This article has been corrected to attribute comments from Myriad Genetics to Johnathan Lancaster, the chief medical officer, instead of Jerry Lanchbury, the chief scientific officer.
SAN ANTONIO (GenomeWeb) – Myriad Genetics is planning to launch a version of its polygenic breast cancer riskScore specific for the Hispanic population next year.
At the San Antonio Breast Cancer Symposium here today Myriad presented data describing the development of an 87-SNP Hispanic-specific Residual Risk Score and compared it to an 86-SNP Residual Risk Score the company developed and is currently marketing only for women with European ancestry.
Last year, at this same meeting, Myriad presented the complete validation data for the 86-SNP riskScore, which involved more than 1,600 women of European ancestry with and without breast cancer. Researchers showed that when combined with the Tyrer-Cuzick model (a risk estimate factoring in various patient characteristics and their family history of breast cancer), the polygenic score was a superior predictor of a woman's five-year and lifetime risk for breast cancer compared to using Tyrer-Cuzick alone.
Myriad's riskScore is an adjunct to its myRisk next-generation sequencing panel, which gauges 28 genes associated with elevated risk for eight hereditary cancers. Specifically, mutations in 11 breast cancer-linked genes in myRisk ― BRCA1, BRCA2, TP53, PTEN, STK11, CDH1, PALB2, CHEK2, ATM, NBN, BARD1 ― confer high and moderate risk but show up in only about 10 percent of women, and can help guide additional screening approaches or in some cases take preventive action. But the 90 percent of women who test negative for mutations in these genes aren't helped.
This is where polygenic risk scores hold promise. Data from genome-wide association studies have identified SNPs that individually confer small risks for diseases, but combined together can quantify risk for all women. Given this potential, polygenic risk scores have garnered the interest of researchers, consumer genetics firms, and commercial labs like Myriad.
Except, to date, the vast majority of GWAS studies, which researchers draw on to identify the SNPs included in these scores, largely involve people of European ancestry. This presents a difficulty for the development of such tests. Unsurprisingly, most commercially available SNP tests and polygenic risk scores, be it for assessing breast cancer or Alzheimer's risk, come with the caveat that the results are relevant for people of European ancestry.
Given this limitation, when Myriad launched riskScore in 2017, it also did so for women with European ancestry. "When we launched that last year, we recognized that this is not okay to just have something for women of European ancestry," said Johnathan Lancaster, Myriad's chief medical officer. "We live in America, frankly, and we don't have a Caucasian population. So, in parallel with the Caucasian risk score we set about how to apply the same technology and approach to the Hispanic population, the African American population, and potentially, even to the Asian population."
The GWAS studies that provided the foundation for the 86-SNP riskScore for women of European ancestry didn't enroll sufficient numbers of women of other ethnicities to enable development of this test for other groups. "So, if the academic GWAS consortia are not going to be able to offer this up, we needed do it ourselves," Lancaster said.
The development of the Hispanic risk score involved approximately 14,000 women who have Hispanic or Latin American ancestry, mostly unaffected and some with breast cancer. The women whose data underlie the development and validation of the Hispanic-specific risk score came from Myriad's database, comprising the "largest polygenic study ever conducted for Hispanic ancestry," according Lancaster. He noted that a cohort of this size would have taken much longer to come together at local academic institutions or cancer center, but the fact that Myriad was able to bring this large dataset together over the period of a year speaks to the capabilities of a large commercial lab with a national presence.
In the development cohort of more than 5,400 women, researchers from Myriad evaluated 93 SNPs, and estimated whether each SNP's association with breast cancer differed for Hispanic versus European women. Drawing on data from the same development cohort, they then devised an 87-SNP risk score optimized for women of Hispanic ancestry, and validated its accuracy for assessing breast cancer risk compared to the European risk score in another 8,500 Hispanic women.
"There are some very specific SNPs that differentiate the Hispanic population from the non-Hispanic population," Lancaster said. "This means that simply picking up the Caucasian score and using it in Hispanic women wouldn't be as accurate. It works, it doesn't do a horrible job, but the new score is superior to the Caucasian one."
According to the data Myriad shared at the meeting, the Hispanic risk score outperformed the European version in this cohort, and was "highly predictive" of breast cancer risk in unaffected Hispanic women with a significant family history of the disease, and who were negative for mutations in 11 high- and moderate-risk genes included in Myriad's myRisk NGS panel. Estimates of the relative risks based on the 87-SNP score ranged from 0.2 to 3.6, demonstrating the ability of the score to discriminate between patients who are at low and high risk of breast cancer.
Next, Myriad is planning to validate the combination of the Hispanic risk score with the Tyrer-Cuzick model in Hispanic women, as it did with the European risk score last year.
The company also has plans to advance risk scores specific for the African American and Asian populations. "Given the demographics of the United States, it just happens we got to the numbers we needed in the Hispanic population first," Lancaster said. Work on the African American risk score is ongoing and involves a similarly large cohort drawn from its own database. Since the Asian population in the US is relatively smaller, the development of a risk score in this population may take longer.
At this same meeting, Gareth Evans from the University of Manchester discussed two large population screening studies (see here and here) in which multi-gene SNP scores were used to screen women in the UK for their breast cancer risk. As in the US, although a minority of women in the general UK population has one of the high-risk or moderate-risk genes for breast cancer, "50 percent [of women] will have a significant change based on SNP testing," Evans said.
"I believe breast cancer risk stratified screening is ready for implementation," he continued. "SNPs are very much part of that. They're extremely accurate and I believe they actually give more information particularly in the general population than going panels looking at high- and moderate-risk genes."
However, even with high-risk genes such as BRCA1 and BRCA2, testing for which has been available for two decades and included in screening guidelines, women who meet requirements for testing still don't receive it. Allison Kurian from Stanford University highlighted research at this meeting showing that doctors often fail to recommend genetic testing to breast cancer patients despite requesting it. In this survey of 2,500 women with early stage breast cancer only 53 percent of high risk women who met guidelines for BRCA testing received it, and among those who didn't get tested, more than half said this was because their doctor didn't recommend it.
This gap in access to testing for well known genes like BRCA1/2 raises questions about the adoption trajectory of newer polygenic risk scores. "We're talking about cutting-edge science with polygenic risk scores, [and] tweaking a score from Caucasians to Hispanics, and at a public health level doctors aren't recommending tests that have been available for over 20 years," Lancaster said. "We need to step back and say, 'What do we need to do in the US to do a better job of ensuring those who need testing get it?'"