Optimizing Mainstreaming of Genetic Testing in Parallel With Ovarian and Endometrial Cancer Tumor Testing: How Do We Maximize Our Impact?

Abstract

PURPOSE

Although germline genetic testing (GT) is recommended for all patients with ovarian cancer (OC) and some patients with endometrial cancer (EC), uptake remains low with multiple barriers. Our center performs GT in parallel with somatic testing via a targeted sequencing assay (MSK-IMPACT) and initiates testing in oncology clinics (mainstreaming). We sought to optimize our GT processes for OC/EC.

METHODS

We performed a quality improvement study to evaluate our GT processes within gynecologic surgery/medical oncology clinics. All eligible patients with newly diagnosed OC/EC were identified for GT and tracked in a REDCap database. Clinical data and GT rates were collected by the study team, who reviewed data for qualitative themes.

RESULTS

From February 2023 to April 2023, we identified 116 patients with newly diagnosed OC (n = 57) and EC (n = 59). Patients were mostly White (62%); English was the preferred language for 90%. GT was performed in 52 (91%) patients with OC (seven external, 45 MSK-IMPACT) and in 44 (75%) patients with EC (three external, 41 MSK-IMPACT). GT results were available within 3 months for 100% and 95% of patients with OC and EC, respectively. Reasons for not undergoing GT included being missed by the clinical team where there was no record that GT was recommended, feeling overwhelmed, financial and privacy concerns, and language barriers. In qualitative review, we found that resources were concentrated in the initial visit with little follow-up to encourage GT at subsequent points of care.

CONCLUSION

A mainstreaming approach that couples somatic and germline GT resulted in high testing rates in OC/EC; however, barriers were identified. Processes that encourage GT at multiple care points and allow self-directed, multilingual digital consenting should be piloted.

INTRODUCTION

Although germline genetic testing (GT) is recommended for all patients with ovarian cancer (OC) and select patients with endometrial cancer (EC), testing rates remain low with marked inequities.^1,2 In OC, GT is universally recommended as germline pathogenic variants (PV) in BRCA1/2 and other genes involved in homologous recombination can help predict benefit for targeted therapies including poly (ADP-ribose) polymerase (PARP) inhibitors and inform at-risk relatives of their cancer risks to facilitate risk-reducing measures/interventions.^3,4 In EC, GT has been recommended for those with mismatch repair deficient (MMR-D) tumors with various algorithms proposed that incorporate methylation testing and microsatellite instability (MSI) assessments.^5-7 However, there are growing data to support universal germline assessment in all patients with EC given high rates of germline PV and actionability for patients and families.^8-10 As a result, national guidelines are evolving with broadening recommendations for GT in EC.¹¹

CONTEXT

• Key Objective

• We performed a quality improvement project to optimize our genetic testing (GT) processes in gynecologic cancer clinics, which uses our paired tumor-normal sequencing panel (MSK-IMPACT) in a mainstreaming approach.

• Knowledge Generated

• From February 2023 to April 2023 during our project period, we observed high rates of GT uptake in ovarian cancer (91%, 52/57 patients) and endometrial cancer (75%, 44/59 patients). However, a portion of patients were missed by the medical team with no documentation of GT being recommended, and many patients who declined recommendations for GT were overwhelmed or had other concerns.

• Relevance

• Our study identified potential barriers to GT and pain points in our institutional processes and supports novel, patient-driven, digital processes for GT that can be performed independently throughout the oncology care timeline.

Multiple barriers to GT have been described,^2,3 and efforts are needed to improve the efficiency of testing processes. Our center performs germline GT for patients with gynecologic cancers in parallel with tumor somatic testing via a targeted sequencing assay (MSK-IMPACT). This process is initiated with the primary oncologist (mainstreaming) rather than referring patients to a separate clinical genetics service. We previously reported on our experiences with mainstreaming in OC and observed an increase in GT rates with the implementation of this process.¹² However, we also identified a small proportion of patients who did not undergo GT and elucidated several barriers. As a result, we sought to study our processes for GT via MSK-IMPACT in gynecologic oncology clinics to optimize our workflows and maximize uptake.

METHODS

Quality Improvement Project and Patient Tracking

We performed a quality improvement (QI) study to evaluate our GT processes within our gynecologic oncology and medical oncology clinics using the Plan-Do-Study-Act method (Fig 1). The study team (Y.L.L., T.Y.S., N.V., M.W. and M.B.) planned the project and trained all the research coordinators in clinic. Although all clinicians underwent standardized training for MSK-IMPACT GT, research coordinators in clinic underwent a dedicated training session reviewing the importance of GT and logistics involved in the MSK-IMPACT process¹³ before initiating this project.

FIG 1.

PDSA model for QI project. This figure describes the PDSA model that served as the framework of the QI project to study the MSK-IMPACT genetic testing process in our GYN oncology clinics. EC, endometrial cancer; GYN, gynecologic; OC, ovarian cancer; PDSA, Plan-Do-Study-Act; QI, quality improvement.

From February 2023 to April 2023, all eligible patients with newly diagnosed OC/EC seen at our institution were identified by the research/clinical teams by prescreening clinic lists. These patients were then flagged electronically in preclinic preparations to be recommended GT by the oncologist. If GT was recommended, interested patients were shown an educational video discussing the importance of germline assessment and the testing process. Patients would then consent to GT after which the order would be placed and a blood sample would be collected. Consent forms and testing information were available in multiple languages, and in-person translators were also available to assist the clinical team. Once the consent process was complete, the GT order was reviewed and released by research supervisors. Germline testing would then occur concurrently with tumor testing (biopsy and surgical samples) in our molecular laboratories. Once results were available, the ordering oncologist would be notified, and the results would also be disclosed directly via the patient portal. Those with a pathogenic germline finding are tracked and referred to our genetics clinic via an expedited pathway.¹⁴

All patients who met inclusion criteria and qualified for GT during the study period were tracked in a REDCap database by the research coordinators. Clinical information and data on testing rates and logistics were abstracted by the research coordinators and study team. For patients recommended GT who did not consent in clinic, research coordinators were asked to select a single, main reason on the basis of patient and provider feedback. Reasons were defined together with the study team on the basis of previous experience and included (1) patients who declined oncologist's recommendations for GT, (2) patients feeling overwhelmed and not wanting to discuss GT, (3) patients with privacy concerns, (4) patients with financial concerns, and (5) language barriers that impaired discussion of GT. The study team then had two members review charts of all patients who did not undergo GT to confirm the above and identify patients who were potentially missed, defined as those without record of being recommended GT by the medical team despite meeting inclusion criteria in our study. The study team met regularly to review and analyze data qualitatively for recurrent themes. Data entry stopped on July 2023 to allow 3 months of follow-up for all patients. This project was approved as a QI initiative under MSK IRB Protocol 12-245.

Inclusion Criteria for Genetic Testing

Germline GT rates were evaluated according to our institutional practices, which follow national guidelines that recommend universal GT in OC and a testing algorithm in EC as previously reported⁷ to evaluate for Lynch Syndrome. Additionally, beginning in 2016, our institution has chosen to uniformly offer GT to all patients with EC treated at our center, irrespective of tumor MMR/MSI status and family history. Using these criteria, we included patients with newly diagnosed, pathologically confirmed EC or epithelial OC who were primarily receiving treatment at our institution's main campus. We excluded patients with other pathologic diagnoses (ie, benign, borderline, or nonepithelial cancers), patients only seen for a one-time visit who had treatment elsewhere, and those with recurrent disease. We also excluded those primarily seen in our regional sites given different workflows for GT at those sites, which may introduce heterogeneity. Additionally, the training and QI project were based at our main Manhattan site.

Analysis

Clinical data were analyzed using summary statistics. Rates of GT completion via MSK-IMPACT were determined. In those who did not undergo germline assessment via MSK-IMPACT, reasons were defined as above and reviewed by the study team. Those who underwent GT before presenting to our institution were also noted. The study team met regularly to continually evaluate the project, review key findings, and identify areas for improvement with our multidisciplinary group.

RESULTS

Patient Characteristics

From February 2023 to April 2023, 182 patients were reviewed. We excluded 66 patients (26 for benign pathology, 36 who were only seen once at MSK and received their care elsewhere, two from regional sites only, and two with recurrent disease). The remaining 116 patients with newly diagnosed OC (n = 57) and EC (n = 59) were included in this study (Fig 2). The median age at diagnosis was 63 (range, 21-92) years overall; the median age was 64 (range, 23-92) years for OC and 60 (range, 21-84) years for EC (Table 1). For self-reported race, 13% of patients identified as African American (AA)/Black, 9% Asian, 2% Other, 62% White, and 14% had missing data. For OC, 7% of patients identified as AA/Black, and for EC, 19% of patients identified as AA/Black. Hispanic ethnicity was reported in 11% of patients overall, 14% in OC and 7% in EC. English was the preferred language for 90% of patients overall with Spanish (5%), Russian (3%), and Chinese (1%) being the next most common languages with similar rates in OC and EC (Table 1).

FIG 2.

Quality improvement project patient tracking and genetic testing initiation and completion rates. Figure depicts results of patient tracking for our quality improvement project and subsequent rates and timing of genetic testing and results for patients with newly diagnosed OC and EC treated in our GYN oncology clinics. EC, endometrial cancer; GYN, gynecologic; OC, ovarian cancer. *Includes any previous outside genetic testing; **one patient had no result as surgery was rescheduled. One patient had insufficient tumor tissue.

TABLE 1.

Patient Demographics

Clinical Characteristic	Overall	OC	EC
	N = 116	N = 57	N = 59
Age at diagnosis, mean, median (range)	61, 63 (21-92)	62, 64 (23-92)	60, 60 (21-84)
Race, No. (%)
African American/Black	15 (13)	4 (7)	11 (19)
Asian	9 (9)	6 (11)	3 (5)
White	71 (62)	38 (68)	33 (57)
Other	3 (2)	2 (3)	1 (2)
Missing/Unknown	16 (14)	6 (11)	10 (17)
Ethnicity, No. (%)
Hispanic	12 (11)	8 (14)	4 (7)
Not Hispanic	89 (77)	41 (73)	48 (81)
Missing/Unknown	14 (12)	7 (13)	7 (12)
Preferred language, No. (%)
English	104 (90)	50 (89)	54 (92)
Spanish	6 (5)	3 (5)	3 (5)
Russian	3 (3)	1 (2)	2 (3)
Chinese	1 (1)	1 (2)	0 (0)
Other	1 (1)	1 (2)	0 (0)

Abbreviations: EC, endometrial cancer; OC, ovarian cancer.

Genetic Testing Rates in OC

GT in the context of paired tumor-normal testing via MSK-IMPACT was performed in 45 (79%) of 57 patients with OC. Of the 12 patients with OC who did not undergo germline assessment via MSK-IMPACT, seven had previous external GT; thus, a total of 52 (91%) of 57 patients with OC had germline assessment overall (Fig 2). However, one patient was not consented due to language concerns, and four were missed by the medical team with no record that GT was discussed/recommended (Fig 3). Germline testing results were available within 3 months of consent for 100% of patients with OC (Fig 2).

FIG 3.

Reasons for not undergoing genetic testing. Figure depicts reasons for not having genetic testing via MSK-IMPACT in those with newly diagnosed ovarian and endometrial cancer treated in our gynecologic oncology clinics. For patients recommended genetic testing who did not consent in clinic, research coordinators were asked to select a reason on the basis of patient and provider feedback. Patients missed by the medical team were defined as those without record of being recommended genetic testing despite meeting inclusion criteria in our study.

Genetic Testing Rates in EC

GT in the context of paired tumor-normal testing was performed in 41 (69%) of 59 patients with EC. Of the 18 patients who did not undergo germline assessment via MSK-IMPACT, three had previous external GT, and a total of 44 (75%) of 59 patients with EC had germline assessment overall (Fig 2). The reasons for not consenting to germline testing included patients declining (n = 1), financial concerns (n = 2), privacy concerns (n = 1), and feeling overwhelmed (n = 5). For six patients, there was no record that GT was recommended by the clinical team, and these patients were considered missed by the medical team (Fig 3). Germline testing results were available within 3 months of consent for 95% of patients with EC with one patient having surgery delayed and another having insufficient tumor tissue for paired tumor-normal testing (Fig 2).

Notably, among the 18 EC tumors from the patients who declined germline assessment via MSK-IMPACT, four (22%) were MMR-D. Of these four MMR-D tumors, all had loss of MLH1 and/or PMS2 by immunohistochemistry. MLH1 hypermethylation (MLH1ph) was identified in one tumor, and this patient was reportedly missed by the medical team with no record of being recommended germline assessment. Another tumor was indeterminate on MLH1ph testing, and this patient did not consent to germline testing because she was reported as feeling overwhelmed by the oncologic discussion. However, two MMR-D ECs did not undergo MLH1ph testing, and these patients were both missed by the medical team for germline assessment with no record that GT was recommended (Data Supplement, Fig S1). In one patient, there was insufficient tumor tissue for methylation testing, and in the other patient, the oncology team misinterpreted the MMR immunohistochemistry results upon further review.

Genetic Testing Timeline

The median time from initial visit at our institution to consent for GT was 7 days (IQR, 0-28). The median time from consent to order being placed and a blood sample for normal control being collected was 0 days. However, time from order being placed to being released, which is performed by a research supervisor and is necessary to initiate the testing process in our laboratory, was 15 days (IQR, 7-26). This represented a bottleneck point in the GT process. The median time to germline result was 44 days (IQR, 37-59) and was similar to the median time for somatic results, 44.5 days (IQR, 37.5-57.5; Fig 4).

FIG 4.

Timeline for various steps of the genetic testing process via MSK-IMPACT. Figure depicts time in days from initial gynecologic oncology clinic visit and MSK-IMPACT consent for the various steps of the genetic testing process until somatic and germline results were resulted.

Qualitative Themes: Barriers to Genetic Testing

In qualitative review of reasons for not undergoing GT in those with OC/EC, the most common reason was patients being missed by the primary team with no record of germline assessment being discussed or recommended by the clinical team. In those offered GT who did not undergo it, common themes included patients feeling overwhelmed, financial and privacy concerns, and language barriers related to discussion of GT (Fig 3).

In review of our GT processes, we found that most resources dedicated to offering GT were concentrated at the initial visit where patients were often overwhelmed by their recent cancer diagnosis, with few attempts made to encourage GT at subsequent points of care. We also found inclusion of many patients with benign pathology in initial review of patients who were being considered for germline GT. As ultimately these patients did not qualify for GT, the inclusion of these patients led to inefficient allocation of resources (Fig 5).

FIG 5.

Identified areas for improvements in the GT process. Figure depicts the identified areas in need of improvement discussed by our multidisciplinary team and corresponding ideas for improvement to optimize the GT process in GYN oncology clinics. EC, endometrial cancer; GT, genetic testing; GYN, gynecologic; OC, ovarian cancer.

In review with our multidisciplinary group, we identified potential strategies to address the challenges of offering universal germline GT to OC and EC patients including leveraging multiple points for contact for GT, checklists of steps in the testing process, and better coordination between the surgical and medical teams. We also discussed ways to optimize our process including approaching patients who had not yet consented to GT at postoperative and follow-up visits and piloting digital processes that would allow patients to independently and remotely self-consent and schedule sample collection for GT (Fig 5).

DISCUSSION

A mainstreaming approach that couples germline and somatic GT was feasible and resulted in high testing rates for patients with OC and EC. However, barriers to testing were identified in our QI project, particularly in EC where Lynch Syndrome testing algorithms were not appropriately deployed in all patients. Processes that encourage GT at multiple points of care and allow self-directed, multilingual, digital consenting may improve uptake and should be investigated in future studies.

Our data add to the growing body of literature exploring QI projects to evaluate GT process and improve uptake. These studies have identified possible interventions including direct referrals and embedding of genetic counselors into medical oncology and surgical clinics as ways to improve GT capacity and decrease costs.^15-20 Many of these highlight mainstreaming as a strategy to improve GT uptake,²¹ and our study further supports mainstreaming efforts within oncology clinics.

Despite our high GT rates, we still encountered challenges. For half of the patients who did not undergo GT, there was no record that GT was discussed or recommended. The reason for this is unclear, and some clinicians may have chosen to defer discussion of GT given other oncologic issues. This should be investigated further in future studies with longer follow-up that capture the clinician and patient perspective. On qualitative review, we found that resources were concentrated at the initial, in-person oncology visit with little evidence of documented follow-up at subsequent visits, which may have contributed to missing patients. Additionally, there were few resources for patients to initiate the GT process independently and remotely. New digital processes that use telegenetics, artificial intelligence chatbots, and other digital resources to facilitate the GT processes have been investigated to help ameliorate this barrier.^22,23

The most common reason for declining GT in those recommended germline assessment was the patient being overwhelmed at the initial oncology visit. This pilot study did not report the patient perspective, and this needs to be investigated in future studies to fully understand the psychological impact of their recent cancer diagnosis. However, other studies have shown mainstreaming strategies to be acceptable to patients without long-term psychological toll.^24,25 Interestingly, Frey et al²⁵ found that although patient-reported stress and anxiety were high at the time of facilitated GT, which occurred around the time of initial diagnosis and gynecologic oncology visit for patients with OC, this improved by 6 months. This suggests that the psychosocial burden of GT may be lower in the months following a cancer diagnosis, and this may represent a better time to approach patients for GT. However, this needs to be balanced with the need for testing results to guide therapy, particularly for PARP inhibitor maintenance in OC.²⁶

Although most of our patients reported English as their preferred language, language was noted as a potential challenge for discussing GT in our qualitative review. Although MSK-IMPACT is available in many translated languages and in-person translators are present in clinic, the GT consenting process is more cumbersome and time-consuming when performed in another language and may be a barrier in busy clinics. Other studies have identified language as a potential challenge and disparity in GT for various cancer types.^27-30 Interventions focused on language and cultural literacy for specific populations, including use of culturally appropriate materials, are needed to expand GT in diverse populations.^31-33

We also found lower rates of GT in EC compared with OC, which may reflect differences in guidelines for germline assessment in these cancers, respectively. In OC, universal GT has been recommended by many national and international organizations since 2014.⁴ In contrast, for EC, GT is historically tied to complicated algorithms of MMR/MSI assessments and MLH1 promoter hypermethylation testing to determine the likelihood of Lynch Syndrome.^7,34 This can lead to misdiagnosis and misinterpretations of these data and inappropriate use of GT.³⁵ Additionally, billing for GT in OC is independent of tumor testing given universal recommendations for germline GT, whereas compensation of EC GT is tied to tumor MSI testing, which may influence germline testing uptake by providers and patients. Although we found that more patients with EC were missed for GT compared with OC, this may simply reflect the importance of GT to the cancer type and providers needing to prioritize their testing given the higher prevalence of germline PV and broader treatment implications in OC. As more studies highlight the potential clinical importance of GT in EC care and germline findings outside of Lynch Syndrome,^8-10 recommendations for GT in EC may broaden. Indeed, in 2024, updated National Comprehensive Cancer Network guidelines now recommend consideration of GT in all patients with EC.¹¹

Finally, this QI project identified several pain points and bottlenecks in GT processes, including a delay in releasing orders due to shortage in personnel reviewing GT orders, which resulted in prolonged turnaround times. These data were presented to our greater multidisciplinary group, and new interventions and processes are being proposed and studied to help optimize GT and timelines.

The limitations of the current study include small sample size, limited follow-up time, and a focus on a very specialized workflow of paired tumor-normal (MSK-IMPACT) sequencing that is used in our Main Campus and may not be generalizable to all oncology clinics. Our workflow uses a single in-house test, whereas many other practices use multiple commercial vendors, each with their own workflow, which may result in different barriers. The diversity of our population was also limited in respect to race/ethnicity and language, and we lack data on other social determinants of health including insurance status and income. Thus, additional barriers and disparities may exist in certain populations that could not be captured in the current study,³⁶ which should be studied further in larger multi-institutional studies. Additionally, we did not formally collect data on the physician perspective or the patient experience, and future studies should incorporate surveys and measures of psychosocial distress and mistrust.³⁷

In conclusion, we demonstrate very high germline GT rates using a mainstreaming model of paired tumor-normal sequencing in our gynecologic oncology clinics for patients with OC and EC. Our QI project has identified important themes to further optimize this process, which should be studied with future interventions.

AUTHOR CONTRIBUTIONS

Conception and design: Ying L. Liu, Tiffany Y. Sia, Nancy Varice, Maureen Byrne, Jesse Galle, Carol Brown, Kara Long Roche, Dennis Chi, David B. Solit

Financial support: David B. Solit

Administrative support: Ying L. Liu, Nancy Varice, Jesse Galle, David B. Solit

Provision of study materials or patients: Ying L. Liu, Nancy Varice, Margaret Sheehan, Carol Brown, David B. Solit

Collection and assembly of data: Ying L. Liu, Tiffany Y. Sia, Nancy Varice, Michelle Wu, Maureen Byrne, Margaret Sheehan, Jesse Galle, David B. Solit

Data analysis and interpretation: All authors

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/po/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).