Collateral damage: the impact on outcomes from cancer surgery of the COVID-19 pandemic.

BACKGROUND: Cancer diagnostics and surgery have been disrupted by the response of health care services to the coronavirus disease 2019 (COVID-19) pandemic. Progression of cancers during delay will impact on patients' long-term survival. PATIENTS AND METHODS: We generated per-day hazard ratios of cancer progression from observational studies and applied these to age-specific, stage-specific cancer survival for England 2013-2017. We modelled per-patient delay of 3 and 6 months and periods of disruption of 1 and 2 years. Using health care resource costing, we contextualise attributable lives saved and life-years gained (LYGs) from cancer surgery to equivalent volumes of COVID-19 hospitalisations. RESULTS: Per year, 94 912 resections for major cancers result in 80 406 long-term survivors and 1 717 051 LYGs. Per-patient delay of 3/6 months would cause attributable death of 4755/10 760 of these individuals with loss of 92 214/208 275 life-years, respectively. For cancer surgery, average LYGs per patient are 18.1 under standard conditions and 17.1/15.9 with a delay of 3/6 months (an average loss of 0.97/2.19 LYGs per patient), respectively. Taking into account health care resource units (HCRUs), surgery results on average per patient in 2.25 resource-adjusted life-years gained (RALYGs) under standard conditions and 2.12/1.97 RALYGs following delay of 3/6 months. For 94 912 hospital COVID-19 admissions, there are 482 022 LYGs requiring 1 052 949 HCRUs. Hospitalisation of community-acquired COVID-19 patients yields on average per patient 5.08 LYG and 0.46 RALYGs. CONCLUSIONS: Modest delays in surgery for cancer incur significant impact on survival. Delay of 3/6 months in surgery for incident cancers would mitigate 19%/43% of LYGs, respectively, by hospitalisation of an equivalent volume of admissions for community-acquired COVID-19. This rises to 26%/59%, respectively, when considering RALYGs. To avoid a downstream public health crisis of avoidable cancer deaths, cancer diagnostic and surgical pathways must be maintained at normal throughput, with rapid attention to any backlog already accrued.

Lymphocyte density determined by computational pathology validated as a predictor of response to neoadjuvant chemotherapy in breast cancer: secondary analysis of the ARTemis trial.

BACKGROUND: We have previously shown lymphocyte density, measured using computational pathology, is associated with pathological complete response (pCR) in breast cancer. The clinical validity of this finding in independent studies, among patients receiving different chemotherapy, is unknown. PATIENTS AND METHODS: The ARTemis trial randomly assigned 800 women with early stage breast cancer between May 2009 and January 2013 to three cycles of docetaxel, followed by three cycles of fluorouracil, epirubicin and cyclophosphamide once every 21 days with or without four cycles of bevacizumab. The primary endpoint was pCR (absence of invasive cancer in the breast and lymph nodes). We quantified lymphocyte density within haematoxylin and eosin (H&E) whole slide images using our previously described computational pathology approach: for every detected lymphocyte the average distance to the nearest 50 lymphocytes was calculated and the density derived from this statistic. We analyzed both pre-treatment biopsies and post-treatment surgical samples of the tumour bed. RESULTS: Of the 781 patients originally included in the primary endpoint analysis of the trial, 609 (78%) were included for baseline lymphocyte density analyses and a subset of 383 (49% of 781) for analyses of change in lymphocyte density. The main reason for loss of patients was the availability of digitized whole slide images. Pre-treatment lymphocyte density modelled as a continuous variable was associated with pCR on univariate analysis (odds ratio [OR], 2.92; 95% CI, 1.78-4.85; P < 0.001) and after adjustment for clinical covariates (OR, 2.13; 95% CI, 1.24-3.67; P = 0.006). Increased pre- to post-treatment lymphocyte density showed an independent inverse association with pCR (adjusted OR, 0.1; 95% CI, 0.033-0.31; P < 0.001). CONCLUSIONS: Lymphocyte density in pre-treatment biopsies was validated as an independent predictor of pCR in breast cancer. Computational pathology is emerging as a viable and objective means of identifying predictive biomarkers for cancer patients. CLINICALTRIALS.GOV: NCT01093235.

PD-L1 protein expression in breast cancer is rare, enriched in basal-like tumours and associated with infiltrating lymphocytes.

BACKGROUND: Expression of programmed death ligand 1 (PD-L1) in solid tumours has been shown to predict whether patients are likely to respond to anti-PD-L1 therapies. To estimate the therapeutic potential of PD-L1 inhibition in breast cancer, we evaluated the prevalence and significance of PD-L1 protein expression in a large collection of breast tumours. PATIENTS AND METHODS: Correlations between CD274 (PD-L1) copy number, transcript and protein levels were evaluated in tumours from 418 patients recruited to the METABRIC genomic study. Immunohistochemistry was used to detect PD-L1 protein in breast tumours in tissue microarrays from 5763 patients recruited to the SEARCH population-based study (N = 4079) and the NEAT randomised, controlled trial (N = 1684). RESULTS: PD-L1 protein data was available for 3916 of the possible 5763 tumours from the SEARCH and NEAT studies. PD-L1 expression by immune cells was observed in 6% (235/3916) of tumours and expression by tumour cells was observed in just 1.7% (66/3916). PD-L1 was most frequently expressed in basal-like tumours. This was observed both where tumours were subtyped by combined copy number and expression profiling [39% (17/44) of IntClust 10 i.e. basal-like tumours were PD-L1 immune cell positive; P < 0.001] and where a surrogate IHC-based classifier was used [19% (56/302) of basal-like tumours were PD-L1 immune cell positive; P < 0.001]. Moreover, CD274 (PD-L1) amplification was observed in five tumours of which four were IntClust 10. Expression of PD-L1 by either tumour cells or infiltrating immune cells was positively correlated with infiltration by both cytotoxic and regulatory T cells (P < 0.001). There was a nominally significant association between PD-L1 and improved disease-specific survival (hazard ratio 0.53, 95% confidence interval 0.26-1.07; P = 0.08) in ER-negative disease. CONCLUSIONS: Expression of PD-L1 is rare in breast cancer, markedly enriched in basal-like tumours and is correlated with infiltrating lymphocytes. PD-L1 inhibition may benefit the 19% of patients with basal-like tumours in which the protein is expressed. NEAT CLINICALTRIALSGOV: NCT00003577.

Evidence for a time-dependent association between FOLR1 expression and survival from ovarian carcinoma: implications for clinical testing. An Ovarian Tumour Tissue Analysis consortium study.

BACKGROUND: Folate receptor 1 (FOLR1) is expressed in the majority of ovarian carcinomas (OvCa), making it an attractive target for therapy. However, clinical trials testing anti-FOLR1 therapies in OvCa show mixed results and require better understanding of the prognostic relevance of FOLR1 expression. We conducted a large study evaluating FOLR1 expression with survival in different histological types of OvCa. METHODS: Tissue microarrays composed of tumour samples from 2801 patients in the Ovarian Tumour Tissue Analysis (OTTA) consortium were assessed for FOLR1 expression by centralised immunohistochemistry. We estimated associations for overall (OS) and progression-free (PFS) survival using adjusted Cox regression models. High-grade serous ovarian carcinomas (HGSC) from The Cancer Genome Atlas (TCGA) were evaluated independently for association between FOLR1 mRNA upregulation and survival. RESULTS: FOLR1 expression ranged from 76% in HGSC to 11% in mucinous carcinomas in OTTA. For HGSC, the association between FOLR1 expression and OS changed significantly during the years following diagnosis in OTTA (Pinteraction=0.01, N=1422) and TCGA (Pinteraction=0.01, N=485). In OTTA, particularly for FIGO stage I/II tumours, patients with FOLR1-positive HGSC showed increased OS during the first 2 years only (hazard ratio=0.44, 95% confidence interval=0.20-0.96) and patients with FOLR1-positive clear cell carcinomas (CCC) showed decreased PFS independent of follow-up time (HR=1.89, 95% CI=1.10-3.25, N=259). In TCGA, FOLR1 mRNA upregulation in HGSC was also associated with increased OS during the first 2 years following diagnosis irrespective of tumour stage (HR: 0.48, 95% CI: 0.25-0.94). CONCLUSIONS: FOLR1-positive HGSC tumours were associated with an increased OS in the first 2 years following diagnosis. Patients with FOLR1-negative, poor prognosis HGSC would be unlikely to benefit from anti-FOLR1 therapies. In contrast, a decreased PFS interval was observed for FOLR1-positive CCC. The clinical efficacy of FOLR1-targeted interventions should therefore be evaluated according to histology, stage and time following diagnosis.

Association between CD8+ T-cell infiltration and breast cancer survival in 12,439 patients.

BACKGROUND: T-cell infiltration in estrogen receptor (ER)-negative breast tumours has been associated with longer survival. To investigate this association and the potential of tumour T-cell infiltration as a prognostic and predictive marker, we have conducted the largest study of T cells in breast cancer to date. PATIENTS AND METHODS: Four studies totalling 12 439 patients were used for this work. Cytotoxic (CD8+) and regulatory (forkhead box protein 3, FOXP3+) T cells were quantified using immunohistochemistry (IHC). IHC for CD8 was conducted using available material from all four studies (8978 samples) and for FOXP3 from three studies (5239 samples)-multiple imputation was used to resolve missing data from the remaining patients. Cox regression was used to test for associations with breast cancer-specific survival. RESULTS: In ER-negative tumours [triple-negative breast cancer and human epidermal growth factor receptor 2 (human epidermal growth factor receptor 2 (HER2) positive)], presence of CD8+ T cells within the tumour was associated with a 28% [95% confidence interval (CI) 16% to 38%] reduction in the hazard of breast cancer-specific mortality, and CD8+ T cells within the stroma with a 21% (95% CI 7% to 33%) reduction in hazard. In ER-positive HER2-positive tumours, CD8+ T cells within the tumour were associated with a 27% (95% CI 4% to 44%) reduction in hazard. In ER-negative disease, there was evidence for greater benefit from anthracyclines in the National Epirubicin Adjuvant Trial in patients with CD8+ tumours [hazard ratio (HR) = 0.54; 95% CI 0.37-0.79] versus CD8-negative tumours (HR = 0.87; 95% CI 0.55-1.38). The difference in effect between these subgroups was significant when limited to cases with complete data (P heterogeneity = 0.04) and approached significance in imputed data (P heterogeneity = 0.1). CONCLUSIONS: The presence of CD8+ T cells in breast cancer is associated with a significant reduction in the relative risk of death from disease in both the ER-negative [supplementary Figure S1, available at Annals of Oncology online] and the ER-positive HER2-positive subtypes. Tumour lymphocytic infiltration may improve risk stratification in breast cancer patients classified into these subtypes. NEAT ClinicalTrials.gov: NCT00003577.

Public health genomics and personalized prevention: lessons from the COGS project.

Using the principles of public health genomics, we examined the opportunities and challenges of implementing personalized prevention programmes for cancer at the population level. Our model-based estimates indicate that polygenic risk stratification can potentially improve the effectiveness and cost-effectiveness of screening programmes. However, compared with 'one-size-fits-all' screening programmes, personalized screening adds further layers of complexity to the organization of screening services and raises ethical, legal and social challenges. Before polygenic inheritance is translated into population screening strategy, evidence from empirical research and engagement with and education of the public and the health professionals are needed.

PREDICT Plus: development and validation of a prognostic model for early breast cancer that includes HER2.

BACKGROUND: Predict (www.predict.nhs.uk) is an online, breast cancer prognostication and treatment benefit tool. The aim of this study was to incorporate the prognostic effect of HER2 status in a new version (Predict+), and to compare its performance with the original Predict and Adjuvant!. METHODS: The prognostic effect of HER2 status was based on an analysis of data from 10 179 breast cancer patients from 14 studies in the Breast Cancer Association Consortium. The hazard ratio estimates were incorporated into Predict. The validation study was based on 1653 patients with early-stage invasive breast cancer identified from the British Columbia Breast Cancer Outcomes Unit. Predicted overall survival (OS) and breast cancer-specific survival (BCSS) for Predict+, Predict and Adjuvant! were compared with observed outcomes. RESULTS: All three models performed well for both OS and BCSS. Both Predict models provided better BCSS estimates than Adjuvant!. In the subset of patients with HER2-positive tumours, Predict+ performed substantially better than the other two models for both OS and BCSS. CONCLUSION: Predict+ is the first clinical breast cancer prognostication tool that includes tumour HER2 status. Use of the model might lead to more accurate absolute treatment benefit predictions for individual patients.

Identification of a novel prostate cancer susceptibility variant in the KLK3 gene transcript.

Genome-wide association studies (GWAS) have identified more than 30 prostate cancer (PrCa) susceptibility loci. One of these (rs2735839) is located close to a plausible candidate susceptibility gene, KLK3, which encodes prostate-specific antigen (PSA). PSA is widely used as a biomarker for PrCa detection and disease monitoring. To refine the association between PrCa and variants in this region, we used genotyping data from a two-stage GWAS using samples from the UK and Australia, and the Cancer Genetic Markers of Susceptibility (CGEMS) study. Genotypes were imputed for 197 and 312 single nucleotide polymorphisms (SNPs) from HapMap2 and the 1000 Genome Project, respectively. The most significant association with PrCa was with a previously unidentified SNP, rs17632542 (combined P = 3.9 × 10(-22)). This association was confirmed by direct genotyping in three stages of the UK/Australian GWAS, involving 10,405 cases and 10,681 controls (combined P = 1.9 × 10(-34)). rs17632542 is also shown to be associated with PSA levels and it is a non-synonymous coding SNP (Ile179Thr) in KLK3. Using molecular dynamic simulation, we showed evidence that this variant has the potential to introduce alterations in the protein or affect RNA splicing. We propose that rs17632542 may directly influence PrCa risk.

Variation in genes required for normal mitosis and risk of breast cancer.

The down-regulation of genes involved in normal cell division can cause aberrant mitoses and increased cell death. Surviving cells exhibit aneuploidy and/or polyploidy. Since mitotic disruption has been linked with tumor development and progression, alterations in the expression or activity of these mitotic regulators may contribute to breast tumor formation. We evaluated associations between common inherited variation in these genes and breast cancer risk. Two hundred and five tagging and candidate functional single nucleotide polymorphisms in 30 genes required for normal cell division were genotyped in 798 breast cancer cases and 843 controls from the Mayo Clinic breast cancer study. Two variants in EIF3A (rs10787899 and rs3824830; P < 0.01) and four variants in SART1 (rs660118, rs679581, rs754532, and rs735942; P(trend) < or = 0.02) were significantly associated with an altered risk of breast cancer along with single variants in RRM2, PSCD3, C11orf51, CDC16, SNW1, MFAP1, and CDC2 (P < 0.05). Variation in both SART1 (P = 0.009) and EIF3A (P = 0.02) was also significant at the gene level. Analyses suggested that SART1 SNPs rs660118 and rs679581 accounted for the majority of the association of that gene with breast cancer. The observed associations between breast cancer risk and genetic variation in the SART1 and EIF3A genes that are required for maintenance of normal mitosis suggest a direct role for these genes in the development of breast cancer.

Prevalent cases in observational studies of cancer survival: do they bias hazard ratio estimates?

Observational epidemiological studies often include prevalent cases recruited at various times past diagnosis. This left truncation can be dealt with in non-parametric (Kaplan-Meier) and semi-parametric (Cox) time-to-event analyses, theoretically generating an unbiased hazard ratio (HR) when the proportional hazards (PH) assumption holds. However, concern remains that inclusion of prevalent cases in survival analysis results inevitably in HR bias. We used data on three well-established breast cancer prognosticators - clinical stage, histopathological grade and oestrogen receptor (ER) status - from the SEARCH study, a population-based study including 4470 invasive breast cancer cases (incident and prevalent), to evaluate empirically the effectiveness of allowing for left truncation in limiting HR bias. We found that HRs of prognostic factors changed over time and used extended Cox models incorporating time-dependent covariates. When comparing Cox models restricted to subjects ascertained within six months of diagnosis (incident cases) to models based on the full data set allowing for left truncation, we found no difference in parameter estimates (P=0.90, 0.32 and 0.95, for stage, grade and ER status respectively). Our results show that use of prevalent cases in an observational epidemiological study of breast cancer does not bias the HR in a left truncation Cox survival analysis, provided the PH assumption holds true.

Tagging single-nucleotide polymorphisms in candidate oncogenes and susceptibility to ovarian cancer.

Low-moderate risk alleles that are relatively common in the population may explain a significant proportion of the excess familial risk of ovarian cancer (OC) not attributed to highly penetrant genes. In this study, we evaluated the risks of OC associated with common germline variants in five oncogenes (BRAF, ERBB2, KRAS, NMI and PIK3CA) known to be involved in OC development. Thirty-four tagging SNPs in these genes were genotyped in approximately 1800 invasive OC cases and 3000 controls from population-based studies in Denmark, the United Kingdom and the United States. We found no evidence of disease association for SNPs in BRAF, KRAS, ERBB2 and PIK3CA when OC was considered as a single disease phenotype; but after stratification by histological subtype, we found borderline evidence of association for SNPs in KRAS and BRAF with mucinous OC and in ERBB2 and PIK3CA with endometrioid OC. For NMI, we identified a SNP (rs11683487) that was associated with a decreased risk of OC (unadjusted P(dominant)=0.004). We then genotyped rs11683487 in another 1097 cases and 1792 controls from an additional three case-control studies from the United States. The combined odds ratio was 0.89 (95% confidence interval (CI): 0.80-0.99) and remained statistically significant (P(dominant)=0.032). We also identified two haplotypes in ERBB2 associated with an increased OC risk (P(global)=0.034) and a haplotype in BRAF that had a protective effect (P(global)=0.005). In conclusion, these data provide borderline evidence of association for common allelic variation in the NMI with risk of epithelial OC.

Validating genetic risk associations for ovarian cancer through the international Ovarian Cancer Association Consortium.

The search for genetic variants associated with ovarian cancer risk has focused on pathways including sex steroid hormones, DNA repair, and cell cycle control. The Ovarian Cancer Association Consortium (OCAC) identified 10 single-nucleotide polymorphisms (SNPs) in genes in these pathways, which had been genotyped by Consortium members and a pooled analysis of these data was conducted. Three of the 10 SNPs showed evidence of an association with ovarian cancer at P< or =0.10 in a log-additive model: rs2740574 in CYP3A4 (P=0.011), rs1805386 in LIG4 (P=0.007), and rs3218536 in XRCC2 (P=0.095). Additional genotyping in other OCAC studies was undertaken and only the variant in CYP3A4, rs2740574, continued to show an association in the replication data among homozygous carriers: OR(homozygous(hom))=2.50 (95% CI 0.54-11.57, P=0.24) with 1406 cases and 2827 controls. Overall, in the combined data the odds ratio was 2.81 among carriers of two copies of the minor allele (95% CI 1.20-6.56, P=0.017, p(het) across studies=0.42) with 1969 cases and 3491 controls. There was no association among heterozygous carriers. CYP3A4 encodes a key enzyme in oestrogen metabolism and our finding between rs2740574 and risk of ovarian cancer suggests that this pathway may be involved in ovarian carcinogenesis. Additional follow-up is warranted.

Predicting the likelihood of carrying a BRCA1 or BRCA2 mutation: validation of BOADICEA, BRCAPRO, IBIS, Myriad and the Manchester scoring system using data from UK genetics clinics.

OBJECTIVES: Genetic testing for the breast and ovarian cancer susceptibility genes BRCA1 and BRCA2 has important implications for the clinical management of people found to carry a mutation. However, genetic testing is expensive and may be associated with adverse psychosocial effects. To provide a cost-efficient and clinically appropriate genetic counselling service, genetic testing should be targeted at those individuals most likely to carry pathogenic mutations. Several algorithms that predict the likelihood of carrying a BRCA1 or a BRCA2 mutation are currently used in clinical practice to identify such individuals. DESIGN: We evaluated the performance of the carrier prediction algorithms BOADICEA, BRCAPRO, IBIS, the Manchester scoring system and Myriad tables, using 1934 families seen in cancer genetics clinics in the UK in whom an index patient had been screened for BRCA1 and/or BRCA2 mutations. The models were evaluated for calibration, discrimination and accuracy of the predictions. RESULTS: Of the five algorithms, only BOADICEA predicted the overall observed number of mutations detected accurately (ie, was well calibrated). BOADICEA also provided the best discrimination, being significantly better (p<0.05) than all models except BRCAPRO (area under the receiver operating characteristic curve statistics: BOADICEA = 0.77, BRCAPRO = 0.76, IBIS = 0.74, Manchester = 0.75, Myriad = 0.72). All models underpredicted the number of BRCA1 and BRCA2 mutations in the low estimated risk category. CONCLUSIONS: Carrier prediction algorithms provide a rational basis for counselling individuals likely to carry BRCA1 or BRCA2 mutations. Their widespread use would improve equity of access and the cost-effectiveness of genetic testing.

Common variants in mismatch repair genes and risk of colorectal cancer.

BACKGROUND AND AIMS: The mismatch repair (MMR) genes are in charge of maintaining genomic integrity. Mutations in the MMR genes are at the origin of a familial form of colorectal cancer (CRC). This syndrome accounts for only a small proportion of the excess familial risk of CRC. The characteristics of the alleles that account for the remainder of cases are unknown. To assess the putative associations between common variants in MMR genes and CRC, we performed a genetic case-control study using a single-nucleotide polymorphism (SNP) tagging approach. PATIENTS AND METHODS: A total of 2299 cases and 2284 unrelated controls were genotyped for 68 tagging SNPs in seven MMR genes (MLH1, MLH3, MSH2, MSH3, MSH6, PMS1 and PMS2). Genotype frequencies were measured in cases and controls and analysed using univariate analysis. Haplotypes were constructed and analysed using logistic regression. We also carried out a two-locus interaction analysis and a global test analysis. RESULTS: Genotype frequencies were found to be marginally different in cases and controls for MSH3 rs26279 with a rare homozygote OR = 1.31 [95% confidence interval (CI) 1.05 to 1.62], p(trend) = 0.04. We found a rare MLH1 (frequency <5%) haplotype, increasing the risk of colorectal cancer: (OR = 9.76; 95% CI, 1.25 to 76.29; p = 0.03). The two-locus interaction analysis has exhibited signs of interaction between SNPs located in genes MSH6 and MSH2. Global testing has showed no evidence of interaction. CONCLUSION: It is unlikely that common variants in MMR genes contribute significantly to colorectal cancer.

The BOADICEA model of genetic susceptibility to breast and ovarian cancers: updates and extensions.

Multiple genetic loci confer susceptibility to breast and ovarian cancers. We have previously developed a model (BOADICEA) under which susceptibility to breast cancer is explained by mutations in BRCA1 and BRCA2, as well as by the joint multiplicative effects of many genes (polygenic component). We have now updated BOADICEA using additional family data from two UK population-based studies of breast cancer and family data from BRCA1 and BRCA2 carriers identified by 22 population-based studies of breast or ovarian cancer. The combined data set includes 2785 families (301 BRCA1 positive and 236 BRCA2 positive). Incidences were smoothed using locally weighted regression techniques to avoid large variations between adjacent intervals. A birth cohort effect on the cancer risks was implemented, whereby each individual was assumed to develop cancer according to calendar period-specific incidences. The fitted model predicts that the average breast cancer risks in carriers increase in more recent birth cohorts. For example, the average cumulative breast cancer risk to age 70 years among BRCA1 carriers is 50% for women born in 1920-1929 and 58% among women born after 1950. The model was further extended to take into account the risks of male breast, prostate and pancreatic cancer, and to allow for the risk of multiple cancers. BOADICEA can be used to predict carrier probabilities and cancer risks to individuals with any family history, and has been implemented in a user-friendly Web-based program (http://www.srl.cam.ac.uk/genepi/boadicea/boadicea_home.html).

Validation of the online prediction tool PREDICT v. 2.0 in the Dutch breast cancer population.

BACKGROUND: PREDICT version 2.0 is increasingly used to estimate prognosis in breast cancer. This study aimed to validate this tool in specific prognostic subgroups in the Netherlands. METHODS: All operated women with non-metastatic primary invasive breast cancer, diagnosed in 2005, were selected from the nationwide Netherlands Cancer Registry (NCR). Predicted and observed 5- and 10-year overall survival (OS) were compared for the overall cohort, separated by oestrogen receptor (ER) status, and predefined subgroups. A >5% difference was considered as clinically relevant. Discriminatory accuracy and goodness-of-fit were determined using the area under the receiver operating characteristic curve (AUC) and the Chi-squared-test. RESULTS: We included 8834 patients. Discriminatory accuracy for 5-year OS was good (AUC 0.80). For ER-positive and ER-negative patients, AUCs were 0.79 and 0.75, respectively. Predicted 5-year OS differed from observed by -1.4% in the entire cohort, -0.7% in ER-positive and -4.9% in ER-negative patients. Five-year OS was accurately predicted in all subgroups. Discriminatory accuracy for 10-year OS was good (AUC 0.78). For ER-positive and ER-negative patients AUCs were 0.78 and 0.76, respectively. Predicted 10-year OS differed from observed by -1.0% in the entire cohort, -0.1% in ER-positive and -5.3 in ER-negative patients. Ten-year OS was overestimated (6.3%) in patients ≥75 years and underestimated (-13.%) in T3 tumours and patients treated with both endocrine therapy and chemotherapy (-6.6%). CONCLUSIONS: PREDICT predicts OS reliably in most Dutch breast cancer patients, although results for both 5-year and 10-year OS should be interpreted carefully in ER-negative patients. Furthermore, 10-year OS should be interpreted cautiously in patients ≥75 years, T3 tumours and in patients considering endocrine therapy and chemotherapy.

Breast and ovarian cancer risks to carriers of the BRCA1 5382insC and 185delAG and BRCA2 6174delT mutations: a combined analysis of 22 population based studies.

A recent report estimated the breast cancer risks in carriers of the three Ashkenazi founder mutations to be higher than previously published estimates derived from population based studies. In an attempt to confirm this, the breast and ovarian cancer risks associated with the three Ashkenazi founder mutations were estimated using families included in a previous meta-analysis of populatrion based studies. The estimated breast cancer risks for each of the founder BRCA1 and BRCA2 mutations were similar to the corresponding estimates based on all BRCA1 or BRCA2 mutations in the meta-analysis. These estimates appear to be consistent with the observed prevalence of the mutations in the Ashkenazi Jewish population.

A population-based validation of the prognostic model PREDICT for early breast cancer.

INTRODUCTION: Predict (www.predict.nhs.uk) is a prognostication and treatment benefit tool developed using UK cancer registry data. The aim of this study was to compare the 10-year survival estimates from Predict with observed 10-year outcome from a British Columbia dataset and to compare the estimates with those generated by Adjuvant! (www.adjuvantonline.com). METHOD: The analysis was based on data from 3140 patients with early invasive breast cancer diagnosed in British Columbia, Canada, from 1989-1993. Demographic, pathologic, staging and treatment data were used to predict 10-year overall survival (OS) and breast cancer specific survival (BCSS) using Adjuvant! and Predict models. Predicted outcomes from both models were then compared with observed outcomes. RESULTS: Calibration of both models was excellent. The difference in total number of deaths estimated by Predict was 4.1 percent of observed compared to 0.7 percent for Adjuvant!. The total number of breast cancer specific deaths estimated by Predict was 3.4 percent of observed compared to 6.7 percent for Adjuvant! Both models also discriminate well with similar AUC for Predict and Adjuvant! respectively for both OS (0.709 vs 0.712) and BCSS (0.723 vs 0.727). Neither model performed well in women aged 20-35. CONCLUSION: In summary Predict provided accurate overall and breast cancer specific survival estimates in the British Columbia dataset that are comparable with outcome estimates from Adjuvant! Both models appear well calibrated with similar model discrimination. This study provides further validation of Predict as an effective predictive tool following surgery for invasive breast cancer.