Health care professionals’ attitudes toward cancer gene panel testing

In 2013, the United States Supreme Court issued a landmark ruling that eliminated the patenting of genes, because they are inventions of nature, not humans.¹ The lawsuit focused on the BRCA1 and BRCA2 gene patents that enabled Myriad Genetics to become the virtual sole provider for testing of hereditary risk for breast and ovarian cancer in the United States. In response to this ruling, a number of laboratories entered the market to test for cancer risks. But rather than limit testing to the BRCA1 and BRCA2 genes, these providers developed panel testing for a number of genes that might convey germline risk of cancer, when mutated. Subsequently, these tests were rapidly introduced into clinical practice without prior clinical validation studies.

An advisory highlighting a number of issues that should be addressed to aid counseling for gene panel testing for cancer risks was published in 2014 by many of the authors of this work. Concerns raised included definition of a target population, content for gene panels, risk assessment, and interpretation and communication of results.² The following year, a separate panel of authors published another advisory reviewing the difficulties in conferring a cancer risk for certain variants as well as assigning risks to other rare variants.³ Satisfactorily addressing the issues would establish the clinical validity of gene panels for cancer risk.

Since that time, gene panel testing has gained a significant foot-hold for cancer risk assessment, including recommendations from professional societies.⁴ Publications comprising test results from case series of hundreds of thousands of individuals have provided risk estimates that were unavailable at the time the panel testing was implemented.^5,6 Despite their widespread adoption of testing, health care providers’ views on the issues that we raised earlier have not been investigated. To fill the gap, we conducted a pilot survey of members of two professional societies, the National Society of Genetic Counselors (NSGC) and the International Society for Cancer Risk Management and Assessment (ISC- RAM).

Participants were provided with an online, 12- question categorical survey to assess their attitude toward cancer gene panel testing (see Supplement). In conducting this research study, the policies and procedures of the Albert Einstein College of Medicine Institutional Review Board were observed.

Description of participants (Questions 1–3). A total of 118 NSGC participants (~4% of those approached) and ISC- RAM 20 participants (~33% of those approached) completed the questionnaire. Of NSGC respondents, 54 (45%) practiced for less than 5 years. Of ISC- RAM respondents, 1 (5%) practiced for less than 5 years; 6 (30%) were medical oncologists; and 12 (60%) were surgical oncologists (χ² = 10.2, p = 1.4 × 10⁻³ for <5 years of practice between groups). Thus, almost half of the genetic counselors practiced exclusively during the time when gene panel sequencing was offered, whereas the remaining genetic counselors and virtually all of the ISC- RAM members practiced before as well as during this period gene panel sequencing period.

Experience with gene panel testing (Questions 4– 9). Several genetic testing laboratories were selected as providers, although with different preferences (χ² = 14.8, p = 0.01). Both groups selected oligo gene panels, multigene panels, and polygenic risk score (PRS) tests, although with different preferences; NSCG members preferred multigene panels (χ² = 7.2, p = 0.007). Both groups chose a variety of different multigene panels indicating that they tried to tailor testing to perceived genetic risks. Genetic counselors were more likely to follow NCCN guidelines, whereas ISC- RAM members were more likely to offer testing to their patients with cancer (36% vs. 10% for “More likely to follow NCCN guidelines,” χ² = 66.8, p = 2.0 × 10⁻¹⁴). At least 46% of each group indicated that testing laboratories provided clear assignment of risk, good customer service, clear reporting, and short reporting time, although frequency of responses varied (χ² = 42.1, p = 3.7 × 10⁻⁹). Members of both groups cited that reports were hard to understand, risk assignment was confusing, reporting takes too long, there were too many VUS and too many insurance hassles, although responses varied (χ² = 11.0, p = 0.05). For patients with VUS, at least 50% of each group indicated that they would wait for reannotation of VUS.

Preferences for a new test (Questions 10– 12). Virtually all members of both groups responded that they were favorable to a new test with higher accuracy, shorter turnaround time, and similar insurance coverage as current tests. Both groups indicated that they would like to see replication studies, review by NCCN, and high sensitivity and specificity as evidence for adopting this test (NS). At least 36% in both groups responded that they would like results to be reported as hazard ratio by decade of life and/or as high, intermediate, or population risk (NS). The groups varied in their requirement for FDA clearance of a new test.

Many of the issues surrounding hereditary cancer gene panel testing that were highlighted in our 2014 advisory remain a concern today, including content for gene panels, risk assessment, and interpreting and communicating results.² Guidance about target populations was provided by NCCN⁷; however, respondents indicated that they frequently do not follow these guidelines, with physicians more likely not to follow than genetic counselors. The target population issue has been compounded by recent publications that have pointed out that many of the patients with pathogenic variants would not have been tested had NCCN guidelines been followed, leaving open the sufficiency of these guidelines.^5,6

The estimates of the germline risks for several genes that were unclear in 2014 have been reported in recent publications, thus filling a gap.^5,6,8 However, other issues about risk assessment and interpreting and communicating results remain unresolved. Although many respondents indicated that risks were clearly communicated in laboratory reports, their responses were often contradictory.

The possibility of a new test provided benchmarks that respondents said should be met prior to introduction. These include high sensitivity and specificity, validation in retrospective and prospective studies, and review by NCCN, although not necessarily by FDA. In so doing, future developments in germline genetic risk assessment must anticipate and attempt to resolve the problems of genetic heterogeneity and phenotypic heterogeneity prior introduction into clinical practice.