Prospective validation of the BOADICEA multifactorial breast cancer risk prediction model in a large prospective cohort study

Validation of the BOADICEA model for epithelial tubo-ovarian cancer risk prediction in UK Biobank

BACKGROUND

The clinical validity of the multifactorial BOADICEA model for epithelial tubo-ovarian cancer (EOC) risk prediction has not been assessed in a large sample size or over a longer term.

METHODS

We evaluated the model discrimination and calibration in the UK Biobank cohort comprising 199,429 women (733 incident EOCs) of European ancestry without previous cancer history. We predicted 10-year EOC risk incorporating data on questionnaire-based risk factors (QRFs), family history, a 36-SNP polygenic risk score and pathogenic variants (PV) in six EOC susceptibility genes (BRCA1, BRCA2, RAD51C, RAD51D, BRIP1 and PALB2).

RESULTS

Discriminative ability was maximised under the multifactorial model that included all risk factors (AUC = 0.68, 95% CI: 0.66-0.70). This model was well calibrated in deciles of predicted risk with calibration slope=0.99 (95% CI: 0.98-1.01). Discriminative ability was similar in women younger or older than 60 years. The AUC was higher when analyses were restricted to PV carriers (0.76, 95% CI: 0.69-0.82). Using relative risk (RR) thresholds, the full model classified 97.7%, 1.7%, 0.4% and 0.2% women in the RR < 2.0, 2.0 ≤ RR < 2.9, 2.9 ≤ RR < 6.0 and RR ≥ 6.0 categories, respectively, identifying 9.1 of incident EOC among those with RR ≥ 2.0.

DISCUSSION

BOADICEA, implemented in CanRisk ( www.canrisk.org ), provides valid 10-year EOC risks and can facilitate clinical decision-making in EOC risk management.

Risk stratification in breast screening workshop

Population screening for breast cancer (BC) is currently offered in the UK for women aged 50 to 71 with the aim of reducing mortality. There is additional screening within the national programme for women identified as having a very high risk of BC. There is growing interest in further risk stratification in breast screening, which would require a whole population risk assessment and the subsequent offer of screening tailored to the individual's risk. Some women would be offered more intensive screening than others or no screening. This might provide a better balance of screening benefits and harms for each individual than the current population age-based programme alone. The UK National Screening Committee (UK NSC) is considering using decision-analytic and other models to evaluate different risk stratification screening strategies and identify remaining gaps in evidence. This paper reports the proceedings of a UK NSC workshop where experts in the field discussed both risk prediction models, as well as decision-analytic models providing a benefit-harm analysis/economic evaluation of risk-stratified screening programmes (see Table 1). The aim of the meeting was to present and discuss the current work of experts, including some data which had not been published at the time of the meeting, to inform the UK NSC. The workshop was not intended to present a balanced evaluation of how to deliver screening in future. Areas for further work identified included methods for comparing models to assess accuracy, the optimum risk assessment tools, the digital screening infrastructure, acceptability of stratification, choice of screening test and reducing inequalities. A move to risk stratification of the whole programme would require a careful phased introduction with continuing assessment of real-world evidence during deployment.

Are women's breast cancer risk appraisals in line with updated clinical risk estimates communicated? Results from a UK Family History Risk and Prevention Clinic

BACKGROUND

The incorporation of breast density and a polygenic risk score (PRS) into breast cancer risk prediction models can alter previously communicated risk estimates. Previous research finds that risk communication does not usually change personal risk appraisals. This study aimed to examine how women from the Family History Risk (FH-Risk) study appraise their breast cancer risk following communication of an updated risk estimate.

METHODS

In the FH-Risk study 323 women attended a consultation to receive an updated breast cancer risk estimate. A subset (n=190) completed a questionnaire, assessing their subjective breast cancer risk appraisals, satisfaction with the information provided and cancer related worry. One hundred and three were notified of a decrease risk, 34 an increase and 53 an unchanged risk.

RESULTS

Women's subjective risk appraisals were in line with the updated risk estimates provided, with age, a PRS and breast density explaining most of the variance in these appraisals. Those notified of an increased risk demonstrated higher subjective risk perceptions compared to those whose risk remained unchanged or decreased.

CONCLUSIONS

Women's subjective breast cancer risk appraisals are amenable to change following updated risk feedback, with new information breast density and a PRS accepted and integrated into existing risk appraisals. Trust in the service, the analogies and visual communication strategies used may have positively influenced the integration of this new information.

IMPACT

Further research is warranted to assess whether similar patterns emerge for other illnesses and in different clinical contexts to determine the best strategies for communicating updated risk estimates.

Improving the communication of multifactorial cancer risk assessment results for different audiences: a co-design process

BACKGROUND

Multifactorial cancer risk prediction tools, such as CanRisk, are increasingly being incorporated into routine healthcare. Understanding risk information and communicating risk is challenging and healthcare professionals rely substantially on the outputs of risk prediction tools to communicate results. This work aimed to produce a new CanRisk report so users can directly access key information and communicate risk estimates effectively.

METHODS

Over a 13-month period, we led an 8-step co-design process with patients, the public, and healthcare professionals. Steps comprised 1) think aloud testing of the original CanRisk report; 2) structured feedback on the original report; 3) literature review; 4) development of a new report prototype; 5) first round of structured feedback; 6) updating the new report prototype; 7) second round of structured feedback; and 8) finalising and publishing the new CanRisk report.

RESULTS

We received 56 sets of feedback from 34 stakeholders. Overall, the original CanRisk report was not suitable for patients and the public. Building on the feedback, the new report has an overview of the information presented: section one summarises key information for individuals; sections two and three present information for healthcare professionals in different settings. New features also include explanatory text, definitions, graphs, keys and tables to support the interpretation of the information.

DISCUSSION

This co-design experience shows the value of collaboration for the successful communication of complex health information. As a result, the new CanRisk report has the potential to better support shared decision-making processes about cancer risk management across clinical settings.

Evaluating Real World Health System Resource Utilization and Costs for a Risk-Based Breast Cancer Screening Approach in the Canadian PERSPECTIVE Integration and Implementation Project

BACKGROUND

A prospective cohort study was undertaken within the PERSPECTIVE I&I project to evaluate healthcare resource utilization and costs associated with breast cancer risk assessment and screening and overall costs stratified by risk level, in Ontario, Canada.

METHODS

From July 2019 to December 2022, 1997 females aged 50 to 70 years consented to risk assessment and received their breast cancer risk level and personalized screening action plan in Ontario. The mean costs for risk-stratified screening-related activities included risk assessment, screening and diagnostic costs. The GETCOST macro from the Institute of Clinical Evaluative Sciences (ICES) assessed the mean overall healthcare system costs.

RESULTS

For the 1997 participants, 83.3%, 14.4% and 2.3% were estimated to be average, higher than average, and high risk, respectively (median age (IQR): 60 [56-64] years). Stratification into the three risk levels was determined using the validated multifactorial CanRisk prediction tool that includes family history information, a polygenic risk score (PRS), breast density and established lifestyle/hormonal risk factors. The mean number of genetic counseling visits, mammograms and MRIs per individual increased with risk level. High-risk participants incurred the highest overall mean risk-stratified screening-related costs in 2022 CAD (±SD) at CAD 905 (±269) followed by CAD 580 (±192) and CAD 521 (±163) for higher-than-average and average-risk participants, respectively. Among the breast screening-related costs, the greatest cost burden across all risk groups was the risk assessment cost, followed by total diagnostic and screening costs. The mean overall healthcare cost per participant (±SD) was the highest for the average risk participants with CAD 6311 (±19,641), followed by higher than average risk with CAD 5391 (±8325) and high risk with CAD 5169 (±7676).

CONCLUSION

Although high-risk participants incurred the highest risk-stratified screening-related costs, their costs for overall healthcare utilization costs were similar to other risk levels. Our study underscored the importance of integrating risk stratification as part of the screening pathway to support breast cancer detection at an earlier and more treatable stage, thereby reducing costs and the overall burden on the healthcare system.

Breast cancer risk assessment for prescription of menopausal hormone therapy in women with a family history of breast cancer: an epidemiological modelling study

BACKGROUND

Menopausal hormone therapy (MHT) can alleviate menopausal symptoms but has been associated with an increased risk of breast cancer. MHT prescription should be preceded by individualised risk/benefit evaluation; however, data outlining the impact of family history alongside different MHT therapeutic approaches are lacking.

AIM

To quantify the risks associated with MHT use in women with varying breast cancer family histories of developing and dying from breast cancer.

DESIGN AND SETTING

An epidemiological modelling study for women in England using the BOADICEA breast cancer prediction model and data relating to MHT use and breast cancer risk taken from research by the Collaborative Group on Hormonal Factors in Breast Cancer.

METHOD

The risk of developing and dying from breast cancer between the ages of 50 and 80 years was modelled in women with four different breast cancer family history profiles: 'average', 'modest', 'intermediate', and 'strong' by using 1) background risks of breast cancer by age and family history, 2) relative risks for breast cancer associated with MHT use, and 3) 10-year breast cancer-specific net mortality rates. This study modelled use of combined oestrogen-progestogen MHT (cyclical or continuous) and oestrogen-only MHT.

RESULTS

For a woman of 'average' family history taking no MHT, the cumulative breast cancer risk (age 50-80 years) is 9.8%, and the risk of dying from the breast cancer is 1.7%. In this model, 5 years' exposure to combined-cyclical MHT (age 50-55 years) was calculated to increase these risks to 11.0% and 1.8%, respectively. For a woman with a 'strong' family history taking no MHT, the cumulative breast cancer risk is 19.6% (age 50-80 years), and the risk of dying from the breast cancer is 3.2%. With 5 years' exposure to MHT (age 50-55 years), this model showed that these risks increase to 22.4% and 3.5%, respectively.

CONCLUSION

In this model, both family history and MHT are associated with increased risk of breast cancer. Estimates of the risks of breast cancer associated with MHT for women with different family histories can be used to support decision making around MHT prescription for women experiencing menopausal symptoms.

Validation of the BOADICEA model in a prospective cohort of BRCA1/2 pathogenic variant carriers

BACKGROUND

No validation has been conducted for the BOADICEA multifactorial breast cancer risk prediction model specifically in BRCA1/2 pathogenic variant (PV) carriers to date. Here, we evaluated the performance of BOADICEA in predicting 5-year breast cancer risks in a prospective cohort of BRCA1/2 PV carriers ascertained through clinical genetic centres.

METHODS

We evaluated the model calibration and discriminatory ability in the prospective TRANsIBCCS cohort study comprising 1614 BRCA1 and 1365 BRCA2 PV carriers (209 incident cases). Study participants had lifestyle, reproductive, hormonal, anthropometric risk factor information, a polygenic risk score based on 313 SNPs and family history information.

RESULTS

The full multifactorial model considering family history together with all other risk factors was well calibrated overall (E/O=1.07, 95% CI: 0.92 to 1.24) and in quintiles of predicted risk. Discrimination was maximised when all risk factors were considered (Harrell's C-index=0.70, 95% CI: 0.67 to 0.74; area under the curve=0.79, 95% CI: 0.76 to 0.82). The model performance was similar when evaluated separately in BRCA1 or BRCA2 PV carriers. The full model identified 5.8%, 12.9% and 24.0% of BRCA1/2 PV carriers with 5-year breast cancer risks of <1.65%, <3% and <5%, respectively, risk thresholds commonly used for different management and risk-reduction options.

CONCLUSION

BOADICEA may be used to aid personalised cancer risk management and decision-making for BRCA1 and BRCA2 PV carriers. It is implemented in the free-access CanRisk tool (https://www.canrisk.org/).

Validation of a clinical breast cancer risk assessment tool combining a polygenic score for all ancestries with traditional risk factors

PURPOSE

We previously described a combined risk score (CRS) that integrates a multiple-ancestry polygenic risk score (MA-PRS) with the Tyrer-Cuzick (TC) model to assess breast cancer (BC) risk. Here, we present a longitudinal validation of CRS in a real-world cohort.

METHODS

This study included 130,058 patients referred for hereditary cancer genetic testing and negative for germline pathogenic variants in BC-associated genes. Data were obtained by linking genetic test results to medical claims (median follow-up 12.1 months). CRS calibration was evaluated by the ratio of observed to expected BCs.

RESULTS

Three hundred forty BCs were observed over 148,349 patient-years. CRS was well-calibrated and demonstrated superior calibration compared with TC in high-risk deciles. MA-PRS alone had greater discriminatory accuracy than TC, and CRS had approximately 2-fold greater discriminatory accuracy than MA-PRS or TC. Among those classified as high risk by TC, 32.6% were low risk by CRS, and of those classified as low risk by TC, 4.3% were high risk by CRS. In cases where CRS and TC classifications disagreed, CRS was more accurate in predicting incident BC.

CONCLUSION

CRS was well-calibrated and significantly improved BC risk stratification. Short-term follow-up suggests that clinical implementation of CRS should improve outcomes for patients of all ancestries through personalized risk-based screening and prevention.

Joint ABS-UKCGG-CanGene-CanVar consensus regarding the use of CanRisk in clinical practice

BACKGROUND

The CanRisk tool, which operationalises the Breast and Ovarian Analysis of Disease Incidence and Carrier Estimation Algorithm (BOADICEA) is used by Clinical Geneticists, Genetic Counsellors, Breast Oncologists, Surgeons and Family History Nurses for breast cancer risk assessments both nationally and internationally. There are currently no guidelines with respect to the day-to-day clinical application of CanRisk and differing inputs to the model can result in different recommendations for practice.

METHODS

To address this gap, the UK Cancer Genetics Group in collaboration with the Association of Breast Surgery and the CanGene-CanVar programme held a workshop on 16th of May 2023, with the aim of establishing best practice guidelines.

RESULTS

Using a pre-workshop survey followed by structured discussion and in-meeting polling, we achieved consensus for UK best practice in use of CanRisk in making recommendations for breast cancer surveillance, eligibility for genetic testing and the input of available information to undertake an individualised risk assessment.

CONCLUSIONS

Whilst consensus recommendations were achieved, the meeting highlighted some of the barriers limiting the use of CanRisk in clinical practice and identified areas that require further work and collaboration with relevant national bodies and policy makers to incorporate wider use of CanRisk into routine breast cancer risk assessments.

Implementing Multifactorial Risk Assessment with Polygenic Risk Scores for Personalized Breast Cancer Screening in the Population Setting: Challenges and Opportunities

Risk-stratified breast screening has been proposed as a strategy to overcome the limitations of age-based screening. A prospective cohort study was undertaken within the PERSPECTIVE I&I project, which will generate the first Canadian evidence on multifactorial breast cancer risk assessment in the population setting to inform the implementation of risk-stratified screening. Recruited females aged 40-69 unaffected by breast cancer, with a previous mammogram, underwent multifactorial breast cancer risk assessment. The adoption of multifactorial risk assessment, the effectiveness of methods for collecting risk factor information and the costs of risk assessment were examined. Associations between participant characteristics and study sites, as well as data collection methods, were assessed using logistic regression; all p-values are two-sided. Of the 4246 participants recruited, 88.4% completed a risk assessment, with 79.8%, 15.7% and 4.4% estimated at average, higher than average and high risk, respectively. The total per-participant cost for risk assessment was CAD 315. Participants who chose to provide risk factor information on paper/telephone (27.2%) vs. online were more likely to be older (p = 0.021), not born in Canada (p = 0.043), visible minorities (p = 0.01) and have a lower attained education (p < 0.0001) and perceived fair/poor health (p < 0.001). The 34.4% of participants requiring risk factor verification for missing/unusual values were more likely to be visible minorities (p = 0.009) and have a lower attained education (p ≤ 0.006). This study demonstrates the feasibility of risk assessment for risk-stratified screening at the population level. Implementation should incorporate an equity lens to ensure cancer-screening disparities are not widened.

Benefits and harms of polygenic risk scores in organised cancer screening programmes: a cost-effectiveness analysis

Background

While polygenic risk scores (PRS) could enable the streamlining of organised cancer screening programmes, its current discriminative ability is limited. We conducted a cost-effectiveness analysis to trade-off the benefits and harms of PRS-stratified cancer screening in China.

Methods

The validated National Cancer Center (NCC) modelling framework for six cancers (lung, liver, breast, gastric, colorectum, and oesophagus) was used to simulate cancer incidence, progression, stage-specific cancer detection, and risk of death. We estimated the number of cancer deaths averted, quality-adjusted life-years (QALY) gained, number needed to screen (NNS), overdiagnosis, and incremental cost-effectiveness ratio (ICER) of one-time PRS-stratified screening strategy (screening 25% of PRS-defined high-risk population) for a birth cohort at age 60 in 2025, compared with unstratified screening strategy (screening 25% of general population) and no screening strategy. We applied lifetime horizon, societal perspective, and 3% discount rate. An ICER less than $18,364 per QALY gained is considered cost-effective.

Findings

One-time cancer screening for population aged 60 was the most cost-effective strategy compared to screening at other ages. Compared with an unstratified screening strategy, the PRS-stratified screening strategy averted more cancer deaths (61,237 vs. 40,329), had a lower NNS to prevent one death (307 vs. 451), had a slightly higher overdiagnosis (14.1% vs. 13.8%), and associated with an additional 130,045 QALYs at an additional cost of $1942 million, over a lifetime horizon. The ICER for all six cancers combined was $14,930 per QALY gained, with the ICER varying from $7928 in colorectal cancer to $39,068 in liver cancer. ICER estimates were sensitive to changes in risk threshold and cost of PRS tools.

Interpretation

PRS-stratified screening strategy modestly improves clinical benefit and cost-effectiveness of organised cancer screening programmes. Reducing the costs of polygenic risk stratification is needed before PRS implementation.

Funding

The Chinese Academy of Medical Sciences, the Jing-jin-ji Special Projects for Basic Research Cooperation, and the Sanming Project of the Medicine in Shenzhen.

Lifestyle Factors and Breast Cancer in Females with PTEN Hamartoma Tumor Syndrome (PHTS)

Females with PTEN Hamartoma Tumor Syndrome (PHTS) have breast cancer risks up to 76%. This study assessed associations between breast cancer and lifestyle in European female adult PHTS patients. Data were collected via patient questionnaires (July 2020-March 2023) and genetic diagnoses from medical files. Associations between lifestyle and breast cancer were calculated using logistic regression corrected for age. Index patients with breast cancer before PHTS diagnosis (breast cancer index) were excluded for ascertainment bias correction. In total, 125 patients were included who completed the questionnaire at a mean age of 44 years (SD = 13). This included 21 breast cancer indexes (17%) and 39 females who developed breast cancer at 43 years (SD = 9). Breast cancer patients performed about 1.1 times less often 0-1 times/week physical activity than ≥2 times (ORtotal-adj = 0.9 (95%CI 0.3-2.6); consumed daily about 1.2-1.8 times more often ≥1 than 0-1 glasses of alcohol (ORtotal-adj = 1.2 (95%CI 0.4-4.0); ORnon-breastcancer-index-adj = 1.8 (95%CI 0.4-6.9); were about 1.04-1.3 times more often smokers than non-smokers (ORtotal-adj = 1.04 (95%CI 0.4-2.8); ORnon-breastcancer-index-adj = 1.3 (95%CI 0.4-4.2)); and overweight or obesity (72%) was about 1.02-1.3 times less common (ORtotal-adj = 0.98 (95%CI 0.4-2.6); ORnon-breastcancer-index-adj = 0.8 (95%CI 0.3-2.7)). Similar associations between lifestyle and breast cancer are suggested for PHTS and the general population. Despite not being statistically significant, results are clinically relevant and suggest that awareness of the effects of lifestyle on patients' breast cancer risk is important.

Differences in polygenic score distributions in European ancestry populations: implications for breast cancer risk prediction

The 313-variant polygenic risk score (PRS313) provides a promising tool for breast cancer risk prediction. However, evaluation of the PRS313 across different European populations which could influence risk estimation has not been performed. Here, we explored the distribution of PRS313 across European populations using genotype data from 94,072 females without breast cancer, of European-ancestry from 21 countries participating in the Breast Cancer Association Consortium (BCAC) and 225,105 female participants from the UK Biobank. The mean PRS313 differed markedly across European countries, being highest in south-eastern Europe and lowest in north-western Europe. Using the overall European PRS313 distribution to categorise individuals leads to overestimation and underestimation of risk in some individuals from south-eastern and north-western countries, respectively. Adjustment for principal components explained most of the observed heterogeneity in mean PRS. Country-specific PRS distributions may be used to calibrate risk categories in individuals from different countries.

Clinical implications of incorporating genetic and non-genetic risk factors in CanRisk-based breast cancer risk prediction

BACKGROUND

Breast cancer (BC) risk prediction models consider cancer family history (FH) and germline pathogenic variants (PVs) in risk genes. It remains elusive to what extent complementation with polygenic risk score (PRS) and non-genetic risk factor (NGRFs) data affects individual intensified breast surveillance (IBS) recommendations according to European guidelines.

METHODS

For 425 cancer-free women with cancer FH (mean age 40·6 years, range 21-74), recruited in France, Germany and the Netherlands, germline PV status, NGRFs, and a 306 variant-based PRS (PRS306) were assessed to calculate estimated lifetime risks (eLTR) and estimated 10-year risks (e10YR) using CanRisk. The proportions of women changing country-specific European risk categories for IBS recommendations, i.e. ≥20 % and ≥30 % eLTR, or ≥5 % e10YR were determined.

FINDINGS

Of the women with non-informative PV status, including PRS306 and NGRFs changed clinical recommendations for 31·0 %, (57/184, 20 % eLTR), 15·8 % (29/184, 30 % eLTR) and 22·4 % (41/183, 5 % e10YR), respectively whereas of the women tested negative for a PV observed in their family, clinical recommendations changed for 16·7 % (25/150), 1·3 % (2/150) and 9·5 % (14/147). No change was observed for 82 women with PVs in high-risk genes (BRCA1/2, PALB2). Combined consideration of eLTRs and e10YRs identified BRCA1/2 PV carriers benefitting from IBS <30 years, and women tested non-informative/negative for whom IBS may be postponed.

INTERPRETATION

For women who tested non-informative/negative, PRS and NGRFs have a considerable impact on IBS recommendations. Combined consideration of eLTRs and e10YRs allows personalizing IBS starting age.

FUNDING

Horizon 2020, German Cancer Aid, Federal Ministry of Education and Research, Köln Fortune.

The benefit of adding polygenic risk scores, lifestyle factors, and breast density to family history and genetic status for breast cancer risk and surveillance classification of unaffected women from germline CHEK2 c.1100delC families

To determine the changes in surveillance category by adding a polygenic risk score based on 311 breast cancer (BC)-associated variants (PRS311), questionnaire-based risk factors and breast density on personalized BC risk in unaffected women from Dutch CHEK2 c.1100delC families. In total, 117 unaffected women (58 heterozygotes and 59 non-carriers) from CHEK2 families were included. Blood-derived DNA samples were genotyped with the GSAMDv3-array to determine PRS311. Lifetime BC risk was calculated in CanRisk, which uses data from the Breast and Ovarian Analysis of Disease Incidence and Carrier Estimation Algorithm (BOADICEA). Women, were categorized into three surveillance groups. The surveillance advice was reclassified in 37.9 % of heterozygotes and 32.2 % of non-carriers after adding PRS311. Including questionnaire-based risk factors resulted in an additional change in 20.0 % of heterozygotes and 13.2 % of non-carriers; and a subanalysis showed that adding breast density on top shifted another 17.9 % of heterozygotes and 33.3 % of non-carriers. Overall, the majority of heterozygotes were reclassified to a less intensive surveillance, while non-carriers would require intensified surveillance. The addition of PRS311, questionnaire-based risk factors and breast density to family history resulted in a more personalized BC surveillance advice in CHEK2-families, which may lead to more efficient use of surveillance.

Integrating a Polygenic Risk Score into a clinical setting would impact risk predictions in familial breast cancer

BACKGROUND

Low-impact genetic variants identified in population-based genetic studies are not routinely measured as part of clinical genetic testing in familial breast cancer (BC). We studied the consequences of integrating an established Polygenic Risk Score (PRS) (BCAC 313, PRS313) into clinical sequencing of women with familial BC in Sweden.

METHODS

We developed an add-on sequencing panel to capture 313 risk variants in addition to the clinical screening of hereditary BC genes. Index patients with no pathogenic variant from 87 families, and 1000 population controls, were included in comparative PRS calculations. Including detailed family history, sequencing results and tumour pathology information, we used BOADICEA (Breast and Ovarian Analysis of Disease Incidence and Carrier Estimation Algorithm) V.6 to estimate contralateral and lifetime risks without and with PRS313.

RESULTS

Women with BC but no pathogenic variants in hereditary BC genes have a higher PRS313 compared with population controls (mean+0.78 SD, p<3e-9). Implementing PRS313 in the clinical risk estimation before their BC diagnosis would have changed the recommended follow-up in 24%-45% of women.

CONCLUSIONS

Our results show the potential impact of incorporating PRS313 directly in the clinical genomic investigation of women with familial BC.

A Systematic Review and Critical Assessment of Breast Cancer Risk Prediction Tools Incorporating a Polygenic Risk Score for the General Population

Single nucleotide polymorphisms (SNPs) in the form of a polygenic risk score (PRS) have emerged as a promising factor that could improve the predictive performance of breast cancer (BC) risk prediction tools. This study aims to appraise and critically assess the current evidence on these tools. Studies were identified using Medline, EMBASE and the Cochrane Library up to November 2022 and were included if they described the development and/ or validation of a BC risk prediction model using a PRS for women of the general population and if they reported a measure of predictive performance. We identified 37 articles, of which 29 combined genetic and non-genetic risk factors using seven different risk prediction tools. Most models (55.0%) were developed on populations from European ancestry and performed better than those developed on populations from other ancestry groups. Regardless of the number of SNPs in each PRS, models combining a PRS with genetic and non-genetic risk factors generally had better discriminatory accuracy (AUC from 0.52 to 0.77) than those using a PRS alone (AUC from 0.48 to 0.68). The overall risk of bias was considered low in most studies. BC risk prediction tools combining a PRS with genetic and non-genetic risk factors provided better discriminative accuracy than either used alone. Further studies are needed to cross-compare their clinical utility and readiness for implementation in public health practices.

Individualizing Breast Cancer Risk Assessment in Clinical Practice

Multiple tools exist to assess a patient's breast cancer risk. The choice of risk model depends on the patient's risk factors and how the calculation will impact care. High-risk patients-those with a lifetime breast cancer risk of ≥20%-are, for instance, eligible for supplemental screening with breast magnetic resonance imaging. Those with an elevated short-term breast cancer risk (frequently defined as a 5-year risk ≥1.66%) should be offered endocrine prophylaxis. High-risk patients should also receive guidance on modification of lifestyle factors that affect breast cancer risk.

Breast Cancer Risk Reduction: Current Status and Emerging Trends to Increase Efficacy and Reduce Toxicity of Preventive Medication

The primary prevention of breast cancer is a worthwhile goal for which the efficacy of antiestrogens is well established. However, implementation has been problematic related to low prioritization by providers and the reluctance of high-risk women to experience medication side effects. Emerging solutions include improved risk estimation through the use of polygenic risk scores and the application of radiomics to screening mammograms; and optimization of medication dose to limit toxicity. The identification of agents to prevent estrogen receptor negative or HER2-positive tumors is being pursued, but personalization of medical risk reduction requires the prediction of tumor subtypes.

Polygenic scores in cancer

Since the publication of the first genome-wide association study for cancer in 2007, thousands of common alleles that are associated with the risk of cancer have been identified. The relative risk associated with individual variants is small and of limited clinical significance. However, the combined effect of multiple risk variants as captured by polygenic scores (PGSs) may be much greater and therefore provide risk discrimination that is clinically useful. We review the considerable research efforts over the past 15 years for developing statistical methods for PGSs and their application in large-scale genome-wide association studies to develop PGSs for various cancers. We review the predictive performance of these PGSs and the multiple challenges currently limiting the clinical application of PGSs. Despite this, PGSs are beginning to be incorporated into clinical multifactorial risk prediction models to stratify risk in both clinical trials and clinical implementation studies.

Robust cross-vendor mammographic texture models using augmentation-based domain adaptation for long-term breast cancer risk

Purpose

Risk-stratified breast cancer screening might improve early detection and efficiency without comprising quality. However, modern mammography-based risk models do not ensure adaptation across vendor-domains and rely on cancer precursors, associated with short-term risk, which might limit long-term risk assessment. We report a cross-vendor mammographic texture model for long-term risk.

Approach

The texture model was robustly trained using two systematically designed case-control datasets. Textural features, indicative of future breast cancer, were learned by excluding samples with diagnosed/potential malignancies from training. An augmentation-based domain adaption technique, based on flavorization of mammographic views, ensured generalization across vendor-domains. The model was validated in 66,607 consecutively screened Danish women with flavorized Siemens views and 25,706 Dutch women with Hologic-processed views. Performances were evaluated for interval cancers (IC) within 2 years from screening and long-term cancers (LTC) from 2 years after screening. The texture model was combined with established risk factors to flag 10% of women with the highest risk.

Results

In Danish women, the texture model achieved an area under the receiver operating characteristic curve (AUC) of 0.71 and 0.65 for ICs and LTCs, respectively. In Dutch women with Hologic-processed views, the AUCs were not different from AUCs in Danish women with flavorized views. The AUC for texture combined with established risk factors increased to 0.68 for LTCs. The 10% of women flagged as high-risk accounted for 25.5% of ICs and 24.8% of LTCs.

Conclusions

The texture model robustly estimated long-term breast cancer risk while adapting to an unseen processed vendor-domain and identified a clinically relevant high-risk subgroup.

Challenges and Opportunities for Data Science in Women's Health

The intersection of women's health and data science is a field of research that has historically trailed other fields, but more recently it has gained momentum. This growth is being driven not only by new investigators who are moving into this area but also by the significant opportunities that have emerged in new methodologies, resources, and technologies in data science. Here, we describe some of the resources and methods being used by women's health researchers today to meet challenges in biomedical data science. We also describe the opportunities and limitations of applying these approaches to advance women's health outcomes and the future of the field, with emphasis on repurposing existing methodologies for women's health.

Clinical Impact of Polygenic Risk Score for Breast Cancer Risk Prediction in 382 Individuals with Hereditary Breast and Ovarian Cancer Syndrome

Single nucleotide polymorphisms are currently not considered in breast cancer (BC) risk predictions used in daily practice of genetic counselling and clinical management of familial BC in Germany. This study aimed to assess the clinical value of incorporating a 313-variant-based polygenic risk score (PRS) into BC risk calculations in a cohort of German women with suspected hereditary breast and ovarian cancer syndrome (HBOC). Data from 382 individuals seeking counselling for HBOC were analysed. Risk calculations were performed using the Breast and Ovarian Analysis of Disease Incidence and Carrier Estimation Algorithm with and without the inclusion of the PRS. Changes in risk predictions and their impact on clinical management were evaluated. The PRS led to changes in risk stratification based on 10-year risk calculations in 13.6% of individuals. Furthermore, the inclusion of the PRS in BC risk predictions resulted in clinically significant changes in 12.0% of cases, impacting the prevention recommendations established by the German Consortium for Hereditary Breast and Ovarian Cancer. These findings support the implementation of the PRS in genetic counselling for personalized BC risk assessment.

Assessing Breast Cancer Risk by Combining AI for Lesion Detection and Mammographic Texture

Background Recent mammography-based risk models can estimate short-term or long-term breast cancer risk, but whether risk assessment may improve by combining these models has not been evaluated. Purpose To determine whether breast cancer risk assessment improves when combining a diagnostic artificial intelligence (AI) system for lesion detection and a mammographic texture model. Materials and Methods This retrospective study included Danish women consecutively screened for breast cancer at mammography from November 2012 to December 2015 who had at least 5 years of follow-up data. Examinations were evaluated for short-term risk using a commercially available diagnostic AI system for lesion detection, which produced a score to indicate the probability of cancer. A mammographic texture model, trained on a separate data set, assessed textures associated with long-term cancer risk. Area under the receiver operating characteristic curve (AUC) analysis was used to evaluate both the individual and combined performance of the AI and texture models for the prediction of future cancers in women with a negative screening mammogram, including those with interval cancers diagnosed within 2 years of screening and long-term cancers diagnosed 2 years or more after screening. AUCs were compared using the DeLong test. Results The Danish screening cohort included 119 650 women (median age, 59 years [IQR, 53-64 years]), of whom 320 developed interval cancers and 1401 developed long-term cancers. The combination model achieved a higher AUC for interval and long-term cancers grouped together than either the diagnostic AI (AUC, 0.73 vs 0.70; P < .001) or the texture risk (AUC, 0.73 vs 0.66; P < .001) models. The 10% of women with the highest combined risk identified by the combination model accounted for 44.1% (141 of 320) of interval cancers and 33.7% (472 of 1401) of long-term cancers. Conclusion Combining a diagnostic AI system and mammographic texture model resulted in improved risk assessment for interval cancers and long-term cancers and enabled identification of women at high risk. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Poynton and Slanetz in this issue.

Exploring the barriers to and facilitators of implementing CanRisk in primary care: a qualitative thematic framework analysis

BACKGROUND

The CanRisk tool enables the collection of risk factor information and calculation of estimated future breast cancer risks based on the multifactorial Breast and Ovarian Analysis of Disease Incidence and Carrier Estimation Algorithm (BOADICEA) model. Despite BOADICEA being recommended in National Institute for Health and Care Excellence (NICE) guidelines and CanRisk being freely available for use, the CanRisk tool has not yet been widely implemented in primary care.

AIM

To explore the barriers to and facilitators of the implementation of the CanRisk tool in primary care.

DESIGN AND SETTING

A multi-methods study was conducted with primary care practitioners (PCPs) in the East of England.

METHOD

Participants used the CanRisk tool to complete two vignette-based case studies; semi-structured interviews gained feedback about the tool; and questionnaires collected demographic details and information about the structural characteristics of the practices.

RESULTS

Sixteen PCPs (eight GPs and eight nurses) completed the study. The main barriers to implementation included: time needed to complete the tool; competing priorities; IT infrastructure; and PCPs' lack of confidence and knowledge to use the tool. Main facilitators included: easy navigation of the tool; its potential clinical impact; and the increasing availability of and expectation to use risk prediction tools.

CONCLUSION

There is now a greater understanding of the barriers and facilitators that exist when using CanRisk in primary care. The study has highlighted that future implementation activities should focus on reducing the time needed to complete a CanRisk calculation, integrating the CanRisk tool into existing IT infrastructure, and identifying appropriate contexts in which to conduct a CanRisk calculation. PCPs may also benefit from information about cancer risk assessment and CanRisk-specific training.

Improved prediction of breast cancer risk based on phenotypic DNA damage repair capacity in peripheral blood B cells

Background

Standard Breast Cancer (BC) risk prediction models based only on epidemiologic factors generally have quite poor performance, and there have been a number of risk scores proposed to improve them, such as AI-based mammographic information, polygenic risk scores and pathogenic variants. Even with these additions BC risk prediction performance is still at best moderate. In that decreased DNA repair capacity (DRC) is a major risk factor for development of cancer, we investigated the potential to improve BC risk prediction models by including a measured phenotypic DRC assay.

Methods

Using blood samples from the Breast Cancer Family Registry we assessed the performance of phenotypic markers of DRC in 46 matched pairs of individuals, one from each pair with BC (with blood drawn before BC diagnosis) and the other from controls matched by age and time since blood draw. We assessed DRC in thawed cryopreserved peripheral blood mononuclear cells (PBMCs) by measuring γ-H2AX yields (a marker for DNA double-strand breaks) at multiple times from 1 to 20 hrs after a radiation challenge. The studies were performed using surface markers to discriminate between different PBMC subtypes.

Results

The parameter F res , the residual damage signal in PBMC B cells at 20 hrs post challenge, was the strongest predictor of breast cancer with an AUC (Area Under receiver-operator Curve) of 0.89 [95% Confidence Interval: 0.84-0.93] and a BC status prediction accuracy of 0.80. To illustrate the combined use of a phenotypic predictor with standard BC predictors, we combined F res in B cells with age at blood draw, and found that the combination resulted in significantly greater BC predictive power (AUC of 0.97 [95% CI: 0.94-0.99]), an increase of 13 percentage points over age alone.

Conclusions

If replicated in larger studies, these results suggest that inclusion of a fingerstick-based phenotypic DRC blood test has the potential to markedly improve BC risk prediction.

A Clinical Risk Model for Personalized Screening and Prevention of Breast Cancer

BACKGROUND

Image-derived artificial intelligence (AI) risk models have shown promise in identifying high-risk women in the short term. The long-term performance of image-derived risk models expanded with clinical factors has not been investigated.

METHODS

We performed a case-cohort study of 8110 women aged 40-74 randomly selected from a Swedish mammography screening cohort initiated in 2010 together with 1661 incident BCs diagnosed before January 2022. The imaging-only AI risk model extracted mammographic features and age at screening. Additional lifestyle/familial risk factors were incorporated into the lifestyle/familial-expanded AI model. Absolute risks were calculated using the two models and the clinical Tyrer-Cuzick v8 model. Age-adjusted model performances were compared across the 10-year follow-up.

RESULTS

The AUCs of the lifestyle/familial-expanded AI risk model ranged from 0.75 (95%CI: 0.70-0.80) to 0.68 (95%CI: 0.66-0.69) 1-10 years after study entry. Corresponding AUCs were 0.72 (95%CI: 0.66-0.78) to 0.65 (95%CI: 0.63-0.66) for the imaging-only model and 0.62 (95%CI: 0.55-0.68) to 0.60 (95%CI: 0.58-0.61) for Tyrer-Cuzick v8. The increased performances were observed in multiple risk subgroups and cancer subtypes. Among the 5% of women at highest risk, the PPV was 5.8% using the lifestyle/familial-expanded model compared with 5.3% using the imaging-only model, p < 0.01, and 4.6% for Tyrer-Cuzick, p < 0.01.

CONCLUSIONS

The lifestyle/familial-expanded AI risk model showed higher performance for both long-term and short-term risk assessment compared with imaging-only and Tyrer-Cuzick models.

Utility of polygenic risk scores in UK cancer screening: a modelling analysis

BACKGROUND

It is proposed that, through restriction to individuals delineated as high risk, polygenic risk scores (PRSs) might enable more efficient targeting of existing cancer screening programmes and enable extension into new age ranges and disease types. To address this proposition, we present an overview of the performance of PRS tools (ie, models and sets of single nucleotide polymorphisms) alongside harms and benefits of PRS-stratified cancer screening for eight example cancers (breast, prostate, colorectal, pancreas, ovary, kidney, lung, and testicular cancer).

METHODS

For this modelling analysis, we used age-stratified cancer incidences for the UK population from the National Cancer Registration Dataset (2016-18) and published estimates of the area under the receiver operating characteristic curve for current, future, and optimised PRS for each of the eight cancer types. For each of five PRS-defined high-risk quantiles (ie, the top 50%, 20%, 10%, 5%, and 1%) and according to each of the three PRS tools (ie, current, future, and optimised) for the eight cancers, we calculated the relative proportion of cancers arising, the odds ratios of a cancer arising compared with the UK population average, and the lifetime cancer risk. We examined maximal attainable rates of cancer detection by age stratum from combining PRS-based stratification with cancer screening tools and modelled the maximal impact on cancer-specific survival of hypothetical new UK programmes of PRS-stratified screening.

FINDINGS

The PRS-defined high-risk quintile (20%) of the population was estimated to capture 37% of breast cancer cases, 46% of prostate cancer cases, 34% of colorectal cancer cases, 29% of pancreatic cancer cases, 26% of ovarian cancer cases, 22% of renal cancer cases, 26% of lung cancer cases, and 47% of testicular cancer cases. Extending UK screening programmes to a PRS-defined high-risk quintile including people aged 40-49 years for breast cancer, 50-59 years for colorectal cancer, and 60-69 years for prostate cancer has the potential to avert, respectively, a maximum of 102, 188, and 158 deaths annually. Unstratified screening of the full population aged 48-49 years for breast cancer, 58-59 years for colorectal cancer, and 68-69 years for prostate cancer would use equivalent resources and avert, respectively, an estimated maximum of 80, 155, and 95 deaths annually. These maximal modelled numbers will be substantially attenuated by incomplete population uptake of PRS profiling and cancer screening, interval cancers, non-European ancestry, and other factors.

INTERPRETATION

Under favourable assumptions, our modelling suggests modest potential efficiency gain in cancer case detection and deaths averted for hypothetical new PRS-stratified screening programmes for breast, prostate, and colorectal cancer. Restriction of screening to high-risk quantiles means many or most incident cancers will arise in those assigned as being low-risk. To quantify real-world clinical impact, costs, and harms, UK-specific cluster-randomised trials are required.

FUNDING

The Wellcome Trust.

Validation of the BOADICEA model for predicting the likelihood of carrying pathogenic variants in eight breast and ovarian cancer susceptibility genes

BOADICEA is a comprehensive risk prediction model for breast and/or ovarian cancer (BC/OC) and for carrying pathogenic variants (PVs) in cancer susceptibility genes. In addition to BRCA1 and BRCA2, BOADICEA version 6 includes PALB2, CHEK2, ATM, BARD1, RAD51C and RAD51D. To validate its predictions for these genes, we conducted a retrospective study including 2033 individuals counselled at clinical genetics departments in Denmark. All counselees underwent comprehensive genetic testing by next generation sequencing on suspicion of hereditary susceptibility to BC/OC. Likelihoods of PVs were predicted from information about diagnosis, family history and tumour pathology. Calibration was examined using the observed-to-expected ratio (O/E) and discrimination using the area under the receiver operating characteristics curve (AUC). The O/E was 1.11 (95% CI 0.97-1.26) for all genes combined. At sub-categories of predicted likelihood, the model performed well with limited misestimation at the extremes of predicted likelihood. Discrimination was acceptable with an AUC of 0.70 (95% CI 0.66-0.74), although discrimination was better for BRCA1 and BRCA2 than for the other genes in the model. This suggests that BOADICEA remains a valid decision-making aid for determining which individuals to offer comprehensive genetic testing for hereditary susceptibility to BC/OC despite suboptimal calibration for individual genes in this population.

Validation of an Abridged Breast Cancer Risk Prediction Model for the General Population

PREVENTION RELEVANCE

In this prospective population-based cohort study, we show the improved performance of a new risk assessment model compared with a gold-standard model (BCRAT). The classification of at-risk women using this new model highlights the opportunity to improve risk stratification and implement existing clinical risk-reduction interventions.

Incorporating Alternative Polygenic Risk Scores into the BOADICEA Breast Cancer Risk Prediction Model

BACKGROUND

The multifactorial risk prediction model BOADICEA enables identification of women at higher or lower risk of developing breast cancer. BOADICEA models genetic susceptibility in terms of the effects of rare variants in breast cancer susceptibility genes and a polygenic component, decomposed into an unmeasured and a measured component - the polygenic risk score (PRS). The current version was developed using a 313 SNP PRS. Here, we evaluated approaches to incorporating this PRS and alternative PRS in BOADICEA.

METHODS

The mean, SD, and proportion of the overall polygenic component explained by the PRS (α2) need to be estimated. $\alpha $ was estimated using logistic regression, where the age-specific log-OR is constrained to be a function of the age-dependent polygenic relative risk in BOADICEA; and using a retrospective likelihood (RL) approach that models, in addition, the unmeasured polygenic component.

RESULTS

Parameters were computed for 11 PRS, including 6 variations of the 313 SNP PRS used in clinical trials and implementation studies. The logistic regression approach underestimates $\alpha $, as compared with the RL estimates. The RL $\alpha $ estimates were very close to those obtained by assuming proportionality to the OR per 1 SD, with the constant of proportionality estimated using the 313 SNP PRS. Small variations in the SNPs included in the PRS can lead to large differences in the mean.

CONCLUSIONS

BOADICEA can be readily adapted to different PRS in a manner that maintains consistency of the model.

IMPACT

: The methods described facilitate comprehensive breast cancer risk assessment.

Rs867228 in FPR1 accelerates the manifestation of luminal B breast cancer

Formyl peptide receptor-1 (FPR1) is a pathogen recognition receptor involved in the detection of bacteria, in the control of inflammation, as well as in cancer immunosurveillance. A single nucleotide polymorphism in FPR1, rs867228, provokes a loss-of-function phenotype. In a bioinformatic study performed on The Cancer Genome Atlas (TCGA), we observed that homo-or heterozygosity for rs867228 in FPR1 (which affects approximately one-third of the population across continents) accelerates age at diagnosis of specific carcinomas including luminal B breast cancer by 4.9 years. To validate this finding, we genotyped 215 patients with metastatic luminal B mammary carcinomas from the SNPs To Risk of Metastasis (SToRM) cohort. The first diagnosis of luminal B breast cancer occurred at an age of 49.2 years for individuals bearing the dysfunctional TT or TG alleles (n = 73) and 55.5 years for patients the functional GG alleles (n = 141), meaning that rs867228 accelerated the age of diagnosis by 6.3 years (p=0.0077, Mann & Whitney). These results confirm our original observation in an independent validation cohort. We speculate that it may be useful to include the detection of rs867228 in breast cancer screening campaigns for selectively increasing the frequency and stringency of examinations starting at a relatively young age.

UK consensus recommendations for clinical management of cancer risk for women with germline pathogenic variants in cancer predisposition genes: RAD51C, RAD51D, BRIP1 and PALB2

Germline pathogenic variants (GPVs) in the cancer predisposition genes BRCA1, BRCA2, MLH1, MSH2, MSH6, BRIP1, PALB2, RAD51D and RAD51C are identified in approximately 15% of patients with ovarian cancer (OC). While there are clear guidelines around clinical management of cancer risk in patients with GPV in BRCA1, BRCA2, MLH1, MSH2 and MSH6, there are few guidelines on how to manage the more moderate OC risk in patients with GPV in BRIP1, PALB2, RAD51D and RAD51C, with clinical questions about appropriateness and timing of risk-reducing gynaecological surgery. Furthermore, while recognition of RAD51C and RAD51D as OC predisposition genes has been established for several years, an association with breast cancer (BC) has only more recently been described and clinical management of this risk has been unclear. With expansion of genetic testing of these genes to all patients with non-mucinous OC, new data on BC risk and improved estimates of OC risk, the UK Cancer Genetics Group and CanGene-CanVar project convened a 2-day meeting to reach a national consensus on clinical management of BRIP1, PALB2, RAD51D and RAD51C carriers in clinical practice. In this paper, we present a summary of the processes used to reach and agree on a consensus, as well as the key recommendations from the meeting.