Validation of an RNA cell cycle progression score for predicting death from prostate cancer in a conservatively managed needle biopsy cohort.

BACKGROUND: The natural history of prostate cancer is highly variable and difficult to predict accurately. Better markers are needed to guide management and avoid unnecessary treatment. In this study, we validate the prognostic value of a cell cycle progression score (CCP score) independently and in a prespecified linear combination with standard clinical variables, that is, a clinical-cell-cycle-risk (CCR) score. METHODS: Paraffin sections from 761 men with clinically localized prostate cancer diagnosed by needle biopsy and managed conservatively in the United Kingdom, mostly between 2000 and 2003. The primary end point was prostate cancer death. Clinical variables consisted of centrally reviewed Gleason score, baseline PSA level, age, clinical stage, and extent of disease; these were combined into a single predefined risk assessment (CAPRA) score. Full data were available for 585 men who formed a fully independent validation cohort. RESULTS: In univariate analysis, the CCP score hazard ratio was 2.08 (95% CI (1.76, 2.46), P<10(-13)) for one unit change of the score. In multivariate analysis including CAPRA, the CCP score hazard ratio was 1.76 (95% CI (1.44, 2.14), P<10(-6)). The predefined CCR score was highly predictive, hazard ratio 2.17 (95% CI (1.83, 2.57), χ(2)=89.0, P<10(-20)) and captured virtually all available prognostic information. CONCLUSIONS: The CCP score provides significant pretreatment prognostic information that cannot be provided by clinical variables and is useful for determining which patients can be safely managed conservatively, avoiding radical treatment.

A telephone reminder intervention to improve breast screening information and access.

OBJECTIVES: In the UK, women aged 50-70 are offered breast cancer screening every three years. Screening participation rates in London have been particularly low. Low rates have been associated with low socio-economic status, and some ethnic groups have been observed to be underserved by cancer screening. This paper reports on a telephone reminder intervention in London Newham, an area of high deprivation and ethnic diversity. STUDY DESIGN: Observational study of planned intervention. METHODS: Women invited for breast screening were telephoned to confirm receipt of the invitation letter, remind invitees of their upcoming appointment, and to provide further information. Aggregate data at general practice level on invitation to and attendance at breast screening and on numbers reached by telephone were analysed by logistic regression. RESULTS: For the 29 participating GP practices (10,928 invitees) overall uptake in 2010 was higher compared to the previous screening round in 2007 (67% vs. 51%; p < 0.001). On average 59% of invitees were reached by the reminder calls. A 10% increase in women reached resulted in an 8% increase in the odds of women attending their screening appointment (95% CI: 5%-11%), after adjusting for 2007 attendance rates. Practices with a higher proportion of South Asian women were associated with a larger uptake adjusted for 2007 uptake and population reached by the telephone intervention, (4% increase in odds of attendance per 10% increase in South Asian population, CI 1%-7%, p = 0.003) while practices with a higher proportion of black women were associated with a smaller uptake similarly adjusted. (11% decrease in odds of attendance per 10% increase in black population, CI 9%-16%, p < 0.001). CONCLUSIONS: A language- and culture-sensitive programme of reminder calls substantially improved breast cancer screening uptake.

A novel, externally validated inflammation-based prognostic algorithm in hepatocellular carcinoma: the prognostic nutritional index (PNI).

BACKGROUND: There is increasing evidence that the presence of an ongoing systemic inflammatory response is a stage-independent predictor of poor outcome in patients with cancer. The aim of this study was to investigate whether an inflammation-based prognostic score, the prognostic nutritional index (PNI), is associated with overall survival (OS) in patients with hepatocellular carcinoma (HCC). METHODS: All patients with a new diagnosis of HCC presenting to the Medical Oncology Department, Hammersmith Hospital between 1993 and 2011 (n=112) were included. Demographic and clinical data were collected. Patients in whom the combined albumin (g l(-1)) × total lymphocyte count × 10(9) l(-1) was ≥45, at presentation, were allocated a PNI score of 0. Patients in whom this total score was <45 were allocated a score of 1. Univariate and multivariate analyses were performed to identify clinicopathological variables associated with OS. Independent predictors of survival identified on multivariate analysis were validated in an independent, stage-matched cohort of 68 patients. RESULTS: Univariate analyses showed that PNI (P=0.003), intrahepatic spread (P<0.001), the presence of extrahepatic disease (P=0.006), portal vein thrombosis (P=0.02), tumour multifocality (P=0.003), alfa-fetoprotein >400 ng ml(-1) (P<0.001) and Barcelona Clinic Liver Cancer score (P<0.01) were all predictors of OS in the training set. Multivariate analysis revealed the PNI (P=0.05), presence of extrahepatic disease (P<0.001) and degree of intrahepatic spread (P<0.001) as independent predictors of worse OS in this population. The PNI retained independent prognostic value in the validation set (P<0.001). CONCLUSION: The presence of a systemic inflammatory response, as measured by the PNI, is an independent and externally validated predictor of poor OS in patients with HCC.

Simple indices of inflammation as predictors of death from cancer or cardiovascular disease in a prospective cohort after two decades of follow-up.

BACKGROUND: Previous studies have identified sub-clinical inflammation as a potential factor in the pathogenesis of cancer and cardiovascular disease (CVD) but the possibility that simple, readily measured indices of sub-clinical inflammation might predict both CVD and cancer has not been tested in the context of a single, prospective analysis. AIM: To evaluate simply measured indices of inflammation as long-term predictors of death from either cancer or CVD. DESIGN: Prospective open cohort study. METHODS: A total of 1192 white males received measurements of a range of risk markers including the inflammation indices white blood cell count (WBC), erythrocyte sedimentation rate (ESR) and serum globulin concentrations. Inflammation marker clustering was quantified as a factor-analysis-derived inflammation score and survival time to death from any cancer or CVD was modeled on baseline measures using the Cox proportional hazards model. RESULTS: A total of 1010 participants met inclusion criteria, of whom 94 died of cancer and 67 of CVD. Mean follow-up times among cases and survivors ranged from 18.2-21.9 years. Independently of established risk factors [age, body mass index (BMI), smoking, alcohol and exercise], WBC, ESR and globulin levels were all individually predictive of both cancer (hazard ratio 1.43, P = 0.002; 1.27, P = 0.02; 1.26, P = 0.02, respectively) and CVD mortality (1.29, P = 0.06; 1.43, P = 0.007; 1.50, P = 0.001). The inflammation score predicted both cancer mortality (1.35, P = 0.003) and CVD mortality (1.46, P = 0.002). Risks associated with high inflammation score were equivalent to and independent of smoking cigarettes for cancer or, for CVD, having a serum cholesterol concentration ≥6.2 mmol/l. CONCLUSIONS: Simple indices of inflammation predict death from cancer or CVD two decades later as strongly as smoking predicts cancer or cholesterol predicts CVD. Their measurement could contribute to evaluation of both cancer and CVD risk.

Investigation of the ability of haplotype association and logistic regression to identify associated susceptibility loci.

While finely spaced markers are increasingly being used in case-control association studies in attempts to identify susceptibility loci, not enough is yet known as to the optimal spacing of such markers, their likely power to detect association, the relative merits of single marker versus multimarker analysis, or which methods of analysis may be optimal. Some investigations of these issues have used markers simulated under different theoretical models of population evolution. However the HapMap project and other sources provide real datasets which can be used to obtain a more realistic view of the performance of these approaches. SNPs around APOE and from two HapMap regions were used to obtain information regarding linkage disequilibrium (LD) relationships between polymorphisms, and these real patterns of LD were used to simulate datasets such as would be obtained in case-control studies were these SNPs to influence susceptibility to disease. The datasets obtained were analysed using tests for heterogeneity of estimated haplotype frequencies and using logistic regression analyses in which only main effects from each marker were considered. All markers surrounding the putative susceptibility locus were analysed, using sets of either 1, 2, 3 or 4 markers at a time. Some markers within 150 kb of the susceptibility locus were able to detect association. At distances less than 100 kb there was no correlation between the distance from the susceptibility locus and the strength of evidence for association. When the average inter-locus spacing is 25 kb many loci would not be detected, while when the spacing is as low as 2 kb one can be fairly confident that at least one marker will be in strong enough LD with the susceptibility locus to enable association to be detected, if the susceptibility locus has a strong enough effect relative to the sample size. With an inter-locus spacing of 4 kb some susceptibility loci did not have a marker locus in strong LD, potentially undermining the ability to detect association. There was little difference in the performance of haplotype-based analysis compared with logistic regression considering effects of each marker as separate. Multimarker analysis on occasion produced results which were much more highly significant than single marker analysis, but only very rarely. Our results support the view that if markers are randomly selected then a spacing as low as 2 kb is desirable. Multimarker analysis can sometimes be more powerful than single marker analysis so both should be performed. However, because it is rare for multimarker analysis to be much more highly significant than single marker analysis one should strongly suspect that when such results occur they may be due to mistakes in genotyping or through some other artefact. Haplotype analysis may be more prone to such problems than logistic regression, suggesting that the latter method might be preferred.

The EIF2AK3 gene region and type I diabetes in subjects from South India.

Mutations in the EIF2AK3 gene underlie susceptibility to the Wolcott-Rallison syndrome, which is a monogenic disease associated with insulin-deficient neonatal diabetes. Furthermore, suggestive evidence of linkage between type 1 diabetes (T1DM) and the EIF2KA3 chromosomal region has been reported in Scandinavian families. We have investigated the hypothesis that polymorphic variants in and around the EIF2AK3 gene might partially account for susceptibility to T1DM in South Indian subjects. Excess transmission of the common alleles of two polymorphic markers (D2S1786 and 15INDEL, located within the gene) downstream of EIF2AK3, either singly (D2S1786, P = 0.01) and 15INDEL (P = 0.02) or as a combination (P < 0.001), were found in 234 families with a T1DM proband. There was also a clear paternal effect for the 15INDEL marker (P = 0.005) on disease susceptibility. The presence of the common allele of both markers was found in decreased frequency in the subjects with normal glucose tolerance compared to probands with T1DM (both P

Further investigation of linkage disequilibrium SNPs and their ability to identify associated susceptibility loci.

There is currently considerable interest in the use of single-nucleotide polymorphisms (SNPs) to map disease susceptibility genes. The success of this method will depend on a number of factors including the strength of linkage disequilibrium (LD) between marker and disease loci. We used a data set of SNP genotypings in the region of the APOE disease susceptibility locus to investigate the likely usefulness of SNPs in case-control studies. Using the estimated haplotype structure surrounding and including the APOE locus, and assuming a codominant disease model, we treated each SNP in turn as if it were a disease susceptibility locus and obtained, for each disease locus and markers, the expected likelihood ratio test (LRT) to assess disease association. We were particularly interested in the power to detect association with the susceptibility polymorphism itself, the power of nearby markers to detect association, and the ability to distinguish between the susceptibility polymorphism and marker loci also showing association. We found that the expected LRT depended critically on disease allele frequencies. For disease loci with a reasonably common allele we were usually able to detect association. However, for only a subset of markers in the close neighbourhood of the disease locus was association detectable. In these cases we were usually, but not always, able to distinguish the disease locus from nearby associated marker loci. For some disease loci, no other loci demonstrated detectable association with the disease phenotype. We conclude that one may need to use very dense SNP maps in order to avoid overlooking polymorphisms affecting susceptibility to a common phenotype.

Assessing optimal neural network architecture for identifying disease-associated multi-marker genotypes using a permutation test, and application to calpain 10 polymorphisms associated with diabetes.

Biallelic markers, such as single nucleotide polymorphisms (SNPs), provide greater information for localising disease loci when treated as multilocus haplotypes, but often haplotypes are not immediately available from multilocus genotypes in case-control studies. An artificial neural network allows investigation of association between disease phenotype and tightly linked markers without requiring haplotype phase and without modelling any evolutionary history for the disease-related haplotypes. The network assesses whether marker haplotypes differ between cases and controls to the extent that classification of disease status based on multi-marker genotypes is achievable. The network is "trained" to "recognise" affection status based on supplied marker genotypes, and then for each multi-marker genotype it produces outputs which aim to approximate the associated affection status. Next, the genotypes are permuted relative to affection status to produce many random datasets and the process of training and recording of outputs is repeated. The extent to which the ability to predict affection for the real dataset exceeds that for the random datasets measures the statistical significance of the association between multi-marker genotype and affection. This permutation test performs well with simulated case-control datasets, particularly when major gene effects are present. We have explored the effects of systematically varying different network parameters in order to identify their optimal values. We have applied the permutation test to 4 SNPs of the calpain 10 (CAPN10) gene typed in a case-control sample of subjects with type 2 diabetes, impaired glucose tolerance, and controls. We show that the neural network produces more highly significant evidence for association than do single marker tests corrected for the number of markers genotyped. The use of a permutation test could potentially allow conditional analyses which could incorporate known risk factors alongside marker genotypes. Permuting only the marker genotypes relative to affection status and these risk factors would allow the contribution of the markers to disease risk to be independently assessed.

A quick and simple method for detecting subjects with abnormal genetic background in case-control samples.

It is important that case-control samples be drawn from a genetically homogeneous population in order to avoid artefactual false positive results and to enhance power to detect disease mutations and markers in linkage disequilibrium with them. Tests which simply compare overall marker allele frequencies between cases and controls will fail to identify a relatively small number of subjects drawn from a different genetic background who could usefully be discarded from the sample. Such subjects can be identified using multilocus tests, but previously described tests have been unnecessarily complex and cumbersome for this simple application. We describe a straightforward test, implemented in the CHECKHET program, which uses a measure of genetic difference and permutation procedures to rapidly identify such subjects using genotypes from multiple unlinked markers. It seems to perform reasonably well on simulated data, and with real data appears to identify two abnormal subjects within a case-control sample. We recommend that such tests be routinely applied to case-control samples once sufficient numbers of markers have been genotyped within them.

Extension of conditional model-free likelihood-based linkage analysis to additive and other models.

We have previously described extending our method of 'model-free' linkage analysis, implemented in the MFLINK program, in order to deal with liability classes. This allows a new form of conditional two-locus linkage analysis, meaning that the genotypes of a known risk locus can be used to define liability classes so that their effects can be incorporated in tests for linkage at additional loci. In this method, relationships between transmission models for different liability classes were constrained so that there was a constant multiplicative effect on penetrance values. Here we present further extensions to the method to allow for different relationships. In particular, rather than only having a multiplicative effect on risk of affection we now allow specification of a multiplicative effect on risk of non-affection, or a combination of both relationships, across liability classes. We now also allow specification of an additive effect on penetrance. By way of example, we apply these methods to genome scan data for Alzheimer's disease using apolipoprotein E genotype to define liability classes. We show that, although in general the different methods produce results which tend to be quite highly correlated, certain markers can produce quite different results according to the method applied and that these could well lead to differences of interpretation. Without knowing a priori which relationship is likely to be most appropriate to describe the overall combined effect of the two loci one might be obliged to apply a number of different methods. This in turn may lead to the familiar difficulties associated with multiple testing. Nevertheless, the new method allows researchers greater flexibility in analysing linkage data for diseases in which one or more risk polymorphisms have already been identified.

Genetic susceptibility to fibrocalculous pancreatic diabetes in Bangladeshi subjects: a family study.

Fibrocalculous pancreatic diabetes (FCPD) is an uncommon cause of diabetes, seen mainly in developing countries. A family-based study was carried out in 67 Bangladeshi families, consisting of a proband with FCPD and both parents, to determine whether an association exists between FCPD susceptibility and either the major histocompatiblity complex (MHC) or insulin gene (INS) loci. HLA-DQB1 typing was done using allele-specific primers, and INS was typed using the restriction enzyme HphI. Three microsatellites (TNFa, TNFc and TNFd), from within and flanking the TNF-LT locus, were used for MHC Class IV typing and a PCR-RFLP assay was used to define the -308G/A TNF promoter polymorphism. The extended transmission disequilibrium test (ETDT) was used for statistical analysis. An overall association was observed between FCPD and HLA-DQB1 (P = 0.003), that was largely due to a positive association with HLA-DQB1*0302 and a negative association with HLA-DQB1*0202. Although no association was found between FCPD and TNF-LT microsatellite markers a trend was observed for TNFc (P = 0.037, Pc = 0.15). No association was found between FCPD and INS (P = 0.26). This study confirms an association between FCPD and the MHC using a family-based study design and the stringent ETDT analysis; a novel protective association was found with HLA-DQB1*0202 in Bangladeshi FCPD subjects. The genetic susceptibility to FCPD has features both similar and dissimilar to T1DM.

Use of an artificial neural network to detect association between a disease and multiple marker genotypes.

Single nucleotide polymorphisms (SNPs) are very common throughout the genome and hence are potentially valuable for mapping disease susceptibility loci by detecting association between SNP markers and disease. However as SNPs are biallelic they may have relatively little power in association studies compared with the information that would be obtainable if marker haplotypes were available and could be used efficiently. Modelling the evolutionary events leading to linkage disequilibrium is very complex and many methods that seek to use information from multiple markers simultaneously need to make simplifying assumptions and may only be applicable when marker haplotypes, rather than genotypes, are available for analysis. We explore the properties of a simple application of a standard artificial neural network to this problem. The pattern-recognition properties of the network are used in the hope that marker haplotypes implicit in the genotypes will differ between cases and controls in a way which will lead to the network being able to classify the subjects correctly, according to their marker genotype. This method makes no assumptions at all regarding population history or the marker map, and can be applied to genotypes, as would be available from a simple case-control sample, without any need to determine haplotypes. Through application to data simulated under a very wide range of assumptions we show that such an analysis produces a useful augmentation in power above that which would be achieved by testing each marker individually, in particular when more than one mutation has occurred in a disease gene at different points in evolution. The application of neural networks to such problems shows considerable promise and further work could usefully be directed towards optimising the design and implementation of such networks.

A novel method of two-locus linkage analysis applied to a genome scan for late onset Alzheimer's disease.

A number of methods have previously been described which carry out linkage analysis considering information for two or more loci simultaneously. Apart from some ad hoc methods such as analysing subsamples, these methods use information regarding linkage at all loci under consideration. However, if the actual genotype-specific effects are known for some loci then it would be preferable to consider the genotypes of these loci directly, rather than the amount of allele-sharing they demonstrate. Here we present an extension to our likelihood-based method of model-free linkage analysis as implemented in the MFLINK program. This allows the incorporation of liability classes. The genotypes of a locus known to affect risk can be used to assign subjects to liability classes prior to carrying out linkage tests at other loci. An example application is presented for genome scan data on Alzheimer's disease with analysis conditional on Apoliprotein E (APOE) genotypes. The results provide support for the existence of additional susceptibility loci linked to D10S1211 and to D12S358.