Oncologist-led germline genetic testing for uveal melanoma

ABSTRACT Purpose To report the genotype and phenotype of a cohort of unselected uveal melanoma (UM) patients who had germline multi-gene panel genetic testing, including the BAP1 gene, from a large multi-ethnic cancer centre. We describe the central role of the medical genetics clinic in collaboration with oncologists in a mainstreaming model to facilitate genetic testing, counselling and streamlining of patients with hereditary cancer predisposition. Methods A retrospective chart review of clinical and genetic findings of unselected UM patients who had germline genetic testing between December 2019 and October 2021 was conducted. Extracted DNA from peripheral blood samples were analyzed with a multi-gene panel that included at least six genes associated with hereditary melanoma. The correlation between the genotype and the phenotype of the cohort was evaluated. Statistical analysis comprised descriptive and comparative statistics with significance assigned at p < .05. The genetics clinic streamlined patients among the relevant oncology clinics for cancer screening in germline BAP1 positive individuals. Results In unselected UM patients, 3.5% (4/114) tested positive for a BAP1 pathogenic variant. Germline BAP1 status was associated with a family history of mesothelioma (p = .0015) and metastatic disease (p = .017). There were no other significant associations between the patient- or tumour-related characteristics and germline BAP1 results. Conclusion A germline BAP1 mutation was detected in 3.5% of unselected UM patients. The oncologist-initiated and genetics-led mainstreaming model is a straightforward process and can be utilized for offering genetic testing to all UM patients.


Introduction
Uveal melanoma (UM) is the most common primary intraocular cancer in adults, with an estimated incidence of 6per million individuals per year (1). Historically, UM has been considered a sporadic cancer with few risk factors suggested, such as fair hair/skin complexion and light iris colour (2), and family history of UM (3). In contrast to other cancer types such as breast and ovarian cancer where hereditary predisposition genes have been known for decades (4,5), studies investigating the genetic contribution to UM are more recent.
Identifying a hereditary cancer germline mutation is important for patients to enroll in screening for early detection of relevant other cancers or guiding risk-reducing options. Highrisk surveillance programs have been demonstrated to reduce cancer-related deaths and cost-effectiveness leading to increase life expectancy for several hereditary cancer conditions, including BAP1 tumour predisposition syndrome (BAP1-TPDS) (19). Moreover, once a gene mutation is identified, cascade genetic testing can be offered to family members to aid in risk stratification. Those who inherited the familial variant can similarly benefit from screening or management options, while those who did not inherit the familial variant can be relieved from screening programs.
As considered a sporadic cancer, UM patients do not traditionally comprise a significant population in cancer genetics clinics (20). Previous guidelines for germline BAP1 genetic testing eligibility included criteria based on age (age of onset<40), personal history of multiple BAP1-TPDSassociated tumours, or family history of one first-or seconddegree relative with≥1 BAP1-TPDS-associated tumours (13,21). In busy oncology clinics, such complex criteria can be confusing to implement, and not all eligible patients may be referred. Other models are needed to mediate genetic testing with increasing demands for genetics services and a shortage of genetic counsellors (20,(22)(23)(24)(25).
Oncologist-led genetic testing, colloquially known as "mainstreaming," is an alternative delivery model developed to improve timely access to genetic services. The oncologist here refers to the clinician who treats patients with specific organ or system cancer, such as an ocular oncologist, cutaneous oncologist, or renal oncologist. This model was first popularized by the Mainstreaming Cancer Genetics Program in the United Kingdom for patients with ovarian cancer (26). Through collaboration with genetic services, oncologists order genetic testing for patients who meet eligibility criteria, with genetics professionals involved only for the disclosure of the results. Studies have shown that this model provides a streamlined alternative to genetic testing with decreased time to genetic testing results, improved cost-effectiveness, and high levels of patient satisfaction (26)(27)(28)(29).
Mainstreaming has been described for several cancer sites, primarily breast and ovarian cancer (29). To our knowledge, oncologist-led genetic testing has not previously been described for patients with uveal melanoma. A recent paper in 2022 however highlighted the need for upskilling of clinicians to provide genetic testing for familial melanoma and describes a similar sample protocol for genetic testing to be ordered by non-genetics clinicians, focusing on cutaneous melanoma (30,31). We present the protocol and initial data describing the genotypic and phenotypic correlation in an unselected cohort of UM patients who underwent germline genetic testing. This may better define the natural history of hereditary UM with impact to patient treatment and followup. We also describe the collaboration model between various oncology clinics and the genetics clinic in integrating mainstreaming for UM. This clinical integration model might be helpful in other centres to adopt a similar model of care.

Study design
Oncologist-mediated genetic testing workflow Beginning in December 2019, we implemented an oncologistled genetic testing workflow in collaboration between the genetics and ocular oncology ( Figure S1). Oncologists began offering multi-gene panel germline testing to all newly diagnosed UM patients. This germline testing was also offered to select patients after UM diagnosis or along the treatment course.
During the consent discussion, patients were provided with a package of an introductory letter, an informational brochure about genetic testing, the genetic testing requisition, and a personal and family history form. Patients were encouraged to return the personal and family history questionnaires to the genetics team for review. Patients who consented to genetic testing were referred at this time to the genetics team for tracking of results.
Genetics testing results were sent to both the genetics team and the ordering oncologist. For patients with negative results, a note was documented in their electronic patient record to summarize the results and limitations, with the option to speak with a genetic counsellor for questions, if requested. Patients with positive or inconclusive results were offered a phone appointment with a genetic counsellor to discuss the results, recommendations, and genetic testing for family members if applicable. Genetics results for all patients were accessible online through the patient portal.

Study participants and recruitment
A retrospective chart review was conducted for clinical and genetics data. We included consecutive patients with a clinically confirmed diagnosis of UM, ≥ 18 years of age, and who were offered germline genetic testing by their oncologist with results received between 1 December 2019 to 1 November 2021. Patients were excluded if they did not meet the inclusion criteria.
Relevant clinical information (personal and family history of cancer, age at diagnosis, ethnicity, tumour and germline genetic testing results, tumour location, tumour characteristics, cancer stage at diagnosis, and history of metastatic disease) was collected from participants' medical records, the genetics database, and the ocular oncology imaging database.
The study was conducted following the Declaration of Helsinki, and the protocol was approved by the Research Ethics Board at the University Health Network (REB 21-6064).

Statistics
Descriptive statistics were utilized to determine the frequency of variables. Group comparisons were assessed using the Mann-Whitney-Wilcoxon test for continuous variables and the Fisher exact test for categorical variables.

Results
A total of 114 unselected UM patients were included. 54% of patients were female and 46% were male. The mean age at diagnosis in the overall cohort was 57.0 (SD = 15.2). Ethnicity data was available for 65 patients from the overall cohort, as not all patients returned the family history package. The majority of patients self-report Caucasian ethnicity including English/Scottish/Irish ethnicity, Other European (ex/Greece, Holland, Portugal, Spain), and Eastern European (ex/ Germany, Hungary, Poland, Russia) ethnicity. 4 patients were treated with enucleation, 3 patients were treated with external beam radiation, and 107 patients were treated with plaque brachytherapy. It is noted that 20 additional patients who were referred to genetics within the study time period did not provide a blood sample.
A BAP1 germline mutation was detected in four patients (3.5%). All four BAP1-positive patients were identified to have a frameshift variant (Table 1). No other pathogenic variants were identified in the other genes tested. Variants of uncertain significance were identified in 15 patients (13.2%) in the POT1 gene (1/114), the BAP1 gene (1/114), the MC1R gene (12/74), and the BRCA2 gene (1/40). The genetic testing panel offered by the genetics laboratory did transition into a new panel as of April 2021, in which the MC1R gene was no longer analyzed, and two other genes, BRCA2 and PTEN were included. Therefore, the frequency of variants of uncertain significance for the BRCA2 gene and the MC1R gene is limited to those subsets of patients tested for these genes.
Patient UM-002 was diagnosed at age 47 and passed away at age 50 from metastatic disease. She was initially treated with plaque brachytherapy. Patient UM-019 was diagnosed at age 31 and is still living. He was treated with external beam radiation. Patient UM-026 was diagnosed at age 40 and is also still living. He was treated with plaque brachytherapy. Patient UM-109 was diagnosed at age 66 is also still living. She was treated with enucleation.
The mean age at diagnosis in the germline BAP1 mutation group was 46.0 (SD = 14.9) years as compared to 57.4 (SD = 15.2) years in the control group (p = 0.14). Patients with a germline BAP1 mutation were not significantly more likely to have a prior diagnosis of another malignancy (p = 1). None of the patients with BAP1 germline mutations had a history of another primary cancer, whereas 17 patients who tested negative had a prior cancer history. Clinical details of the examined patients are indicated in Table 2.
Germline BAP1 mutation status was associated with a family history of malignant mesothelioma (MMe) (p = .00015) ( Table 3). Three of the four patients with BAP1 pathogenic variants reported a family history of other BAP1-associated malignancies (Table 1), however this was not statistically significant. Germline BAP1 mutation status was also associated with metastatic disease (p = .017) (Table 4). Otherwise, there were no statistically significant associations with demographic or clinical tumour characteristics.
BAP1 tumour (somatic) mutation testing was conducted on a subset of 10 patients either on a biopsy of metastatic disease or on tumour tissue from enucleation as part of clinical care. The incidence of a BAP1 germline mutation when a variant was detected on tumour tissue was 30% (3/10).
MC1R data is displayed in Table 5. The polymorphisms were seen at a higher frequency in the uveal melanoma study population than in the general population (gnomAD). These results were disclosed to patients by a genetic counsellor due to preference of the ordering oncologist and patient confusion when receiving a non-negative result. American College of Medical Genetics classification for Mendelian disease variants were not applied by the genetics laboratory as some variants within this gene have been associated with increased risk for melanoma, but with low penetrance. This data is therefore limited as the variant reporting changed significantly over the course of the study time period. As different criteria were applied at different times for these variants, not all variants were reported. A complete dataset is therefore not available to conduct appropriate statistical analysis.

Discussion
In this study, the detection rate of germline BAP1 pathogenic variants was 3.5% in an unselected UM population, as part of a clinically integrated model, which is consistent with rates from the previous literature ranging from 1-2% (9-11). Our data also supports observations from other research groups showing that UM in BAP1-TPDS is more aggressive and carries a higher risk for metastasis compared to UM in BAP1 negative patients (49,50). Median age of UM diagnosis in the literature in persons with BAP1-TPDS (53 years) is younger than UM in the general population (62 years), which is also a trend demonstrated in our population of 43.5 years and 58.5 years, respectively (51). Our full UM cohort displays equal distribution by gender and a vast representation of this sample   Ethnicity data was only available for 65 patients from the full sample. Total percentage is greater than 100 due to inclusion of both maternal and paternal ancestry. reports Caucasian ethnicity following the population trend (52). Interestingly, BAP1 germline mutation status was associated with a family history of MMe but not with a family history of the other BAP1-associated tumours. Previous research varies with which tumours of the BAP1 spectrum have significantly higher rates in families with BAP1 mutations, with some literature showing association with MMe and RCC (13), while others just with UM and cutaneous melanoma (10). While mesotheliomas are the second most frequent BAP1-TPDS-related cancer, specifically inquiring about the family history of BAP1-associated cancers can be a quick screening strategy towards patients with a new diagnosis of UM by clinicians discussing genetic testing. Positive germline BAP1 genetic results, whether detected in UM patients or in other patients with cancers related to BAP1-TPDS, have important implications for screening, prognosis, and overall life expectancy. In our study, patients who tested BAP1-positive were offered genetic counselling with a genetic counsellor in the genetics clinic, and were provided with appropriate screening recommendations and specialty referrals according to current guidelines ( Figure 1) (51). When applicable, carriers were referred to dermatology for surveillance every 6-12 months, a renal specialist for annual abdominal imaging, a mesothelioma specialist for annual chest imaging, and to ocular oncology for annual ophthalmic surveillance. In addition, patients were provided with their genetic testing report, a family letter, and our clinic referral form, being encouraged to share their genetic testing results with relatives to help arrange cascade testing.
For the individuals identified to have variants of uncertain significance (VUS), patients were offered a genetic counselling appointment to review the results. The majority of VUS were identified in the MC1R gene (12/15), which has been recently removed from our genetic testing panels due to limited clinical utility (22). Polymorphisms in the MC1R gene are extremely common with greater than half of the general population carrying one or more variants (53). Many variants demonstrate reduced penetrance and evidence is controversial for association with cutaneous melanoma (35)(36)(37)53). Mainstreaming models differ across genetics centres in whether patients with a VUS are seen for genetic counselling or reviewed by e-consult; our process involved contacting these patients as per the preference of the ordering oncologists. This also provided a central point of contact for patients to stay in touch with, regarding variant reclassification updates in keeping with shared responsibility for re-contact (38).
No other pathogenic variants were identified in the other genes included on the hereditary melanoma panel for the 114 UM patients. However, as noted that the genetic testing panel offered by the genetics laboratory transitioned in April 2021 as a result of changes in governmental funding of hereditary cancer testing, not all patients were tested for the same subset of genes. In particular, 74 patients with the initial panel were not tested for BRCA2, which has been reported in several patients with UM (14). As genetic testing continues to become more accessible and more patients have comprehensive testing, further research may identify additional hereditary cancer genes involved in the pathogenesis of uveal melanoma. Although BAP1 tumour mutation results were only available for a small subset of patients, the germline yield for BAP1 was considered high at 30%. The three germline BAP1 patients who had tumour testing completed for BAP1 (UM-002, UM-019, UM-109) all showed concordant tumour-normal results. The fourth germline BAP1 patient did not have the BAP1 gene included on the tumour testing panel, due to variability with the tumour testing panels offered and the quality of DNA available for testing. Approximately 40% of all UM patients and greater than 80% of metastatic UM patients have a BAP1 tumour mutation identified in the literature (39,40). Comparatively, our results matched the previously published European Society of Medical Oncology (ESMO) guidelines suggesting a > 10% germline conversion rate when BAP1 mutations are detected in the on-tumour context. The importance of this recommendation relies on the fact that it provides support for offering germline testing to confirm if somatic or germline (41).
Mainstreaming has been described in detail for patients with breast and ovarian cancer and is shown to be effective, efficient, and patient-centred (26)(27)(28)(29). This model was acceptable by patients, with high satisfaction regarding genetic care and the method of receiving results (26)(27)(28)(29)42,43). McCuaig et al. reported differences with lower median knowledge in the mainstreaming patient population compared to the traditional route; however, the overall mainstreaming cohort agreed that the information provided about genetic testing was helpful, and 74% agreed that the process worked well (29). Wait time for patients to receive genetic testing results is also significantly decreased with mainstreaming and studies show a reduction of 128-212 days between oncologist-initiated genetic testing versus traditional genetic testing (28,44,45,54).
Our study further exemplifies the success of mainstreaming UM, which is not a commonly referred patient population in genetics clinics (20). Despite the limited genetics research available compared to other hereditary cancer sites, oncologists were able to safely order germline genetic testing for all UM patients. In agreement with Sculco et al., less stringent genetic testing criteria may allow for the identification of BAP1 carriers from non-typical BAP1-TPDS families or when patients may not know their family history (e.g. UM-019) (55). Although cost-effectiveness was not assessed in our study, previous research in the breast and ovarian cancer population shows appreciable cost-effectiveness per discounted quality-adjusted life years (55). Despite higher testing volumes, reduced genetics appointment volumes in combination with identification of affected individuals to facilitate cascade testing and preventative measures can save overall healthcare expenditures. (26,55) Overall, this mainstreaming model seems to be useful for ocular oncologists and could be implemented for UM at other centres to offer genetic testing for all uveal melanoma casesCancer hospitals with embedded genetics clinics have the advantage to establish relationships for these mainstreaming models; however, other ocular oncology clinics or underserved areas independent from genetics clinics may also be able to integrate similar models through partnerships with expert centres. Furthermore, the BAP1-TPDS hub and spokes model for mainstreaming genetic testing could be expanded to include oncologist-ordering within the other spoke specialist sites.