Molecular characterization of hereditary breast and ovarian cancer patients from a public precision medicine service in the Southeast Brazilian population

Abstract

To address the need for specialized care in hereditary cancer, we developed a Hereditary Cancer Predisposition Assessment and Family Monitoring Program in the southeast of Brazil. The program was designed to identify suspected cases and provide molecular diagnostic testing through a structured care flow that included genetic counseling, psychological support, and clinical follow-up. This initiative targeted individuals within the public healthcare system and aimed to implement accessible precision medicine approaches for hereditary cancer syndromes. As part of this initiative, we performed systematic genetic screening using both Sanger sequencing and a next-generation sequencing panel to investigate cancer susceptibility genes in a cohort of 210 patients with suspected Hereditary Breast and Ovarian Cancer Syndrome (HBOC). Variants were classified according to the American College of Medical Genetics and Genomics (ACMG) guidelines. Statistical analyses were conducted to compare clinical characteristics between patients carrying pathogenic or likely pathogenic variants and those with benign findings. Pathogenic or likely pathogenic mutations were identified in 33.3% (70/210) of patients, with 14.3% (30/210) involving non-BRCA genes. BRCA2 was the most frequently mutated gene contrasting with most reports from the country, in which BRCA1 predominates. The most frequent pathogenic mutation was c.4829_4830del in BRCA2, present in 8.57% of positive cases and apparently rare in other Brazilian cohorts. Additionally, 12.8% of the patients with pathogenic or likely pathogenic variants had mutations in genes associated with hereditary cancer syndromes other than HBOC. A total of 35 variants of uncertain significance were identified, most commonly in the ATM gene. By identifying pathogenic mutations in these individuals, precision medicine strategies can be implemented to improve outcomes for patients and their families.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-16870-0.

Introduction

The impact of cancer in the world, based on estimates from Global Cancer Observatory (GLOBOCAN), indicated almost 20 million new cases worldwide (excluding cases of non-melanoma skin neoplasm) in 2022. Female breast cancer was the second most frequently diagnosed, responsible for nearly 2.3 million new occurrences (11.6%)¹. In Brazil, breast cancer is the first cause of death in women, and the Southeast and South regions have the highest recorded death rates (12.43 and 12.69/100 thousand women, respectively). The incidence of ovarian cancer is lower compared to breast cancer. However, it ranks eighth among the most common female cancers in Brazil, with a higher concentration rate in the South and Southeast regions. Additionally, ovarian cancer is known for its severity and difficulty in diagnosis².

Hereditary Breast and Ovarian Cancer syndrome (HBOC) is characterized by an increased risk in the development of breast and ovarian cancer due to the presence of inherited pathogenic mutations³. Although most cases of this syndrome are caused by mutations in the BRCA1/2 genes, other genes have also been associated with an increased risk. Recent studies have shown that around 12% of BRCA1 and BRCA2 negative patients with clinical criteria for HBOC harbor a pathogenic variant in another gene^4,5. Therefore, it is essential to understand the genetic variations in specific ethnic groups related to the syndrome to identify high-risk individuals and implement appropriate population-based preventive measures⁶.

The Brazilian population is highly diverse, with ethnic composition varying across different states. The primary groups that contributed to the formation of the Brazilian population include Amerindians, Europeans, and Africans, alongside other specific groups. Understanding this diversity is crucial for analyzing the frequencies of genetic mutations, as it provides insights into regional profiles, heredity, genetic susceptibility, ancestry, and disease segregation⁶.

In recent years, next-generation sequencing (NGS) has emerged as a powerful tool for analysing multiple genes, significantly enhancing the efficiency and accessibility of molecular tests. This technology enables the identification of individuals at high risk for breast and/or ovarian cancer and supports the molecular diagnosis of various other hereditary syndromes in affected individuals⁷. However, managing cancer within public health systems remains a challenge, particularly in middle- and low-income countries, where financial constraints and inadequate infrastructure may prevent effective diagnosis and care for cancer patients⁸. According to the 2019 National Health Survey (Pesquisa Nacional de Saúde, PNS), 71.5% of Brazilians, or more than 150 million people, rely exclusively on the public Unified Health System (Sistema Único de Saúde, SUS) for their healthcare. However, in a country of continental dimension such as Brazil, the primary healthcare model under SUS often falls short due to the lack of molecular diagnostic centers and insufficient support in many states^9–11.

Despite the considerable allelic and locus heterogeneity in HBOC patients in Brazil, most studies have paid more attention to the BRCA1, BRCA2, and TP53 genes¹². The first studies of patients meeting the criteria for HBOC using next-generation sequencing in Brazil were only published in 2016^13–15. This scenario has changed in the last few years, although the number of studies is still low to show the broad spectrum of mutations in different regions of Brazil¹².

The state of Minas Gerais (MG), in southeast Brazil, with almost 21 million inhabitants, is demographically the second of the country. In the Midwest region of Minas Gerais, Brazil, we implemented a Hereditary Cancer Predisposition Assessment and Family Monitoring Program to support patients at risk for hereditary cancer syndromes. This is a collaborative program between the Oncology Unit of Hospital São João de Deus, the Molecular Biology Laboratory of the Universidade Federal de São João del-Rei (UFSJ), and the ACOM (Associação de Combate ao Câncer), Cancer Support Association of the Midwest Region of Minas Gerais. The program was based on the development of a patient care protocol to define follow-up strategies, employing a highly specialized team, which included genetic counseling, psychological support, and the use of advanced techniques for molecular diagnosis. The program provides care and monitoring for families affected by various hereditary cancer syndromes, with HBOC being the most prevalent. The entire care service was structured using resources from research projects funded by various funding agencies, as well as donations raised by ACOM. Collaborative research efforts with these institutions have focused on patients monitored by the public health system. In the initial years of the program, molecular characterization of the BRCA1 and BRCA2 genes was conducted on 44 patients using Sanger sequencing, along with the analysis of point mutations in the CHEK2, TP53, and PALB2 genes^16,17. Through the Minas Gerais Network for Population Genomics and Precision Medicine, funded by FAPEMIG (Fundação de Amparo à Pesquisa do Estado de Minas Gerais), it was possible to expand the service, and a NGS panel for genes associated with HBOC was implemented to support the program. Despite all patients being assisted by the SUS, some are able to afford the test, which is conducted in private laboratories. Here, we present all these results collectively to ensure a more comprehensive understanding of the molecular profile in the state.

Results

Variants found

This is the first study to evaluate mutations using an NGS panel in patients (n = 166) assisted by the public health system with clinical criteria for HBOC in Minas Gerais. The results provide information on the mutation profile across more than 20 genes in the studied group, in conjunction with findings from Sanger sequencing (n = 44). Pathogenic and likely pathogenic (P/LP) mutations were found in 33.3% of patients (70/210), all in the heterozygous state, while variants of uncertain significance (VUS) were identified in 16.7% of patients (35/210). All pathogenic and likely pathogenic variants are presented below (Tables 1 and 2). It was observed that 71.5% (50/70) of the pathogenic and likely pathogenic variants were found in high penetrance genes^18,19 related to HBOC: 23 (32.9%) in BRCA2; 17 (24.3%) in BRCA1; 6 (8.6%) in TP53, 5 (7.1%) in PALB2. Genes with moderate penetrance^18,19, such as ATM (3), CHEK2 (3) and RAD51C (4) represented 14.3% (10/70) of the mutations in the patients analyzed. It is worth mentioning that 12.8% of the probands with pathogenic or likely pathogenic variants had mutations in genes not typically associated with HBOC, but rather with other hereditary cancer syndromes: MSH2 (n = 3), BRIP1 (n = 1), CTC1 (n = 1), MITF (n = 1), PTCH1 (n = 1), RECQL4 (n = 1), and NTHL1 (n = 1). The most frequently mutated gene in this study was BRCA2 (23/210; 11.0%). The distribution of pathogenic or likely pathogenic variants by genes and description of the most frequent pathogenic mutations found in Minas Gerais state can be seen in Figs. 1 and 2.

Table 1.

Pathogenic and likely pathogenic variants identified in BRCA genes (n = 40).

Variant (HGVS)		Variant ID	n	ACMG Classification	ACMG Criteria
BRCA1
c.68_69del	p.Glu23Valfs*17	rs80357914	1	P	PVS1 (Very Strong), PS4 (Moderate), PM2 (Moderate)
c.112_113del	p.Lys38Valfs*2	rs80357949	1	P	PVS1 (Very Strong), PS4 (Moderate), PM2 (Moderate)
c.441 + 2 T > A	__	rs397509173	1	P	PVS1 (Very Strong), PS4 (Moderate), PM2 (Moderate)
c.2037delGinsCC	p.Lys679Asnfs*4	rs397508932	1	P	PVS1 (Very Strong), PS4 (Moderate), PM2 (Moderate)
c.3328_3329del	p.Lys1110Alafs*4	__	1	LP	PVS1 (Very Strong), PM2 (Moderate)
c.3331_3334del	p.Gln1111Asnfs*5	rs80357701	1	P	PVS1 (Very Strong), PS4 (Moderate), PM2 (Moderate), PP1 (Supporting)
c.3756_3759del	p.Ser1253Argfs*10	rs80357868	2	P	PVS1 (Very Strong), PS4 (Moderate), PM2 (Moderate)
c.4484G > T	p.Arg1495Met	rs80357389	1	P	PS4 (Strong), PM2 (Moderate), PM5 (Moderate), PP3 (Supporting), PP5 (Supporting)
c.4689_4694del	pTyr1563_Leu1564delinsTer	__	2	LP	PVS1 (Very Strong), PM2 (Moderate)
c.5072 C > T	p.Thr1691Ile	rs80357034	1	P	PS4 (Strong), PM1 (Moderate), PM2 (Moderate), PM5 (Moderate), PP1 (Supporting), PP3 (Supporting)
c.5266dupC	p.Gln1756Profs*74	rs80357906	5	P	PVS1 (Very Strong), PS4 (Moderate), PM2 (Moderate)
BRCA2
c.2T>C	p.Met1Thr	rs80358547	2	P	PVS1 (Moderate), PS1 (Strong), PS4 (Moderate), PM2 (Moderate)
c.2T>G	p.Met1Arg	rs80358547	4	P	PVS1 (Moderate), PS1 (Strong), PS4 (Moderate), PM2 (Moderate)
c.156_157insAlu	p.Lys53Alafs*9	rs2138704192	4	P	PVS1 (Very strong), PS4 (strong), PM2 (moderate), PP1 (supporting), PP5 (supporting)
c.1310_1313del	p.Lys437Ilefs*22	rs80359277	1	P	PVS1 (Very Strong), PM2 (Moderate), PP5 (Strong)
c.4829_4830del	p.Val1610Glyfs*4	rs80359468	6	P	PVS1 (Very Strong), PS4 (Moderate), PM2 (Moderate)
c.5985delC	p.Asn1995Lysfs*9	rs2137522955	1	P	PVS1 (Very Strong), PM2 (Moderate), PP5 (Supporting)
c.6405_6409del	p.Asn2135Lysfs*3	rs80359584	4	P	PVS1 (Very Strong), PS4 (Moderate), PM2 (Moderate)
c.9154C>T	p.Arg3052Trp	rs45580035	1	P	PS3 (Strong), PS4 (Moderate), PM2 (Moderate), PM5 (Moderate), PP1 (Supporting), PP3 (Supporting)

Table 2.

Pathogenic and likely pathogenic variants identified in non-BRCA genes (n = 30).

Gene	Variant (HGVS)		Variant ID	n	ACMG Classification	ACMG Criteria
TP53	c.1010G > A	p.Arg337His	rs121912664	5	P	PS4 (Strong), PM1 (Moderate), PM2 (Moderate), PM5 (Moderate), PP1 (Supporting), PP3 (Supporting)
	c.524G > A	p.Arg175His	rs28934578	1	LP	PS4 (Moderate), PM1 (Moderate), PM2 (Moderate), PM5 (Moderate), PP1 (Moderate), PP3 (Supporting)
PALB2	c.3027delT	p.Glu1010Argfs*5	rs876659378	4	P	PVS1 (Very Strong), PM2 (Moderate), PP5 (Supporting)
	c.1036 C > T	p.Arg346Cys	rs201206424	1	LP	PM2 (Moderate), PM5 (Moderate), PM1 (Supporting), PP3 (Supporting)
CHEK2	c.409 C > T	p.Arg137T*	rs730881701	2	P	PVS1 (Very Strong), PS4 (Moderate), PM2 (Moderate)
	c.485 A > G	p.Asp162Gly	rs587781652	1	LP	PM1 (Moderate), PM2 (Moderate), PP3 (Moderate), PP5 (Supporting)
RAD51C	c.890_899del	p.Leu297Hisfs*2	rs1555602141	3	P	PVS1 (Very Strong), PS4 (Moderate), PM2 (Moderate)
	c.709 C > T	p.Arg237*	rs770637624	1	P	PVS1 (Very Strong), PS4 (Moderate), PM2 (Moderate)
ATM	c.3485T > G	p.Leu1162*	rs1591636613	1	P	PVS1 (Very Strong), PS4 (Strong), PM2 (Supporting)
	c.3802del	p.Val1268*	rs587779834	2	P	PVS1 (Very Strong), PM3 (Strong), PP5 (Supporting)
MSH2	c.2152 C > T	p.Gln718*	rs587779139	3	P	PVS1 (Very Strong), PS4 (Moderate), PM2 (Moderate)
BRIP1	c.2990_2993del	p.Thr997Argfs*61	rs771028677	1	P	PVS1 (Very Strong), PS4 (Moderate), PM2 (Moderate)
NTHL1	c.526-1G > A	__	rs779757251	1	LP	PVS1 (Very Strong), PM2 (Moderate)
MITF	c.952G > A	p.Glu318Lys	rs149617956	1	P	PP1 (Very Strong), PM2 (Moderate), PS3 (Supporting), PP2 (Supporting)
CTC1	c.2831del	p.Pro944Leufs*7	rs199473677	1	LP	PVS1 (Very Strong), PM2 (Moderate)
PTCH1	c.1511 C > A	p.Pro504Gln	rs1588598694	1	LP	PP3 (Strong), PM2 (Moderate), PM5 (Supporting)
RECQL4	c.2412_2420del	p.Ala805_Arg807del	rs766312203	1	LP	PM4 (Moderate), PM2 (Moderate), PP1 (Supporting), PP5 (Supporting)

Fig. 1.

Distribution of mutations by pathogenicity in a cohort of 210 patients and gene-specific frequencies of pathogenic and likely pathogenic variants.

Fig. 2.

Most frequently pathogenic variants found in patients with clinical criteria for HBOC in Minas Gerais state.

It’s worth noting that some variants presented here have been identified for the first time in Brazil as c.112_113del, c.3328_3329del, c.4689_4694del and c.5072 C > T in the BRCA1 gene; c.409 C > T (CHEK2); c.2990_2993del (BRIP1); c.526-1G > A (NTHL1); c.2831del (CTC1); c.1511 C > A (PTCH1) and c.2412_2420del (RECQL4).

Variants of uncertain clinical significance

Thirty five VUS were found in this study and distributed in 22 different genes. Most of these variants (7/35, 20%) were observed in the ATM gene (Supplementary Table S1). In Brazil, few studies have evaluated this gene in the HBOC population, and further research is needed to better understand the pathogenicity of these variants¹².

Patients and clinical characteristics

A total of 210 female patients who met the clinical criteria for HBOC as recommended by NCCN were included in this study. Among the patients with identified pathogenic mutations, 30 were diagnosed with primary breast cancer before the age of 40 (42.86%) (Table 3). However, a significant number of patients with benign variants were also diagnosed before the age of 40 (39 out of 105; 37.14%), and no statistically significant difference was observed between these groups (p = 0.4486). Furthermore, no statistically significant difference was found between the groups with respect to tumor type (Breast p = 0.1272; Ovarian p = 0.3090). The analysis of hormonal receptors revealed a high prevalence of triple-negative tumors, which are typically associated with BRCA1 gene mutations; however, in this study, such tumors were frequently observed in both groups.

Table 3.

Clinical characteristics of patients harboring benign (n = 105), VUS (n = 35) and pathogenic variants (n = 70).

Clinical features	Patients with benign variants (n and %)	Patients with pathogenic variants (n and %)	χ²	Patients with VUS (n and %)
Age of diagnosis
20–40	n = 39 (37.14%)	n = 30 (42.86%)		n = 17 (48.57%)
41–60	n = 58 (55.24%)	n = 27 (38.57%)	p < 0.05	n = 14 (40%)
61–80	n = 8 (7.62%)	n = 9 (12.86%)		n = 2 (5.71%)
NI	n = 0 (0%)	n = 4 (5.71%)	p < 0.05	n = 2 (5.71%)
Tumors
Breast	n = 94 (89.52%)	n = 57 (81.43%)		n = 24 (68.57%)
Ovarian	n = 5 (4.76%)	n = 6 (8.57%)		n = 2 (5.71%)
Others	n = 6 (5.71%)	n = 2 (2.86%)		n = 2 (5.71%)
NI	n = 0 (0%)	n = 5 (7.14%)	p < 0.05	n = 7 (20%)
Triple negative receptors	n = 21 (20%)	n = 15 (21.42%)		n = 6 (17.14%)
Family History	n = 89 (84.76%)	n = 62 (88.57%)		n = 29 (82.86%)

Discussion

Studied population

Brazil is the seventh most inhabited country in the world, and it has a complex pattern of ethnic diversity. Among their five regions, the Southeast is the most populous with about 85 million people²⁰. The majority of the HBOC molecular studies in Brazil are concentrated on the Southeast and South regions of the country. However, most patients came from the states of São Paulo, Rio de Janeiro, and Rio Grande do Sul, where the most extensive molecular diagnosis and care centers for cancer patients are located¹². The literature review shows a lack of studies using NGS technology in patients with HBOC in Minas Gerais, the largest state in the Southeast in territorial extension (occupying 63% of the area)²¹.

Minas Gerais has 20.539.989 inhabitants, the second most populous state in the Southeast region. Its territory was inhabited by indigenous people when the Portuguese arrived in Brazil. Therefore, most of the population of Minas Gerais are descendants of Portuguese settlers from northern Portugal and African slaves, mainly from West Africa. According to a study carried out on genetic ancestry, the composition of the Minas Gerais population is: 75.4% European, 18.3% African and 5.8% indigenous²². In a comparative study of Brazilian geographic regions, it was seen that the Northeast has the most significant African ancestry. In contrast, the Southeast/South of Brazil has the greatest European ancestry²³. Furthermore, other populations arrived in Minas Gerais at different times, such as Italians, Spaniards, Japanese, Germans, Lebanese, Syrians, and others²⁴.

Few studies now have focused attention on HBOC in the Minas Gerais population. The first three studies in the state used screening mutation methodologies concentrated in specific genes or punctual mutations, while two of them are from our research group, and the results are synthesized here^16,17,25. Two more recent studies from Belo Horizonte, capital of Minas Gerais state, described the mutational profile found in a medical service at a private genetic referral center in the city and in individuals with health insurance tested by private labs in Minas Gerais, both by NGS panel^26,27. In the Carvalho et al. (2023) study, asymptomatic individuals with a familial history of cancer (169/382) were included, and it was not possible to know the real frequency of mutations as they included relatives in the research. However, in Carvalho’s study, the BRCA1 c.470_471delCT mutation was identified in five different families but was not detected in our cohort. Similarly, the BRCA1 c.2808_2811del mutation was more frequent in Faria’s research, yet absent in both Carvalho et al. (2023) and the present study. Conversely, BRCA2 variants such as c.4829_4830del and c.6405_6409del were more prevalent in our cohort and were not reported by Faria et al. (2024). Regarding the TP53 gene, Carvalho et al. (2023) found no carriers of the c.1010G > A mutation, one of the most common in the Brazilian population, while Faria et al. (2024) identified it in only one patient. In contrast, this mutation ranked as the third most frequent in our study. Therefore, certain variants in Brazil appear to be regionally clustered, particularly within areas of large states such as Minas Gerais. Other mutations, such as c.5266dupC, c.2T > G, and c.156_157insAlu, were present in all studies^26,27.

Mutational profile

A recent Brazilian review provides a comprehensive overview of the broad variability in molecular profiles related to hereditary breast and ovarian cancer in the country. Certain mutations stand out in the Brazilian population: c.5266dupC, c.156_157insAlu, and c.1010G > A in the BRCA1, BRCA2, and TP53 genes, respectively¹². In the present work, pathogenic and likely pathogenic mutations have been found in 33.3% of patients (70/210), and a higher frequency of probands harboring the mutations in non-BRCA genes (30/210, 14.3%) was found when compared with other Brazilian research. Only ten studies in Brazil up to now evaluated breast cancer probands using multigene genetic panel tests, and the frequency of patients with pathogenic mutations in non-BRCA genes varied from 1.5 to 12.5%^26–35. It is important to note that differences in the panels used may contribute to the variation in frequency.

Another recent study presents a geographical distribution of the most frequent BRCA1/2 mutations in Brazil, established by BRCA genetic testing results from 1267 unrelated individuals investigated routinely in a private laboratory, but it was not possible to identify the distribution of samples through the southeast states of Brazil³⁶.

Considering the 210 patients tested in Minas Gerais in the present work, it was seen that the frequency of pathogenic mutations in the BRCA2 gene was higher than in the BRCA1 gene, contrary to most studies published in the country¹². This result is in accordance with the two previous studies from Minas Gerais^26,27. Four BRCA2 mutations have a significant impact on this data as they represent 78.3% (18/23) of all BRCA2 mutations and 25.7% (18/70) of all P/LP mutations found in this study (Fig. 2). One of these mutations, as c.4829_4830del, the most common mutation in our cohort, seems to be very rare in other cohorts of Brazilian breast/ovarian patients. The pathogenic frameshift variant c.4829_4830del results in the deletion of two nucleotides in exon 11 of the BRCA2 gene.

Records of this mutation have been identified in populations such as Korean, Pakistani, Moroccan, Israeli, and Ashkenazi Jewish groups^37–41. Interestingly, immigration from the Middle East and Asia, regions where this mutation is most prevalent, accounted for only 2% of Brazil’s total immigration flow until 1972²². In Brazil, this variant has been reported in only four other studies, all from the South and Southeast regions, with just seven documented cases.

These regions exhibit more prominent European and African ancestries, reflecting historical European colonization and the legacy of slavery^22,26,42,43. Mechanisms such as genetic drift and founder effect may partially explain the high frequency of this mutation in the Midwest region of MG, as we see many rural populations in this region of the state.

The second most frequent mutation found in Minas Gerais state was c.5266dupC in the BRCA1 gene present in 2.4% (5/210) of patients. The frequency of this variant varies among regions of Brazil and had already been reported in a frequency of 11.6% (11/95) in the Carvalho et al., study with lower frequencies in other works^{28,32,34,35,44}. This is an ancestral mutation in the Ashkenazi Jewish population⁴⁵, considered the most frequent mutation among Brazilian patients, representing 26.8% of all germline mutations identified in the BRCA1 gene and detected across all geographic regions¹². It was responsible for 7% of the germline pathogenic mutations found here.

The c.1010G > A mutation was detected in five women, all diagnosed under 45 years old, and two of them had a family history of other cancers, such as prostate, pancreatic, esophageal, and leukemia, beyond breast cancer. The c.1010G > A variant is a Brazilian founder mutation and was found in 0.3% of the general population in southern Brazil⁴⁶. This mutation is associated with Li-Fraumeni syndrome, an inherited cancer predisposition disease caused by a germline mutation in the TP53 gene. People with this syndrome have an increased risk for several types of cancer, such as childhood sarcoma, breast cancer, central nervous system tumors, leukemia, melanoma, prostate and pancreatic cancer^47,48. It is very common to find Li-Fraumeni families filling clinical criteria for HBOC, which demonstrates that NGS panels give a precise molecular diagnosis, contributing to the follow-up of the patients. In Brazil, TP53 is the most mutated gene after BRCA in breast cancer patients, with the c.1010G > A variant representing more than 75% of the identified variants inside the gene. In the recent Brazilian review, it has been identified in several HBOC Brazilian studies with frequencies ranging from 0.8 to 7.1%. Considering all works that screened for this specific mutation, the frequency of c.1010G > A in patients who met clinical criteria for HBOC from Brazil was estimated at 1.83% (61/3336)¹². In the present study, the c.1010G > A was found in 2.3% of patients from Minas Gerais, and it was responsible for 7% of the germline pathogenic mutations found.

The c.156_157insAlu mutation accounted for 5.7% (4/70) of the pathogenic mutations identified in this study and warrants particular attention. This is a Portuguese founder mutation, and its high occurrence in Brazil is probably the result of Portuguese immigration during centuries of colonization¹². It is frequently found in the predisposition genes BRCA2 for breast cancer and causes a jump in exon 3 that leads to splicing errors and, consequently, in the transcription and translation of the tumor suppressor protein. This mutation originated from families with cases of HBOC in the northeast and central regions of Portugal, representing 27–38% of all pathogenic BRCA2 mutations. Brazil is a country of Portuguese colonization, and until 1991, 2.2 million of these immigrants were received, which makes this mutation an interesting target of study^49,50. This rearrangement was reported most frequently in populations from the south and southeast regions^32,49,51 and has been seen with low frequency in HBOC families from the central-western region 1/224 (0.4%)³³. To date, the Portuguese founding mutation BRCA2 c.156_157insAlu has not been identified in populations from the north and northeast of Brazil. This differential loading of Alu elements across the BRCA2 locus in many regions is likely due to differences in structure between populations.

The c.2T > G mutation is also frequently reported in Portuguese families and has been previously described in Brazilian patients with hereditary breast and/or ovarian cancer, particularly in the Southeast and South regions of Brazil^13,14,28,31.

Related to the variants of uncertain significance, the ATM gene had the highest variant frequency. This gene is associated not only with Ataxia-telangiectasia Syndrome and breast cancer but also with several other types of cancer, such as ductal adenocarcinoma of the pancreas, colorectal, prostate, endometrial, kidney, liver, ovarian, esophageal, salivary gland, gastric, thyroid and urinary tract^52,53. Among seven patients with VUS in this gene, four had family members with these cancers, especially prostate, breast, liver, endometrial, and pancreatic. Other VUS have been identified in the BRCA2, CHEK2, MLH1, MSH2, NF1 and RAD50 genes in patients who had a strong family history of several types of cancer. Future studies are essential to elucidate the pathogenicity of these variants.

The Hereditary Cancer Predisposition Assessment and Family Monitoring Program in Minas Gerais, Brazil, has enabled access to comprehensive hereditary cancer services within the public health system. More than 250 family members of patients carrying pathogenic mutations have been attended through the program, with all services extended to them as well. This initiative facilitates the identification of individuals at high risk for cancer development, along with the implementation of preventive measures to reduce risks, enhance surveillance, enable early diagnosis, personalize patient prognoses, and explore the potential use of targeted therapies.

This study has some limitations. The sample size may not fully represent the genetic diversity of the Minas Gerais state. The gene panel used was limited and might have missed some relevant pathogenic variants. Additionally, 44 patients were analyzed by Sanger sequencing, which only assessed BRCA genes, potentially overlooking mutations in other relevant genes. Future studies with larger cohorts in Minas Gerais, expanded panels, and full NGS analysis are needed. Finally, since the data were collected over several years, the PM2 criterion was retained as moderate rather than supporting, to avoid inconsistencies.

Conclusions

This study aimed to evaluate the molecular profile of hereditary breast and ovarian cancer in the state of Minas Gerais and represents the first research in the region to use a multigene panel via next-generation sequencing, focusing on patients from the Brazilian public health system. By identifying pathogenic mutations in individuals at high genetic risk for breast and/or ovarian cancer, precision medicine care can be implemented, providing better assistance to patients and their families. There is often overlap in clinical criteria among different hereditary syndromes, and next-generation sequencing technology in precision medicine is a valuable method for distinguishing between these conditions and ensuring appropriate clinical care for patients and family members.

Methods

Patients and clinical data

Overall, 210 female patients enrolled in the Hereditary Cancer Predisposition Assessment Program in Minas Gerais underwent genetic testing for germline mutations. Despite all being assisted by the SUS, 84 patients were able to afford the test, which was performed in private laboratories. Meanwhile, 82 patients, who lacked the financial means to pay for the test, were tested at the UFSJ, following the methodologies described herein. It is important to note that the number of genes included in NGS panels from private laboratories differs from those covered in this study; however, all panels include at least the 22 genes presented here. All results from tests conducted in private laboratories, as well as those conducted in the University, are evaluated by the medical team for patient and family support. These results are presented together to ensure greater accuracy in the mutation frequency in the state. The mutations identified in the forty-four patients who had the BRCA1 and BRCA2 genes sequenced by Sanger sequencing during the initial years of the program are also included here for the same purpose^16,17. This study received approval from the Ethics and Research Committee of the São João de Deus Hospital (45662921.9.0000.5545). Written informed consent was obtained from all participants. All methods were performed in accordance with relevant guidelines and regulations. To be included in the study, the individuals had a prior breast or ovarian cancer diagnosis and fulfilled the ‘NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines^®) for Genetic/Familial High-Risk Assessment: Breast, Ovarian, and Pancreatic’ (version 2022.1)¹⁸. Clinical data were collected from patient medical records. Pedigrees were constructed using the Progeny Pedigree Tool (https://pedigree.progenygenetics.com/).

DNA samples

Peripheral blood samples (3–5 mL) were collected in vacutainer tubes with EDTA. Genomic DNA was extracted through the Salting Out method and the Qiagen MiniAmp DNA Kit. The concentration and purity of the DNA samples obtained were analyzed through NanoDrop™ 2000/2000c Spectrophotometer and Qubit 4.0 fluorometer (Thermo Fisher) with the kit QubitTM 1X dsDNA HS Assay (Thermo Fisher).

Genetic screening

A NGS multi-gene panel composed of 22 genes, ATM, BARD1, BRCA1, BRCA2, CHEK2, CDH1, EPCAM, MLH1, MSH2, MSH6, NBN, NF1, PALB2, PTEN, RAD50, RAD51, RAD51B, RAD51C, RAD51D, SMARCA4, STK11 and TP53, was designed. The primers were designed through the Sequence Assay Designer Illumina tool, and two primer pools were created for the amplification of 803 amplicons with a 100% horizontal coverage of exons, 5’ and 3’ UTRs and splicing sites of all the 22 genes, targeting a total of 177.873 bp (Supplementary Table S2).

The kit AmpliSeq for Illumina Custom DNA Panel was used to prepare the DNA libraries according to the procedures established in the protocol AmpliSeq for Illumina On-Demand, Custom, and Community Panels. The sequencing was performed on the MiSeq equipment using the kit MiSeq Regent v2 Micro (Illumina), with an input of 11 pM. This kit enables paired-end sequencing of 150 bp short reads, generating up to 1 GB of data and 6.6 million reads per run, with 96.7% of bases having a quality score above Q30.

Bioinformatic analysis

The primary analysis was performed using FastQC (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) and MultiQC (https://github.com/MultiQC/MultiQC) to evaluate the quality of the data, and Trimmomatic (http://www.usadellab.org/cms/?page=trimmomatic) to trim reads. Two distinct bioinformatics strategies were adopted for the secondary analysis of the trimmed FASTQ files. The first pipeline utilized the DNA Amplicon tool (v2.1.1), available in Illumina’s environment, while the second was developed based on the Genome Analysis Toolkit (GATK) best practices for data pre-processing and germline short variant discovery workflows (https://gatk.broadinstitute.org/hc/en-us). Quality control checkpoints were established throughout the analysis pipelines to monitor and control of the obtained data. To achieve this, tools such as Qualimap (http://qualimap.conesalab.org/), Samtools (https://www.htslib.org/), BEDtools (https://bedtools.readthedocs.io/en/latest/), BCFtools (https://samtools.github.io/bcftools/bcftools.html) and VariantQC (https://github.com/BimberLab/DISCVRSeq/), among others, were employed.

The annotation process (tertiary analysis) of the VCF files generated by the DNA Amplicon pipeline was performed using the GEMINI framework, which employs databases such as RefSeq, ENCODE, OMIM, dbSNP, KEGG, and HPRD. In the GATK pipeline, annotation was conducted by the TAPES tool (https://github.com/a-xavier/tapes), based on the ANNOVAR tool, utilizing databases such as ClinVar, gnomAD, dbSNP, dbscSNV, among others.

Variant classification

Quality values showed a GC content of 41%, an average coverage of approximately 230X, and 97.6% of the target regions covered at 20X, ensuring the quality of the generated files. The identified variants were filtered based on a minimum coverage of 20 reads (DP > 20) and an allele frequency greater than 30% for heterozygous variants. Only the variants detected by both pipelines were considered for the results. A visual analysis of the BAM and VCF files was also performed using the Integrative Genomics Viewer (IGV, https://igv.org/) tool. Variants considered as likely pathogenic, pathogenic, or of uncertain significance by the annotation process were verified by four independent researchers using the ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/), Varsome (https://varsome.com/) and Franklin (Genoox) databases (https://franklin.genoox.com), in accordance with the guidelines of the American College of Medical Genetics (ACMG).

Statistical analysis

Statistical analyses were performed to compare clinical characteristics between patients with pathogenic or likely pathogenic variants and those with benign variants. Categorical variables, including categorized age at diagnosis, tumor type, and variant pathogenicity, were compared using the Chi-square test or Fisher’s exact test, as appropriate. All statistical analyses were conducted using R software (version 4.3.2). A p-value < 0.05 was considered statistically significant.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

A.A.F.R. drafted the manuscript, analyzed and interpreted the data, and performed sequencing (NGS); T.Q.L. conducted the sequencing, analyzed and interpreted the data, and edited the manuscript; M.V.G.A. developed in silico analysis code, contributed to the molecular interpretation of the data, and edited the manuscript; C.P.N., F.C.F., and F.C.F. were responsible for patient recruitment, genetic counseling, data analysis, and clinical interpretation; D.O.L. provided research facilities; E.M.T.S. assisted with bioinformatics analyses; L.L.S. conducted a critical review of the manuscript and conceived the original idea for this article. All authors contributed to data interpretation, read, and approved the final manuscript.

Funding 

This article is funded by Foundation for Research Support of the State of Minas Gerais (FAPEMIG, Grant Number, RED-00314-16 and RED-00089-23), Foundation for the Coordination of Higher Education Personnel Improvement (CAPES) and Federal University of São João del-Rei.

Data availability

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests

Contest for publication

The authors declare no competing interests. The authors have given their consent for the publication of this manuscript. All patients voluntarily agreed to participate in this study by signing an informed consent form, which was approved by the ethics committee of our institution.

Ethical approval and consent to participate

This study received approval from the Ethics Committee of São João de Deus Hospital (approval number 45662921.9.0000.5545). All participants provided written informed consent prior to inclusion, and all procedures were carried out in compliance with applicable ethical standards and regulations.