Causative Role for a BRCA2 Germline Pathogenic Variant in External Auditory Canal Squamous Cell Carcinoma.

1
Introduction
External auditory canal squamous cell carcinoma (EACSCC) is among the rarest types of squamous cell cancers, with an estimated prevalence of 1–6 per million [1]. The average age of onset is around 64 years old [2]. These tumors initially present as otorrhea, ear pain, or facial nerve palsy [1]. Treatment consists of surgical resection, sometimes followed by adjuvant radiotherapy and/or radiochemotherapy [3].
Available genomic studies have reported recurrent somatic pathogenic variants in TP53 and NOTCH1/2, with substantial genetic heterogeneity [4, 5, 6]. To date, no germline pathogenic variants (GPVs) have been described in EACSCC patients. BRCA2 GPVs predispose to breast, ovarian, pancreas, and prostate cancer [7], and have been associated with melanoma [8], but not non‐melanoma skin cancer [8]. Here, we present a case of a BRCA2 GPV carrier who developed an EACSCC at age 66 and investigated whether this GPV contributed to tumor development.

2
Materials and Methods
2.1
Patient Samples and Whole‐Exome Sequencing
Peripheral blood and Formalin‐Fixed Paraffin‐Embedded tumor blocks were collected. Genomic DNA extraction and whole‐exome sequencing (WES) analyses were performed as previously reported [9]. Variant filtering and prioritization were conducted using an approach adapted from a previous study [9], and Tumor Mutational Burden (TMB) was calculated as the number of somatic non‐synonymous coding mutations per megabase of coding sequence (see Supporting Information).

2.2
Copy‐Number, Genomic Scar and Mutational Signature Analysis
Copy‐number analysis was performed using Sequenza [10]. Genomic scar analysis was conducted with ScarHRD [11]. Single base substitution (SBS) and INDEL mutational signatures were analyzed using SigProfilerAssignment (v0.1.9) [12] and SigProfilerMatrixGenerator (v1.2.31) [13]. Somatic INDELs were inspected in the Integrative Genomics Viewer to exclude artifacts. SBS3 was additionally evaluated using SigMA [14].

3
Results
3.1
Case Presentation
At age 54, the patient was diagnosed with infiltrating ductal carcinoma of the right breast (estrogen receptor positive, progesterone receptor negative, HER2 negative, no lymph node involvement). She underwent right partial mastectomy under wire localization and sentinel lymph node biopsy during the same year. She was subsequently treated with adjuvant radiotherapy and anastrozole for 10 years. At age 55, following her sister's diagnosis of breast cancer, associated with the BRCA2 GPV c.8537_8538del; p.Glu2846GlyfsTer22 (a founder variant in the French‐Canadian population), genetic testing confirmed that she was also a carrier.
She subsequently underwent total hysterectomy and bilateral salpingectomy at age 56, as well as prophylactic bilateral total mastectomy at age 57 with no signs of recurrence of her breast cancer (at the time of writing this report, the patient is recurrence free, 16 years following her partial mastectomy). Her daughter tested negative for the BRCA2 GPV (Figure 1A).
At age 66, she presented with right ear bleeding. Examination revealed a polypoid lesion on the posterior aspect of the external auditory canal. Pathology revealed an invasive keratinizing squamous cell carcinoma (Figure 1B). Mastoid computed tomography (CT) scan showed a focal nodular soft tissue thickening along the posterior wall of the bony right external auditory canal (Figure 1C,D). The tumor was resected by sleeve resection without radiation therapy.
The surgery had negative resection margins, and the patient is recurrence‐free, 3 years following surgery. She had no history of recurrent otitis, immune suppression, or psoriasis. (Figure 1A).

3.2
Molecular Findings
Germline WES analysis confirmed the BRCA2 GPV (c.8537_8538del), while tumor copy‐number analysis revealed loss of heterozygosity (LOH) at the BRCA2 locus with copy‐number gain (Figure 2A). VAF of the BRCA2 GPV in the tumor was 76%, consistent with amplification of the mutant allele across a large segment of chromosome 13q spanning about 88 Mb (27.3–115.03 Mb).
Tumor WES analysis identified 7 potentially disease‐associated somatic variants prioritized through exome analysis, including a somatic TP53 splicing variant (c.782 + 2T>C), together with a somatic nonsense variant in the homologous recombination repair gene RAD54B (c.C2138G: p.S713X) and a somatic nonsense variant in the cancer susceptibility gene SMAD4 (c.C1454G: p.S485X), among others (Table 1A). Given the presence of a BRCA2 GPV with LOH in the tumor, we hypothesized that homologous recombination repair may be impaired. Genomic scar analysis exhibited an HRD score of 66, well‐above the established cutoff of > 42, consistent with an HRD+ status. TMB was 12.7 somatic mutations per megabase of coding sequence, indicating a moderate mutational load.
A total of 818 somatic single‐base substitutions were identified in the tumor and used for mutational signature analysis, which revealed a dominant contribution of the APOBEC‐related signatures SBS2 and SBS13, together accounting for approximately 87% of the mutational profile, followed by the clock‐like signatures SBS1 and SBS5 with a combined contribution of 12% (Figure 2C and Table 1B). SBS3 was not detected by this tool. The overall cosine similarity between the observed and reconstructed SBS profile was 0.989. INDEL mutational signature analysis showed the HRD‐associated ID6 as the unique contributing INDEL signature with 7 INDELs assigned (Figure 2B,C). The cosine similarity between the observed and reconstructed INDEL profile was 0.814. Additional analysis using SigMA detected the HRD‐associated signatures SBS3 and SBS8 (12% combined), alongside a dominant contribution of the APOBEC signatures SBS2 and SBS13 (84%).

4
Discussion
To the best of our knowledge, this case represents the first report of an individual carrying a GPV in BRCA2 developing EACSCC. In our study, copy‐number and tumor VAF analysis showed LOH at the BRCA2 locus, consistent with biallelic inactivation. Independent genomic features further supported an HRD phenotype, including an elevated genomic scar‐based HRD score and the presence of the HRD‐associated mutational signatures SBS3, SBS8, and ID6. ID6 is characterized by ≥ 5 base pairs deletions with microhomology at breakpoint junctions, consistent with repair via microhomology‐mediated end‐joining, an alternative mechanism in tumors with BRCA1/2 biallelic inactivation [15].
APOBEC signatures showed a highly dominant contribution, suggesting that combined effects of HRD and APOBEC activity contributed to tumorigenesis, with BRCA2 as the germline predisposition factor. Prior work showed that replication stress in BRCA2‐deficient experimental models promotes APOBEC‐mediated DNA damage, which generates double‐strand breaks contributing to HRD‐associated mutational signatures [16]. Three studies have collectively investigated the molecular characteristics of about 40 EACSCC cases [4, 5, 6], which found TP53 as the gene most frequently mutated in the tumor, consistent with our findings. Other somatic mutations reported were not consistently observed across these studies (and absent in our study), highlighting the genetic heterogeneity of EACSCC. Notably, none of them reported alterations in canonical HRD genes or provided genomic evidence of HRD. In one cohort, APOBEC signatures were detected with moderate contributions [4], but HRD signatures were not observed.
Although SBS3 was detected in our study with low contribution, a prior study reported tumors with elevated genomic scar‐based HRD scores exhibiting low SBS3 [17]. HRD is characterized by multiple genomic alterations beyond SBS3, optimally captured by whole‐genome sequencing (WGS) data [18]. Accordingly, WGS‐based integrative classifiers have shown improved accuracy in HRD detection compared with reliance on single HRD measures [19]. Our findings have therapeutic relevance. While HRD predicts sensitivity to poly (ADP‐ribose) polymerase (PARP) inhibitors in breast, ovarian, prostate, and pancreatic cancer [20], preliminary clinical data further support responses of head and neck squamous cell carcinoma patients to PARP inhibitors [21].
Our study has limitations. Analysis of a single EACSCC case carrying a BRCA2 GPV limits broader comparisons, and the use of WES rather than WGS precluded the use of WGS‐based integrative classifiers. Nevertheless, multiple genomic features supported an HRD phenotype, including an elevated HRD score, SBS3, SBS8, and ID6. Altogether, we report the first case of EACSCC harboring a BRCA2 GPV with genomic evidence of HRD and APOBEC activity, supporting a potential causative role of BRCA2‐associated HRD in EACSCC tumorigenesis, emphasizing the importance for further molecular analysis in EACSCC.

Author Contributions

José Camacho‐Valenzuela: data curation; formal analysis; writing – original draft. Dylan Pelletier: investigation; visualization; writing – original draft. Paz Polak: supervision. Lili Fu: investigation, resources. Nancy Hamel: supervision; conceptualization. Carla Daniela Robles‐Espinoza: supervision; conceptualization. William D. Foulkes: supervision; conceptualization; investigation; project administration; funding acquisition.

Funding
This work was supported by the Canadian Institutes of Health Research (grant FDN‐148390).

Ethics Statement
Study approved by the McGill University Health Centre Research Ethics Board (Project No. MP‐37‐2019‐4865).

Consent
Written consent was obtained from the patient.

Conflicts of Interest
The authors declare no conflicts of interest.

Supporting information

Data S1: gcc70123‐sup‐0001‐Supinfo.docx.