Advertisement

Annals of Surgical Oncology

, Volume 25, Issue 10, pp 2925–2931 | Cite as

Underdiagnosis of Hereditary Breast and Ovarian Cancer in Medicare Patients: Genetic Testing Criteria Miss the Mark

  • Shan Yang
  • Jennifer E. Axilbund
  • Erin O’Leary
  • Scott T. Michalski
  • Robbie Evans
  • Stephen E. Lincoln
  • Edward D. EsplinEmail author
  • Robert L. Nussbaum
Breast Oncology

Abstract

Background

An estimated 5–10% of breast and ovarian cancers are due to hereditary causes such as hereditary breast and ovarian cancer (HBOC) syndrome. Medicare, the third-party payer that covers 44 million patients in the United States, has implemented a set of clinical criteria to determine coverage for the testing of the BRCA1 and BRCA2 genes. These criteria, developed to identify carriers of BRCA1/2 variants, have not been evaluated in the panel testing era. This study investigated a series of Medicare patients undergoing genetic testing for HBOC to determine the efficacy of genetic testing criteria in identifying patients with hereditary risk.

Methods

This study retrospectively examined de-identified data from a consecutive series of Medicare patients undergoing genetic testing based on personal and family history of breast and gynecologic cancer. Ordering clinicians indicated whether patients did or did not meet established criteria for BRCA1/2 genetic testing. The genetic test results were compared between the group that met the criteria and the group that did not. Patients in families with known pathogenic (P) or likely pathogenic (LP) variants were excluded from the primary analysis.

Results

Among 4196 unique Medicare patients, the rate of P/LP variants for the patients who met the criteria for genetic testing was 10.5%, and for those who did not, the rate was 9% (p = 0.26).

Conclusions

The results of this study indicate that a substantial number of Medicare patients with clinically actionable genetic variants are being missed by current testing criteria and suggest the need for significant expansion and simplification of the testing criteria for HBOC.

Breast or ovarian cancer is diagnosed for more than 275,000 patients every year. An estimated 5–10% of these cancers are due to hereditary causes such as hereditary breast and ovarian cancer (HBOC) syndrome.1,2 Studies have estimated that less than 10% of all BRCA1 and BRCA2 carriers have been identified and informed.3 Moreover, 50–80% of individuals at risk for HBOC have not received appropriate genetic testing, in part because they do not meet the family history criteria of current testing guidelines4,5 and in part because insurance seldom reimburses testing for those outside these criteria. In addition to enhanced cancer screening, surgical treatment, and risk reduction, germline genetic results are increasingly relevant to systemic therapy.68

Medicare, the third-party payer covering 44 million patients in the United States, uses a set of clinical criteria to determine coverage for the testing of BRCA1/2. The policies of other insurance providers often mimic those of Medicare. The current Medicare BRCA1/2 genetic testing criteria require a personal diagnosis of cancer and, for many patients, a family history of cancer.9 However, the efficacy of these widely used clinical testing criteria has not been established.

Recent studies have demonstrated the clinical actionability associated with expanded multi-gene panels,10 but current testing criteria were developed to identify carriers of BRCA1/2 variants and have not been evaluated in the panel testing era. In a series of patients insured by Medicare and undergoing genetic testing, we retrospectively evaluated the efficacy of established genetic testing criteria in identifying patients with hereditary risk for breast and ovarian cancer.

Methods

We reviewed de-identified data under an institutional review board (IRB)-approved protocol for a consecutive series of Medicare patients whose genetic test indication was a personal or family history of breast or gynecologic cancer and whose test order included at least the BRCA1 and BRCA2 genes. In addition to the BRCA1 and BRCA2 genes, the number of hereditary cancer syndrome genes (the panel size) analyzed for each patient was at the discretion of the ordering clinician in accordance with Invitae’s flexible ordering policy, but generally ranged from 40 to 80 genes.

Ordering clinicians completed a brief checklist indicating whether patients did or did not meet criteria for BRCA1/2 genetic testing. These criteria were derived from the National Comprehensive Cancer Network (NCCN) guidelines but were slightly more inclusive than the Medicare criteria. Consistent with the Medicare criteria, a personal history of breast, ovarian, or a related cancer was required, even when the family had a known BRCA1/2 pathogenic variant or a strong history of these cancers. For patients not meeting the testing criteria, a detailed description of personal and/or family history of cancer was not required. Genetic test outcomes were compared for various sets of genes between the group that met the criteria and the group that did not.

Clinical reports were categorized as positive when results contained pathogenic (P) or likely pathogenic (LP) variants, as uncertain when results contained one or more variants of uncertain significance (VUS) without positive findings, and as negative when results contained findings of only benign or likely benign variants or no variants at all. Patients in families with known P/LP variants were excluded from the primary analysis.

Testing was performed with next-generation sequencing (NGS), as previously described,11 and variant interpretation was carried out based on an expansion of the American College of Medical Genetics and Genomics guidelines.12 We used Fisher’s’ exact test to calculate the p value of the difference between the group that met the criteria and the group that did not. The z test was used to calculate the difference in ordered panel size between the two groups.

Results

Among the 4196 unique Medicare patients in this cohort, 3549 (85%) met the established genetic testing criteria for coverage of genetic testing, and 647 (15%) did not. The age, gender, ethnicity, and personal and family history of cancer in the group that met the criteria and the group that did not are described in Table 1. The rate for the P/LP variants among all the genes ordered was 10.5% for the patients who met the criteria and 9% for those who did not (p = 0.26; Table 2). The rate for the P/LP variants among the 19 genes associated with elevated risk of breast cancer, ovarian cancer, or both, as designated by NCCN guidelines, was 8.4% for the patients who met criteria and 6.2% for those who did not (Table 2). When only the results from BRCA1 and BRCA2 testing were considered, the positive rate for the group that met the criteria was 1.75-fold that of the group that did not (3.2 vs. 1.9%), a difference that was not statistically significant (p < 0.05; Table 2).
Table 1

Demographics of the patients who met and those who did not meet the genetic testing criteria

Characteristic

Met criteria

Did not meet criteria

Gender

Female

3316

564

Male

233

83

Age

Average testing age

70

68

Age range

24–100

28–92

Ethnicity

White/Caucasian

2513

436

Ashkenazi Jewish

278

45

Black/African-American

206

28

Hispanic

121

32

Asian

76

16

Other, mixed, unknown

355

90

Cancer history

Personal history only

1354

189

Family history only

17

72

Personal and family history only

2178

386

Information on personal and family history of cancer was based on ICD-10 codes available for patients as provided by clinicians

Table 2

Positive genetic test results for in-criteria versus out-of-criteria Medicare patients

Positive result

BRCA1/2 alone (%)

Breast cancer management genes (%)

Breast or GYN cancer management genes (%)

Large HBOC panels (%)

In-criteria patients

3.2

6.9

8.4

10.5

Out-of-criteria patients

1.9

5.3

6.2

9.0

BRCA1/2 alone—results for testing only BRCA1 and BRCA2. Breast cancer management genes have recommendations for modifying breast cancer risk management. Breast or GYN cancer management genes have recommendations for modifying breast or gynecologic cancer risk management. Large HBOC panels: large hereditary breast and ovarian cancer panels including results for multi-gene panels averaging 34–37 genes

GYN gynecologic, HBOC hereditary breast and ovarian cancer

The VUS rates also did not differ significantly between the two groups. Compared with testing ordered by clinicians for the patients who did not meet the criteria, the clinician-ordered genetic testing for the patients who met the criteria included, on the average, slightly fewer genes (average panel size, 37 genes vs. 34 genes), but the difference was not significant (p = 0.47, one-tailed z test).

The Medicare patients in this cohort had the option to self-pay for testing, which obviated the need for their health care professionals to submit documentation verifying that they met the criteria. This exception could have introduced bias in the event that any of these patients met the testing criteria but were assigned to the group that did not because their health care professionals did not submit documentation that they met the criteria. However, review of the billing records allowed us to identify the self-paying patients. When these patients were excluded, the positive and negative result ratios were almost identical to those in the original analysis for both the group that met the criteria and the group that did not (Table 3).
Table 3

Positive and negative result rates excluding self-pay patients

Test result

In-criteria patients

Out-of-criteria patients

Total

In-criteria rate (%)

Out-of-criteria rate (%)

Positive

373

40

413

10.53

8.99

Negative

3170

405

3575

89.47

91.01

Negative: no pathogenic or likely pathogenic (P/LP) variant identified on genetic test result; positive: P/LP variant identified

The P/LP variants identified for both Medicare patient groups occurred in genes associated specifically with breast and ovarian cancer, as well as in known cancer genes traditionally associated with other hereditary cancer syndromes (Fig. 1). With few exceptions, the genes in which P/LP variants were identified have published management recommendations and are clinically actionable in the context of acute treatment for a patient’s cancer, post-treatment surveillance and prophylaxis, or cascade testing for at-risk family members (Table 4).
Fig. 1

Clinically actionable variants identified by gene, including BRCA1, BRCA2, and other genes associated with breast cancer (orange bracket); genes associated with breast and/or gynecologic (GYN) cancers (green bracket); and genes associated with other cancers (all remaining genes listed). The number of variants per gene is normalized to the total number of patients tested, by cohort

Table 4

Clinical actionability by gene

Gene

Management guidelines

Clinical trial eligibility for breast cancerc

ATM

Annual mammographya

Consider breast MRI, genetic counseling, and cascade FVT

NCT02401347

BRCA1

Annual breast MRIa

Consider RRM

Recommend RRSO, genetic counseling, and cascade FVT

NCT03286842 (treatment)

NCT03495609 (prevention)

BRCA2

Annual breast MRIa

Consider RRM

Recommend RRSO, genetic counseling, and cascade FVT

NCT03286842 (treatment)

NCT03495609 (prevention)

CHEK2

Annual mammographya

Consider breast MRI, Colonoscopy screening at 40 years, genetic counseling, and cascade FVT

NCT02401347

MSH6

Colonoscopy screening at 20 yearsb and upper endoscopy at 35 years

Consider RRHyst, RRSO, genetic counseling, and cascade FVT

NCT03495544

MUTYH

Consider colonoscopy at 40 yearsb per family history of CRC, genetic counseling, and cascade FVT

NCT03495544

RAD50

No current guidelines

NCT02401347

RAD51D

Recommend RRSO,a genetic counseling, and cascade FVT

NCT02401347

List of genes with variants in more than two Medicare patients and associated with established management guidelines. The clinical trials listed indicate a selection of actively recruiting trials for which patients with breast cancer may be eligible and in which a germline pathogenic/likely pathogenic variant in the specified gene is among the inclusion criteria for enrollment

MRI magnetic resonance imaging, FVT family variant testing, RRM risk-reducing mastectomy, RRSO risk-reducing salpingo oophorectomy, RRHyst risk-reducing hysterectomy, CRC colorectal cancer

aNational Comprehensive Cancer Network. Genetic/Familial High-Risk Assessment: Breast and Ovarian. Version 1.2018, accessed January 2018

bNational Comprehensive Cancer Network. Genetic/Familial High-Risk Assessment: Colorectal. Version 3.2017, accessed January 2018

cwww.clinicaltrials.gov, accessed April 2018

Discussion

The rate of P/LP variants was similar between the Medicare patients who did meet the criteria for BRCA1/2 genetic testing and those who did not, and the difference was not statistically significant. These data suggest that the current genetic testing criteria for HBOC are not significantly better at identifying individuals with increased risk for breast or ovarian cancer than a strategy that simply tests all patients with a personal history of breast or ovarian cancer.13 In fact, the results of our study suggest that such a strategy would increase by almost 90% the number of patients identified to have a clinically actionable genetic test result.

The rate of BRCA1/2 P/LP variants was lower among both groups of Medicare patients than the 6–9% rate estimates of previous reports.1416 This is not surprising because Medicare patients are older, on the average, than groups historically undergoing analysis of these two genes. It could be reasonably surmised that patients with the highest likelihood of a BRCA1/2 P/LP variant were ascertained and tested positive before becoming eligible for Medicare, and thus would not have been included in the cohorts of this study. Also, patients previously testing negative for P/LP variants in BRCA1/2 who currently are presenting for expanded gene testing would potentially reduce the observed prevalence of BRCA1/2 variants in this study. Finally, Medicare patients who are older at their diagnosis of breast cancer would be expected to have a lower incidence of variants in genes associated with young age of disease onset (i.e., BRCA1/2), further contributing to the BRCA1/2 variant rate observed in our study cohorts.

It should be noted that in contrast to the current NCCN guidelines, the Medicare criteria for HBOC genetic testing exclude coverage for unaffected individuals, even in the setting of a previously identified P/LP family variant in a close blood relative. We excluded unaffected individuals with known family variants from this analysis to avoid artifactual inflation of the positive result rate observed in the Medicare patients who did not meet the criteria because each of these individuals had up to a 50% chance of a P/LP genetic test result. However, the group that met the criteria did include a small number of patients with no previously identified pathogenic family variant whose affected status could not be confirmed by clinician-provided information. The P/LP rate among these individuals was not statistically different from the rate for the patients meeting the criteria whose clinician documentation confirmed them as being affected (12.5 vs. 8.5%; p = 0.27).

The current genetic testing criteria reflect the historical strategy of their original development. At a time when BRCA1/2 genetic testing cost $2000 to $5000 per test, the original criteria were designed to limit the population of cancer patients being tested to those with the most severe clinical presentation of highly penetrant BRCA1/2 variants. This approach was intended to maximize the yield of patients with positive test results and minimize the number of tests needed to find each positive case. However, advances in NGS technologies have dramatically reduced the cost of BRCA1/2 testing, and these NGS advances currently enable simultaneous sequencing and deletion/duplication testing of BRCA1/2 concomitantly with dozens of additional risk genes for breast cancer, ovarian cancer, or both (e.g., PALB2, PTEN, ATM, CHEK2), in some cases for as little as $250 per patient (https://www.ncbi.nlm.nih.gov/gtr/). Furthermore, the current criteria do not adequately account for the full range of clinical presentations described to date as associated with HBOC, and carriers of clinically actionable variants in genes other than BRCA1/2 are likely to fall outside the current criteria. For example, the P/LP variants in RAD51C, RAD51D, and BRIP1 are associated with increased risk for the development of ovarian cancer.

Because ovarian cancer is not amenable to early detection, which results in a high mortality rate, the NCCN guidelines recommend that women who are carriers of variants in these genes undergo risk-reducing salpingo-oophorectomy. However, these genes are not known to increase the risk for breast cancer notably. As such, the only indication of a P/LP variant in one of these genes is a family history of ovarian cancer. Because family history alone does not meet the current Medicare criteria for genetic testing, a female carrier must have a highly fatal malignancy diagnosed before becoming eligible for genetic testing.

Accordingly, a dramatic modification of the scope and intent of the existing genetic testing guidelines is critically overdue. In contrast to preventable diagnoses as the prerequisite for genetic analysis, the criteria must, at a minimum, be expanded to allow for genetic testing of unaffected, at-risk individuals, thereby alerting them to early detection and risk-reducing options. Additionally, the criteria should consider the varying penetrance and expressivity of clinically actionable genes beyond BRCA1/2 and acknowledge that a more moderate family history often is still indicative of an increased cancer risk necessitating enhanced surveillance.

The prevalence of CHEK2 variants (Fig. 1) in both the group that met the criteria (1.6%) and the group that did not (2.3%) was consistent with the 1.5–2.5% prevalence reported in other studies.10,17,18 Overall, the rate of CHEK2 variants in the group that did not meet the criteria was marginally higher than in the group that met the criteria. However, CHEK2-truncating variants (which can confer higher risk than missense variants) were a greater percentage of the CHEK2 findings (61%) in the group that met the criteria than in the group that did not (27%). In the group that did not meet the criteria, the total number of CHEK2 variants nominally exceeded the number of BRCA1 and BRCA2 variants observed. These findings suggest that individuals who do not meet the testing criteria lack clinical characteristics, such as a strong family history and a young age at diagnosis, which would be predictive of higher risk variants. The overall increased rate in CHEK2 is reflective of the milder clinical picture for the cohort of individuals not meeting the clinical criteria.

The estimated life expectancy for men and women currently 65 years of age is respectively 84.3 and 86.6 years (https://www.ssa.gov/planners/lifeexpectancy.html, accessed 28 May 2018). Thus, therapeutic options for mutation carriers with cancer notwithstanding, a large number of Medicare patients would presumably benefit from increased surveillance and risk-reducing measures. To illustrate this point, BRCA2 was the second most commonly involved gene among the patients who did not meet the criteria in this study. A 60-year-old woman with a BRCA2 P/LP variant has almost a 10% chance for the development of ovarian cancer by the age of 70 years.19 Therefore, assuming a life expectancy of 20+ years, risk-reducing bilateral salpingo-oophorectomy is an appropriate consideration. Additionally, although most Americans become insured under Medicare at 65 years of age, 17% (approximately 8 million subscribers) are currently listed as “under 65 and disabled” (https://www.medicareresources.org/basic-medicare-information/what-is-medicare/, accessed 28 May 2018), representing a population covered by Medicare for which aggressive surveillance has the potential to confer additional clinical benefit.

In an effort to lower the barriers to clinical genetic testing, the current study did not require a detailed personal and family history as part of the testing process. The information provided by ordering clinicians was variable and often limited for the patients who did not meet the criteria. Accordingly, a definitive and comprehensive clinical description of the group that did not meet the criteria was not feasible for our cohort, which was a limitation of this observational study. It might be surmised that the patients who did not meet the criteria had some characteristic or characteristics that raised suspicion (Table 1), and, as such, it is possible that trends exist among the group that did not meet the criteria with P/LP variants. Therefore, focused collection and analysis of these factors are recommended. In the meantime, our current data suggest that clinical judgment outperforms the existing testing criteria for identification of high-risk patients.

Our results suggest that identification of up to half of Medicare patients with actionable variants is not taking place because testing is restricted to patients meeting the current criteria and because testing strategies are limited to panels containing only BRCA1/2. This underdiagnosis of HBOC is likely occurring for patients of other insurance providers nationwide who follow the genetic testing criteria applied to Medicare patients. At a minimum, we propose that the testing criteria should be expanded to include unaffected Medicare patients with a family history in line with the NCCN guidelines. The testing criteria should also incorporate patients with a less severe phenotypic presentation, in keeping with moderate-risk genes, which still require intervention.

The current study suggests that BRCA1/2 testing of all patients with a personal history of breast or ovarian cancer would substantially increase the discovery of clinically actionable findings in Medicare patients. Further research is needed to determine whether the clinical benefits justify a strategy in which all patients with a personal history of breast or ovarian cancer are eligible to undergo genetic testing for BRCA1/2 (i.e., elimination of additional testing criteria altogether), including multi-gene panel testing of additional breast and ovarian cancer risk genes, to mitigate the unnecessary underdiagnosis of HBOC patients, to implement genotype-based precision medicine effectively for these patients, and to apply family cascade testing to identify at-risk family members.

Notes

Acknowledgment

We appreciate the manuscript editing assistance of Nancy Jacoby and graphic design assistance of Carole Leung and Alla Zarifyan.

References

  1. 1.
    American Cancer Society: Breast Cancer Facts & Figures 2017–2018. American Cancer Society, Atlanta, GA. 2017. https://www.cancer.org/content/dam/cancer-org/research/cancer-facts-and-statistics/breast-cancer-facts-and-figures/breast-cancer-facts-and-figures-2017-2018.pdf. Accessed April 2018.
  2. 2.
    Norquist BM, Harrell MI, Brady MF, et al. Inherited mutations in women with ovarian carcinoma. JAMA Oncol. 2016;2(4):482–90.CrossRefPubMedPubMedCentralGoogle Scholar
  3. 3.
    Drohan B, Roche CA, Cusack Jr JC, Hughes KS. Hereditary breast and ovarian cancer and other hereditary syndromes: using technology to identify carriers. Ann Surg Oncol. 2012;19(6):1732–7.CrossRefGoogle Scholar
  4. 4.
    Levy-Lahad E, Lahad A, King M-C. Precision medicine meets public health: population screening for BRCA1 and BRCA2. J Natl Cancer Inst. 2015;107(1):420.CrossRefGoogle Scholar
  5. 5.
    Childers CP, Childers KK, Maggard-Gibbons M, Macinko J. National estimates of genetic testing in women with a history of breast or ovarian cancer. J Clin Oncol. 2017;35(34):3800–6.CrossRefPubMedPubMedCentralGoogle Scholar
  6. 6.
    Guo F, Hirth JM, Lin Y-L, et al. Use of BRCA mutation test in the US, 2004–2014. Am J Prev Med. 2017;52(6):702–9.CrossRefPubMedPubMedCentralGoogle Scholar
  7. 7.
    Evans DGR, Ingham SL, Baildam A, et al. Contralateral mastectomy improves survival in women with BRCA1/2-associated breast cancer. Breast Cancer Res Treat. 2013;140(1):135–42.CrossRefPubMedPubMedCentralGoogle Scholar
  8. 8.
    Robson M, Im S-A, Senkus E, et al. Olaparib for metastatic breast cancer in patients with a germline BRCA mutation. N Engl J Med. 2017;377(6):523–33.CrossRefPubMedPubMedCentralGoogle Scholar
  9. 9.
    Local coverage determination: MolDx BRCA1 and BRCA2 genetic testing. Noridian Medicare. https://med.noridianmedicare.com/documents/10546/6990981/MolDX+BRCA1+and+BRCA2+Genetic+Testing+LCD/220c55c8-8bc8-4319-80a1-66c4a1d51dd8. Accessed April 2018.
  10. 10.
    O’Leary E, Iacoboni D, Holle J, et al. Expanded gene panel use for women with breast cancer: identification and intervention beyond breast cancer risk. Ann Surg Oncol. 2017;24(10):3060–6.CrossRefPubMedPubMedCentralGoogle Scholar
  11. 11.
    Lincoln SE, Kobayashi Y, Anderson MJ, et al. A systematic comparison of traditional and multigene panel testing for hereditary breast and ovarian cancer genes in more than 1000 patients. J Mol Diagn. 2015;17(5):533–44.CrossRefGoogle Scholar
  12. 12.
    Nykamp K, Anderson M, Powers M, et al. Sherloc: a comprehensive refinement of the ACMG–AMP variant classification criteria. Genet Med. 2017;19:1105.CrossRefPubMedPubMedCentralGoogle Scholar
  13. 13.
    Grindedal EM, Heramb C, Karsrud I, et al. Current guidelines for BRCA testing of breast cancer patients are insufficient to detect all mutation carriers. BMC Cancer. 2017;17(1):438.CrossRefPubMedPubMedCentralGoogle Scholar
  14. 14.
    Tung N, Lin NU, Kidd J, et al. Frequency of germline mutations in 25 cancer-susceptibility genes in a sequential series of patients with breast cancer. J Clin Oncol. 2016;34(13):1460–8.CrossRefPubMedPubMedCentralGoogle Scholar
  15. 15.
    Tung N, Battelli C, Allen B, et al. Frequency of mutations in individuals with breast cancer referred for BRCA1 and BRCA2 testing using next-generation sequencing with a 25-gene panel. Cancer. 2015;121(1):25–33.CrossRefGoogle Scholar
  16. 16.
    Desmond A, Kurian AW, Gabree M, et al. Clinical actionability of multigene panel testing for hereditary breast and ovarian cancer risk assessment. JAMA Oncol. 2015;1(7):943–51.CrossRefPubMedPubMedCentralGoogle Scholar
  17. 17.
    Hauke J, Horvath J, Groß E, et al. Gene panel testing of 5589 BRCA1/2-negative index patients with breast cancer in a routine diagnostic setting: results of the German Consortium for Hereditary Breast and Ovarian Cancer. Cancer Med. 2018;7(4):1349–58.CrossRefPubMedPubMedCentralGoogle Scholar
  18. 18.
    CHEK2 Breast Cancer Case-Control Consortium. CHEK2*1100delC and susceptibility to breast cancer: a collaborative analysis involving 10,860 breast cancer cases and 9065 controls from 10 studies. Am J Hum Genet. 2004;74(6):1175–82.CrossRefGoogle Scholar
  19. 19.
    Chen S, Parmigiani G. Meta-analysis of BRCA1 and BRCA2 penetrance. J Clin Oncol. 2007;25(11):1329–33.CrossRefPubMedPubMedCentralGoogle Scholar

Copyright information

© Society of Surgical Oncology 2018

Authors and Affiliations

  • Shan Yang
    • 1
  • Jennifer E. Axilbund
    • 1
  • Erin O’Leary
    • 1
  • Scott T. Michalski
    • 1
  • Robbie Evans
    • 1
  • Stephen E. Lincoln
    • 1
  • Edward D. Esplin
    • 1
    Email author
  • Robert L. Nussbaum
    • 1
  1. 1.InvitaeSan FranciscoUSA

Personalised recommendations