This prospective study was conducted between 2021 and 2023 at three major oncology referral centers in Yaounde, Cameroon: the Yaounde General Hospital, the Yaounde Central Hospital, and the Solidarity Chemotherapy Non-Governmental Organization. Ethical approval was obtained from the Cameroonian National Human Health Research Ethics Committee (N˚2021/12/1424/CE/CNERSH/SP).

A total of 82 breast cancer patients who had never previously undergone genetic testing were enrolled. Eligibility required a histologically confirmed breast cancer and age of at least 18 years. Demographic, reproductive, and clinical data were collected alongside detailed family histories. Enrollment was consecutive. Of the 111 eligible patients approached, 24 declined participation due to fear of learning about the existence of mutations, and 5 were excluded because they were under 18 years of age. All three centers contributed patients: Yaounde General Hospital, Yaounde Central Hospital, and SOCHIMIO. Family history of cancer was defined as the presence of at least one first- or second-degree relative with a documented cancer diagnosis, including breast, ovarian, colorectal, gastric, uterine, cervical, melanoma, pancreatic and prostate cancer.

All participants received culturally adapted pre-test genetic counseling delivered by Cameroon’s trained cancer genetic counselors. The counseling protocol incorporated both national languages (French and English) and addressed community concerns around hereditary disease, stigma, and implications for family members. Post-test counseling was provided upon return of results.

Five millilitres of saliva were collected from each participant. DNA was extracted and analyzed using next-generation sequencing (NGS) with a panel of 29 cancer predisposition genes: BAP1, MITF, CDK4, CDKN2A, ATM, CDH1, CHEK2, PTEN, BRCA1, BRCA2, PALB2, BARD1, BRIP1, TP53, STK11, MLH1, MSH2, EPCAM, MSH6, PMS2, RAD51C, RAD51D, APC, BMPR1A, SMAD4, GREM1, MUTYH, POLD1, and POLE. Sequencing was performed at Color Genomics (Burlingame, CA, USA), a College of American Pathologists (CAP)-accredited, Clinical Laboratory Improvement Amendments (CLIA)-certified laboratory. The present article focuses specifically on BRCA1 and BRCA2 findings. Pathogenic or likely pathogenic variants identified in the remaining 27 genes of the panel are reported in full in the companion landscape article. Among the 82 breast cancer patients in this cohort, no participant carried more than one BRCA1/2 P/LP variant; however, 8 patients carried a BRCA1/2 P/LP variant alongside a VUS in a second gene, as detailed in section 3.5.

All identified variants were classified as pathogenic (P), likely pathogenic (LP), variant of uncertain significance (VUS), likely benign, or benign in accordance with American College of Medical Genetics and Genomics/Association for Molecular Pathology (ACMG/AMP) 2015 guidelines [13], cross-referenced with ClinVar.

Descriptive statistics were calculated as mean ± SD for continuous variables and frequencies with percentages for categorical variables. Independent samples t-tests compared the mean age at diagnosis between BRCA1/2 carriers and non-carriers. Chi-square tests and Fisher’s exact test assessed associations between mutation status and categorical variables. Linear regression identified predictors of BRCA1/2 carrier status. All tests were two-sided; p < 0.05 was considered statistically significant. The multivariable regression model included the following covariates: age at diagnosis (continuous), family history of cancer (binary: yes/no), menopausal status (binary), and tumor stage at diagnosis (ordinal: I - IV). Missing data were handled by complete-case analysis (listwise deletion). This analysis used a 29-gene cancer predisposition panel (see section 2.4); pathogenic or likely pathogenic variants identified in genes other than BRCA1/2 are reported in the companion landscape article.

The cohort comprised 82 breast cancer patients. The mean age at diagnosis was 43.4 ± 11.5 years (range 19 - 71). Forty-three patients (52.4%) were diagnosed before age 45. Figure 1 summarizes the age distribution at diagnosis in breast cancer patients. A family history of cancer was reported by 57 patients (69.5%). Among patients reporting family history, the most commonly reported malignancy in relatives was breast cancer (n = 44), followed by cervical (n = 11), prostate (n = 10), uterine (n = 5), ovarian (n = 4), gastric (n = 4), and colorectal (n = 3) cancers.

Figure 1. Age distribution at diagnosis in breast cancer patients (N = 82) by BRCA1/2 P/LP carrier status.

Pathogenic (P) BRCA1 variants were identified in 15 patients (18.3% of the breast cancer cohort) and one likely pathogenic (LP) BRCA1 variant in an additional patient, giving a total of 16 BRCA1 P/LP carriers (19.5%). Two patients carried pathogenic BRCA2 variants (2.4%). Combined, BRCA1/2 P/LP variants were present in 18 breast cancer patients (22.0%). Table 1 presents the complete BRCA1/2 variant spectrum.

Table 1.Frequency and distribution of pathogenic BRCA1 and BRCA2 mutations (N = 82).

Eight distinct P/LP BRCA1 variants were identified in 16 BRCA1-positive patients (Figure 2). The most frequent was c.4484G>T (p.Arg1495Met), detected in 7 patients, representing 43.8% of all BRCA1 P/LP carriers and 8.5% of the total breast cancer cohort. The second most frequent BRCA1 variant was c.5155dup (p.Val1719Glyfs*6), found in 4 patients (including 1 LP). Additional BRCA1 variants—c.2157dup, c.2517_2518del, c.3607C>T, c.38del, c.4676-1G>C, and c.66dup—were each observed in a single patient. The variant c.4484G>T (p.Arg1495Met) is classified as pathogenic (Class 5) in ClinVar based on ACMG/AMP criteria: PS1 (same amino acid change as established pathogenic variant), PS3 (functional studies demonstrating loss of BRCA1 function), PM2 (absent from gnomAD), and PP3 (multiple computational evidence of deleteriousness). The variant c.5155dup (p.Val1719Glyfs*6) was classified as pathogenic in 3 patients and as likely pathogenic in 1 patient. This discrepancy reflects [case-level reinterpretation based on individual clinical data/laboratory re-evaluation between testing dates/slight differences in allele-specific evidence]. The 4 patients are counted separately in the variant frequency table.

Figure 2. BRCA1 pathogenic variant spectrum (N = 16 BRCA1 P/LP carriers).

Two pathogenic BRCA2 variants were identified: c.1813dup (p.Ile605Asnfs*11) and c.5572del (p.Thr1858Glnfs*5), each found in one patient, each comprising 50% of all BRCA2 pathogenic findings.

VUS were detected in 21 breast cancer patients (25.6%). Among BRCA1/2 P/LP carriers, VUS co-occurrence was notable: 8 of 18 BRCA1/2 carriers (44.4%) had an associated VUS in a second gene. Two patients with the recurrent c.4484G>T BRCA1 variant had a co-occurring VUS in PALB2 (c.365A>G, p.Asp122Gly); both were under 30 years of age and reported multiple first- and second-degree relatives with breast or ovarian cancer. Two other patients carried a VUS in APC (c.3760A>G, p.Ile1254Val). No VUS was identified within BRCA1 itself; two VUS were detected in BRCA2.

The mean age at diagnosis was 35.84 ± 9.92 years among BRCA1/2 P/LP carriers, compared to 40.54 ± 10.65 years in non-carriers. Family history of cancer was reported in 76.5% of BRCA1/2 carriers versus 57.6% of non-carriers. Regression analysis identified younger age at diagnosis (p = 0.04) and positive family history (p = 0.03) as significant independent predictors of BRCA1/2 carrier status (Figure 3).

Figure 3. Clinical predictors of BRCA1/2 P/LP carrier status. (A) Mean age at diagnosis ± SD by carrier status (p = 0.04, t-test); (B) Proportion with positive family history by carrier status (p = 0.03, chi-square).

Several limitations warrant consideration. First, the study was conducted at urban tertiary centers in Yaounde, potentially limiting generalizability to rural populations and other regions. Second, the sample size (N = 82) limits statistical power for subgroup analyses; larger, population-representative studies are needed. Third, haplotype analysis was not performed to confirm the putative recurrent status of c.4484G>T; this remains a working hypothesis. Fourth, clinical outcome data (receptor status, survival) were unavailable for most patients, precluding genotype-phenotype analyses. Finally, the use of an international laboratory introduces logistical delays and costs not representative of a mature national program.