDG in metastatic UC is significantly associated with OS. This data may assist for prognosis and trial selection of patients with metastatic UC but should be confirmed in a multicentre trial. © 2012 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 187Issue 4SApril 2012Page: e764 Peer Review Report Advertisement Copyright & Permissions© 2012 by American Urological Association Education and Research, Inc.MetricsAuthor Information Maria Carmen Mir Melbourne, Australia More articles by this author Nathan Papa Melbourne, Australia More articles by this author Andrew Scott Melbourne, Australia More articles by this author Rustom Manecksha Melbourne, Australia More articles by this author Damien Bolton Melbourne, Australia More articles by this author Ian Davis Melbourne, Australia More articles by this author Nathan Lawrentschuk Melbourne, Australia More articles by this author Expand All Advertisement Advertisement PDF downloadLoading ...

Misplaced criticism of breast screening research

Dear Sir, The paper by Gotzsche and Jorgensen1 contains a number of inaccuracies and omissions in its criticism of my work and that of my colleagues.2 Firstly, Gotzsche and Jorgensen take exception to the quoted 50% improvement in breast cancer survival in screen-detected cancers. They omit to mention that the approximate 50% improvement is after correction for lead time and length bias – before correction, the figure was a 70% improvement.3,4 Secondly, they claim that the 28% reduction in breast cancer mortality in England in the screened ages compared to other ages did not occur. They are mistaken. Table 1 shows breast cancer mortality by epoch in ages 50–69 and in all other age groups. While mortality rose by 2% in the latest period compared to the earliest for all other age groups, it fell by 27% in the age group 50–69. The relative risk of breast cancer mortality is therefore: Table 1 Breast cancer mortality in England by age group and epoch That, is a 28% reduction compared to other age groups. In the published analysis, the estimate was age-adjusted, but I give the crude analysis here so that readers can see where the estimate comes from. Gotzsche and Jorgensen's criticisms are particularly error-prone on the subject of over-diagnosis. It might be illuminating to contrast our approach2 with that of Jorgensen and Gotzsche.5 Both teams attempted to estimate overdiagnosis by calculation of expected incidence of breast cancer in the screening epoch based on trends observed in the pre-screening epoch. However, the methods differed at each stage, as follows: Data Sources: Duffy et al. used data on numbers of cases and populations at risk in England from Cancer Registry data.2 Jorgensen and Gotzsche estimated rates for England and Wales from a published graph. Changes in Incidence Independent of Screening: Duffy et al. took full account of these changes by correcting for the 7% increase above expected values at ages below the target age group for screening.2 Jorgensen and Gotzsche failed to do so.5 Method of analysis: Duffy et al. used poisson regression, as is the correct procedure for rate data.2,6 Jorgensen and Gotzsche used linear regression, which is incorrect.5 Data selection: Duffy et al. used all pre-screening and screening epoch data available.2 Jorgensen and Gotzsche excluded the three years of highest incidence in the pre-screening period, and only included the year of highest incidence in their screening epoch data, 1999.5 This inflated their estimate of overdiagnosis. Adjustment for lead time: Duffy et al. subtracted the deficit in incidence above the screening age from the excess observed in screening ages.2 Jorgensen and Gotzsche failed to do so.5 To be fair, they claimed not to observe a deficit. This is partly because in 1999 too few women above the screening age range had been screened in the past, but also because of their failure to fully adjust for changes in incidence independent of screening, as noted in point 2 above. Ductal Carcinoma in situ (DCIS): In the absence of data, Duffy et al. restricted estimation to invasive disease, although in the same paper, they estimated overdiagnosis including DCIS in a randomized trial.2 Like Duffy et al, Jorgensen and Gotzsche had no data on DCIS in the UK, so they assumed a result which was not observed.5 From the above, it can be seen that our modest estimate of overdiagnosis has more reliability than the implausibly high estimate of Jorgensen and Gotzsche. Gotzsche and Jorgensen make a number of further errors in defence of their estimate, including: failure to acknowledge that in the 1990's in the age range for screening, a full paper (not an abstract as stated by Gotzsche and Jorgensen),7 has shown that around 40% of tumours were screen-detected; citation of figures from 2006 to justify their estimate for 1999; and misinterpretation of those figures from 2006, as pointed out previously.8 More importantly, one should not lose sight of the benefit of the NHS Breast Screening Programme and the fact that the only randomized trial with more than 25 years of follow-up, shows that the quoted benefit of one breast cancer death prevented for every 400 women screened is accurate and may even underestimate the benefit.9

Length and GC-Biases During Sequencing Library Amplification: A Comparison of Various Polymerase-Buffer Systems with Ancient and Modern DNA Sequencing Libraries

High-throughput sequencing technologies frequently necessitate the use of PCR for sequencing library amplification. PCR is a sometimes enigmatic process and is known to introduce biases. Here we perform a simple amplification-sequencing assay using 10 commercially available polymerase-buffer systems to amplify libraries prepared from both modern and ancient DNA. We compare the performance of the polymerases with respect to a previously uncharacterized template length bias, as well as GC-content bias, and find that simply avoiding certain polymerase can dramatically decrease the occurrence of both. For amplification of ancient DNA, we found that some commonly used polymerases strongly bias against amplification of endogenous DNA in favor of GC-rich microbial contamination, in our case reducing the fraction of endogenous sequences to almost half.

Manchester Motorway Box: Post-Survey Research of Induced Traffic Effects

In the 1990s, the usual assumption for an appraisal of road schemes in the UK was that total volumes of traffic were not affected by the capacity provided by the schemes. This assumption was questioned by the influential SACTRA committee in 1994, which also recommended that Before and After studies be undertaken to quantify the scale of traffic that would be “induced” by the provision of road capacity. An opportunity to investigate this issue arose with the completion of the M60 Manchester Motorway Box, one of the last major links in the UK's national road network, and a large program of Before and After data collection was undertaken. The paper describes the analysis that was made of the Before and After data, to which household interview records were added to form a large database linked to modeled level-of-service data and land-use data. This combined dataset has been used to estimate disaggregate models that represent frequency, mode, destination and time-of-day choice decisions within a hierarchical structure. Time-of-day choice has been represented by distinguishing four time periods that cover a day, and modeling the choice between those four time periods. The use of a hierarchical structure allows the scale of the different behavioral effects to be measured in a parametric form and also allows the construction of a detailed (market segmented) travel demand model. A further aim of the analysis was to distinguish the induced traffic effects from any other changes that may have occurred. Analysis of the level-of-service data showed that the conventional assignment procedures used were not able to reproduce the observed changes in journey times between the Before and After situations. Models including mode, destination and time-of-day choices were estimated separately, using observed journey times where available, for intercept surveys (correcting for the trip length bias in that data), for household interview data and then for combined data. The values of time and elasticities implied by the models were found to be reasonable. Application of the models took into account the relevant changes in the population in the period between the Before and After observations. The models indicated that the M60 Scheme is likely to have induced traffic at the level of a 15–17% increase across the most relevant screenline counts, of which the majority were due to destination switching and less to mode shift. Time-of-day effects were found to be negligible, although in the M60 situation, journey time changes across time periods were broadly similar.

A stochastic model for survival of early prostate cancer with adjustments for leadtime, length bias, and over‐detection

To compare the survival between screen-detected and clinically detected cancers, we applied a series of non-homogeneous stochastic processes to deal with leadtime, length bias, and over-detection by using full information on detection modes obtained from the Finnish randomized controlled trial for prostate cancer screening. The results show after 9-year follow-up the hazard ratio of prostate cancer death for screen-detected cases against clinically detected cases increased from 0.24 (95% CI: 0.16-0.35) without correction for these biases, to 0.76 after correction for leadtime and length biases, and finally to 1.03 (95% CI: 0.79-1.33) for a further adjustment for over-detection. Adjustment for leadtime and length bias but no over-detection led to a 24% reduction in prostate cancer death as a result of prostate-specific antigen test. The further calibration of over-detection indicates no gain in survival of screen-detected prostate cancers (excluding over-detected case as stayer considered in the mover-stayer model) as compared with the control group in the absence of screening that is considered as the mover. However, whether the model assumption on over-detection is robust should be validated with other data sets and longer follow-up.

A bias-reducing pathway enrichment analysis of genome-wide association data confirmed association of the MHC region with schizophrenia

BackgroundAfter the recent successes of genome-wide association studies (GWAS), one key challenge is to identify genetic variants that might have a significant joint effect on complex diseases but have failed...

MIQuant – Semi-Automation of Infarct Size Assessment in Models of Cardiac Ischemic Injury

BackgroundThe cardiac regenerative potential of newly developed therapies is traditionally evaluated in rodent models of surgically induced myocardial ischemia. A generally accepted key parameter for determining the success of the applied therapy is the infarct size. Although regarded as a gold standard method for infarct size estimation in heart ischemia, histological planimetry is time-consuming and highly variable amongst studies. The purpose of this work is to contribute towards the standardization and simplification of infarct size assessment by providing free access to a novel semi-automated software tool. The acronym MIQuant was attributed to this application.Methodology/Principal FindingsMice were subject to permanent coronary artery ligation and the size of chronic infarcts was estimated by area and midline-length methods using manual planimetry and with MIQuant. Repeatability and reproducibility of MIQuant scores were verified. The validation showed high correlation (r midline length = 0.981; r area = 0.970 ) and agreement (Bland-Altman analysis), free from bias for midline length and negligible bias of 1.21% to 3.72% for area quantification. Further analysis demonstrated that MIQuant reduced by 4.5-fold the time spent on the analysis and, importantly, MIQuant effectiveness is independent of user proficiency. The results indicate that MIQuant can be regarded as a better alternative to manual measurement.ConclusionsWe conclude that MIQuant is a reliable and an easy-to-use software for infarct size quantification. The widespread use of MIQuant will contribute towards the standardization of infarct size assessment across studies and, therefore, to the systematization of the evaluation of cardiac regenerative potential of emerging therapies.

A Study of Gate Length and Source-Drain Bias on Electron Transport Properties in SiC Based MOSFETs Using Monte Carlo Method

Ensemble Monte Carlo simulations have been carried out to investigate the effects of Gate length and different source-drain bias on the characteristics of wurtzite SiC MOSFETs. Electronic states within the conduction band valleys are represented by non-parabolic ellipsoidal valleys centred on important symmetry points of the Brillouin zone. The following scattering mechanisims, i.e, impurity, polar optical phonon, acoustic phonon, alloy and piezoelectric are inculded in the calculation. Ionized imurity scattering has been treated beyound the Born approximation using the phase-shift analysis. Two transistors with gate lengths of 200 and 400 nm are simulated. Simulations show that with a fixed channel length, when the gate length is decreased, the output drain current is increased, and therefore the transistor transconductance increases. Moreover, with increasing temperature the drain current is reduced, which results in the reduced drain barrier lowering. The simulated device geometries and doping are matched to the nominal parameters described for the experimental structures as closely as possible, and the predicted drain current and other electrical characteristics for the simulated device show much closer agreement with the available experimental data. DOI: http://dx.doi.org/10.11591/ijece.v1i1.18

Mortality reduction in the Swedish mammography screening program.

29 Background: The introduction of a nationwide mammography screening program in Sweden in 1986 was associated with a 50% persistent increase of breast cancer incidence (including carcinoma in situ). Whether screening has been instrumental in a recent decline in the breast cancer mortality remains an open question. Methods: We compared the 18 years breast cancer mortality for two groups of women; those diagnosed in the first 6 years of the screening program and those diagnosed during a 6 year period that included the last 4 years before screening started and the first 2 years of the screening program. The test group (n + 328,927) included women aged 40 to 69 at the first invitation to screening. On average the women in the age-matched control group (n + 317,404) were born 4 years earlier then the women in the test group. The numbers of incident invasive breast cancers in each county were obtained from the Swedish Cancer Registry (in-situ cancers were excluded) and the corresponding breast cancer deaths were obtained from the Swedish Causes of Death Registry. Results: After 18 years follow up, the mortality reduction was 14% in the test group. However, most of this difference (10%) occurred in the first five years after diagnosis reflecting improved medical treatment of breast cancer. Most of the 14% difference also disappeared if we adjusting for period effects. Conclusions: About 83% of women in the test group were also included in the control group making the two groups almost identical with respect to risk variables. By screening the control group at the end of a 6-year period, the results are adjusted for differences due to lead-time bias and by studying mortality instead of survival overdiagnosis (length time bias) is also adjusted for. We conclude that organized mammography screening is not an important cause for the reduction in the breast cancer mortality in Sweden.

Digital gene expression for non-model organisms

Next-generation sequencing technologies offer new approaches for global measurements of gene expression but are mostly limited to organisms for which a high-quality assembled reference genome sequence is available. We present a method for gene expression profiling called EDGE, or EcoP15I-tagged Digital Gene Expression, based on ultra-high-throughput sequencing of 27-bp cDNA fragments that uniquely tag the corresponding gene, thereby allowing direct quantification of transcript abundance. We show that EDGE is capable of assaying for expression in >99% of genes in the genome and achieves saturation after 6-8 million reads. EDGE exhibits very little technical noise, reveals a large (10(6)) dynamic range of gene expression, and is particularly suited for quantification of transcript abundance in non-model organisms where a high-quality annotated genome is not available. In a direct comparison with RNA-seq, both methods provide similar assessments of relative transcript abundance, but EDGE does better at detecting gene expression differences for poorly expressed genes and does not exhibit transcript length bias. Applying EDGE to laboratory mice, we show that a loss-of-function mutation in the melanocortin 1 receptor (Mc1r), recognized as a Mendelian determinant of yellow hair color in many different mammals, also causes reduced expression of genes involved in the interferon response. To illustrate the application of EDGE to a non-model organism, we examine skin biopsy samples from a cheetah (Acinonyx jubatus) and identify genes likely to control differences in the color of spotted versus non-spotted regions.

Optimal estimating functions in incomplete data and length biased sampling data problems

AbstractIt is well known that the score function is the optimal estimating function among all regular unbiased estimating functions (Godambe, 1960). In the presence of incomplete data such as missing data or length biased sampling data, Horvitz and Thompson's (1952) method is an effective way of eliminating the possible bias induced by using complete data only. In this article, we show that the inverse weighted Horvitz and Thompson score estimating function is not optimal in the presence of incomplete data. By using Godambe's estimating function theory, we can identify the optimal estimating function in this situation. In the case of the accelerated failure time model with length bias sampling data, the optimal estimating function can produce an unbiased estimator for the slope parameter even when the underlying density function is misspecified. Simulation studies show that the estimate derived from the optimal estimating function can be substantially better than the estimate derived from the inverse weighted score estimating function. The Canadian Journal of Statistics 39: 510–518; 2011 © 2011 Statistical Society of Canada

Optical bistability in InGaN-based multisection laser diodes

Optical bistability is observed in cw-operating InGaN-based laser diodes including a saturable absorber (SA) section. The dependence of the light-current hysteresis on the SA section length and reverse bias (VSA) has been studied. An analytical approach is developed to estimate the carrier lifetime τa in the SA section from the measurements of the hysteresis width, which leads to τa=1.9 ns at zero bias. τa is found to decrease rapidly for higher reverse biases and a minimum of τa=0.4 ns is interpolated for flatband conditions. We explain the dependence of the carrier lifetime on VSA via the modification of the quantum-confined Stark effect.

Statistical approaches for the analysis of DNA methylation microarray data

Following the rapid development and adoption in DNA methylation microarray assays, we are now experiencing a growth in the number of statistical tools to analyze the resulting large-scale data sets. As is the case for other microarray applications, biases caused by technical issues are of concern. Some of these issues are old (e.g., two-color dye bias and probe- and array-specific effects), while others are new (e.g., fragment length bias and bisulfite conversion efficiency). Here, I highlight characteristics of DNA methylation that suggest standard statistical tools developed for other data types may not be directly suitable. I then describe the microarray technologies most commonly in use, along with the methods used for preprocessing and obtaining a summary measure. I finish with a section describing downstream analyses of the data, focusing on methods that model percentage DNA methylation as the outcome, and methods for integrating DNA methylation with gene expression or genotype data.

The proportion of oesophageal adenocarcinoma patients with prior barrett's oesophagus: results from a large population based cohort

IntroductionBarrett's oesophagus (BO) is a premalignant condition that predisposes to oesophageal adenocarcinoma (OAC). BO is difficult to detect due to the asymptomatic nature of the condition. Prognosis from OAC is...