Introduction

Health disparities in breast cancer have been defined in multiple ways, englobing aspects related to socioeconomic inequities, lack of access to screening protocols, poor quality of cancer care, along with differences in the presentation of several risk factors among populations, ultimately leading to differences in breast cancer incidence and mortality rates, as well as in the disease outcome among women with different race/ethnicities [1–3]. The latest 2022 American Cancer Society report revealed the highest breast cancer incidence in non-Hispanic White women (132.5 per 100.000), followed by non-Hispanic Blacks (127.1 per 100.000), American Indian/Alaska Natives (110.5 per 100.000), Asian/Pacific Islanders (98.8 per 100.000), and in the last place Hispanic/Latina women (96.3 per 100.000) [4]. In terms of mortality, non-Hispanic Black women presented the highest mortality rates among all population groups (28.0 per 100.000), followed by non-Hispanic Whites (19.9 per 100.000), American Indian/Alaska Natives (17.8 per 100.000), Hispanic/Latinas (13.7 per 100.000), and lastly Asian/Pacific Islanders women (11.7 per 100.000).

Recent studies have explored genetic ancestry as a biological factor that has been associated with differences in breast cancer risk and its phenotype, contributing this way to health disparities [5,6]. For instance, it has been reported that women with higher European component present higher odds for the disease [7,8], compared to women with higher Indigenous American (IA) ancestry [9], whereas women with greater African ancestry have higher odds for breast cancer-specific mortality [10,11]. On the other hand, differences in the prevalence of intrinsic subtypes between population groups have also been reported [12,13]. For example, luminal A breast tumors are more prevalent in women with higher European ancestry, whereas the triple-negative (TN) subtype is more frequently observed in women with a higher African component, and greater fractions of IA ancestry have been associated with higher odds for HER2-positive tumors [6,10,14]. These differences have also been associated to the differences in the disease outcome between women with different genetic ancestry fractions [15].

Genetic ancestry, as a variable that is objectively measured by assessing the differences in genome-allele frequencies of multiple polymorphisms among ancestral populations [15,16], might be associated to breast cancer health disparities. However, genetic ancestry can also be understood as a reflection of the human migration process and the Americas’ colonization history, and therefore, it can mirror the degree of poverty, racial/ethnic discrimination, and other socioeconomic determinants that also contribute to breast cancer health disparities among populations [17,18]. Other factors related to breast cancer health disparities include etiology, tumor biology, and treatment-related clinical-pathological variables. The correlation between socioeconomic and clinical-pathological determinants entails complex interactions, and adding genetic ancestry as a biological contributor could impact this interplay and affect the presentation of health disparities between populations. However, the relationship between these factors with genetic ancestry in Colombian women, a highly admixed population from Latin America, has not been explored. For that matter, we aimed to assess contributing factors to breast cancer health disparities according to genetic ancestry in Colombian patients from a national cancer reference center.

Materials and methods

Results

Patients’ sociodemographic and clinical-pathological characteristics

Patients’ clinical-pathological and sociodemographic information is described in Table 1. Patients came mostly from Colombia’s capital city, Bogotá D.C (52.4%), and 8% came from other main cities around the country. A total of 143 patients (39.6%) came from little towns and more scattered places around the country, known as provinces, where there is less access to oncological services; consequently, these patients are referred to main cities, like Bogotá D.C, for health attention. The majority were either married or in free union (45.4%) and have had at least one child. As a socioeconomic status indicator, we used the socioeconomic stratum [19]. Most of breast cancer patients included in this study live in low-income neighborhoods (stratums I and II: 53.3%). This is consistent with the fact that only 13% of the study population had a technician or university education.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Sociodemographic and clinical-pathological characteristics of the study population.

https://doi.org/10.1371/journal.pone.0306037.t001

Regarding Colombia’s available health insurance schemes, contributive and subsidized [20], we found that half of the patients included belonged to the latter group (48.8%). Moreover, among individuals within the contributive scheme, we observed that 13% of our patients are beneficiaries, leaving only 35.2% of the study population under a non-subsidized insurance model (Table 1).

Regarding clinical-pathological variables, most patients were diagnosed over the age of 50 (77.0%) with ductal invasive carcinomas (100%), at advanced clinical stages (AJCC Clinical stage III: 43.8% and IV: 1.9%), with moderately differentially tumors (Scarff-Bloom Richardson II: 53.2%) and sizes that ranged between 21–49 millimeters (mm) (40.4%). Furthermore, over half of the study population presented histological invasion (51.5%), lymph-node involvement (61.3%), and a high Ki67 index (≥20%: 59.0%) at diagnosis. Intrinsic subtype distribution showed that luminal/HER2- tumors were the most prevalent subtypes (65.4%), followed by the TN (16.1%), luminal/HER2+ (10.2%), and HER2-enriched subtypes (5.8%). At the end of the study, 24.1% of the patients had presented recurrence, and 23% had already died.

Contributing factors to breast cancer health disparities according to genetic ancestry

Genetic ancestry data was available for 85.3% of the cases (308/361). The average genetic ancestry proportions for the European, IA, and African components were 50.9%, 40.6%, and 7.0%, respectively (Fig 1A). According to the 2018 Colombian census, almost 90% of the population self-identify as mestizo [27,28], which is consistent with our results where most of the patients included showed a highly admixed genetic profile, mainly between the European and the IA components (Fig 1B). We analyzed genetic ancestry as a continuous variable, in order to measure differences in genetic ancestry fractions according to different etiology, tumor-specific, treatment administration, and socioeconomic factors, all of which might contribute to breast cancer health disparities (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1.

A. Median values of genetic ancestry (European, IA and African) among breast cancer patients from Colombia. B. Genetic ancestry distribution for 308 breast cancer patients from Colombia. Each patient is represented by a vertical bar at the x-axis.

https://doi.org/10.1371/journal.pone.0306037.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Distribution of European, IA, and African ancestry according to potential sources of breast cancer health disparities.

https://doi.org/10.1371/journal.pone.0306037.t002

Among etiology-related factors, such as age of diagnosis and body mass index (BMI), we observed a higher median of IA ancestry fraction in overweight patients compared to the lean group (>25: 0.42 vs. ≤25: 0.38, respectively, p = 0.034). No statistically significant differences in genetic ancestry fractions were observed by age at diagnosis. Regarding tumor-specific factors, a higher median of African ancestry was observed in patients diagnosed at higher clinical stages (III/IV: 0.08 vs. I/II: 0.06, p = 0.036). On the other hand, patients with larger tumors (>20 mm) presented a lower median of European ancestry (0.50 vs. 0.54, p = 0.009) and a higher IA| ancestry fraction (0.42 vs. 0.39, p = 0.022), compared to patients with smaller tumors (≤ 20 mm). No statistically significant differences in any of the genetic ancestry fractions were found according to ER, PR, HER2 status, and intrinsic subtype.

Concerning treatment-related factors, a lower median of European ancestry fraction and a tendency for higher IA ancestry was observed in patients who received neoadjuvant therapy, compared to patients who did not receive this treatment (for European ancestry: 0.48 vs. 0.53, respectively, p = 0.002; for IA ancestry: 0.42 vs. 0.40, respectively, p = 0.055). Interestingly, a significantly higher African ancestry was observed in patients who achieved a complete clinical or pathological response (0.12) or presented a stable clinical response (0.10), compared to patients who either progressed (0.06) or had a partial response (0.05) to neoadjuvant treatment (p = 0.006). In addition, a higher European ancestry fraction was observed in patients who underwent conservative surgeries, compared to patients who received mastectomies; however, this difference did not reach statistically significant values (0.52 vs. 0.50, p = 0.075).

Regarding socioeconomic and demographic factors, our results showed that patients who live in middle-income households (III/IV stratums) present higher European ancestry levels compared to patients in the low stratum group (0.55 vs. 0.49, p = 0.047); patients in high-income stratums were not included in the analysis given the low number of samples in this group (n = 3). Likewise, significantly higher European ancestry (0.53 vs. 0.48, p<0.001) and lower IA median ancestry (0.39 vs. 0.42, p = 0.017) were observed in patients from the contributory insurance regime, compared to patients from the subsidized scheme. Lastly, we also observed higher European ancestry fractions in patients that had never had children before (0.53), compared to patients with one (0.47) or more than 2 children (0.52) (p = 0.048). No statistically significant differences in genetic ancestry fractions were found according to civil status, educational level, and provenance region. We conducted a second analysis where genetic ancestry was categorized as high and low, according to the median value, and found similar results (S1 Table).

Furthermore, a regression analysis showed an association between a higher IA ancestry and being overweight (BMI>25: OR = 1.73, 95% CI, 1.05–2.91, p = 0.0321), having larger tumors (>20 mm: OR = 1.91, 95% CI, 1.13–3.32, p = 0.0173), and being in a subsidized insurance regime (OR = 1.73, 95% CI, 1.07–2.85, p = 0.0254). On the other hand, patients with higher European ancestry showed lower odds for larger tumors (≥20 mm: OR = 0.45, 95% CI, 0.25–0.77, p = 0.00496), for receiving neoadjuvant therapy (OR = 0.49, 95% CI, 0.30–0.80, p = 0.00487), and being in a subsidized health regime (OR = 0.42, 95% CI, 0.25–0.69, p = 0.0008). At the same time, higher odds were observed in patients with higher European ancestry fractions to live in higher socioeconomic stratums (III/IV: OR = 1.93, 95% CI, 0.99–3.87, p = 0.055).

Discussion

It has been observed that individuals among population groups not only share a heritage and many cultural traditions, but they also seem to have comparable socioeconomic conditions, such as similar monetary income and education level [29,30]. It is well known that socioeconomic features are well-established contributing factors to health disparities, nonetheless, biological factors can also contribute to the differences reported in incidence, prevalence, and mortality among these population groups [12,31,32]. High genomic resolution studies have identified disease-related loci enriched for a particular ancestry component in the genome of highly admixed populations from Mexico [33], Puerto Rico [34], and Colombia [35]. These investigations have led to the hypothesis that the genetic footprint left by the colonization process in Latin America might contribute to health disparities among admixed populations, affecting several health determinants and epidemiological measures such as incidence, prevalence, and mortality [35].

In this study, we focused on genetic ancestry as one of the potential biological contributors to health disparities, and assessed its relationship with important socioeconomic factors, but also with other etiologic, tumor, and treatment-administration-related factors, in a cohort of breast cancer patients from Colombia. In terms of genetic complexity, Latin American countries like Colombia are of great interest, as a result of the genetic admixture that began in the Americas with the arrival of the Spanish colonization and trans-Atlantic slave trade around 500 years ago [36]. This process entailed a massive genetic exchange between three continental groups (Europeans, Native Americans, and Africans) that until that point, had never been in contact before. This brought a particularly high level of genetic diversity among Latin American countries, including Colombia [35].

Unlike the United States (US) population, Colombia counts with just a few officially recognized minority ethnic groups, including the Indígenas, Roms, Raizales, Palanqueros, and Afro-Colombians. However, according to the 2018 Colombian census, almost 90% of the people do not self-identify with any of these ethnic groups, and instead, self-identify as mestizos [27,28]. As reported by Chande et al [37], these ethnic categories are strongly correlated with genetic ancestry, but they do not capture the complexity of ancestry and admixture seen in the Colombian population. Therefore, studying health disparities-contributing factors according to ethnic groups is not useful in a country like Colombia. Instead, we evaluated genetic ancestry and found consistent associations with what has been reported before in other population groups [38,39].

Our results showed that patients with higher European ancestry presented smaller tumors, received less neoadjuvant therapy, and belonged to higher socioeconomic stratums. In contrast, patients with higher IA ancestry fractions were more frequently classified as obese, presented larger tumors, and received neoadjuvant chemotherapy. A previous study by Fejerman et al. [40], which investigated the association between genetic ancestry and clinical outcome in U.S. Latina women with breast cancer, hypothesized that the relationship between genetic ancestry and this clinical feature might be related to the strong association between self-identified ethnicity and socioeconomic status [40]. Epidemiology studies conducted in the US have reported that self-identified White people, who are primarily of European descent, often have a higher socioeconomic status, whereas other population groups like African Americans, with a greater African ancestry component, and Latinos, with an important IA ancestry fraction in their genome, allegedly live under inferior socioeconomic conditions [41,42]. For example, a study conducted in Mexican and Colombian patients with type-2 diabetes reported a positive association between IA ancestry and lower socioeconomic stratums [43]. Certainly, having a low socioeconomic status will lead to fewer educational opportunities, along with less healthy lifestyle habits and lower access and awareness to screening programs [44,45], which will finally correlate to the presentation of bad prognosis breast cancer clinical-pathological features [46,47].

Furthermore, a higher proportion of women with higher IA ancestry was found among the subsidized health insurance group, whereas a greater European ancestry fraction was observed for the contributory health insurance scheme. Reports for gastric cancer in Colombia have stated that patients in the subsidized regime have lower survival times and a higher mortality risk, compared to patients in the contributory regime, as well as those affiliated to the special insurance scheme [48]; this corresponds to a special program for government workers in the fields of education, military, policy, and petrochemical industry [20]. Likewise, an extensive review of the latest annuals reports from the National Health Institute of Colombia stated that patients in the contributory regime often present cancer at early stages at diagnosis, compared to those affiliated to the subsidized regime [49]. It is likely that this might be related to the better healthcare services provided by the contributory and special regimes, however, other reports have found that, regardless of their type of insurance affiliation, patients from higher socioeconomic stratums present better survival probabilities [48]. Understanding the role of these health determinants in the context of breast cancer, a multifactorial disease, can be challenging, as these variables are often highly correlated with each other. In that sense, we hypothesized that having better healthcare services and other aspects related to living in a higher socioeconomic stratum, like a healthy diet and a higher education level, might explain why in our study, patients with higher European ancestry presented better clinical-pathological features, and also, that women with higher IA ancestry presented less favorable clinical-pathological traits like obesity and larger tumors.

On the other hand, higher African ancestry fractions were found among patients with poor clinical-pathological features, like advanced clinical stages and larger tumors. Several studies in the US have reported these associations, claiming that this could be due to both socioeconomic and genetic factors [50–52]. The latter includes the presence of population-specific genetic variants that predispose women with high African ancestry to develop more aggressive breast cancer characteristics and intrinsic subtypes like the TN [12,52,53]. However, in our study, we did not find a statistically significant association between a higher fraction of African ancestry and a greater prevalence of TN tumors. This might be possibly related to the lower contribution of this particular component within Colombian patients. Compared to studies where this association has been reported [54,55], the African ancestry contribution in our study is considerably low, at approximately 7%. This poses a challenge in drawing conclusions regarding the potential contributions of the African component to disease phenotype and to breast cancer health disparities in Colombian women.

We certainly encountered some limitations during the conduction of this study. The majority of the patients’ socioeconomic data was retrieved from medical records and government sources, however, there were still many cases with missing information, especially for socioeconomic stratum and education level, which were some of the most relevant variables explored in the analyses. Along the same line, almost all the patients that are admitted at the Colombian NCI are from the low and medium socioeconomic stratums; just 3 of our patients belonged to the high (V and VI) category. It is possible that this fact did not allow us to detect even further differences between socioeconomic stratums by genetic ancestry. Other limitations include the small sample size among several categories and the generally low contribution of the African ancestry component among the included patients. It is possible that this might have reduced the study’s statistical power and limited the opportunity to find biological associations. We also highlight the potential impact on AIMs genotyping quality resulting from the use of FFPE blocks for DNA extraction and genetic ancestry estimation, as these samples are known to be difficult to handle due to their high degree of nucleic acid degradation [56]. Even so, our study also had several strengths. All cases were recruited from a single institution, whereby, there were no inclusion biases. In addition, a single pathologist assessed all immunohistochemistry biomarkers, and a single person collected clinical and pathological information from medical records, all of which contributed to the homogeneity of our data.

Our results show important differences in several aspects of the disease according to genetic markers. Considering the current Colombian Health Ministry guidelines for breast cancer screening and early detection [57], our work could potentially be used to implement a more guided and personalized approach for future breast cancer screening and management protocols in the country. For instance, health determinants and genetic biomarkers, such as genetic ancestry, could potentially be incorporated as covariates in future artificial intelligence models to enhance the decision-making process in clinical settings.

Conclusions

Overall, our findings show that patients with higher IA ancestry present clinical-pathological characteristics of poor prognosis like obesity and larger tumors and are more frequently affiliated to a subsidized healthcare regime. Furthermore, we found that patients with higher European ancestry present smaller tumors, do not receive as much neoadjuvant chemotherapy, and usually belong to higher socioeconomic stratums and to the contributory healthcare regime. Our results reflect complex interactions between socioeconomic, etiological, tumor-related, and genetic factors, that ultimately affect the presentation of health disparities reported in breast cancer.

Supporting information