Abstract GS3-06: Dynamics of breast cancer relapse reveal molecularly defined late recurring ER-positive subgroups: Results from the METABRIC study

Background: Recent studies have demonstrated that women with early stage ER-positive (ER+) and HER2-negative (HER2-) breast cancer have a persistent risk of recurrence and cancer related death up to 20 years post diagnosis, highlighting the chronic nature of ER+ breast cancer and critical need to identify tumor characteristics that are more predictive of risk of recurrence than standard clinical covariates. However, progress in delineating the dynamics of breast cancer relapse and biomarkers of late recurrence has been hindered by the lack of large cohorts with long-term clinical follow-up and molecular information.

Methods: We report the results of a cohort of 3,240 breast cancer patients from the United Kingdom and Canada with 20 years of follow-up (median 9.75 years), including 1,980 with accompanying molecular data from the primary breast tumor. Information for each patient on loco-regional recurrence (LR), distant recurrence (DR), and site(s) of metastases was collected. We developed a non-homogenous Markov chain model that accounted for different clinical endpoints and timescales, as well as competing risks of mortality and the distinct baseline hazards that characterize different molecular subgroups. This approach enabled robust analysis of the spatio-temporal dynamics of breast cancer recurrence across the clinical subgroups, PAM50 subgroups and the integrative clusters, while also enabling individual risk of relapse predictions.

Results: We employed our multistate model to compute the probability of experiencing a LR or DR, as well as the baseline transition probabilities from surgery, LR or DR at various time intervals for average individuals in each of the clinical/molecular subgroups. These analyses reveal four late-recurring ER+ (predominantly HER2-) subgroups, together accounting for 26% of all ER+ tumors, with high (median 42-55%) risk of recurrence up to 20 years post-diagnosis. Each of these four subgroups maps to one of the Integrative Clusters, defined based on genomic copy number alterations and gene expression, and is enriched for a characteristic copy number amplification events: 11q13 (CCND1, RSF1), 8p12 (FGFR1, ZNF703), 17q23 (RPS6KB1) and 8q24 (MYC). These four molecular subgroups are superior in predicting late DR than standard clinical variables.

Conclusions: A detailed understanding of the rates and routes of metastasis and their variability across the distinct molecular subtypes is essential for devising personalized approaches to breast cancer care. We describe a molecularly characterized breast cancer cohort with long-term clinical follow-up and a statistical modeling framework, enabling delineation of the dynamics of breast cancer recurrence at unprecedented resolution. These analyses reveal four late recurring ER+ subgroups and accompanying biomarkers that collectively define the quarter of ER+ cases at highest risk of recurrence. Our findings highlight opportunities for improved patient stratification and biomarker-driven clinical trials directed at the subset of breast cancer patients with persistent risk of recurrence.

Citation Format: Curtis C, Rueda OM, Sammut S-J, Chin S-F, Caswell-Jin JL, Seoane JA, Callari M, Batra R, Pereira B, Bruna A, Ali HR, Provenzano E, Liu B, Parisien M, Gillett C, McKinney S, Green A, Murphy L, Purushotham A, Ellis I, Pharoah P, Rueda C, Aparicio S, Caldas C. Dynamics of breast cancer relapse reveal molecularly defined late recurring ER-positive subgroups: Results from the METABRIC study [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr GS3-06.

An evaluation of the prognostic model PREDICT using the POSH cohort of women aged ⩽40 years at breast cancer diagnosis

BACKGROUND

Breast cancer is the most common cancer in younger women (aged ⩽40 years) in the United Kingdom. PREDICT (http://www.predict.nhs.uk) is an online prognostic tool developed to help determine the best available treatment and outcome for early breast cancer. This study was conducted to establish how well PREDICT performs in estimating survival in a large cohort of younger women recruited to the UK POSH study.

METHODS

The POSH cohort includes data from 3000 women aged ⩽40 years at breast cancer diagnosis. Study end points were overall and breast cancer-specific survival at 5, 8, and 10 years. Evaluation of PREDICT included model discrimination and comparison of the number of predicted versus observed events.

RESULTS

PREDICT provided accurate long-term (8- and 10-year) survival estimates for younger women. Five-year estimates were less accurate, with the tool overestimating survival by 25% overall, and by 56% for patients with oestrogen receptor (ER)-positive tumours. PREDICT underestimated survival at 5 years among patients with ER-negative tumours.

CONCLUSIONS

PREDICT is a useful tool for providing reliable long-term (10-year) survival estimates for younger patients. However, for more accurate short-term estimates, the model requires further calibration using more data from young onset cases. Short-term prediction may be most relevant for the increasing number of women considering risk-reducing bilateral mastectomy.

The malignant phenotype in breast cancer is driven by eIF4A1-mediated changes in the translational landscape

Human mRNA DeXD/H-box helicases are ubiquitous molecular motors that are required for the majority of cellular processes that involve RNA metabolism. One of the most abundant is eIF4A, which is required during the initiation phase of protein synthesis to unwind regions of highly structured mRNA that would otherwise impede the scanning ribosome. Dysregulation of protein synthesis is associated with tumorigenesis, but little is known about the detailed relationships between RNA helicase function and the malignant phenotype in solid malignancies. Therefore, immunohistochemical analysis was performed on over 3000 breast tumors to investigate the relationship among expression of eIF4A1, the helicase-modulating proteins eIF4B, eIF4E and PDCD4, and clinical outcome. We found eIF4A1, eIF4B and eIF4E to be independent predictors of poor outcome in ER-negative disease, while in contrast, the eIF4A1 inhibitor PDCD4 was related to improved outcome in ER-positive breast cancer. Consistent with these data, modulation of eIF4A1, eIF4B and PCDC4 expression in cultured MCF7 cells all restricted breast cancer cell growth and cycling. The eIF4A1-dependent translatome of MCF7 cells was defined by polysome profiling, and was shown to be highly enriched for several classes of oncogenic genes, including G-protein constituents, cyclins and protein kinases, and for mRNAs with G/C-rich 5'UTRs with potential to form G-quadruplexes and with 3'UTRs containing microRNA target sites. Overall, our data show that dysregulation of mRNA unwinding contributes to the malignant phenotype in breast cancer via preferential translation of a class of genes involved in pro-oncogenic signaling at numerous levels. Furthermore, immunohistochemical tests are promising biomarkers for tumors sensitive to anti-helicase therapies.

Implementing risk-stratified screening for common cancers: a review of potential ethical, legal and social issues

BACKGROUND

The identification of common genetic variants associated with common cancers including breast, prostate and ovarian cancers would allow population stratification by genotype to effectively target screening and treatment. As scientific, clinical and economic evidence mounts there will be increasing pressure for risk-stratified screening programmes to be implemented.

METHODS

This paper reviews some of the main ethical, legal and social issues (ELSI) raised by the introduction of genotyping into risk-stratified screening programmes, in terms of Beauchamp and Childress's four principles of biomedical ethics--respect for autonomy, non-maleficence, beneficence and justice. Two alternative approaches to data collection, storage, communication and consent are used to exemplify the ELSI issues that are likely to be raised.

RESULTS

Ultimately, the provision of risk-stratified screening using genotyping raises fundamental questions about respective roles of individuals, healthcare providers and the state in organizing or mandating such programmes, and the principles, which underpin their provision, particularly the requirement for distributive justice.

CONCLUSIONS

The scope and breadth of these issues suggest that ELSI relating to risk-stratified screening will become increasingly important for policy-makers, healthcare professionals and a wide diversity of stakeholders.

Stratified cancer screening: the practicalities of implementation

BACKGROUND

Improving understanding of the genetic basis of disease susceptibility enables us to estimate individuals' risk of developing cancer and offer them disease prevention, including screening, stratified to reflect that risk. Little attention has so far been given to the implementation of stratified screening. This article reviews the issues that would arise in delivering such tailored approaches to prevention in practice.

RESULTS

Issues analysed include the organisational context within which implementation of stratified prevention would occur, how the offer of screening would be made, making sure consent is adequately informed, how individuals' risk would be assessed, the age at which risk estimation should occur, and the potential use of genetic data for other purposes. The review also considers how management might differ depending on individuals' risk, how their results would be communicated and their follow-up arranged, and the different issues raised by modification of an existing screening programme, such as that for breast cancer, and the establishment of a new one, for example for prostate cancer.

CONCLUSION

Stratified screening based on genetic testing is a radically new approach to prevention. Various organisational issues would need to be considered before it could be introduced, and a number of questions require further research.

Prognosis by breast cancer subtypes in patients treated with adjuvant chemotherapy in a clinical trial

12

Background: Breast cancer can be classified into molecular subtypes that have distinct survival patterns. The purpose of this study was i) to evaluate the prognostic significance of breast cancer subtypes in a cohort of women taking part in the NEAT and BR9601 clinical trials comparing CMF with ECMF, and ii) to evaluate whether the subtypes were predictive of the added benefit of epirubicin in these trials.

Methods: Tumor tissue microarrays were stained and scored for ER, PR, HER2, EGFR and CK5/6. These were used to classify the tumors into six intrinsic subtypes (1). We used Cox regression to compare overall survival (OS), breast cancer specific survival (BSS) and relapse free survival (RFS) in the different subgroups. We also compared the effect of ECMF with CMF by subgroup.

Results: IHC data were available for 1725 cases of whom 805 were Luminal 1-basal negative, 153 were Luminal 1-basal positive, 174 were Luminal 2, 192 were HER2-like, 230 were core basal phenotype and 171 were 5-negative phenotype. Median follow-up time was 7 years. The prognostic effects of the subtypes were similar to those reported for unselected breast cancer cases irrespective of adjuvant therapy (Blows FM, et al. PLoS Med 2010;75:e1000279.). In particular, the luminal 1-basal negative tumors were associated with the best prognosis in five years after surgery and the HER2-like tumors were associated with the poorest prognosis. ECMF has previously shown to be associated with a 33% relative risk reduction for OS compared to CMF (Poole CJ et al. N Engl J Med 2006;35518:1851-62.). There was little evidence for significant heterogeneity of effect by tumor subtype for any end point (OS P= 0.40, BSS P=0.53 RFS P=0.50). However, there was an observed trend towards the largest additional benefit from ECMF being in women with tumors of the 5-negative phenotype (OS HR=0.39 95% CI 0.21-0.73) and the smallest being in Luminal 1-basal negative tumors (OS HR=0.86 95% CI 0.64-1.16).

Conclusions: In a clinical trial in which all patients received chemotherapy, we confirmed that breast cancer subtypes show distinct behaviour with differences in short and long term survival. The benefit of ECMF over CMF was statistically similar in all disease subtypes.

Patient and tumour characteristics, management, and age-specific survival in women with breast cancer in the East of England

BACKGROUND

Breast cancer relative survival (BCRS), which compares the observed survival of women with breast cancer with the expected survival of women for the whole population of the same age, time period, and geographical region, tends to be poorer in older women, but the reasons for this are not clear. We examined the influence of patient and tumour characteristics, and treatment on BCRS to see whether these could explain the age-specific effect.

METHODS

Data for 14,048 female breast cancer patients diagnosed from 1999 to 2007, aged 50 years or over were obtained from the Eastern Cancer Registration and Information Centre. We estimated relative 5- and 10-year survival for patients in four age groups (50-69, 70-74, 75-79, and 80+ years). We also modelled relative excess mortality (REM) rate using Poisson regression adjusting for patient characteristics and treatment. The REMs derived from these models quantify the extent to which the hazard of death differs from the hazard in the reference category, after taking into account the background risk of death in the general population. We compared the results with those obtained for breast cancer-specific mortality, analysed using multivariate Cox regression.

RESULTS

Median follow-up time was 4.7 years. Relative 5-year survival was 89, 81, 76, and 70% for patients aged 50-69, 70-74, 75-79, and 80+ years, respectively. Corresponding relative 10-year survival was 84, 77, 67, and 66%. Unadjusted REM was 1.93, 2.74, and 3.88 for patients aged 70-74, 75-79, and 80+ years, respectively, (50-69 years as reference). The equivalent hazard ratios from the Cox model were 1.88, 2.45, and 3.81. These were attenuated after adjusting for confounders (REM - 1.49, 1.36, and 1.23; Cox - 1.47, 1.50, and 1.76).

CONCLUSION

We confirmed poorer BCRS in older women in our region. This was partially explained by known prognostic factors. Further research is needed to determine whether biological differences or suboptimal management can explain the residual excess mortality.

BCL2 in breast cancer: a favourable prognostic marker across molecular subtypes and independent of adjuvant therapy received

BACKGROUND

Breast cancer is heterogeneous and the existing prognostic classifiers are limited in accuracy, leading to unnecessary treatment of numerous women. B-cell lymphoma 2 (BCL2), an antiapoptotic protein, has been proposed as a prognostic marker, but this effect is considered to relate to oestrogen receptor (ER) status. This study aimed to test the clinical validity of BCL2 as an independent prognostic marker.

METHODS

Five studies of 11 212 women with early-stage breast cancer were analysed. Individual patient data included tumour size, grade, lymph node status, endocrine therapy, chemotherapy and mortality. BCL2, ER, progesterone receptor (PR) and human epidermal growth factor receptor 2 (HER2) levels were determined in all tumours. A Cox model incorporating the time-dependent effects of each variable was used to explore the prognostic significance of BCL2.

RESULTS

In univariate analysis, ER, PR and BCL2 positivity was associated with improved survival and HER2 positivity with inferior survival. For ER and PR this effect was time dependent, whereas for BCL2 and HER2 the effect persisted over time. In multivariate analysis, BCL2 positivity retained independent prognostic significance (hazard ratio (HR) 0.76, 95% confidence interval (CI) 0.66-0.88, P<0.001). BCL2 was a powerful prognostic marker in ER- (HR 0.63, 95% CI 0.54-0.74, P<0.001) and ER+ disease (HR 0.56, 95% CI 0.48-0.65, P<0.001), and in HER2- (HR 0.55, 95% CI 0.49-0.61, P<0.001) and HER2+ disease (HR 0.70, 95% CI 0.57-0.85, P<0.001), irrespective of the type of adjuvant therapy received. Addition of BCL2 to the Adjuvant! Online prognostic model, for a subset of cases with a 10-year follow-up, improved the survival prediction (P=0.0039).

CONCLUSIONS

BCL2 is an independent indicator of favourable prognosis for all types of early-stage breast cancer. This study establishes the rationale for introduction of BCL2 immunohistochemistry to improve prognostic stratification. Further work is now needed to ascertain the exact way to apply BCL2 testing for risk stratification and to standardise BCL2 immunohistochemistry for this application.

Common germ-line polymorphism of C1QA and breast cancer survival

BACKGROUND

A synonymous single nucleotide polymorphism (SNP) rs172378 (A>G, Gly->Gly) in the complement component C1QA has been proposed to be associated with distant breast cancer metastasis. We previously reported overexpression of this gene to be significantly associated with better prognosis in oestrogen-receptor-negative tumours. The purpose of this study was to investigate the association of rs172378 with expression of C1QA and breast cancer survival.

METHODS

We analysed the gene expression pattern of rs172378 in normal and tumour tissue samples, and further explored its involvement in relation to mortality in 2270 women with breast cancer participating in Studies of Epidemiology and Risk factors in Cancer Heredity, a population-based case-control study.

RESULTS

We found that although rs172378 showed differential allelic expression significantly different between normal (preferentially expressing the G allele) and tumour tissue samples (preferentially expressing the A allele), there was no significant difference in survival by rs172378 genotype (per allele hazard ratio (HR) 1.02, 95% CI: 0.88-1.19, P=0.78 for all-cause mortality; HR 1.03, 95% CI: 0.87-1.22, P=0.72 for breast-cancer-specific mortality).

CONCLUSION

Our study results show that rs172378 is linked to a cis-regulatory element affecting gene expression and that allelic preferential expression is altered in tumour samples, but do not support an association between genetic variation in C1QA and breast cancer survival.

Adjuvant UK: A UK Prognostic Model for Early Breast Cancer That Includes Mode of Detection

AIM: To develop and validate a prognostication model to predict overall survival for women treated for early breast cancer in the UK based on cancer registry data. Unlike SEER, this dataset includes accurate information on mode of detection as well as systemic therapy.METHOD: Using the Eastern Cancer Registration & Information Centre (ECRIC) dataset, information was collated for 5818 women diagnosed with invasive breast cancer in East Anglia from 1999-2003. All patients underwent surgery, had records of pathological staging (tumour size, grade, lymph node status, ER status), systemic treatment (chemotherapy, hormone therapy, both), mode of detection (screen-detected, symptomatic) and follow up, censored on 31 December 2007. A model was derived from these data using Cox proportional hazards, with ER positive and ER negative tumours modelled separately, and this was subsequently validated in an external dataset of 5468 patients from the West Midlands Cancer Intelligence Unit (WMCIU). Validation was performed by comparing (a) observed and expected mortality (overall & breast cancer specific) at 5 & 8 years and (b) receiver operating characteristic (ROC) curves in both ECRIC & WMCIU datasets.RESULTS:ECRIC datasetDifference in overall observed/expected mortality &lt;1% at 5 years (14.8% vs 14.9%) and 8 years (18.9% vs 18.9%). Area under ROC curve (AUC) was 0.81.Difference in breast cancer specific observed/expected mortality &lt;1% at 5 years (10.6% vs 11.0%) and 8 years (12.9% vs 13.5%). AUC was 0.84.WMCIU dataset:Difference in overall observed/expected mortality &lt; 1% at 5 years (15.8% vs 16.5%) and 8 years (17.5% vs 17.8%). AUC was 0.79.Difference in breast cancer specific observed/expected mortality &lt;2% at 5 years (11.0% vs 12.6%) and 8 years (12.2% vs 13.6%). AUC was 0.82.Overall model fit was good across all subgroups although the ER positive model provided better discrimination (AUC 0.82) than ER negative (AUC 0.75). There was no significant difference between the ROC curves generated with ECRIC and WMCIU data (ER positive X2 = 0.17, p=0.68; ER negative X2 =0.00, P=0.95).CONCLUSION: We have developed a prognostication model for early breast cancer based on data from a UK cancer registry that has included mode of detection for the first time. The model is well calibrated, provides a high degree of discrimination and has been validated in a second patient cohort. This model will underpin a new web-based prognostication and treatment benefit tool for early breast cancer based on UK data (Adjuvant UK).

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 4033.

Adjuvant UK: Validation of a UK prognostic model for early breast cancer

11086

Background: To develop and validate a prognostication model to predict overall survival for women treated for early breast cancer in the UK based on cancer registry data. The main advantages are that unlike SEER, this dataset includes accurate information on mode of detection as well as systemic therapy. Methods: Using the Eastern Cancer Registration & Information Centre (ECRIC) dataset, information was collated for 5,818 women diagnosed with invasive breast cancer in East Anglia from 1999–2003. All patients underwent surgery, had records of pathological staging (tumour size, grade, lymph node status, ER status) and systemic treatment (chemotherapy, hormone therapy, both), mode of detection (screen-detected, symptomatic) and were followed up to December 2007. A model was derived from these data using Cox proportional hazards and this was subsequently validated in an external dataset of 5,468 patients from the West Midlands Cancer Intelligence Unit (WMCIU). Validation was performed by comparing (a) observed and predicted mortality - overall and by risk group decile - and (b) receiver operating characteristic (ROC) curves in both ECRIC & WMCIU datasets. Results: ECRIC dataset: Observed/predicted deaths were: 2 years (262/267), 5 years (868/867), 8 years (1108/1093). Area under ROC curve was 0.800WMCIU dataset: Observed/predicted deaths were: 2 years (317/298), 5 years (862/919), 8 years (955/999). Area under ROC curve was 0.783. The model and validation worked equally well at all time points (2, 5, 8 years) and in all risk group deciles. In the validation dataset the observed and predicted 8-year OS was within 1% (17.5% versus 18.3%). There was no significant statistical difference between the ROC curves (X2 = 2.35, P=.12) for prediction at 8 years in ECRIC and WMCIU data. Conclusions: We have developed a prognostication model for early breast cancer based on data from a UK cancer registry and validated the model using data from another UK registry. The model was well calibrated and provides a high degree of discrimination. This model will underpin the first web-based prognostication and treatment benefit tool for early breast cancer in the UK (Adjuvant UK).

No significant financial relationships to disclose.

Mean sojourn time, overdiagnosis, and reduction in advanced stage prostate cancer due to screening with PSA: implications of sojourn time on screening

This study aimed to assess the mean sojourn time (MST) of prostate cancer, to estimate the probability of overdiagnosis, and to predict the potential reduction in advanced stage disease due to screening with PSA. The MST of prostate cancer was derived from detection rates at PSA prevalence testing in 43 842 men, aged 50–69 years, as part of the ProtecT study, from the incidence of non-screen-detected cases obtained from the English population-based cancer registry database, and from PSA sensitivity obtained from the medical literature. The relative reduction in advanced stage disease was derived from the expected and observed incidences of advanced stage prostate cancer. The age-specific MST for men aged 50–59 and 60–69 years were 11.3 and 12.6 years, respectively. Overdiagnosis estimates increased with age; 10–31% of the PSA-detected cases were estimated to be overdiagnosed. An interscreening interval of 2 years was predicted to result in 37 and 63% reduction in advanced stage disease in men 65–69 and 50–54 years, respectively. If the overdiagnosed cases were excluded, the estimated reductions were 9 and 54%, respectively. Thus, the benefit of screening in reducing advanced stage disease is limited by overdiagnosis, which is greater in older men.

Proliferation markers and survival in early breast cancer: a systematic review and meta-analysis of 85 studies in 32,825 patients

We have performed a systematic review and meta-analysis of proliferation markers (Ki-67, mitotic index (MI), proliferating cell nuclear antigen (PCNA) and thymidine or bromodeoxyuridine labelling index (LI)) with respect to survival in early breast cancer. Eighty-five studies involving 32,825 patients were analysed. Ki-67 (43 studies, 15,790 patients), MI (20 studies, 7021 patients), and LI (11 studies, 7337 patients) were associated with significantly shorter overall and disease free survival, using results from univariate and multivariate analyses from the individual studies. PCNA (11 studies, 2677 patients) was associated with shorter overall survival by multivariate analysis only, because of lack of data. There was some evidence for publication bias, but all markers remained significant after allowing for this. Ki-67, MI, PCNA and LI are associated with worse survival outcomes in early breast cancer. However, whether these proliferation markers provide additional prognostic information to commonly used prognostic indices remains unclear.

Issues of consent and feedback in a genetic epidemiological study of women with breast cancer

Women (N=21) who had had breast cancer and had been enrolled in a large genetic breast cancer epidemiological study were interviewed about their experience of participation in the study, their attitudes to the confidentiality of data, and the feedback of personal and general research results. Collection of family history information seemed more salient in indicating the genetic nature of the study than the enrolment information sheet. There were no concerns about confidentiality. While participants would have welcomed general feedback about the results of the study and were critical that this had not been provided, the feedback of personal information proved complicated and, sometimes, difficult. It is suggested that individual feedback of genetic test information in epidemiological studies should be undertaken only when there are specific reasons.

Interleukin 1B and interleukin 1RN polymorphisms are associated with increased risk of gastric carcinoma

BACKGROUND & AIMS

Interleukin (IL)-1 gene cluster proinflammatory polymorphisms have been associated with development of gastric atrophy and with increased risk of gastric carcinoma. We aimed to determine the association between IL-1 loci polymorphisms and increased risk of gastric carcinoma in samples from a Portuguese population, and to find whether there was any relationship with the histologic types of gastric carcinoma.

METHODS

In a case-control study including 220 controls and 152 patients with gastric carcinoma (intestinal, 76; diffuse, 37; and atypical, 39), both the IL-1B-511 biallelic polymorphism and the IL-1RN penta-allelic variable number of tandem repeats were genotyped.

RESULTS

We found a significant association between the IL-1 polymorphisms and increased risk for tumor development in patients with intestinal-type gastric carcinoma. A trend towards an increased risk of tumor development was also observed in patients with diffuse-type gastric carcinoma. No significant relationship was observed in patients with atypical carcinoma. Carriers of IL-1B-511T and IL-1RN*2 homozygotes had increased risk for developing intestinal-type gastric carcinoma with odds ratios of 2.7 (95% confidence interval [CI], 1.5-4.9) and 3.1 (95% CI, 1.5-6.5), respectively. Statistical analysis showed an interaction between the 2 loci with the risk conferred by the IL-1B-511T allele substantially increased (odds ratio, 9.0; 95% CI, 3.5-23.0) in individuals homozygous for the IL-1RN*2 allele.

CONCLUSIONS

Our results provide further support to the association between IL-1 gene cluster proinflammatory polymorphisms and increased risk of gastric carcinoma. Furthermore, we found evidence pointing to the existence of a synergistic interaction between the IL-1B and IL-1RN polymorphisms.

An analysis of ovarian tumor diameter and survival

Residual disease following primary debulking surgery is a recognized prognostic factor in ovarian cancer. Few studies have looked at the effect of initial ovarian tumor diameter on survival. As larger tumors are more likely to be detected by ultrasound, this information may be important in determining a survival benefit in screen-detected cancers. We reviewed the case notes and pathology of 168 consecutive cases of primary debulking surgery in epithelial ovarian cancer. We examined the influence of ovarian tumor diameter on survival and its relationship to CA125 levels and stage. For the purposes of analyses, we divided subjects into two groups: those with tumors < 6 cm and those with tumors > 6 cm. There were significant differences between the groups, with smaller tumors having more advanced stage disease compared to larger tumors (chi23 = 15.7, P = 0.0013) The median survival for tumors less than or equal to 6 cm was 17months (95% confidence interval [95% CI], 12 to 22), while for tumors greater than 6cm, the median survival was 36 months (95% CI, 13 to 59; logrank test = 8.61, P = 0.003). However, stage is also an important predictor of survival, and in a multivariate analysis, tumor size was not found to be an independent prognostic factor. There was no significant difference between the groups for CA125 levels. As larger diameter ovaries are more likely to be detected by ultrasound, it may be that screen-detected ovarian cancers will show a survival benefit simply because they detect a subset of ovarian cancers that are associated with a better prognosis.

The frequency of germ-line mutations in the breast cancer predisposition genes BRCA1 and BRCA2 in familial prostate cancer. The Cancer Research Campaign/British Prostate Group United Kingdom Familial Prostate Cancer Study Collaborators

Predisposition to prostate cancer has a genetic component, and there are reports of familial clustering of breast and prostate cancer. Two highly penetrant genes that predispose individuals to breast cancer (BRCA1 and BRCA2) are known to confer an increased risk of prostate cancer of about 3-fold and 7-fold, respectively, in breast cancer families. Blood DNA from affected individuals in 38 prostate cancer clusters was analyzed for germ-line mutations in BRCA1 and BRCA2 to assess the contribution of each of these genes to familial prostate cancer. Seventeen DNA samples were each from an affected individual in families with three or more cases of prostate cancer at any age; 20 samples were from one of affected sibling pairs where one was < or = 67 years at diagnosis. No germ-line mutations were found in BRCA1. Two germ-line mutations in BRCA2 were found, and both were seen in individuals whose age at diagnosis was very young (< or = 56 years) and who were members of an affected sibling pair. One is a 4-bp deletion at base 6710 (exon 11) in a man who had prostate cancer at 54 years, and the other is a 2-bp deletion at base 5531 (exon 11) in a man who had prostate cancer at 56 years. In both cases, the wild-type allele was lost in the patient's prostate tumor at the BRCA2 locus. However, intriguingly, in neither case did the affected brother also carry the mutation. Germ-line mutations in BRCA2 may therefore account for about 5% of prostate cancer in familial clusters.

Mammographic screening in women with a family history of breast cancer: some results from the Swedish two-county trial

BACKGROUND

The objective of this study is to compare the effectiveness of mammographic screening in women with a family history of breast cancer to those without. In the invited arm of a randomised trial of breast cancer screening, data on family history of breast cancer were available on 29.179 women aged 40-74 attending for screening. Among those women, 358 were diagnosed with breast cancer during the trial.

METHODS

Those with and without a family history were compared with respect to mammographic parenchymal pattern, interval cancer rates, mean sojourn time and sensitivity of screening. In the 358 cancers, the effect of family history was estimated on survival, incidence of advanced cancers and their relationship to screen detection.

RESULTS

A significantly higher proportion of high risk mammographic patterns was observed in association with family history among women aged 40-49. Interval cancer rates were higher in women with a family history, and in older women at least, mean sojourn time was shortened in women with a family history (1.89 years compared to 2.70). Survival was better (although not significantly so) in cancers in women with a family history (relative hazard=0.52) independently of detection mode and was significantly poorer in interval cancers then screen detected cancers (relative hazard=2.72) independently of family history. Similarly, interval cancers tended to be larger, and worse malignancy grade in those with and without a family history of breast cancer.

CONCLUSIONS

These results suggest that the policy often adopted of annual screening for woman aged 40-49, with a family history of breast cancer, is a reasonable one, and that it may also be necessary to shorten the inter-screening interval to one year in women aged over 50 but with a positive family history.

Hospital outbreak of Salmonella virchow possibly associated with a food handler

A foodborne outbreak of salmonella infection at a private hospital in London in 1994 was found to be associated with eating turkey sandwiches prepared by a food handler. One patient, nine staff, and a foodhandler's baby were confirmed to have Salmonella enterica serotype virchow, phage type 26 infection. The attack rate was estimated to be 5% among the approximately 200 patients and staff at risk. A food handler reportedly became ill days after, but her baby days before, the first hospital case. Although it appeared to be a single outbreak, antibiogram analysis, supplemented by plasmid profile typing, demonstrated that there were two strains of S. virchow involved, one with resistance to sulphonamides and trimethoprim and a second sensitive to these antimicrobial drugs. Mother and child had different strains. The investigation demonstrated the importance of full phenotypic characterization of putative outbreak strains including antimicrobial susceptibility testing. Outbreaks of foodborne infection in hospitals are preventable and are associated with high attack rates and disruption of services. There is a need for good infection control policies and training of all staff involved in patient care as well as in catering services. Consultants in Communicable Disease (CCDCs) should include private hospitals in their outbreak control plans. Good working relations between Infection Control Doctors (ICDs) in the private health sector and their local CCDCs are important if outbreaks are to be properly investigated.

Comparative trends in cause-specific fetal and neonatal mortality in twin and singleton births in the North of England, 1982-1994

OBJECTIVE

To examine trends in cause- and birthweight-specific fetal and neonatal mortality rates in twins and singletons.

DESIGN

Descriptive analysis based on a regional register.

SETTING

The Northern Health Region of England, 1982-1994.

SAMPLE

Two hundred and thirty-six fetal and 356 neonatal twin deaths; 2,687 fetal and 2,301 neonatal singleton deaths from a population of 10,734 twins and 505,477 singletons.

MAIN OUTCOME MEASURES

Fetal and neonatal autopsy rates, cause- and birthweight-specific fetal and neonatal mortality rates in twins and singletons.

RESULTS

The extended perinatal mortality (including stillbirths and neonatal deaths) rate (EPMR) was 55.2 per 1,000 in 1982-1994 in twins compared with 9.9 per 1,000 in singletons. The relative risk for twin compared with singleton deaths was 5.6 (95% CI 5.1-6. 1) being highest for immaturity (12.9, 95% CI 11.1-15.0). A significant decrease in the EPMR occurred in both twins and singletons in 1988-1994 compared with 1982-1987. The EPMR decreased mainly due to a reduction of deaths from antepartum asphyxia in twins and intrapartum asphyxia and trauma in singletons, as well as a reduction in congenital malformations in both groups. In both twins and singletons, birthweight-specific mortality rates improved between 1982-1987 and 1988-1994.

CONCLUSION

The higher relative risk for twin deaths remained stable due to a similar decrease in the EPMR for both twins and singletons. The cause-specific relative risk in twins declined for antepartum asphyxia. The mortality rate resulting from lethal congenital malformations decreased in twins and singletons mainly due to earlier detection and subsequent termination of pregnancy.

A systematic review and meta-analysis of family history and risk of ovarian cancer

OBJECTIVE

To estimate the relative risk and lifetime risk of ovarian cancer in women with various categories of family history.

DESIGN

A meta-analysis of all published case-control and cohort studies.

METHODS

Pooled relative risk estimates were calculated for the case control studies, using the Mantel-Haenzel method. These estimates were combined with the relative risks from the cohort studies. The pooled estimates of relative risk were used to estimate lifetime risks of ovarian cancer from age 15 up to age 75, for various categories of family history.

MAIN OUTCOME MEASURES

Relative risks and lifetime risks of developing ovarian cancer were calculated for the categories of women with 1. an affected first degree relative; 2. an affected mother; 3. an affected sister; and 4. women with more than one affected relative.

RESULTS

The relative risk to first degree relatives is 3.1 (95% CI 2.6-3.7). There is some evidence that this relative risk declines with age. The relative risk to mothers of cases 1.1 (95% CI 0.8-1.6) was lower than the relative risks to sisters: 3.8 (95% CI 2.9-5.1), and daughters: 6.0 (95% CI 3.0-11.9); the explanation of this difference is unclear.

CONCLUSIONS

Women with a family history of ovarian cancer have a substantially higher risk of developing ovarian cancer compared with women without such a history. However the risk is small for most categories of family history, except for the small number of individuals who have more than one affected relative.

Maternal thyroid function and motor competence in the child

Serum thyroid hormone levels were measured in women living in an area of severe dietary iodine deficiency in a remote region of Papua New Guinea. Levels of maternal hormone were found to be related to the motor competence of the women's offspring, who were aged between six years and 11 years 10 months at the time of assessment. Motor competence in the children was measured by two tests of manual dexterity which demanded speed and accuracy in performance. A significant correlation was obtained between the children's motor performance and maternal serum thyroxine levels, but not with maternal serum triiodothyronine levels. The implications of the findings for the pathogenesis of some forms of cerebral palsy are considered.

Outcome of pregnancies following the use of oral contraceptives

The use of oral contraception by women doctors on the Medical Register for England and Wales in 1975 was studied and related to the outcome of their pregnancies. Nearly two thirds had used oral contraception at some time, but less than one third of their pregnancies had followed its use. Users tended to be younger at the time of the survey and to smoke in pregnancy more often. After allowing for these associations a poor outcome was just slightly more common in pregnancies occurring after OC use, particularly in conceptions occurring within a month or after a year of cessation. In contrast conceptions occurring in the 2nd or 3rd month after cessation had an unusually favourable outcome. It therefore seems likely that at least some of this weak association between oral contraception and poor outcome may be explained by factors that determine the length of time it takes to conceive. Overall it seems that any real risk to pregnancies conceived after cessation of oral contraception must be very small and outweighed by the undoubted advantages of its use.

Fetal loss, gravidity, and pregnancy order

An investigation of the reproductive history of 3068 women doctors showed that the risk of fetal loss at a given pregnancy order varied with their gravidity--that is, the total number of pregnancies that has occurred before the survey. Fetal loss rates in even the first pregnancy varied with eventual gravidity in a J-shaped manner. They fell from 12.4% in women with only one pregnancy at the time of the study, to 5.7% in women with two, and then increased steadily to 36.8% in those with six pregnancies. This variation in risk remained when allowance was made for the incomplete nature of some of the reproductive histories. When gravidity was held constant, fetal loss rates decreased with each successive pregnancy. This finding conflicts with previous suggestions that the risk of fetal loss increase with pregnancy order and age.