The clinical and molecular landscape of breast cancer in women of African and South Asian ancestry

The clinical and molecular landscape of breast cancer in women of African and South Asian ancestry | medRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-P4HH5NV'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search The clinical and molecular landscape of breast cancer in women of African and South Asian ancestry View ORCID Profile GJ Thorn , View ORCID Profile E Gadaleta , AZM Dayem Ullah , LGE James , M Abdollahyan , View ORCID Profile R Barrow-McGee , JL Jones , View ORCID Profile C Chelala doi: https://doi.org/10.1101/2024.05.15.24307435 GJ Thorn 1 Centre for Cancer Biomarkers and Biotherapeutics, Barts Cancer Institute, Queen Mary University of London , London EC1M 6BQ, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for GJ Thorn E Gadaleta 1 Centre for Cancer Biomarkers and Biotherapeutics, Barts Cancer Institute, Queen Mary University of London , London EC1M 6BQ, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for E Gadaleta AZM Dayem Ullah 1 Centre for Cancer Biomarkers and Biotherapeutics, Barts Cancer Institute, Queen Mary University of London , London EC1M 6BQ, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site LGE James 1 Centre for Cancer Biomarkers and Biotherapeutics, Barts Cancer Institute, Queen Mary University of London , London EC1M 6BQ, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site M Abdollahyan 1 Centre for Cancer Biomarkers and Biotherapeutics, Barts Cancer Institute, Queen Mary University of London , London EC1M 6BQ, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site R Barrow-McGee 2 Centre for Tumour Biology, Barts Cancer Institute, Queen Mary University of London , London EC1M 6BQ, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for R Barrow-McGee JL Jones 2 Centre for Tumour Biology, Barts Cancer Institute, Queen Mary University of London , London EC1M 6BQ, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site C Chelala 1 Centre for Cancer Biomarkers and Biotherapeutics, Barts Cancer Institute, Queen Mary University of London , London EC1M 6BQ, UK Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for C Chelala For correspondence: c.chelala{at}qmul.ac.uk Abstract Full Text Info/History Metrics Supplementary material Data/Code Preview PDF ABSTRACT Our study evaluates the molecular and clinical effects of genetic ancestry in African and South Asian women compared to European using a combined cohort of 7,253 breast cancer patients available from four data sources: the 100,000 Genomes Project (UK), The Cancer Genome Atlas (US), the Breast Cancer Now Biobank (UK) and Genes & Health (UK). We find that non-European breast cancer patients present significantly earlier and die at a younger age. The African ancestry group has an increased prevalence of higher grade and hormone receptor negative disease. The South Asian group shows tendency towards lower stage at diagnosis and lower tumour mutational burden. We observe significant differences and similarities in the somatic mutational landscape. We identify differences in germline mutation rates in genes used in genetic testing and others implicated in breast cancer predisposition. Potential therapeutic targets are differentially mutated. We report a higher propensity for homologous recombination deficiency serving as a potential therapy response indicator. Although precision oncology holds the promise of revolutionising healthcare, it could exacerbate the disparities it seeks to eradicate unless research biases are addressed. We harness multimodal data to improve our understanding of ancestry-associated differences in breast cancer and highlight opportunities to advance health equity. INTRODUCTION Breast cancer (BC) is the most commonly diagnosed cancer globally and the leading cause of cancer death in women, with an estimated 2.3 million new cases annually and accounting for about 685,000 deaths in 2020. 1 Racial disparities are observed in these figures, with higher mortality rates reported in ethnic minority populations. While social determinants of health have been shown to contribute to population-based differences in mortality, they do not fully explain the disparity observed. 2 Precision oncology is transforming the landscape of healthcare by offering a future in which the one-size-fits-all approach is superseded by tailored diagnosis and treatment. However, the under-representation of patients from ethnic minority populations in research studies and clinical trials limits the impact of translating these findings to non-white patients, exacerbating racial gaps in care delivery. 3 , 4 Observed disparities in outcomes between ethnic groups are likely perpetuated by differences in the distributions of germline pathogenic variants, the incidence of different subtypes, unique somatic mutations, pharmacokinetic behaviour, and tumour biology and behaviour. 5 – 11 Additionally, reports have shown that disease risk models and polygenic risk scores used for disease stratification can exhibit lower predictive accuracy in non-white populations. 11 – 17 To improve our understanding of how the range of genomic and clinical features contribute to BC in different ethnic groups, we analysed data from four large-scale UK and US projects – Genomics England 100,000 Genomes Project, 18 The Cancer Genome Atlas (TCGA), 19 the Breast Cancer Now Biobank (BCN Biobank) 20 and Genes & Health (G&H) 21 . The Genomics England 100,000 Genomes Project, 22 established in 2013, aimed to sequence 100,000 whole genomes from patients within the National Health Service (NHS) in England to facilitate incorporating genomic medicine into routine healthcare, delivered through 14 Genomic Medicine Centres across England. Collections were based on two themes: rare disease and cancer, with sequenced data being linked to its associated clinical data and then being made available to researchers within a dedicated trusted research environment (TRE). The US TCGA 23 is a comprehensive programme comprising molecular and clinical data from over 20,000 samples spanning 33 cancer types. The multi-omics data generated from sequencing and array-based technologies is publicly available to the research community for use in projects as a research or validation dataset. The BCN Biobank 24 is the UK’s largest unique BC disease-specific collection of high-quality specimens, comprising tissues, serial liquid biopsies and bespoke cell lines, and longitudinal clinical data derived from primary and secondary electronic healthcare records (EHRs). Finally, G&H 25 is a community-based general health study that recruits individuals of Pakistani and Bangladeshi origin within the East London, Bradford and Greater Manchester areas, with germline genomic and clinical data available within its own TRE. These participants are consented for lifelong access to their primary and secondary EHRs and a saliva sample for genetic studies. Pan-cancer and breast cancer-specific sequencing initiatives, such as TCGA, the International Cancer Genome Consortium (ICGC) 26 and METABRIC, 27 represent invaluable data assets for BC research. While information pertaining to ethnicity is not available for the METABRIC cohort, those of the TCGA and ICGC data are comprised of the dominant ethnicities of their sampled cohorts White, Black and East Asian. There is a significant underrepresentation of data from collections within South Asian populations. The paucity of data on a group that comprises up to 20% of the world’s population is now recognised, 28 with initiatives attempting to reduce this research inequality. The collection of data from South Asian patients in this study represents one of the largest UK South Asian BC cohorts currently available. Our study harnesses multimodal data available from large cohorts to improve our understanding of ancestry-associated differences in BC and highlight opportunities to address inequalities and achieve more equitable clinical outcomes ( Figure 1 ). Download figure Open in new tab Figure 1. Graphical summary of the analytical approach. A . Four multi-modal clinical and sequenced cohorts used in this study. B. Data analytics and linkage, focusing on developing a framework for harmonising and linking a defined set of EHR data and deriving a research-ready dataset alongside the sequencing data. Stratification is performed based on the threshold-inferred genetic ancestry of patients. C. Data processing workflow to explore clinical data alongside sequence data. For a given genetic ancestry, we explore the relationship between BC clinical variables, mutational profiles and regulators of drug response. D. This framework supports the use of multi-modal longitudinal EHRs and sequencing in BC to inform care for under-represented superpopulations. RESULTS Cohort Characteristics The study dataset comprised 7,253 BC patients from four cohorts – 3,334 from Genomics England, 2,479 from Barts Health NHS Trust patients within BCN Biobank (BCN Biobank-Barts cohort), 1,076 from TCGA and 364 from G&H ( Figure 2 ). To minimise biological variability between the cohorts, the inclusion criteria were restricted to female BC patients presenting with primary disease. For Genomics England and TCGA, the analytical cohort was further restricted to those patients for whom their primary tumours were sequenced and genetic ancestry (gAncestry), determined from germline sequencing, was available. For G&H, only the gAncestry was available from germline sequencing. For the BCN Biobank-Barts cohort, sequencing data, and gAncestry, were only available for 231 patients that were dually consented to BCN Biobank and Genomics England. For the rest of the BCN Biobank-Barts cohort, self-reported ethnicity (SRE) was used. Download figure Open in new tab Figure 2. Study schema of the four cohorts – Genomics England, TCGA, BCN Biobank-Barts cohort, G&H. For Genomics England, 2,840 patients passed initial inclusion criteria, of whom 2,781 also had gAncestry available (see Methods). This analytical cohort comprised European (EUR, n=2,343 (84.3%)), African (AFR, n=138 (5.0%)), South Asian (SAS, n=123 (4.4%)), East Asian (EAS, n=37 (1.3%)), American (AMR, n=12 (0.4%)) and Admix (Admix, n=128 (4.6%)) populations (Supplementary Figure 1a). The gAncestry profile of the TCGA analytics cohort (n=1,064) differs in its composition of Asian gAncestry groups to Genomics England and comprises EUR (n=821 (77.2%)), AFR (n=125 (11.7%)), SAS (n=4 (0.4%)), EAS (n=56 (5.3%)), AMR (n=5 (0.5%)) and Admix (n=125 (11.7%)) gAncestry populations (Supplementary Figure 1b). As G&H is an initiative that looks at more general health concerns within communities of South Asian British heritage, the participants without a history of BC and without inferred gAncestry information were excluded from initial analyses (56,707 out of 57,047), leaving 340 patients with an inferred gAncestry of SAS in the analysis cohort (Supplementary Figure 1c). Following initial filtering for women with primary disease, the ethnic distribution of the BCN Biobank analytical cohort (n=2,126) was as follows: White (n=1,332 (62.7%)), Black/Black British (n=321 (15.1%)), Asian/Asian British (n=234 (11.0%)), Other Ethnic group (n=142 (6.7%)), Mixed (n=56 (2.6%)), and not stated/unknown (n=41 (1.9%)) (Supplementary Figure 2). Patients within the AMR and EAS gAncestry groups were removed from the final analytical cohorts due to the small sizes of the cohorts, which would significantly reduce the power of any statistical inference on this cohort. Furthermore, the limited number of participants in the TCGA SAS population (n=4), precludes the use of this population as a validation for SAS v EUR comparisons. The clinical and molecular analyses in this study focus on the EUR, AFR and SAS gAncestry superpopulations within all cohorts. To determine the ethnic representativeness of the cohorts relative to the geographical region of collection, SRE was compared to that of the UK Census data for Genomics England and the BCN Biobank-Barts cohort. The ethnic composition of the Genomics England cohort broadly recapitulates that of England (2021 UK Census data – Supplementary Figure 2) with 81.0% White, 4.2% Black/Black British, 9.6% Asian/Asian British, 3.0% Mixed, 2.2% Other. However, it is not comparable to that of London, which comprises a higher proportion of ethnic minority groups (53.8% White, 13.5% Black/Black British and 20.7% Asian/Asian British). The BCN Biobank-Barts cohort closely resembles the London statistics, likely representing the ethnic composition of the area around the key collection site in North East London (Supplementary Figure 2a). The gAncestry composition of the TCGA analysis cohort (Supplementary Figure 2b) presents a higher proportion of AFR gAncestry and a much lower proportion of SAS gAncestry relative to Genomics England (AFR, 11.5% in TCGA v 5.0% in Genomics England; SAS, 0.4% TCGA v 4.4% Genomics England) recapitulating the diversity of the geographically diverse TCGA data collection centres. Concordance Between Self-Reported Population Descriptors and gAncestry There is high, almost identical, concordance between population descriptors and gAncestry in Genomics England (92.9%) and TCGA (92.3%) (Supplementary Figure 1). Concordance between the BCN Biobank SRE and gAncestry for dually-consented patients is 88.5%. However, this figure improves to 94.5% if the Discovery Data Service 29 is used to improve stated ethnicity. G&H exhibited the highest concordance between SRE and gAncestry (99.7%), however this was expected due to the recruitment criteria of this cohort (Supplementary Figure 1c). Patients assigned to the Admix cohort in Genomics England represent the greatest source of discordance between SRE and gAncestry, with 50% concordance between a mixed SRE and Admix gAncestry (Supplementary Figure 1). Descriptive Statistics of Clinical and Molecular Features We explored the associations between gAncestry and key clinical or molecular variables by applying linear regression for numeric variables (such as age at diagnosis and age at death), logistic regression for categorical variables (such as receptor status, tumour stage and grade) and a scaled negative binomial regression for the tumour mutational burden (TMB). The group comprising the largest number of individuals, EUR, was used as a reference against which all other gAncestry groups were compared (Supplementary Methods). The results for Genomics England indicate that AFR and SAS women present to the clinic with BC significantly earlier (5.28 and 6.91 years, respectively, p<0.001) and die at a younger age (8.94 and 13.20 years earlier, respectively, p<0.05) than their EUR counterparts ( Figure 3a , Supplementary Table 1). Patients in the AFR population also present with higher grade tumours (OR 1.88, p<0.05), with a higher incidence of hormone receptor negative (HR-) disease relative to EUR (ER- OR 2.06; PR- OR 2.07, p<0.05). Similar findings were reported in the AFR group of the TCGA, with these patients presenting at a younger age (4.08 years earlier, p<0.0015) and with increased HR- disease compared to the EUR group (ER- OR 2.9, PR- OR 2.37, p<0.001) (Supplementary Figure 4a, Supplementary Table 2). There is a tendency for AFR patients to present with slightly lower ancestry-corrected TMB relative to their EUR counterparts, with this being significantly lower in the SAS population (p<0.05). Download figure Open in new tab Figure 3. a. Forest plots of the clinical and molecular features of the non-EUR cohorts relative to EUR. b. Age of diagnosis distributions for EUR (Genomics England/TCGA)/White (BCN Biobank), AFR (Genomics England/TCGA)/Black (BCN Biobank), SAS (Genomics England)/Asian (BCN Biobank and G&H). c. Visualisation of the gAncestry-derived screening windows in the Genomics England cohort. A relationship was observed between gAncestry and quintiles of index of multiple deprivation (IMD) within Genomics England, with AFR and SAS patients more likely to reside in areas within the more deprived quintiles, an association that is more pronounced in the former ( Figure 3a , Supplementary Table 1). Tests to determine whether IMD was confounding the analyses for the other variables showed no significant difference when adding IMD as the second variable (Supplementary Table 3), bar the age at diagnosis (likelihood-ratio test X 2 = 14.0 (4 df), p<0.01). A test of clinical variables versus IMD alone for the largest gAncestry group (EUR) showed that these patients in the most deprived quintile present 2.38-3.12 years earlier than those in the other four quintiles, where age at diagnosis was similar. No other factor was significantly associated with IMD (Supplementary Table 3, Supplementary Table 4). Associations between these clinical variables and SRE were investigated in the BCN Biobank dataset. In agreement with our Genomics England gAncestry-based findings, Black and South Asian patients in the BCN Biobank presented earlier (2.54 years, p<0.01 and 2.75 years, p<0.01, respectively) and died at a younger age (6.48 years, p<0.01, and 6.21 years, p<0.05, respectively) relative to their White counterparts (Supplementary Figure 4b, Supplementary Table 5), although the difference between the populations is smaller. Examining the IMD distribution from dually consented EUR patients within Genomics England shows that this particular subset derives from areas of higher deprivation, which may account for the reduction in effect size if this occurred across the whole BCN Biobank. Furthermore, patients within these ethnic groups also tended to present with aggressive – significantly higher frequencies of high-grade tumours and lymph node involvement – HR- disease. We examined each gAncestry group within Genomics England split based on a 50-years-old cut-off, to represent the age at which the NHS breast screening is initially offered to individuals, however this resulted in small numbers of participants in these sub-groups, increasing the uncertainty in our point estimates (Supplementary Figure 5a). Patients within the <50-year-old AFR group show a propensity to present with higher grade tumours, HR- disease and a HER2+ receptor status. In addition, the TMB of this younger cohort tends to be higher than their EUR counterparts (Supplementary Figure 5a). The clinical features of the <50-year-old SAS group are like those of the corresponding EUR group, bar a potential trend towards PR- disease, a trend that appears inverted in the ≥50-year-old group. Further examination of TMB trends stratified by ER status across the cohort identified higher median TMB in ER- patients (2.54 muts/Mb) compared to ER+ patients (1.33 muts/Mb; p<2.2e-16). Finally, as observed in the unstratified analysis, non-EUR patients in the ≥50-year-old cohort died at a younger age (5.80 years and 9.47 years earlier for AFR and SAS, respectively, p<0.05) (Supplementary Figure 5a). The age at diagnosis curves for each gAncestry show differences between the three populations. The EUR cohort distribution curve of age at diagnosis in the Genomics England analytical cohort is flatter relative to the non-EUR groups, with two modes of similar heights at 51 and 70 years ( Figure 3b , Supplementary Figure 5b). There is a unimodal age distribution for patients of AFR and SAS gAncestries, with prominent peaks at 47 and 49 years, respectively. The age distribution of the EUR cohort of the TCGA exhibits a mild bimodal distribution, with the highest peak presenting at 63 years and the mean age at diagnosis being similar to that of the Genomics England analytical cohort (59.42 years and 61.7 years, respectively) ( Figure 3b , Supplementary Table 2). The White cohort from BCN Biobank, whilst showing a larger IQR than non-White cohorts, appears to shift towards a younger age at presentation (58.41 years) ( Figure 3b , Supplementary Table 5). As mentioned, this could be attributable to the BCN Biobank recruitment centre being located close to areas of higher deprivation in North East London. Similar trends in age distribution profiles are observed between the AFR/Black cohorts of Genomics England, TCGA and BCN Biobank, and the SAS cohorts of Genomics England and G&H ( Figure 3b , Supplementary Tables 2 and 5). The South Asian cohort of the BCN Biobank exhibits a wider peak than that from both Genomics England and G&H, although the mean age of diagnosis is similar (BCN Biobank 55.2 years; Genomics England 54.8 years). This may be due to the differences between ethnic subgroups within BCN Biobank: Bangladeshi patients (n=82) had a mean age at diagnosis of 53.99 (s.d.=11.94 years); Pakistani patients (n=65) had a mean age of 54.37 (s.d.=12.54); but the Indian patient group (n=87), had a higher mean age of 57.20 (s.d.=10.03). We stratified each gAncestry cohort within Genomics England based on the central 60% of the age distribution, to improve equity between the cohorts in breast screening. Current guidelines apply at the 20 th percentile in the EUR age distribution, so we placed the lower bound on the intervals at that percentile in the other gAncestries. We also increased the upper bound to the upper 20 th percentile of the age distribution. This resulted in gAncestry-specific intervals: 50-75 years (EUR), 47-69 years (AFR) and 45-68 years (SAS) ( Figure 3c ). With gAncestry not readily available in the clinic, we isolated patients with concordant SRE and gAncestry assignments and confirmed the applicability of these modified windows (Supplementary Figure 5b). Applying the screening windows to BCN Biobank patients not dually-consented with Genomics England, gives better coverage for all three major ethnic groups: 67.4% (793/1176) within White patients, 64.8% (190/293) within Black patients and 70.5% (148/210) within Asian patients, and applying the suggested Asian age window to the G&H cohort would improve coverage from 26.4% (96/363) for the standard 50-70 window to 59.2% (215/363) within this group. gAncestry-Associated Somatic Variants within the Genomics England Cohort The variant landscape of the TCGA cohorts has been described previously. 6 To determine gAncestry-associated somatic variant profiles within the Genomics England cohort, two models were applied to identify differentially mutated genes and variants relative to the EUR comparator: logistic regression (all gAncestries jointly, with EUR as reference), and a threshold criterion. The latter was implemented to identify variants exclusively present in one cohort, where the application of a logistic regression model would fail to converge (Supplementary Methods). Significant differences in the mutational frequencies of 481 genes in AFR group and 275 genes in the SAS group were observed (Supplementary Table 6). Of these, the logistic classifier identified 2 and 19 genes ( Figure 4 ) and the threshold classifier 479 and 269 genes in the AFR and SAS cohorts respectively (Supplementary Table 6). Six genes ( RBM5 , OTOF , FBXW7 , NCKAP5 , NOTCH3 and GPR158 ) were found commonly differentially mutated between the AFR and EUR populations of the Genomics England and TCGA cohorts, with RBM5 also enriched in the SAS gene list for Genomics England. Download figure Open in new tab Figure 4. Somatic genes and variants identified as differentially mutated in non-EUR versus EUR gAncestry (in the Genomics England cohort) by the logistic regression model: a. EUR v AFR: i. significant genes; ii. significant variants; b. EUR v SAS: i. significant genes; ii. significant variants. We identified 71 variants with significant differences in mutational frequencies in the AFR cohort and 60 variants in the SAS cohort (Supplementary Table 7). Of these, the logistic classifier identified 7 and 4 variants ( Figure 4 ) and the threshold classifier 70 and 52 variants in the AFR and SAS cohorts respectively (Supplementary Table 7). Running the logistic regression on AFR and SAS populations (excluding the EUR population) determined that the computed effect sizes were equal to the differences between the EUR and non-EUR comparisons, resulting in no novel observations. Germline Variant Profiles of Cancer Susceptibility Genes We applied logistic regression, using the EUR population as baseline, to identify gAncestry-associated germline variants within the Genomics England (EUR, AFR and SAS) and the TCGA (EUR and AFR) cohorts. This analysis focussed on variants within 180 germline genes: those tested in the clinic (as reported in the NHS Genomic Medicine Service Panels v80); 30 and those conferring susceptibility to cancer (as reported in the using the PanelApp gene panel. 31 Variants within these genes were prioritised based on the level of evidence supporting a ClinVar classification of pathogenicity (see Methods). The same analytical parameters were applied to germline variants in the G&H cohort to allow for case:control observations within the SAS population. Differences in the prevalence of germline variants of cancer susceptibility genes were observed between the gAncestries ( Figure 5a , Supplementary Table 8). The mean number of both pathogenic variants and those of unknown significance (VUS) per participant was higher in the non-EUR populations, with participants in the AFR group exhibiting the greatest number of variants in both classifications (6.45 pathogenic variants per participant and 21.23 VUS per participant) relative to the reference EUR population (5.27 and 19.54, respectively). Download figure Open in new tab Figure 5. a. Pathogenic classification of variants across the gAncestry groups within Genomics England identify more pathogenic variants and VUS per patient in non-EUR populations. b. 45.9% (17/37) of differentially mutated genes in Genomics England were also differentially mutated in TCGA. ORs for Genomics England and TCGA shown (when difference is significant in either cohort) Of the pathogenic (n=37) and VUS (n=43) variants identified as significantly mutated between EUR and AFR populations in Genomics England, a 36.3% overlap was observed with those reported in the TCGA between the same populations (pathogenic, 45.9%; VUS 27.9%) ( Figure 5b , Supplementary Table 8). Only one of the 23 pathogenic variants identified as significantly mutated between SAS and EUR populations, BRCA1, was also identified in the case:control study using G&H, supporting that this gene confers susceptibility to BC ( Figure 5b , Supplementary Table 8). The AFR and SAS groups in Genomics England were both significantly enriched for BRCA1 (AFR, OR 6.21; SAS, OR 4.27) mutations as well as those commonly associated with mechanisms of DNA damage: nucleotide excision repair; mismatch repair; Fanconi anaemia germline mutations (Supplementary Table 8). In addition, the AFR population also exhibited a higher frequency of additional variants associated with breast cancer predisposition: BRCA2 (OR 2.24) and PALB2 (OR 2.92). To test whether ER status could be confounding this analysis specifically for germline BRCA mutations, we split the Genomics England analytic cohort by ER status and found a small difference between ER- disease (7.5% of patients, n=22/292) and ER+ disease (7.1% of patients, n=116/1621), which was not significant (Fisher’s exact p-value=0.8061). Mutational Signatures and Homologous Recombination Deficiency To determine the contribution of environmental exposures to differences in mutation frequencies between the gAncestry groups within Genomics England, we implemented the signature.tools.lib R library 32 to perform mutational signature analyses confined to breast-specific parameters. Our dataset comprised 32,558,096 substitutions, 339,207 double substitutions, 15,384,289 indels, and 489,339 rearrangements ( Table 1 ). There are significantly more substitutions, double substitutions, and rearrangements per patient in the AFR group, and fewer indels, whereas the SAS group tended to mirror this trend with fewer substitutions, double substitutions and rearrangements but more indels. View this table: View inline View popup Download powerpoint Table 1. Types of somatic variation present in each gAncestry, with comparisons made against EUR (* <0.05, ** <0.01, *** <0.001). Three common signatures were identified as occurring in significantly differing proportions in the AFR or SAS group relative to EUR ( Figure 6a ). SBS1 (associated with increasing age) and SBS3 (associated with homologous recombination deficiency, HR-deficiency) were reported in high proportions across all gAncestry groups (SBS1 EUR 95.2%, AFR 99.3%, SAS 94.3%; SBS3 EUR 90.3%, AFR 97.8%, SAS 87.8%, Supplementary Table 9a), with a significant increase in frequency observed in the AFR population. SBS2 (APOBEC) was significantly lower in the SAS group relative to EUR. The rare SBS57 (identified as a possible sequencing artifact) was observed at a significantly higher frequency in the SAS group. Download figure Open in new tab Figure 6. Mutational signatures and Homologous Recombination Deficiency. a. Mutational signatures within each cohort with significance computed using EUR as reference. b. Relative contributions of each component of the HRDetect scores used to determine BRCA1/BRCA2 deficient tumours in the gAncestry groups. Stars above each split violin indicate significant differences compared to the EUR baseline (* <0.05, *** <0.001). The double base substitution signature DBS12 (correlated with SBS105) was more prevalent in the AFR group (Supplementary Table 9b), with the other double base signatures showing similar proportions across all three groups. Rearrangement signatures RefSig R4 (dominated by clustered translocation patterns, associated with CDK4 driver mutations) and RefSig R5 (characterised by unclustered deletions <100kb) were more prevalent in AFR group, with a trend towards lower prevalence of both RefSig R4 and RefSig R5 in the SAS population (Supplementary Table 9c). The supervised lasso logistic regression model HRDetect was used to determine HR-deficient tumours ( Figure 6b ). There is a trend for more HR deficiency within the AFR population (17/138 – 12.3%) and lower HR deficiency within the SAS population (10/123 – 8.1%) against the EUR population baseline (225/2343 – 9.6%) (Supplementary Table 10a). While the AFR population exhibits higher proportions of SBS3, this trend is reversed in the stratified HR- deficient groups, with SBS3 presentation lower in the AFR HR-deficient group relative to that of the EUR population ( Figure 6 ). Similarly, the RefSig R5 rearrangement contribution is also significantly higher in the AFR HR-deficient group. This suggests that the mechanism of HR deficiency could be different for different gAncestry groups. In HR-deficient SAS tumours, the contribution from deletions with microhomology is significantly higher compared to the HR- deficient EUR tumour baseline. Trends towards lower contributions from deletions with microhomology and loss-of-heterozygosity are seen in HR-deficient AFR tumours, although this is not significant. HR-deficient SAS tumours tended to have lower RefSig R5 contributions, but a similar loss-of-heterozygosity score to those of HR-deficient EUR tumours (Supplementary Table 10b). Survival analyses 33 splitting the cohort into HR-deficient and HR-proficient groups showed that, although the mechanisms of HR deficiency are different between the cohorts, overall survival for HR-deficient tumours was significantly lower than HR-proficient tumours (p<0.001, HR 2.4, Supplementary Figure 6). However, when further stratifying the cohort by gAncestry, only the EUR remained significant (p<0.001, HR 2.5). Pharmacological potential of genomic differences observed between the gAncestry groups Somatic Eight significantly mutated somatic genes ( GPR158, FBXW7, BRIP1, CDKN2A, CHEK2, KDM6A, RET, STK11 ) in the non-EUR populations were identified as candidates for therapeutic targeting in the OncoKB (accessed 10/10/2024), 34 within clinical trials 35 and the literature ( Figure 7a ). 36 , 37 Download figure Open in new tab Download figure Open in new tab Figure 7. Pharmacogenomic potential of the Genomics England cohort. a. Somatic variation. i. Alluvial plot of clinically-actionable targets. Ii. Identification of windows of opportunity within a patient therapeutic journey and b. Germline variation. i. Alluvial plot of clinically-actionable targets. Ii. Identification of windows of opportunity within a patient therapeutic journey. (l) BRCA1/2 variation only; (r) including PALB2 variants. HR deficiency has also been associated with response to PARPi as well as a propensity for resistance to taxane therapies. 38 , 39 We assessed this on a specific cohort of BCN Biobank patients for whom complete longitudinal healthcare data were available from our local Barts NHS Trust and that best represented the EUR, AFR and SAS comparisons presented in this study (n=160) (see Methods). In this BCN Biobank pharmacogenomic exploratory cohort, we identified 17 (10.6%) patients as HR-deficient none of whom were administered PARPi. Furthermore, 70.6% of patients in the HR-deficient cohort were administered therapeutic regimens inclusive of a taxoid agent, of which 47.1% had a secondary event indicative of progression (e.g., recurrence, metastasis or death) compared to no events indicative of progression recorded in the non-taxane group ( Figure 7a ). Germline Eight of the germline genes ( BRAF, BRCA1, BRCA2, DPYD, KIT, MET, PALB2, VHL ) tested were identified as enriched in non-EUR cohorts and as potential pharmacogenomic candidates ( Figure 7b ). BRCA1, BRCA2 and PALB2 , known genes in BC gene testing panels, germline mutations have been shown to confer susceptibility to PARPi. 40 – 43 An eighth of participants in the BCN Biobank pharmacogenic cohort (12.5%) that were not administered PARPi as a part of their clinical journey had pathogenic germline mutations in BRCA1/2 ( Figure 7b ). Furthermore, when including PALB2 , a gene mutated in significantly higher proportions in the AFR population relative to EUR, the fraction of patients that may have benefitted from inclusion of PARPi as a part of the treatment regimen increased to 31.3% ( Figure 7b ). Survival analyses of the differentially mutated somatic and germline variants on our Genomics England cohorts split by gAncestry did not show differences in survival for those mutations (data not shown). 33 However, the analysis is limited by the short length of follow-up data on several participants, leading to a dominant right-censored data effect that adds bias to the survival estimates. METHODS Data Collection and Collation Genomics England clinical data For the clinical data for our Genomics England cohort, the following tables within the Genomics England TRE (v17; March 2023) were queried using LabKey via the R/LabKey API. The relevant participant IDs were identified from the cancer_participant_disease table, with other primary and secondary tables queried using these unique Participant IDs as primary keys (Extended Methods). This gave an initial BC Genomics England cohort of 3,336 participants. One participant was in the database twice under two participant IDs, flagged under a duplicate_participant_id entry; one further participant withdrew consent before the v17 data release, leaving 3,334 BC participants in the initial investigative cohort. For each selected data item, a single entry was consolidated for each participant, using the primary data source. In the absence of primary data, data from the secondary source was used, with information extracted from the record closest to diagnosis date, unless explicitly stated otherwise (Extended Methods). Upon collation, the clinical information of the patient with duplicated IDs was assessed for discrepancies and the duplicated information was removed, with the earliest date selected to represent date of diagnosis. The final analytical cohort was determined following clinical and genomic criteria for exclusion as described in the previous section ( Figure 2 ), with three further samples removed following QC checks based on TMB calculations. BCN Biobank clinical data The BCN Biobank extract (accessed 27/04/2023) comprised 2,479 patients from Barts Health (BH) NHS Trust. BH is the largest NHS trust in London with five hospitals (St Bartholomew’s Hospital, The Royal London Hospital, Mile End Hospital, Newham Hospital and Whipps Cross University Hospital) serving 2.5 million people across the diverse population of North East London. BCN Biobank-Barts data are linked to EHRs from BH via the NHS or hospital number. 2,126 were female and had complete clinical data associated with incidence of unilateral primary BC. Of these, the genomic data from 195 samples were available from the Genomics England TRE as these patients were consented by both BCN Biobank and Genomics England. Basic demographic statistics (age, ethnic group) was updated using linkage between this dataset and the primary care data set from Discovery East London programme (extract date 16/11/2023). 29 Concordance between reported ethnicity (within BCN Biobank) and gAncestry for the dually-consented patients is 88.5%, for the primary care data alone the concordance is 86.8%. Using primary care ethnicity data when not specified in BCN Biobank improves concordance between the reported ethnicity and gAncestry to 94.5%. These dually-consented participants (BCN Biobank and Genomics England) form the basis of our BCN Biobank pharmacogenomic exploratory cohort. With the South Asian ethnic group being a focus of the study, participants from “Any Other Asian Background” were moved to the “Other” ethnic category to prevent dilution with possible Middle-Eastern, Chinese or other South-East Asian ethnic groups. The calculated BCN Biobank age at diagnosis was used within the Genomics England cohort for those patients who were dually consented. TCGA clinical data TCGA clinical data for the TCGA-BRCA study (n=1,098) was downloaded from the Genomic Data Commons Data Portal, 44 and filtered for all those cases with neoplastic breast disease (n=1,076). This was further filtered to remove male patients, leaving 1,064 cases. G&H clinical data We used the G&H clinical dataset (December 2023 release), selecting 57,047 volunteers with basic demographic information available (55.4% female; 44.6% male) through primary care records from Discovery East London programme or secondary care records from Barts Health NHS Trust and NHS Digital. We filtered the available data to define individuals with BC (n=364), if the EHR contains a clinical code (ICD-10 or SNOMED CT diagnosis) indicative of the present or past malignancy of breast. Genomics England genomic data All genomic data is accessible from the Genomics England TRE. Samples of germline genetic material and tumour tissue were sequenced, mapped to the hg38 genome assembly, and variants called: germline variants using the Isaac single sample short variant caller and somatic short variants using Strelka2 45 on the somatic and matched germline. Larger copy number changes and structural variants were called using the Manta (structural variant calling) 46 and Canvas (copy number changes) 47 pipelines. Genomics England reports of small somatic variants (SNVs and indels) were obtained from the cancer_tier_and_domain_variants table, with locations for the raw germline vcf data held in the cancer_analysis table. Ancestry inference was conducted by Genomics England as described previously. 48 In brief, a random forest classifier was applied on the first 8 principal components derived from 188,382 good quality SNP locations within the 1000G phase3 data. The classifier generated 400 trees, with the proportion of trees classifying a germline vcf into a gAncestry superpopulation (AFR, EUR, EAS, SAS, AMR). Assignment of a sample into a gAncestry group was performed using a ≥0.8 proportion threshold, with individuals in the Admix group defined as those without a gAncestry proportion exceeding the defined threshold. The superpopulation proportions for each vcf are available from the aggregate_gvcf_sample_stats table in Genomics England. Up-to-date information from EHRs, comprising structured and unstructured data, was accessible for the patients dually consented by both Genomics England and the BCN Biobank, with clinical data from other resources showing propensity for being relatively sparse and/or fragmented. A BCN Biobank pharmacogenomic exploratory cohort (n=160) was compiled from this dually-consented group, taken from those participants which were in the Genomics England EUR, AFR and SAS gAncestry cohorts. TCGA genomic data Simple somatic variation data for all TCGA-BRCA patients was downloaded from the GDC data commons, and then filtered using the TCGA participant ID for our TCGA cohort. This file arises from paired-sample consensus calling using MuSE, 49 VarScan2, 50 Pindel 51 and Mutect2 from GATK. 52 The gAncestry calls were accessed from a previous study 6 that defined the consensus ancestry call from five calling methods. Germline data for the TCGA-BRCA cohort was downloaded from the PanCanAtlas 53 from the analysis of 10,389 TCGA cancers and filtered for those participants which overlap our TCGA analyses cohorts. This file is based on consensus variants arising from multiple variant calling methods (MuSE, 49 VarScan2, 50 Pindel 51 and HaplotypeCaller from GATK 52 ). G&H genomic data Germline data for the G&H cohort was available within its Trusted Research Environment for 44k participants (July 2021). This data release, arising from GATK HaplotypeCaller with hard post-calling filtering on WES data was used to derive for our breast cancer cases (n=296) and gAncestry-matched control (n=17626) participants. Data Analysis Demographic and clinical features Demographic data from the non-White/non-EUR cohorts were compared against the White/EUR reference using methods described previously. 6 Where age at diagnosis and age at death were available, a linear model was applied; similarly, a scaled negative binomial model was applied to the tumour-mutation burden. For other variables, a logistic regression was performed, with upper levels of the factor compared against the reference level. For Genomics England, upper levels with small numbers (particularly in the AFR and SAS gAncestry groups) were grouped together. For receptor statuses, positive receptor status was taken as the reference level. In all clinical comparisons, for the Genomics England cohort, the gAncestry groups were jointly modelled using unordered gAncestry as a factor. To confirm the effect sizes in a direct AFR against SAS comparison, we split these two groups and ran the regressions separately. In Genomics England, because of the potential confounding effect of IMD on demographic data, we performed two tests. Firstly, a nested likelihood ratio test of the demographic factors was conducted, comparing a reduced model (gAncestry as sole predictor) with a full model (gAncestry and IMD as predictor variables), using the lmtest R package (v 0.9-40). Secondly, the association between IMD and other demographic factors were tested within the Genomics England EUR gAncestry group, using the arsenal R package (v 3.6.3). Tumour mutational burden Tumour mutational burden for the Genomics England cohort was calculated following the method previously described, 54 correcting for gAncestry. The chosen gAncestry reference population within gnomAD 55 was AFR for AFR, SAS for SAS and NFE (non-Finnish European) for EUR. Non-synonymous somatic variants within exonic regions were filtered based on the relevant gnomAD population frequency, the COSMIC database 56 and the TOPMED (v3) database. 57 A mutation was retained if it satisfied one of the following criteria: (i) gnomAD population frequency was ≤ 0.1% (rare) and the VAF was ≥ 3% or if the gnomAD population frequency ≤ 0.1% and the VAF was 0.1% but ≤ 10 and annotated by at least two COSMIC identifiers ( Figure 8 ). The variant was discarded if it appeared in the TOPMED database at a frequency > 0.1%. The count of retained mutations per sample was divided by the size of the exome (≈ 35.4 Mb) and multiplied by 10 6 to calculate the TMB per Mb of exome. Download figure Open in new tab Figure 8. Method for determining retained somatic variants for TMB calculation. Differential somatic variation determination (Genomics England cohort) Once the filtered variant list from the TMB calculation was determined, both mutated genes and the individual variants were collated separately. For each gene or variant in the lists, a logistic regression was performed of the following form (comparing the three gAncestry groups jointly as an unordered factor): where the gAncestry reference level was EUR. If the adjusted p-value for the gAncestry odds ratio was less than 0.1, the gene or variant was considered differentially present. However, in the cases where variants were not present in one of the two comparative cohorts, the algorithm did not converge, and so a second method was employed. In this second method, mutated genes or variants were marked as present or absent in a cohort if present at above or below 2% in the gAncestry group (a threshold of 47/2343 in EUR, 3/138 in AFR and 3/123 in SAS). The thresholded lists of genes and variants present in each cohort were intersected to give seven sets of intersections (EUR only, AFR only, SAS only, EUR+AFR, EUR+SAS, AFR+SAS, EUR+AFR+SAS). For our purposes, we examined those genes or variants present at above 2% in AFR only and above 2% in SAS only, with the implication that they are present at <2% in the other two cohorts. To control for the potential confounding factors of germline BRCA1 and BRCA2 status and IMD in the logistic model, we added extra terms and restricted to those participants for which all the data was available: For the logistic regression model, given the sample sizes for each gAncestry within Genomics England, we have 90% power to detect an OR 5.86 (SAS)/5.12 (AFR) down to 1.97 (SAS)/1.89 (AFR) as mutation frequency increases from 5% to 25% in the minority gAncestry. Germline variations (Genomics England cohort, TCGA and G&H) The genes we selected for analysis were combined from a list of those tested in the clinic (as reported in the NHS Genomic Medicine Service Panels v8 30 ) and those conferring susceptibility to cancer (as reported in the Genomics England PanelApp gene panel 31 ), leaving 180 genes in total. For Genomics England and TCGA, germline vcf files were passed through a pipeline to collate and annotate variants in selected genes using bcftools and VEP (version 107) 58 . For determining pathogenicity, we restricted to the canonical transcript, and filtered further using the pathogenicity classification and confidence from ClinVar ( Figure 9 ). Download figure Open in new tab Figure 9. Method for determining germline pathogenic/LoF mutations. For Genes and Health, the joint germline WES vcf, which had been previously annotated using VEP version 107, was filtered on the genes and the participants in our case and control and passed through the selection criteria ( Figure 9 ). The G&H WES release did not have variants on chromosome X because of ploidy issues when using GATK HaplotypeCaller, so genes from our list located on this chromosome were not included for the case:control study. As per the demographic calculation, a logistic model was fitted for the forest plots (with EUR in Genomics England/TCGA or control as the reference level for the factor ): Note that in the Genomics England cohort, the factor is the unordered gAncestry with EUR as the reference level. Signature and HRDetect calculation (Genomics England cohort) The HRDetect 32 predictor was used to detect homologous recombination-deficient tumours within our Genomics England analytical cohort. Here, the somatic short-variant vcfs available within the Genomics England TRE were split into SNVs and indels without filtering and the somatic copy number vcfs were split into SVs and CNVs. The R library was used to process the SNVs into SNV and DNV catalogues, and the SVs into SV catalogues. Finally, the HRDetect function was applied on the four categories of somatic variants (SNPs, indels, SVs and CNVs) to compute the homologous recombination deficiency score using their pretrained classifier. Survival analyses (Genomics England cohort) Survival analysis was performed as in the recent pan-cancer Genomics England paper, 33 using the code and methods provided, on our cohort, both as a whole, and split between the three gAncestry groups EUR, AFR and SAS. DISCUSSION We present a comprehensive analysis of clinical and molecular features associated with African and South Asian genetic ancestry from a cohort of 7,253 breast tumours. Our findings highlight the importance of addressing the racial disparity in research to optimise precision oncology, ameliorating outcome, and guiding the patient clinical journey. Each cohort examined in this study uses different methods for calling germline variants (Methods). The callers applied display similar sensitivities for SNVs and short indels on tissue samples, but TCGA’s approach of taking a consensus call increases confidence in their final variants. 59 Similarly, for somatic small variants, both Strelka2 (Genomics England) and Mutect2 (TCGA) use a statistical model for calling variants from paired samples but differ in their application of the model: Mutect2 uses a Bayesian classifier and Strelka2 uses a trained machine learning method. Mutect2 trades accuracy for lower sensitivity, but Strelka2 increases sensitivity at the risk of false positives, but on paired tissue samples, their results are comparable. 60 It is not recommended to use self-identified population descriptors as proxies for gAncestry groups due to their categorical and continuous properties, respectively. 61 So, while we report a high concordance between self-reported ethnicity or race and gAncestry, our study focusses on use of gAncestry as an analytical variable across all cohorts, bar those patients in the BCN Biobank for whom genomic data and associated gAncestry derivations are not available. 61 As such, SRE derived from primary and secondary EHRs were used to inspect overarching trends observed in analyses focussing on the gAncestry groups studied but were not used an analytical determinants per se . Use of the latter for the stratification of populations not only promotes equity, clarity and reproducibility in research methods but also allows for greater examination into the genetic diversity of a population and how this diversity influences disease pathology and response to treatment. 3 , 14 , 62 When considering clinical associations whose effect sizes are biologically meaningful, we report non-EUR/non-White populations to present at younger ages, experience more clinically aggressive HR- disease and suffer from a higher mortality rate relative to the EUR/White population. 5 , 7 , 63 , 64 For the Genomics England cohort, where IMD is available, the distribution of IMD within non-EUR participants within our dataset is skewed to higher deprivation quintiles compared to EUR participants. Testing IMD as a confounding factor did not reduce significance, aside from patients presenting on average 3 years younger when comparing within the EUR cohort. Despite this, it is important that patient cohorts are socio-economically matched as closely as possible to reduce this potentially confounding effect. Even after this matching, it has been shown that long-term activation of the stress response – the cumulative biological burden extracted on the body’s systems due to repeated adaptation to stressors over time (allostatic load) – may also significantly influence disparities in presentation, progression, and treatment. 65 Similarly, systemic barriers to accessing health care also contributes to differing outcomes. 66 , 67 It is debatable as to whether current screening guidelines, in which a single age-based window is applied, benefit all women equally. 5 , 68 – 72 The AgeX and UK Age trials examined the feasibility of lowering the age at which patients are first invited for routine BC screening. While our study did not examine mortality statistics, results from the AgeX trial, which aimed to amend the screening window to 47-73 years, reported a 24% reduction in mortality. 69 , 70 Similarly, the UK Age trial, which investigated the benefits of initiating mammography screening at the age of 40 or 41, demonstrated a 25% reduction in BC mortality after 10 years. 28 However, these findings did not examine ethnicity as an independent factor, thus the long-term benefits of these strategies on reducing BC mortality in women from ethnic minority groups remain largely unknown. Our findings support the premise that current screening guidelines detect imbalanced proportions of BC between the gAncestry groups and that ethnicity–adapted screening windows would allow for greater equity in the screening process. This includes adjusting the screening range for non-EUR cohorts to start at a younger age. In the Genomics England cohort, these revised windows would be EUR 50-75 years (to detect 59.47% of BCs), AFR 47-69 years (to detect 61.31% of BCs) and SAS 45-68 years (to detect 59.35% of BCs). Applying these screening windows to BCN Biobank patients not dually-consented with Genomics England gives increased coverage – White 67.4% (793/1176), Black 64.8% (190/293) and Asian 70.5% (148/210) – as does applying these new windows to the G&H data improving from 26.4% (96/363) to 59.2% (215/363) within this group. The extra coverage in BCN Biobank patients could be explained by the potential effect of the socioeconomic factors (as represented by estimated IMD) for these groups compared to Genomics England. The appropriate designation of a bespoke screening model, which would provide an indicative screening window and guide the modality to be used during the screen, requires the interplay of multiple complex factors, such as demographics, family history of inheritable cancers, mammographic density, previous benign breast conditions and the results of genetic testing for gAncestry-based susceptibility genes. Input from patient advocates and representation from social groups with low current uptake are also necessary to increase the penetrance of earlier intervention, working to reduce reluctance in seeking medical assistance or taking part in screening. Literature associating TMB with gAncestry or ethnicity reports that AFR patients tend to have a higher TMB relative to EUR. 73 , 74 The corresponding trend in Asian patients is equivocal, with reports of TMB being both higher and lower in this group relative to EUR. 73 – 75 This is likely due to the combination of the presence of EAS and SAS gAncestry within Asian study groups, with no distinction made between the two. We found TMB to be significantly lower in the SAS population, with a similar trend in the AFR population, compared to that in the EUR group. When patients are stratified based on screening age (50 years), TMB is not significantly different between non-EUR populations and their corresponding EUR age group, but there is a trend towards higher TMB in the <50 years AFR subgroup. These observations could be due to the association of increased TMB in ER- breast cancer, of which a higher proportion are present in the <50 AFR subgroup compared to the EUR counterpart. Although, overall, higher TMB is associated with poorer survival in many cancers including BC, 76 very high TMB (≥50 mut/Mb) has been reported to have a protective effect in immunotherapy-naïve patients, 77 where it is attributed to increased cell lethality from extreme genetic instability. TMB, both in isolation and in combination with expression markers, has also been reported as a potential predictive biomarker of response to immune checkpoint inhibitors (ICIs) and has been associated with better treatment response and improved outcomes. 78 , 79 Estimating TMB from tumour-only sequencing (whether WGS, WES or targeted gene panels) is commonly used to determine suitability for ICI treatment, but this may exacerbate disparities between ethnic groups. This is because reference data is predominantly based on EUR gAncestry groups, meaning non-EUR gAncestries are likely under-represented. This reference data is used to identify and remove potential germline variants and true non-EUR gAncestry germline variants can persist after filtering, inflating the estimation of TMB in ethnic minority groups and the use of ICI treatments. Correcting the estimated TMB for gAncestry in this case is imperative. 54 The present study used somatic variants called from paired sequencing, allowing the patients’ own germlines to be used for accurate calculation of the TMB. Six genes with somatic variants differentially mutated between EUR and AFR in the Genomics England cohort ( RBM5, OTOF, FBXW7, NCKAP5, NOTCH3, GPR158 ) were also identified in the TCGA for the same comparison. These genes have been associated previously with key roles in tumourigenesis and progression, 6 , 80 , 81 and as diagnostic predictors poor prognosis in other solid tumours. 82 – 84 Mutations in FBXW7 are found across multiple cancer types in AFR populations 6 , suggesting this alteration is not limited to BC. Mutations in FBXW7 and GPR158 have been identified as therapeutic candidates and have been associated with sensitivity to lunresertib and camonsertib in ovarian cancer, 85 , 86 sensitivity to bevacizumab in osteosarcoma 37 and resistance to paclitaxel in BC. 87 Although both the AFR and SAS cohorts presented with higher grade tumours compared to EUR, none of the somatic differentially-mutated genes were found in the literature have any association to the aggressiveness of the disease. Higher grade tumours are associated with more ER-/HER2-cancers, and also with germline mutations in DNA damage repair genes, which both these cohorts exhibit higher proportions of when compared to the EUR cohort. In addition, six other somatic genes significantly enriched in non-EUR populations in the Genomics England cohort ( BRIP1, CDKN2A, CHEK2, KDM6A, RET, STK11 ) were also identified as pharmacogenomic candidates. Mutations in these genes have been reported to confer sensitivity to PARPi, 88 CDK4/6 inhibitors, 89 EZH2 inhibitor, 90 RET inhibitors 91 and combination tyrosine kinase inhibitors with immunotherapy, 92 respectively, in studies on BC as well as other solid tumours. These findings, suggest repurposing opportunities of some of these drugs to treat breast tumours in non-EUR ancestries. Mutation signatures and HR deficiency inferred from somatic genomic alterations (SNVs/indels and CNVs/SVs) show differences between the three gAncestry Genomics England cohorts. SBS signature 8, associated with HR deficiency, and SBS signatures 2 and 13, related to APOBEC3A and APOBEC3B function, are enriched in AFR (SBS2 and 13) and SAS (SBS8) populations, and have also been reported enriched previously in non-EUR BC. 93 There is a trend towards more HR deficiency in the AFR population and less HR deficiency in the SAS population compared to the EUR population baseline. Splitting the cohort into HR- deficient and HR-proficient tumours shows that overall survival for HR-deficient tumours is significantly lower than HR-proficient tumours, although following stratification by gAncestry only the EUR remained significant. This is attributable to two factors: the small numbers of HR-deficient patients with survival data within the non-EUR populations, leading to large confidence intervals; and that other clinical features may be colliding with HR deficiency and influencing outcome. For instance, our analysis may exhibit bias as patients in non-EUR gAncestry populations tended to present with higher tumour grade (potentially due to increased ER- and/or enrichment of BRCA1 / 2 mutations), which is associated with decreased overall survival. With larger cohorts, allowing for more stratification on tumour grade and HRD status, we would be able to determine how much of the effect on survival was due to HR deficiency, and how much was due to presentation at higher grades in non-EUR participants. HR deficiency classifications inferred from whole genome sequencing attempt to classify BRCA1 / 2 deficiency and identify those patients that exhibit mutational landscapes similar to those with BRCA1/2 defects i.e., BRCAness. The HR deficiency findings support our germline observations in which a higher proportion of germline BRCA1 , BRCA2 and PALB2 mutations are present within AFR and SAS populations. Identification of these genomic alterations provides opportunities by highlighting potential therapeutic vulnerabilities and could be exploited for patient benefit. For instance, research has reported on the clinical efficacy of PARPi, either as a single agent or in combination with existing treatment regimens, in ameliorating outcome in BC patients with germline mutations in BRCA1 /2 and PALB2, as well as BRCAness. 41 – 43 Similarly, BRCA1/2- mutated tumours have also been reported to respond better to platinum therapies. 94 However, BRCAness is known to confer resistance to taxane-based chemotherapies. 95 We identified a subset of patients in our BCN Biobank pharmacogenomic exploratory cohort with drug responses associated with specific germline mutations and/or somatic mutation patterns associated with HR deficiency. Almost 40% of these patients presented with a potential sensitivity to PARPi to whom this was not administered as a part of their therapeutic regime(s). Similarly, we also identified a subset of patients potentially resistant to common therapies in the HR-deficient group. Our findings also show that the landscape of BC-associated susceptibility genes differs between the gAncestry groups. With current clinical genetic panels developed from research that would have been ethnically biased towards a White population, the panel may not accurately represent the mutations, frequency of mutations or risk of cancer in ethnic minority populations. In agreement with previous studies, our findings show that the landscape of BC- associated susceptibility genes differs between the gAncestry groups, suggesting that germline screening protocols modified based on ethnicity could be more informative. 11 , 96 , 97 The penetrance of BRCA1 and BRCA2 is known to be modified by SNPs that influence risk in general population, but the landscape of germline diversity within African populations has yet to be explored, and so there are potentially other ancestry-related germline risk modifiers that are yet to be found. 98 A similar issue arises for germline diversity within Asian populations, with SAS and EAS gAncestries each having their own distinct BRCA1/2 mutation patterns, which are similarly relatively little examined. 96 , 99 Germline mutations in DNA damage repair genes, such as BRCA1, BRCA2, CHEK2 and ATM , are associated with lower efficacy of CDK4/6 inhibitors and endocrine therapy in advanced BC: 100 , 101 This is particularly relevant within the non-EUR populations that tend to present at a later stage. The clinical implications of the benefits of PARPi and CDK4/6 inhibitors discussed above are not specific to a single ethnicity but, instead, are of benefit to all patients. As such, increased genetic testing of patients, using routinely genomic tests from BC clinics and the somatic profiling of tumours could guide the personalised therapeutic management of disease. Our findings highlight the broader issue of under-representation of ethnic minority groups in research and the historical merging of East Asian and South Asian populations biasing the translational implications and resources developed as a result. Use of references developed from research based on predominantly White populations to categorise mutations and define pharmacogenomic implications has been shown to have broader implications in terms of ethnicity-associated response to treatment. For instance, it has been shown that variations in BRCA1/2 in resources such as ClinVar, are not representative of global ethnic populations. 96 Another study based on the G&H cohort identified ethnicity-driven pharmacogenomic consequences unique to British South Asian patients with cardiometabolic disease. Common polymorphisms in CYP2C19 in this population have been associated with poor activation of clopidogrel and recurrent myocardial infarction. 102 We report an imbalance in our research data from Genomics England, TCGA and BCN Biobank. These three data sources are not representative of the countries they derive from, due to the limited number of collection sites in each. The largest estimate of the South Asian American population in the US is 2% of the total US population (US Census Bureau estimate, 2021), which is exemplified by the low percentage of SAS patient data available not only from TCGA but also from the ICGC Pan-Cancer Analysis of Whole Genomes project (PCAWG). 6 This reduces its utility as validation of our results for the SAS gAncestry group if the population is sampled randomly. Thus, while the TCGA dataset can only be used to validate the molecular and clinical data differences between EUR and AFR ancestries, the BCN Biobank clinical data can be used to validate clinical features across all three of our gAncestry cohorts, notwithstanding that the lead collection site of BCN Biobank is in North East London, where the IMD distribution is skewed to more deprived quintiles. Our cohort has the largest proportion of SAS ancestry patients among studies of similar size, reflecting the population structure of England (Supplementary Figure 2a). In fact, METABRIC, 27 examining 2000 tumours, focused more on molecular subtypes and did not report ethnicity or gAncestry differences, leaving our study as the largest study which specifically examines these. The under-representation of the SAS ancestry group in international cancer-associated studies does, however, mean that the variants we report are difficult to validate in publicly available datasets. The scarcity of the stratification of Asian ethnicities in cancer research means that any study of genomic variation using these merged population data would likely be underpowered to identify alterations unique to the South Asian population. In fact, data from large-scale US-led initiatives are likely biased towards East Asian ethnicities (and ancestries) due to them recapitulating their sampling populations. Our genomic findings are limited by the relatively small sizes of the AFR and SAS groups (120-140 patients) within our Genomics England analytic cohort of 2,781 patients. However, this is still larger than most studies which aim to determine gAncestry-related genomic differences within the BC landscape. Nevertheless, this could be improved through the inclusion of more somatic tissue sequencing from our other British-based cohorts, but the overall ratio (of EUR to AFR or SAS patients) would continue to exhibit bias towards the EUR gAncestry, unless the non-EUR groups within the UK (or worldwide) were specifically targeted for collection. However, our study was still powerful enough to find significant differences between EUR and the other two groups. Any further study would need to specifically target the non-EUR groups either within the UK or worldwide to move towards equipoise and extract the more subtle variations that may not currently be apparent. Tackling disparities in research has become a public health priority. One of the Cancer Grand Challenges 103 is focused on cancer inequities; Genomics England has recently implemented the Diverse Data initiative to bridge the ethnic data gap in genomics-driven personalised medicine; and the BCN Biobank is focusing on increasing recruitment from ethnic minority populations. Furthermore, the Breast Cancer Now’s Inequalities Funding Scheme was implemented to encourage applications for research into increasing health equity within BC. These shifts in future data and sample collections will help ensure that findings from research will be more powerful due to better equipoise in ethnic groups ensuring equitability in translational implications. There are evident health disparities in BC diagnosis, therapeutic management, and outcome globally. While determining genetic diversity is important for the advancement of precision oncology, disparities in healthcare cannot be attributed to a single factor but rather stem from a complex web of interlinking clinical, social and genetic factors. To ensure that precision oncology benefits all patients equally, regardless of their ethnic background, it is imperative to foster trans-disciplinary co-operation and conduct multi-modal studies, that incorporate data from primary and secondary healthcare in addition to genomics. Data Availability Genomic and phenotypic data for the 100KGP study participants are available through the Genomics England Research Environment via application at https://www.genomicsengland.co.uk/research/academic/join-gecip . The clinical data and donor ancestry calls for the TCGA cohort used in this study are available from the supplemental information (Table S1) of Carrot-Zhang et al. 6 Additional clinical and genomic data for the TCGA BRCA cohort were accessed from the Genomics Data Commons Data Portal. 44 Clinical data and specimens of dually consented patients are available from the BCN Biobank on application to the Tissue Bank, and clinical and molecular data from G&H is available through the Genes and Health Research Environment via application at https://www.genesandhealth.org/research/scientists-using-genes-health-scientific-research . Code Availability Although clinical and molecular data within the Genomics England TRE have been anonymised, the data are sufficiently detailed that the data could be deanonymised through data linkage, and thus cannot be extracted directly from the TRE. Codes for the analyses from this manuscript are available from the authors on request, but will need importing into the TREs for approved Genomics England Research Network and Genes & Health Research Network participants. Acknowledgements This work was supported by Barts Charity (grant code MGU0504) and Barts NIHR BRC (grant code BTXH1A1R). BCN Biobank is funded by the Breast Cancer Now charity (grant code TB2022BAR). This work forms part of the research portfolio of the National Institute for Health and Care Research Barts Biomedical Research Centre (NIHR203330); a delivery partnership of Barts Health NHS Trust, Queen Mary University of London, St George’s University Hospitals NHS Foundation Trust and St George’s University of London. This research was made possible through access to data in the National Genomic Research Library, which is managed by Genomics England Limited (a wholly owned company of the Department of Health and Social Care). The National Genomic Research Library holds data provided by patients and collected by the NHS as part of their care and data collected as part of their participation in research. The National Genomic Research Library is funded by the National Institute for Health Research and NHS England. The Wellcome Trust, Cancer Research UK and the Medical Research Council have also funded research infrastructure. The results published here are in part based upon data generated by the TCGA Research Network. 19 The authors wish to acknowledge the roles of the Breast Cancer Now Biobank in collecting and making available the samples and/or data, and the patients who have generously donated their tissues and shared their data to be used in the generation of this publication. We are especially grateful to members of the BCN Biobank-BCI (Catherine McMaster-Christie, Rachel Nelan, Jennifer McGuinness, Jenny Gomm and Iain Goulding) for their help in setting up the framework for data collection. We thank Barts Health NHS Trust for their help with the collection of secondary and tertiary care data for BCN Biobank. We thank members of the Discovery East London Programme Board, Discovery Data Service/Endeavour Health Charitable Trust and Voror Health Technologies Ltd for their support in facilitating collection of BCN Biobank primary care patient records. Genes & Health has recently been core-funded by Wellcome (WT102627, WT210561), the Medical Research Council (UK) (M009017, MR/X009777/1, MR/X009920/1), Higher Education Funding Council for England Catalyst, Barts Charity (845/1796), Health Data Research UK (for London substantive site), with research delivery support from the NHS National Institute for Health Research Clinical Research Network (North Thames). Most of all, we thank all of the individuals participating in the Genomics England 100,000 Genomes Project, The Cancer Genome Atlas, Genes & Health and the Breast Cancer Now Biobank. Footnotes ↵ † The first 2 authors should be regarded as joint First Authors. Update the metadata of the supplementary information. No text updated from previous version REFERENCES ↵ Sung , H. et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries . CA Cancer J Clin 71 , 209 – 249 ( 2021 ). doi: 10.3322/caac.21660 OpenUrl CrossRef PubMed ↵ Sadigh , G. et al. Assessment of Racial Disparity in Survival Outcomes for Early Hormone Receptor-Positive Breast Cancer After Adjusting for Insurance Status and Neighborhood Deprivation: A Post Hoc Analysis of a Randomized Clinical Trial . JAMA Oncol 8 , 579 – 586 ( 2022 ). doi: 10.1001/jamaoncol.2021.7656 OpenUrl CrossRef PubMed ↵ Aldrighetti , C. M. , Niemierko , A. , Van Allen , E. , Willers , H. & Kamran , S. C . Racial and Ethnic Disparities Among Participants in Precision Oncology Clinical Studies . JAMA Netw Open 4 , e2133205 ( 2021 ). doi: 10.1001/jamanetworkopen.2021.33205 OpenUrl CrossRef PubMed ↵ Guerra , C. E. & Viswanath , C . Advancing Equity in Cancer Clinical Trials: Lessons From the Evidence . JCO Oncol Pract 18 , 633 – 634 ( 2022 ). doi: 10.1200/OP.22.00390 OpenUrl CrossRef PubMed ↵ Bowen , R. L. , Duffy , S. W. , Ryan , D. A. , Hart , I. R. & Jones , J. L . Early onset of breast cancer in a group of British black women . Br J Cancer 98 , 277 – 281 ( 2008 ). doi: 10.1038/sj.bjc.6604174 OpenUrl CrossRef PubMed Web of Science ↵ Carrot-Zhang , J. et al. Comprehensive Analysis of Genetic Ancestry and Its Molecular Correlates in Cancer . Cancer Cell 37 , 639 – 654 e636 ( 2020 ). doi: 10.1016/j.ccell.2020.04.012 OpenUrl CrossRef PubMed ↵ Copson , E. et al. Ethnicity and outcome of young breast cancer patients in the United Kingdom: the POSH study . Br J Cancer 110 , 230 – 241 ( 2014 ). doi: 10.1038/bjc.2013.650 OpenUrl CrossRef PubMed Web of Science Gathani , T. , Reeves , G. , Broggio , J. & Barnes , I . Ethnicity and the tumour characteristics of invasive breast cancer in over 116,500 women in England . Br J Cancer 125 , 611 – 617 ( 2021 ). doi: 10.1038/s41416-021-01409-7 OpenUrl CrossRef PubMed Goel , N. et al. Racial Differences in Genomic Profiles of Breast Cancer . JAMA Netw Open 5 , e220573 ( 2022 ). doi: 10.1001/jamanetworkopen.2022.0573 OpenUrl CrossRef PubMed Vazquez , E. D. et al. Chemokine receptors differentially expressed by race category and molecular subtype in the breast cancer TCGA cohort . Sci Rep 12 , 10825 ( 2022 ). doi: 10.1038/s41598-022-14734-5 OpenUrl CrossRef PubMed ↵ Yadav , S. et al. Racial and Ethnic Differences in Multigene Hereditary Cancer Panel Test Results for Women With Breast Cancer . J Natl Cancer Inst 113 , 1429 – 1433 ( 2021 ). doi: 10.1093/jnci/djaa167 OpenUrl CrossRef PubMed Evans , D. G. et al. The importance of ethnicity: Are breast cancer polygenic risk scores ready for women who are not of White European origin? Int J Cancer 150 , 73 – 79 ( 2022 ). doi: 10.1002/ijc.33782 OpenUrl CrossRef PubMed Ho , W. K. et al. Polygenic risk scores for prediction of breast cancer risk in Asian populations . Genet Med 24 , 586 – 600 ( 2022 ). doi: 10.1016/j.gim.2021.11.008 OpenUrl CrossRef PubMed ↵ Khor , S. et al. Racial and Ethnic Bias in Risk Prediction Models for Colorectal Cancer Recurrence When Race and Ethnicity Are Omitted as Predictors . JAMA Netw Open 6 , e2318495 ( 2023 ). doi: 10.1001/jamanetworkopen.2023.18495 OpenUrl CrossRef Martin , A. R. et al. Clinical use of current polygenic risk scores may exacerbate health disparities . Nat Genet 51 , 584 – 591 ( 2019 ). doi: 10.1038/s41588-019-0379-x OpenUrl CrossRef PubMed Park , J. I. , Bozkurt , S. , Park , J. W. & Lee , S . Evaluation of race/ethnicity-specific survival machine learning models for Hispanic and Black patients with breast cancer . BMJ Health Care Inform 30 ( 2023 ). doi: 10.1136/bmjhci-2022-100666 OpenUrl Abstract / FREE Full Text ↵ Kantor , O. et al. Racial and Ethnic Disparities in Locoregional Recurrence Among Patients With Hormone Receptor-Positive, Node-Negative Breast Cancer: A Post Hoc Analysis of the TAILORx Randomized Clinical Trial . JAMA Surg 158 , 583 – 591 ( 2023 ). doi: 10.1001/jamasurg.2023.0297 OpenUrl CrossRef PubMed ↵ Genomics England Limited. https://www.genomicsengland.co.uk/initiatives/100000-genomes-project ( 2024 ). ↵ National Cancer Institute Center for Cancer Genomics . The Cancer Genome Atlas Program (TCGA) , https://www.cancer.gov/ccg/research/genome-sequencing/tcga ( 2024 ). ↵ Breast Cancer Now . Breast Cancer Now Tissue Bank , https://breastcancernow.org/breast-cancer-research/breast-cancer-now-tissue-bank ( 2024 ). ↵ East London Genes & Health . Genes & Health , https://www.genesandhealth.org/ ( 2024 ). ↵ The National Genomic Research Library v5.1 Genomics England. doi: 10.6084/m9.figshare.4530893.v7 (n.d.). OpenUrl CrossRef ↵ The Cancer Genome Atlas Research Network et al. The Cancer Genome Atlas Pan-Cancer analysis project . Nat Genet 45 , 1113 - 1120 ( 2013 ). doi: 10.1038/ng.2764 OpenUrl CrossRef PubMed ↵ Abdollahyan , M. et al. Dynamic Biobanking for Advancing Breast Cancer Research . J Pers Med 13 ( 2023 ). doi: 10.3390/jpm13020360 OpenUrl CrossRef ↵ Finer , S. et al. Cohort Profile: East London Genes & Health (ELGH), a community-based population genomics and health study in British Bangladeshi and British Pakistani people . Int J Epidemiol 49 , 20 – 21i ( 2020 ). doi: 10.1093/ije/dyz174 OpenUrl CrossRef ↵ Zhang , J. et al. The International Cancer Genome Consortium Data Portal . Nat Biotechnol 37 , 367 – 369 ( 2019 ). doi: 10.1038/s41587-019-0055-9 OpenUrl CrossRef PubMed ↵ Curtis , C. et al. The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups . Nature 486 , 346 – 352 ( 2012 ). doi: 10.1038/nature10983 OpenUrl CrossRef PubMed Web of Science ↵ Genomics England Limited . Diverse Data https://www.genomicsengland.co.uk/initiatives/diverse-data ( 2024 ). ↵ Discovery Health and Care Data Service . Discovery Data Service , https://www.discoverydataservice.org/Content/Home.htm ( 2023 ). ↵ NHS England . National Genomic Test Directory , https://www.england.nhs.uk/genomics/the-national-genomic-test-directory/ ( 2024 ). ↵ Genomics England Limited . Genomics England PanelApp , https://panelapp.genomicsengland.co.uk/ ( 2024 ). ↵ Davies , H. et al. HRDetect is a predictor of BRCA1 and BRCA2 deficiency based on mutational signatures . Nat Med 23 , 517 – 525 ( 2017 ). doi: 10.1038/nm.4292 OpenUrl CrossRef PubMed ↵ Sosinsky , A. et al. Insights for precision oncology from the integration of genomic and clinical data of 13,880 tumors from the 100,000 Genomes Cancer Programme . Nat Med 30 , 279 – 289 ( 2024 ). doi: 10.1038/s41591-023-02682-0 OpenUrl CrossRef PubMed ↵ Chakravarty , D. et al. OncoKB: A Precision Oncology Knowledge Base . JCO Precis Oncol 2017 ( 2017 ). doi: 10.1200/PO.17.00011 OpenUrl CrossRef PubMed ↵ National Library of Medicine . ClinicalTrials.gov , https://clinicaltrials.gov ( 2024 ). ↵ Youssef , A. , Haskali , M. B. & Gorringe , K. L . The Protein Landscape of Mucinous Ovarian Cancer: Towards a Theranostic . Cancers (Basel) 13 ( 2021 ). doi: 10.3390/cancers13225596 OpenUrl CrossRef ↵ Wang , B. D. et al. Bevacizumab attenuates osteosarcoma angiogenesis by suppressing MIAT encapsulated by serum-derived extracellular vesicles and facilitating miR-613-mediated GPR158 inhibition . Cell Death Dis 13 , 272 ( 2022 ). doi: 10.1038/s41419-022-04620-3 OpenUrl CrossRef PubMed ↵ Chopra , N. et al. Homologous recombination DNA repair deficiency and PARP inhibition activity in primary triple negative breast cancer . Nat Commun 11 , 2662 ( 2020 ). doi: 10.1038/s41467-020-16142-7 OpenUrl CrossRef PubMed ↵ Lord , C. J. & Ashworth , A. BRCAness revisited . Nat Rev Cancer 16 , 110 – 120 ( 2016 ). doi: 10.1038/nrc.2015.21 OpenUrl CrossRef PubMed ↵ Ragupathi , A. , Singh , M. , Perez , A. M. & Zhang , D . Targeting the BRCA1/2 deficient cancer with PARP inhibitors: Clinical outcomes and mechanistic insights . Front Cell Dev Biol 11 , 1133472 ( 2023 ). doi: 10.3389/fcell.2023.1133472 OpenUrl CrossRef PubMed ↵ Tutt , A. N. J. et al. Adjuvant Olaparib for Patients with BRCA1- or BRCA2-Mutated Breast Cancer . N Engl J Med 384 , 2394 – 2405 ( 2021 ). doi: 10.1056/NEJMoa2105215 OpenUrl CrossRef PubMed National Library of Medicine . NCT03344965 , https://clinicaltrials.gov/study/NCT03344965 ( 2024 ). ↵ Tung , N. M. et al. TBCRC 048: Phase II Study of Olaparib for Metastatic Breast Cancer and Mutations in Homologous Recombination-Related Genes . J Clin Oncol 38 , 4274 – 4282 ( 2020 ). doi: 10.1200/JCO.20.02151 OpenUrl CrossRef PubMed ↵ Genomics Data Commons Data Portal . National Cancer Institute GDC Data Portal , https://portal.gdc.cancer.gov/ ( 2024 ). ↵ Kim , S. et al. Strelka2: fast and accurate calling of germline and somatic variants . Nat Methods 15 , 591 – 594 ( 2018 ). doi: 10.1038/s41592-018-0051-x OpenUrl CrossRef PubMed ↵ Chen , X. , et al. Manta: rapid detection of structural variants and indels for germline and cancer sequencing applications . Bioinformatics 32 , 1220 - 1222 ( 2016 ). doi: 10.1093/bioinformatics/btv710 OpenUrl CrossRef PubMed ↵ Roller , E. , Ivakhno , S. , Lee , S. , Royce , T. & Tanner , S . Canvas: versatile and scalable detection of copy number variants . Bioinformatics 32 , 2375 – 2377 ( 2016 ). doi: 10.1093/bioinformatics/btw163 OpenUrl CrossRef PubMed ↵ Genomics England Research Consortium . AggV2 ancestry inference , https://re-docs.genomicsengland.co.uk/ancestry_inference/ ( 2024 ). ↵ Fan , Y. et al. MuSE: accounting for tumor heterogeneity using a sample-specific error model improves sensitivity and specificity in mutation calling from sequencing data . Genome Biol 17 , 178 ( 2016 ). doi: 10.1186/s13059-016-1029-6 OpenUrl CrossRef PubMed ↵ Koboldt , D. C. , Larson , D. E. & Wilson , R. K . Using VarScan 2 for Germline Variant Calling and Somatic Mutation Detection . Curr Protoc Bioinformatics 44 , 15 14 11 – 17 ( 2013 ). doi: 10.1002/0471250953.bi1504s44 OpenUrl CrossRef ↵ Ye , K. , Schulz , M. H. , Long , Q. , Apweiler , R. & Ning , Z . Pindel: a pattern growth approach to detect break points of large deletions and medium sized insertions from paired-end short reads . Bioinformatics 25 , 2865 – 2871 ( 2009 ). doi: 10.1093/bioinformatics/btp394 OpenUrl CrossRef PubMed Web of Science ↵ Van der Auwera , G. A. & O’Connor , B. D . Genomics in the Cloud: Using Docker, GATK, and WDL in Terra . 1st Edition edn, ( O’Reilly Media , 2020 ). ↵ Huang , K. L. et al. Pathogenic Germline Variants in 10,389 Adult Cancers . Cell 173 , 355 – 370 e314 ( 2018 ). doi: 10.1016/j.cell.2018.03.039 OpenUrl CrossRef PubMed ↵ Nassar , A. H. et al. Ancestry-driven recalibration of tumor mutational burden and disparate clinical outcomes in response to immune checkpoint inhibitors . Cancer Cell 40 , 1161 – 1172 e1165 ( 2022 ). doi: 10.1016/j.ccell.2022.08.022 OpenUrl CrossRef ↵ Chen , S. et al. A genome-wide mutational constraint map quantified from variation in 76,156 human genomes . bioRxiv , 2022.2003.2020.485034 ( 2022 ). doi: 10.1101/2022.03.20.485034 OpenUrl Abstract / FREE Full Text ↵ Tate , J. G. et al. COSMIC: the Catalogue Of Somatic Mutations In Cancer . Nucleic Acids Res 47 , D941 – D947 ( 2019 ). doi: 10.1093/nar/gky1015 OpenUrl CrossRef PubMed ↵ Taliun , D. et al. Sequencing of 53,831 diverse genomes from the NHLBI TOPMed Program . Nature 590 , 290 – 299 ( 2021 ). doi: 10.1038/s41586-021-03205-y OpenUrl CrossRef PubMed ↵ McLaren , W. et al. The Ensembl Variant Effect Predictor . Genome Biol 17 , 122 ( 2016 ). doi: 10.1186/s13059-016-0974-4 OpenUrl CrossRef PubMed ↵ Warden , C. D. , Adamson , A. W. , Neuhausen , S. L. & Wu , X . Detailed comparison of two popular variant calling packages for exome and targeted exon studies . PeerJ 2 , e600 ( 2014 ). doi: 10.7717/peerj.600 OpenUrl CrossRef PubMed ↵ Xu , H. , DiCarlo , J. , Satya , R. V. , Peng , Q. & Wang , Y . Comparison of somatic mutation calling methods in amplicon and whole exome sequence data . BMC Genomics 15 , 244 ( 2014 ). doi: 10.1186/1471-2164-15-244 OpenUrl CrossRef PubMed ↵ Feero , W. G. et al. Guidance on use of race, ethnicity, and geographic origin as proxies for genetic ancestry groups in biomedical publications . Nat Genet 56 , 555 – 556 ( 2024 ). doi: 10.1038/s41588-024-01708-8 OpenUrl CrossRef PubMed ↵ Krainc , T. & Fuentes , A . Genetic ancestry in precision medicine is reshaping the race debate . Proc Natl Acad Sci U S A 119 , e2203033119 ( 2022 ). doi: 10.1073/pnas.2203033119 OpenUrl CrossRef PubMed ↵ Ansari-Pour , N. et al. Whole-genome analysis of Nigerian patients with breast cancer reveals ethnic-driven somatic evolution and distinct genomic subtypes . Nat Commun 12 , 6946 ( 2021 ). doi: 10.1038/s41467-021-27079-w OpenUrl CrossRef PubMed ↵ Prakash , O. et al. Racial Disparities in Triple Negative Breast Cancer: A Review of the Role of Biologic and Non-biologic Factors . Front Public Health 8 , 576964 ( 2020 ). doi: 10.3389/fpubh.2020.576964 OpenUrl CrossRef PubMed ↵ Mathew , A. et al. Allostatic Load in Cancer: A Systematic Review and Mini Meta-Analysis . Biol Res Nurs 23 , 341 – 361 ( 2021 ). doi: 10.1177/1099800420969898 OpenUrl CrossRef PubMed ↵ Moore , J. X. , Andrzejak , S. E. , Bevel , M. S. , Jones , S. R. & Tingen , M. S. Exploring racial disparities on the association between allostatic load and cancer mortality: A retrospective cohort analysis of NHANES, 1988 through 2019 . SSM Popul Health 19 , 101185 ( 2022 ). doi: 10.1016/j.ssmph.2022.101185 OpenUrl CrossRef PubMed ↵ Obeng-Gyasi , S. et al. Association of Allostatic Load With All-Cause Mortality in Patients With Breast Cancer . JAMA Netw Open 6 , e2313989 ( 2023 ). doi: 10.1001/jamanetworkopen.2023.13989 OpenUrl CrossRef PubMed ↵ Miller , A. B . Final results of the UK Age trial on breast cancer screening age . Lancet Oncol 21 , 1125 – 1126 ( 2020 ). doi: 10.1016/S1470-2045(20)30428-9 OpenUrl CrossRef ↵ Duffy , S. W. et al. Effect of mammographic screening from age 40 years on breast cancer mortality (UK Age trial): final results of a randomised, controlled trial . Lancet Oncol 21 , 1165 – 1172 ( 2020 ). doi: 10.1016/S1470-2045(20)30398-3 OpenUrl CrossRef ↵ Moser , K. et al. Extending the age range for breast screening in England: pilot study to assess the feasibility and acceptability of randomization . J Med Screen 18 , 96 – 102 ( 2011 ). doi: 10.1258/jms.2011.011065 OpenUrl CrossRef PubMed Newsome , M . We Must Improve Equity in Cancer Screening . Nature ( 2021 ). doi: 10.1038/d41586-021-03403-8 OpenUrl CrossRef ↵ Chen , T. , Kharazmi , E. & Fallah , M . Race and Ethnicity-Adjusted Age Recommendation for Initiating Breast Cancer Screening . JAMA Netw Open 6 , e238893 ( 2023 ). doi: 10.1001/jamanetworkopen.2023.8893 OpenUrl CrossRef PubMed ↵ Brawley , O. W. et al. Disparities in Tumor Mutational Burden, Immunotherapy Use, and Outcomes Based on Genomic Ancestry in Non-Small-Cell Lung Cancer . JCO Glob Oncol 7 , 1537 – 1546 ( 2021 ). doi: 10.1200/GO.21.00309 OpenUrl CrossRef PubMed ↵ Carson , K. R. et al. in ASCO Annual Meeting Vol. 40 ( 2022 ). ↵ De La Vega , F. , Rhead , B. , Pouliot , Y. & Guinney , J. in ASCO Annual Meeting Vol. 40 ( 2022 ). ↵ Valero , C. et al. The association between tumor mutational burden and prognosis is dependent on treatment context . Nat Genet 53 , 11 – 15 ( 2021 ). doi: 10.1038/s41588-020-00752-4 OpenUrl CrossRef PubMed ↵ Riviere , P. et al. High Tumor Mutational Burden Correlates with Longer Survival in Immunotherapy-Naive Patients with Diverse Cancers . Mol Cancer Ther 19 , 2139 – 2145 ( 2020 ). doi: 10.1158/1535-7163.MCT-20-0161 OpenUrl Abstract / FREE Full Text ↵ Alva , A. S. et al. Pembrolizumab in Patients With Metastatic Breast Cancer With High Tumor Mutational Burden: Results From the Targeted Agent and Profiling Utilization Registry (TAPUR) Study . J Clin Oncol 39 , 2443 – 2451 ( 2021 ). doi: 10.1200/JCO.20.02923 OpenUrl CrossRef PubMed ↵ Marabelle , A. et al. Association of tumour mutational burden with outcomes in patients with advanced solid tumours treated with pembrolizumab: prospective biomarker analysis of the multicohort, open-label, phase 2 KEYNOTE-158 study . Lancet Oncol 21 , 1353 – 1365 ( 2020 ). doi: 10.1016/S1470-2045(20)30445-9 OpenUrl CrossRef PubMed ↵ Sutherland , L. C. , Wang , K. & Robinson , A. G . RBM5 as a putative tumor suppressor gene for lung cancer . J Thorac Oncol 5 , 294 – 298 ( 2010 ). doi: 10.1097/JTO.0b013e3181c6e330 OpenUrl CrossRef PubMed ↵ Leontovich , A. A. et al. NOTCH3 expression is linked to breast cancer seeding and distant metastasis . Breast Cancer Res 20 , 105 ( 2018 ). doi: 10.1186/s13058-018-1020-0 OpenUrl CrossRef PubMed ↵ Cox , A. et al. Otoferlin is a prognostic biomarker in patients with clear cell renal cell carcinoma: A systematic expression analysis . Int J Urol 28 , 424 – 431 ( 2021 ). doi: 10.1111/iju.14486 OpenUrl CrossRef PubMed Chen , K. et al. Immune infiltration patterns and identification of new diagnostic biomarkers GDF10, NCKAP5, and RTKN2 in non-small cell lung cancer . Transl Oncol 29 , 101618 ( 2023 ). doi: 10.1016/j.tranon.2023.101618 OpenUrl CrossRef PubMed ↵ Patel , N. , Itakura , T. , Gonzalez , J. M. , Jr. , Schwartz , S. G. & Fini , M. E . GPR158, an orphan member of G protein-coupled receptor Family C: glucocorticoid-stimulated expression and novel nuclear role . PLoS One 8 , e57843 ( 2013 ). doi: 10.1371/journal.pone.0057843 OpenUrl CrossRef PubMed ↵ National Library of Medicine . NCT04855656 , https://clinicaltrials.gov/study/NCT03344965 ( 2024 ). ↵ Yap , T. A. et al. Abstract PR008: MYTHIC: First-in-human (FIH) biomarker-driven phase I trial of PKMYT1 inhibitor lunresertib (lunre) alone and with ATR inhibitor camonsertib (cam) in solid tumors with CCNE1 amplification or deleterious alterations in FBXW7 or PPP2R1A Molecular Cancer Therapeutics 22 , Abstract PR008 ( 2023 ). doi: 10.1158/1535-7163.TARG-23-PR008 OpenUrl CrossRef ↵ Gasca , J. et al. Loss of FBXW7 and accumulation of MCL1 and PLK1 promote paclitaxel resistance in breast cancer . Oncotarget 7 , 52751 – 52765 ( 2016 ). doi: 10.18632/oncotarget.10481 OpenUrl CrossRef PubMed ↵ (!!! INVALID CITATION !!! 41,43). ↵ Elvin , J. A. et al. Clinical Benefit in Response to Palbociclib Treatment in Refractory Uterine Leiomyosarcomas with a Common CDKN2A Alteration . Oncologist 22 , 416 – 421 ( 2017 ). doi: 10.1634/theoncologist.2016-0310 OpenUrl Abstract / FREE Full Text ↵ Wang , Q. et al. Elevating H3K27me3 level sensitizes colorectal cancer to oxaliplatin . J Mol Cell Biol 12 , 125 – 137 ( 2020 ). doi: 10.1093/jmcb/mjz032 OpenUrl CrossRef PubMed ↵ Drilon , A. et al. Efficacy of Selpercatinib in RET Fusion-Positive Non-Small-Cell Lung Cancer . N Engl J Med 383 , 813 – 824 ( 2020 ). doi: 10.1056/NEJMoa2005653 OpenUrl CrossRef PubMed ↵ Li , H. et al. AXL targeting restores PD-1 blockade sensitivity of STK11/LKB1 mutant NSCLC through expansion of TCF1(+) CD8 T cells . Cell Rep Med 3 , 100554 ( 2022 ). doi: 10.1016/j.xcrm.2022.100554 OpenUrl CrossRef PubMed ↵ Pitt , J. J. et al. Characterization of Nigerian breast cancer reveals prevalent homologous recombination deficiency and aggressive molecular features . Nat Commun 9 , 4181 ( 2018 ). doi: 10.1038/s41467-018-06616-0 OpenUrl CrossRef PubMed ↵ Balmana , J. , Diez , O. , Rubio , I. T. , Cardoso , F. & Group , E. G. W . BRCA in breast cancer: ESMO Clinical Practice Guidelines . Ann Oncol 22 Suppl 6 , vi31 - 34 ( 2011 ). doi: 10.1093/annonc/mdr373 OpenUrl CrossRef PubMed Web of Science ↵ Liu , L. et al. BRCAness as a prognostic indicator in patients with early breast cancer . Sci Rep 10 , 21173 ( 2020 ). doi: 10.1038/s41598-020-78016-8 OpenUrl CrossRef PubMed ↵ Bhaskaran , S. P. et al. Germline variation in BRCA1/2 is highly ethnic-specific: Evidence from over 30,000 Chinese hereditary breast and ovarian cancer patients . Int J Cancer 145 , 962 – 973 ( 2019 ). doi: 10.1002/ijc.32176 OpenUrl CrossRef PubMed ↵ Wang , S. et al. Germline variants and somatic mutation signatures of breast cancer across populations of African and European ancestry in the US and Nigeria . Int J Cancer 145 , 3321 – 3333 ( 2019 ). doi: 10.1002/ijc.32498 OpenUrl CrossRef PubMed ↵ Lakeman , I. M. M. et al. The predictive ability of the 313 variant-based polygenic risk score for contralateral breast cancer risk prediction in women of European ancestry with a heterozygous BRCA1 or BRCA2 pathogenic variant . Genet Med 23 , 1726 – 1737 ( 2021 ). doi: 10.1038/s41436-021-01198-7 OpenUrl CrossRef PubMed ↵ Bhaskaran , S. P. et al. Ethnic-specific BRCA1/2 variation within Asia population: evidence from over 78 000 cancer and 40 000 non-cancer cases of Indian, Chinese, Korean and Japanese populations . J Med Genet 58 , 752 – 759 ( 2021 ). doi: 10.1136/jmedgenet-2020-107299 OpenUrl Abstract / FREE Full Text ↵ Bruno , L. et al. Cyclin-Dependent Kinase 4/6 Inhibitor Outcomes in Patients With Advanced Breast Cancer Carrying Germline Pathogenic Variants in DNA Repair-Related Genes . JCO Precis Oncol 6 , e2100140 ( 2022 ). doi: 10.1200/PO.21.00140 OpenUrl CrossRef PubMed ↵ Collins , J. M. et al. A Real-World Evidence Study of CDK4/6 Inhibitor Treatment Patterns and Outcomes in Metastatic Breast Cancer by Germline BRCA Mutation Status . Oncol Ther 9 , 575 – 589 ( 2021 ). doi: 10.1007/s40487-021-00162-4 OpenUrl CrossRef PubMed ↵ Magavern , E. F. et al. CYP2C19 Genotype Prevalence and Association With Recurrent Myocardial Infarction in British-South Asians Treated With Clopidogrel . JACC Adv 2 , None ( 2023 ). doi: 10.1016/j.jacadv.2023.100573 OpenUrl CrossRef ↵ Cancer Research UK and National Institutes of Health . Cancer Inequities , https://cancergrandchallenges.org/challenges/cancer-inequities ( 2024 ). View the discussion thread. Back to top Previous Next Posted January 24, 2025. Download PDF Supplementary Material Data/Code Email Thank you for your interest in spreading the word about medRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following The clinical and molecular landscape of breast cancer in women of African and South Asian ancestry Message Subject (Your Name) has forwarded a page to you from medRxiv Message Body (Your Name) thought you would like to see this page from the medRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share The clinical and molecular landscape of breast cancer in women of African and South Asian ancestry GJ Thorn , E Gadaleta , AZM Dayem Ullah , LGE James , M Abdollahyan , R Barrow-McGee , JL Jones , C Chelala medRxiv 2024.05.15.24307435; doi: https://doi.org/10.1101/2024.05.15.24307435 Share This Article: Copy Citation Tools The clinical and molecular landscape of breast cancer in women of African and South Asian ancestry GJ Thorn , E Gadaleta , AZM Dayem Ullah , LGE James , M Abdollahyan , R Barrow-McGee , JL Jones , C Chelala medRxiv 2024.05.15.24307435; doi: https://doi.org/10.1101/2024.05.15.24307435 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Genetic and Genomic Medicine Subject Areas All Articles Addiction Medicine (573) Allergy and Immunology (865) Anesthesia (303) Cardiovascular Medicine (4457) Dentistry and Oral Medicine (445) Dermatology (383) Emergency Medicine (610) Endocrinology (including Diabetes Mellitus and Metabolic Disease) (1517) Epidemiology (15244) Forensic Medicine (30) Gastroenterology (1132) Genetic and Genomic Medicine (6620) Geriatric Medicine (669) Health Economics (1002) Health Informatics (4557) Health Policy (1372) Health Systems and Quality Improvement (1615) Hematology (543) HIV/AIDS (1272) Infectious Diseases (except HIV/AIDS) (15936) Intensive Care and Critical Care Medicine (1106) Medical Education (624) Medical Ethics (147) Nephrology (670) Neurology (6634) Nursing (346) Nutrition (999) Obstetrics and Gynecology (1148) Occupational and Environmental Health (957) Oncology (3348) Ophthalmology (980) Orthopedics (369) Otolaryngology (421) Pain Medicine (436) Palliative Medicine (130) Pathology (665) Pediatrics (1696) Pharmacology and Therapeutics (693) Primary Care Research (714) Psychiatry and Clinical Psychology (5463) Public and Global Health (9257) Radiology and Imaging (2210) Rehabilitation Medicine and Physical Therapy (1371) Respiratory Medicine (1198) Rheumatology (598) Sexual and Reproductive Health (716) Sports Medicine (532) Surgery (714) Toxicology (99) Transplantation (289) Urology (265) (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'a031b0e6aca84193',t:'MTc4MDAxNzc3Nw=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();