An Oncology Provider-Driven Model to Implement Germline Genetic Testing Among Diverse Patients With Advanced Prostate Cancer

J Clin Pathways. 2025;11(1):20-26. doi:10.25270/jcp.2025.11.02

Prostate cancer (PCa) remains one of the most prevalent cancers across the globe and accounts for 14% of new cancer diagnoses annually.^1-3 Genomic testing, through multigene panels, is now the standard of care for patients with high-risk or advanced PCa because of its familial and therapeutic consequences.^4-6 Previous research demonstrates that approximately 12% to 17% of patients with advanced PCa who undergo germline genetic testing harbor germline pathogenic or likely pathogenic (P/LP) variants within many relevant homologous recombination repair (HRR) genes.^7-9 The most common identified mutations of DNA-damage repair genes seen in patients with PCa include BRCA2, BRCA1, ATM, CHEK2, PALB2, and mismatch repair genes (eg, MLH1, MSH2, PMS2, and MSH6). Based on these data, the National Comprehensive Cancer Network (NCCN) and American Urological Association guidelines recommend germline genetic testing in all high-risk and advanced patients with PCa, regardless of age or family history, as well as somatic tumor testing in all patients with advanced PCa.^2,4 Despite these recommendations having been implemented in 2018, comprehensive genetic testing continues to be underutilized, with estimates that only 5% to 30% of eligible patients receive appropriate testing due to various systematic barriers, including, but not limited to, the burden of appointments, cost, and transportation.^10-12

It is well established that significant differences in PCa incidence and mortality contribute to disparate outcomes for men from racial and ethnic minoritized groups.^1,11,13 Prior studies also reveal that Black and Hispanic men with PCa are less likely to complete genetic testing, and, when they do undergo testing, they are found to have lower rates of pathogenic variants detected compared with non-Hispanic White men.^6-8 Inequitable access to germline and somatic testing likely exacerbates and perpetuates racial and ethnic disparity in PCa outcomes, particularly as more targeted therapies become available that require identification of genomic alterations. Recent innovations in precision medicine, such as the approval of poly (ADPribose) polymerase (PARP) inhibitors for patients with PCa who harbor a germline or somatic mutation in DNA-damage repair genes, underscore the importance of genetic testing to guide clinical management of PCa.^5,14-16 It is notable that the pivotal phase 3 trials leading to PARP inhibitor approvals lacked adequate representation of men from racial and ethnic minoritized groups. For example, in the study that led to the approval of olaparib as monotherapy, only 3% of participants self-identified as Black or African American.^5,11,14,15 Our study sought to increase the uptake of guideline-concordant germline genetic testing among patients with PCa in our diverse patient population through the development of a streamlined genetic-care delivery model in a safety-net hospital.

A multidisciplinary team comprising genitourinary medical oncologists, advanced practice registered nurses, and genetic counselors collaborated to develop a tailored pre- and post-test decision-making protocol for patients with PCa undergoing germline testing. Germline genetic testing was designed to be performed concurrently with somatic tissue testing when available. The protocol included the purpose and benefits of testing and discussed the heritable and therapeutic consequences of results; it also offered an opportunity to discuss results at an advanced level with certified genetic counselors, as illustrated in Figure 1. Oncology providers following patients with PCa for their routine oncologic care screened patients at follow-up appointments to identify those who qualified for germline and somatic testing based on national guidelines.⁴ Germline genetic testing was performed using the Tempus xG prostate panel that was powered by GeneDx, which comprised 52 genes, as outlined in the Supplemental Table.¹⁷ Pre-test education for germline genetic testing was incorporated into standard-of-care counseling for somatic tumor genomic testing that is routinely performed to personalize therapeutics in the advanced PCa stage. Once informed consent was completed, a blood specimen was collected and forwarded to a singular commercial laboratory for analysis. Concurrently, patients completed a financial assistance form through Tempus Labs to assess eligibility for complete financial coverage for germline and somatic testing.

Oncology providers delivered post-test counseling for all patients. All men with germline P/LP variant results were referred to certified genetic counselors for comprehensive posttest genetic counseling. For individuals found to have variants of uncertain significance (VUS) or negative results, these results were disclosed during routine follow-up visits for their oncologic care. For patients with negative results or VUS, all patients were offered the opportunity to discuss their results further with genetic counselors.

We conducted quarterly audits that included collecting sociodemographic and clinical information from EMR. The race category was self-reported and included the following: Black/African American (henceforth referred to as Black), White, Pacific Islander, Chinese, Other, and Not Reported. The ethnic category was also self-reported and included the following: Hispanic, non-Hispanic, and Not Reported. Electronic record review could not differentiate between non-Hispanic White and Hispanic White patients, as these data were not collected through self-reporting. Multiple study personnel reviewed and cross-checked the data through manual chart review metrics. Retrospective data collection and analysis were approved by the University of Illinois at Chicago IRB (STUDY2023-0664).

Data analysis was done using Analytics Software & Solutions (SAS). Descriptive data analysis was used to describe the mean age, standard deviation (SD), and range, and to describe frequency distributions for all categorical variables. Chi-square tests were conducted for categorical variables to determine the association between various predictors and the presence of VUS and VUS numbers. Multiple logistic regression analyses were performed to explore the effects of various predictors on the likelihood of multiple VUS occurrences and VUS presence. The different predictors were included as independent variables, and the presence or number of VUS was included as a dependent variable in the respective models. Odds ratios and confidence intervals were calculated to quantify the strength of associations.

Between August 28, 2021, and December 28, 2023, a total of 132 patients with PCa underwent germline testing at the University of Illinois at Chicago, with 100% of tests resulted within the study timeframe. Each patient underwent germline genetic testing using the Tempus xG 52 gene prostate panel powered by GeneDx.¹⁷ The demographics of our patient cohort were diverse, as shown in Table 1. Of the 132 patients, 91 (68.9%) self-identified as Black, 32 (24.2%) as White, 27 (20.5%) as Hispanic, and 102 (77.3%) as non-Hispanic. The mean age at diagnosis was 64.1 years old (± 8.1). There were 38 (28.8%) patients who reported a positive family history of PCa, whereas 48 (36.4%) reported a positive family history of other cancers. Patients had various types of insurance coverage, with 67 (50.8%) patients having Medicare and 47 (35.6%) having Medicaid (Table 1). At the time of testing, 61 (46.2%) patients had an advanced disease diagnosis (stage IV) according to NCCN guidelines, and 108 (81.8%) were considered advanced disease (stage IV) at the time of paired germline/somatic testing. Of patients who self-identified as Black, 48 men (52.7%) had advanced-stage disease at diagnosis, and 78 men (85.7%) had advanced-stage disease (stage IV) at the time of genetic testing.

Most patients had negative test results (n = 57, 43.2%) as shown in Table 2. A smaller subset of patients was found to have P/LP germline variants (n = 11, 8.3%), as shown in Figure 2. Among individuals with P/LP germline variants, the incidence of actionable HRR alterations was lower than previously reported in the literature, accounting for 4.7% of those tested (n = 6). Of patients found to have HRR alterations, three patients were found to have CHEK2 mutations, two had ATM mutations, and one had a BRCA2 mutation. A notably low rate of alterations, 6.3% (n = 8), was found in autosomal dominant genes, including MSH2 and PMS2, in combination with actionable HRR alterations listed. Other monoallelic pathogenic variants occurred in the autosomal recessive genes FANCC and MUTYH, accounting for 2.3% of the total study population (n = 3). About half of the patients (n = 76, 57.6%) were found to have a VUS, the majority of whom self-identified as Black (n = 55, 60.4%). The most frequently identified VUS were in FANCM (n = 12), POLE (n = 9), and ATM (n = 9), of which eight individuals also reported a positive family history of cancer.

There were no statistically significant associations found between the presence of VUS, stage of disease at the time of testing, or family history. When comparing VUS rates according to race and ethnicity, VUS were modestly more frequent in patients who self-identified as Black (68.9% of the total study population) with 60.4% (n = 55) compared with 56.3% (n = 18) of White men and 51.9% (n = 14) of Hispanic men. The unadjusted odds ratio (OR) for detecting a VUS in Black participants compared with non-Hispanic White participants (the referent group) was 1.19 (95% CI, 0.53-2.69). In multivariable models adjusted for other factors, this association was also not found to be significant (OR, 1.44; 95% CI, 0.28-7.49). Black participants were more likely to have multiple VUS present (OR, 2.00; 95% CI, 0.58-6.91). However, there was no association between race and multiple VUS in multivariable models adjusted for all other variables (OR, 0.28; 95% CI, 0.01-5.86). Of the patients found to have pathogenic variants, 100% of patients (n = 11) had follow-up care with genetic counselors to discuss and interpret their results.

This study demonstrates that providing guideline-concurrent germline testing with streamlined pretest counseling delivered by medical oncology providers as part of standard oncologic care for patients with PCa, in conjunction with somatic tumor testing, is feasible and effective. Comprehensive genetics consults with dedicated genetic specialists were reserved for the subset of patients with P/LP germline variants and represented an efficient and sustainable way to provide guideline-directed genetic services for patients with PCa. Establishing a collaborative oncology directed model helped to address barriers to patients receiving timely testing and relevant treatment options. In our diverse patient population, this model proved effective at providing guideline-concordant genetic testing.

Despite national guideline recommendations, routine germline genetic testing across eligible patients with PCa is not consistently implemented.^9,10,14,15 Many systemic barriers prevent patients from receiving germline testing promptly, including, but not limited to, the burden of appointments, cost, and transportation, as well as provider knowledge of genetic testing and insufficient genetics workforce that may limit timely referral to genetic counselors.^18-20 Thus, there is a need to find innovative solutions to address barriers to receiving recommended genetic testing. Increased wait times disproportionately impact racial and ethnic minoritized patients, those in rural areas, as well as those from socioeconomically disadvantaged backgrounds.^1,13

We developed a sustainable model implementing guidelinedirected testing for our patients with advanced PCa to adapt to the increasing volume of patients with PCa. Recent alternative models incorporate telehealth or streamlined pretest education and mainstream genetic testing performed by oncology providers, with genetic counselors conducting post-test counseling.^18,21-23 However, many patients in our study population faced barriers to reliable phone and internet connections for telehealth visits. Moreover, the process involves navigating through a cumbersome multistep process, leading to decreased rates of test completion. In our model, oncology providers discussed results with the 91.6% (n = 121) of tested patients with negative and clinically negative results, while those individuals with P/LP results were referred for genetic counseling. By integrating a streamlined pretest education process, point-of-care genetic testing, and a comprehensive post-test counseling strategy into the patient’s oncology visits, we addressed several systemic barriers to genetic testing. By condensing testing into fewer visits, this approach helped reduce the burden of additional visits and alleviated potential transportation concerns.

Our study’s P/LP variant detection rate was 8.3%, lower than previous reports of 12% to 17%.6-8 Only 4.6% of the P/LP variants identified were in relevant HRR genes and thus have relevant treatment implications for their PCa in the form of PARP inhibitor therapy.^5,14-16 The lower incidence of actionable alterations in this diverse patient population reduces the number of patients eligible to receive targeted therapy options, potentially exacerbating disparities in racial and ethnic minority groups in a treatment landscape that is increasingly focused on personalized treatment strategies.

Similar to prior studies, we saw frequent VUS rates in our diverse PCa patient population.^8,13 Our data demonstrated that Black and White men had similar rates of VUS presence (60.4% vs 56.3%, respectively). Although the odds of having multiple VUS were higher for Black participants in the univariable analysis, the association between race and multiple VUS was not significant after adjustment for relevant covariates. These results illustrate variations in the prevalence of VUS among different racial groups and highlight the importance of including information regarding VUS reclassification in pre- and post-test germline genetic counseling. Currently, there is no clinical utility of VUS results, and they are designated as clinically negative. It is important to note that approximately 80% of VUS will be reclassified as benign or likely benign after approximately 1 to 10 years of data.^24,25 Thus, while most will be reclassified as negative, there is a proportion of patients whose results will be reclassified as positive and may have a clinical impact on their risk or treatment, or both. Thus, a systematic strategy must be implemented to track and share results with patients as they are reclassified.

Reports of higher rates of VUS are likely due to the underrepresentation of individuals from racial and ethnic minoritized backgrounds in genetic testing cohorts and the large number of genes included for testing with contemporary multigene panel tests. However, this underscores the critical need for systematic delivery of genetic testing to improve equity in genetic testing and diversity in genomic research, which should lead to improved variant classification.^8,24,26 The frequency of actionable mutations in our population was reduced, as has been seen in other reports of the literature, and reflects limitations of current germline testing and bioinformatics platforms and the need for greater inclusion of diverse populations in future studies.^8,11,26-28

Our study had several limitations. This retrospective study examined clinical practice in a safety-net hospital, presenting the possibility of selection bias and reliance on preexisting and possibly confounding data. The retrospective and observational nature of the study limits our ability to establish causality. The use of a commercial laboratory, as it provided financial coverage for testing, may also limit the generalizability of this analysis.

As the treatment landscape rapidly evolves in the era of precision medicine, it is crucial to improve health equity for racial and ethnic minority groups by expanding access to genetic testing and addressing its familial and therapeutic implications.

This article has supplementary material, which can be accessed here.

Affiliations:

¹University of Illinois Cancer Center, Chicago, IL; ²University of Illinois Health Division of Hematology/Oncology, Department of Medicine, University of Illinois College of Medicine and University of Illinois Health, Chicago, IL; ³University of Chicago Health Division of Academic Internal Medicine, Department of Medicine, University of Illinois, Chicago

Correspondence:

Natalie Reizine, MD

840 South Wood St, Suite 1020N M/C 787

Chicago, IL 60612

Email: nreizi2@uic.edu

Disclosures:

K.H. has received nonfinancial research support from Agendia and research support to the institution from Genentech/Roche, Merck, Novartis, and Pfizer, all outside of the submitted work. K.T. has served on an advisory board for Pfizer and Astellas outside of the submitted work. N.R. has served on advisory boards for Exelixis, Janssen, and Sanofi, and received compensation from

AstraZeneca, EMD Merck, Serono, and Tempus outside the submitted work. All other authors have no disclosures to report.