A systematic review and meta-analysis of germline BRCA mutations in pancreatic cancer patients identifies global and racial disparities in access to genetic testing

Background Germline BRCA1 and BRCA2 mutations (gBRCAm) can inform pancreatic cancer (PC) risk and treatment but most of the available information is derived from white patients. The ethnic and geographic variability of gBRCAm prevalence and of germline BRCA (gBRCA) testing uptake in PC globally is largely unknown. Materials and methods We carried out a systematic review and prevalence meta-analysis of gBRCA testing and gBRCAm prevalence in PC patients stratified by ethnicity. The main outcome was the distribution of gBRCA testing uptake across diverse populations worldwide. Secondary outcomes included: geographic distribution of gBRCA testing uptake, temporal analysis of gBRCA testing uptake in ethnic groups, and pooled proportion of gBRCAm stratified by ethnicity. The study is listed under PROSPERO registration number #CRD42022311769. Results A total of 51 studies with 16 621 patients were included. Twelve of the studies (23.5%) enrolled white patients only, 10 Asians only (19.6%), and 29 (56.9%) included mixed populations. The pooled prevalence of white, Asian, African American, and Hispanic patients tested per study was 88.7%, 34.8%, 3.6%, and 5.2%, respectively. The majority of included studies were from high-income countries (HICs) (64; 91.2%). Temporal analysis showed a significant increase only in white and Asians patients tested from 2000 to present (P < 0.001). The pooled prevalence of gBRCAm was: 3.3% in white, 1.7% in Asian, and negligible (<0.3%) in African American and Hispanic patients. Conclusions Data on gBRCA testing and gBRCAm in PC derive mostly from white patients and from HICs. This limits the interpretation of gBRCAm for treating PC across diverse populations and implies substantial global and racial disparities in access to BRCA testing in PC.


INTRODUCTION
Germline aberrations in BRCA1 and BRCA2 genes are prevalent and clinically relevant for treatment and potential screening for pancreatic cancer (PC). [1][2][3] They are associated with increased progression-free survival in metastatic PC patients treated with poly-ADP-ribose polymerase inhibitor olaparib in the maintenance setting 4 and confer better response to platinum-based chemotherapy. [5][6][7][8] Germline BRCA1 and/or BRCA2 mutations (gBRCAm) can also inform cancer risk and drive preventative strategies in healthy relatives of patients with PC who are at high risk of BRCAassociated cancers (i.e. breast cancer, ovarian cancer, prostate cancer, and PC). [9][10][11][12] Given the above implications, and the lack of significant family history of cancer in up to 60% of mutation carriers, [13][14][15] the American Society of Clinical Oncology and the National Comprehensive Cancer Network now recommend BRCA germline (gBRCA) testing for all individuals with PC irrespective of personal or family history, age, or ethnicity, to maximize the opportunity for targeted treatments for patients and screening for families. 5,16 Although such progress provides promising opportunities for a subset of patients with PC, most data that define our current understanding of the role and clinical relevance of gBRCAm in PC have been generated within the context of mainly European ancestry (commonly termed as 'white' or 'Caucasians'). 15,17 Lack of data from non-white populations, likely due to disparities in access to genetic testing in routine and research settings, limits our understanding of the clinical significance of gBRCAm in broad populations with PC. This limitation is consistent with the general underrepresentation of racial minorities in precision oncology studies broadly as well as genetic and genomic research. [18][19][20][21][22] This is a major public health concern as it precludes access to innovative and potentially active oncological treatments, limits scientific progress by underappreciating genetic diversity, and translates into significant health disparities of already marginalized racial minorities. [21][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37] Studies of ethnic and geographic variability of gBRCA testing uptake and gBRCAm prevalence in PC are currently lacking. To the best of our knowledge, this aspect has only been described in the phase III POLO trial of maintenance olaparib in patients with advanced, platinum-sensitive PC carrying a gBRCAm. 38 However, this study included a small number of non-white patients, making conclusions about the prevalence of gBRCAm in diverse populations challenging.
Here we carried out a systematic review and metaanalysis of the ethnic and geographic variability of gBRCA testing uptake and gBRCAm prevalence in patients with PC. Data concerning gBRCAm in PC are mostly from white patients and high-income countries (HICs), limiting interpretation of gBRCAm testing and associated therapy across diverse populations and highlighting disparities in access to testing and clinical trials.

Search strategy, selection, and inclusion criteria
The study protocol and data extraction for the systematic review and meta-analysis were designed according to the updated version of the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines. 39 The research protocol was registered at the International Prospective Register of Systematic Reviews (PROSPERO, https://www.crd.york.ac.uk/prospero/; registration number: #CRD42022311769). This systematic review was arranged as follows: (i) identification: to search studies on germline BRCA1 and BRCA2 testing in patients with PC; (ii) screening: to search all studies reporting the ethnicity of patients included; (iii) eligibility: to include all data reporting the prevalence of pathogenic and/or likely pathogenic gBRCAm stratified as per patients' ethnicity; (iv) inclusion: to analyze the extracted information. The PRISMA flowchart is reported in Figure 1 Two authors independently screened the titles and abstracts of all identified articles (SP, RC). Articles were included if the study cohort was composed of at least 20 patients, regardless of study type and design. The authors worked independently, and each selected manuscript was double-checked. After the initial set of articles were identified, four authors resolved discrepancies through consensus (SP, RC, AVB, RS). Study selection and data extraction are presented in Supplementary Material S1, available at https:// doi.org/10.1016/j.esmoop.2023.100881.
All authors had access to the study data and reviewed and approved the final manuscript.

Outcomes of interest and data extraction
The main outcome was the pooled proportion of gBRCA testing uptake across all patients, with a focus on diverse ethnicities. Secondary outcomes were the pooled proportion of gBRCAm stratified as per ethnicity, the temporal analysis of gBRCA testing uptake in diverse populations, and the geographic distribution of gBRCA testing utilized worldwide.
For each study, the country of origin, study period, sample size, and patients' ethnicity were collected. Only pathogenic or likely pathogenic mutations were extracted. Studies reporting gBRCA testing and results stratified as per patient's ethnicity were considered for the prevalence analysis. Given the granularity of the data reported in literature, only the following ethnicities were considered: Caucasian/white, African American, Asian, Hispanic/Latino, mixed, minorities (e.g. Hawaiian/Pacific Islander). For the same reason, information on gBRCAm was collected overall, without distinction between familial and sporadic studies.
When ethnicity of the patient was not reported, raw data were requested from the corresponding authors. If no response was obtained, the manuscript was not entered into the systematic review. Studies with only partial information on ethnicity were not included in the systematic review. Studies focusing on Ashkenazi Jewish ancestry patients only were excluded as the prevalence of gBRCAm in this subgroup was previously described in our recent systematic review and meta-analysis. 15 Countries are classified according to the World Bank as lower-middle-income country (LMIC), upper-middle-income country (UMIC), and HIC. 40

Research ethics
The concepts of race, ethnicity, or ancestry reported in the manuscript are not intended to discriminate patients/individuals, and they are merely considered for biomedical research purposes. The possible clinical and social consequences of this research were also evaluated.
The authors acknowledge race as a social construct rather than a biological category but given the lack of scientific validity of race for human categorization and the social implications associated to this term, the word ethnicity has been adopted instead.

Statistical analysis
Descriptive statistics were used to quantify the frequencies of gBRCAm by patient ethnicity and frequencies of patient ethnicity tested, regardless of the testing result. A randomeffects meta-analysis (DerSimonian and Laird model) was carried out on the prevalence data to calculate the pooled event rate using the FreemaneTukey transformation. 41,42 The studyspecific and the pooled prevalence with the 95% confidence intervals (CIs) were graphically represented in a forest plot. The Cochran's Q test for heterogeneity was carried out, reporting the I 2 statistic, which indicates the percentage of variation across studies due to heterogeneity rather than chance. 43 Heterogeneity values of >75% were classified considerable. 44 A temporal analysis was made, dividing the study period into three groups: 2000-2009, 2009-2019, and after 2019 to present. The random-effects meta-regression models and subgroup analyses were estimated accounting for the study period before 2009, 2009-2019, and after 2019 as a possible effect modifier. Statistical analysis was carried out using R (R Foundation for Statistical Computing, Vienna, Austria, v. 4.02) 45 with the metafor 2.4-0 46 and FactoMineR packages. 47

Publication bias and study bias
The publication bias has been graphically inspected using a funnel plot representation where the inverse of standard error against log odds is shown. This representation has been considered because the literature demonstrated that the conventional funnel plots representing log odds versus the inverse of standard error could be asymmetric despite no publication bias for the event rate outcome. 48 The risk of bias was assessed by using Translating-ROB (i.e. TRANSLATIonal caNcer Genomic Risk Of Bias), specifically developed for this kind of studies. 15 Table 2). Less represented ethnicities were excluded from the meta-analysis due to granularity of data ( Table 2).
The temporal analysis showed a statistically significant increase in the percentage of white and Asian patients

DISCUSSION
This is the first systematic review and meta-analysis describing the global landscape of gBRCAm in PC with particular focus on ethnic and geographic variability in utility of genetic testing and overview of pathogenic mutation prevalence. Our findings show that, similar to all genetic knowledgebases, data on gBRCAm in PC derive mostly from white patients (74.7%) and from HICs (91.2%), while those from minority populations and limited-resource countries are extremely low. This is in line with recent reports on the ethnicity of PC patients participating in clinical trials, showing that 84.7% of patients enrolled are of white origin, 21 and with the findings from the POLO trial. 17 Poor representation of minorities in clinical trials and genomic research also applies to other common cancers 18,49,50 reflecting substantial inequalities in accessing medical research. 33,34 The reasons for this are multifactorial, and explanation is beyond this study's aims; however, some considerations can be made according to the included geographic regions of our included studies. Thirty studies originated from North America (the USA and Canada, and one collaboration USA/ Central-Latin America), including 10 528 PC patients overall. The prevalence of non-Hispanic white patients tested was 74.7%, leaving the remaining 25.3% for all other ethnicities. This is despite the higher PC rate in non-Hispanic black patients, compared to non-Hispanic white patients (16.3 versus 14.1 per 100 000, respectively). 51 The population in the United States is multiracial, where many genetic ancestries coexist. 52 According to the United States Census Bureau 2020 results, African Americans, Asians, and

ESMO Open
Hispanics account for approximately 12.4%, 6%, and 16.3% of the population, respectively (including other ethnicities, they constitute 39% of the population). 53 This ethno-racial heterogeneity is not reflected in the present study findings, where the rate of gBRCA testing uptake in non-white patients was extremely low. In fact, the disproportionate bias of higher rates in the non-Hispanic black population exposes a severe under-engagement of the actual patient pool/catchments. The payor coverage of genetic testing in the United States, which is based on health insurance plans, is certainly one of the main reasons for ethnic disparities in access to BRCA testing here observed. Eight studies originated from Europe, where the ethnogeographic makeup is more complex, granular, and country-specific. 54 It is estimated that 2% of the European population has African ancestry, 55 and according to the World Health Organization, almost 10% of the European population are international migrants. 56 Based on the results of the current study, we estimated that thousands of PC patients of non-white origin are realistically excluded from medical studies. Of note, none of the studies included in the present analysis originated from any Eastern European nation. 57 Asian countries and Asian patients were also underrepresented. While Asians represented the second most common ethnicity (17.4% of the patients included), this number is extremely low when considering the entire population of countries such as China or the Indian subcontinent (that are not represented at all). Sociocultural, structural, and economic barriers may explain this extremely low access to gBRCA testing in Asian countries, and no clear conclusions can be drawn. Notably, 7 out of 10 studies including Asian patients were from Japan, an HIC, further suggesting economic factors as a key driver of disparities in genomic research between same continent countries with different income.
Another important aspect that emerged from the present study is the identification of poor reporting of ethnicity amongst the studies included. Twenty-six out of 67 potentially eligible studies, corresponding to 38.8%, did not provide any information about participant ethnicity. Excluding those conducted in mono-ethnic populations only (n ¼ 22; 12 in white, 10 in Asians), this rate rises to 57.7%. When it comes to reporting gBRCAm prevalence, only 8 of 29 studies enrolling mixed populations reported ethnicity information (27.6%). This indicates significant bias in reporting methods and results and lesser importance attributed to ethnicity in medical research. The general poor reporting of studies has also been pointed out by the risk of bias   Last, we identified substantial discrepancies in the pooled prevalence of gBRCAm in different ethnic groups (3.3% in white, 1.7% in Asians, 0.2% in African Americans, and 0% in Hispanics). However, this needs to be interpreted with caution. Indeed, given the overall low prevalence of gBRCAm in unselected PC patients, and the scarce inclusion of non-white patients in the current study, no clear conclusions can be drawn on any real difference in the prevalence of gBRCAm across populations. Enhancing diversity and equity in genomic research will be essential to define the real prevalence of gBRCAm in non-white populations and to assess any significant differences in mutation rate across ethnicities, all as an effect to more effectively mitigate cancer risk and to optimize personalized treatment of patients.
The current lack of diversity identified in BRCA research is of concern as it does not assess human variability, 31,32 thus limiting the generalizability of research findings that may not account for differences in biological and sociocultural factors across populations impacting PC susceptibility and treatment outcomes. [58][59][60][61][62][63][64][65][66] For BRCA1 and BRCA2 genes, founder mutations have been identified across multiple ethnicities, 67,68 including African Americans and Hispanics, only seldom profiled in PC due to global and racial disparities in BRCA testing uptake. 26,27 This will likely compound the already existing inequities in cancer care described for patients with PC, [23][24][25] which translates to poorer survival amongst non-white populations, especially African Americans. 26,[28][29][30] Multi-level interventions are encouraged to enhance inclusion of multiple ethnicities in PC genomic studies to better understand cancer susceptibility across diverse populations and improve tailored early detection strategies, to allow minority groups access to innovative targeted treatments and, more in general, to mitigate health disparities in cancer screening and treatment.
This study has some limitations, including: (i) general heterogeneity among studies in terms of sample size, study period, patients enrolled (familial or sporadic cases), and study design; (ii) inconsistency of definitions used to categorize human populations in different studies (race, ethnicity, or ancestry, often interchangeably used 33 ), making results only partially comparable; (iii) general low numbers of non-white PC patients, limiting the validity of the pooled estimates of gBRCAm in these subgroups; (iv) lack of investigation of factors that may influence access to genetic testing in included studies; (v) considerable heterogeneity in some meta-analysis results; and (vi) inherent publication bias and study bias (e.g. overlapping study populations coming from the same institution). Altogether, these factors hamper analysis and interpretation.
In conclusion, although gBRCA testing is relevant for precision PC treatment and more broad cancer screening strategies, this study suggests that information on gBRCAm in PC derives mostly from white patients and from studies conducted in HICs. This implies racial and global disparities in access to BRCA testing for patients with PC and translates into missed opportunities: (i) to study ethnic and racial minorities in terms of impact of social determinants of cancer risk and survival outcomes; (ii) to expand effective precision medicine strategies globally, including PC prevention and treatment; and (iii) to enhance the understanding of cancer biology in minority populations and pharmacoethnicity. 69 Real-world data will be important to see if this figure reflects access to genetic testing as standard of care, outside the research setting. Our findings suggest that there is an urgent need for a concerted effort to improve global access to BRCA testing in PC and, in general, address genomic diversity as well as geographic and racial disparities in research and health care delivery.

DATA SHARING
The data that support the findings of this study are available from the corresponding author (AVB), upon reasonable request.