“NEW YORK – Personalized genomics company MyOme’s cross-ancestry, integrated risk score (caIRS) improved breast cancer prediction over the standard Tyrer-Cuzick (T-C) model in a retrospective validation study published with Natera.
The study, published Tuesday in JCO Precision Oncology, tested the association between the caIRS and breast cancer in retrospective data from two validation cohorts with longitudinal follow-up amounting to over 130,000 women. The results were then compared to five-year and lifetime breast cancer risk prediction by T-C alone.
The caIRS combines MyOme’s cross-ancestry polygenic risk score (caPRS) with the T-C model.
To ensure representation from diverse genetic backgrounds, MyOme obtained its data from multiple cohorts, including the Women’s Health Initiative, the Multiethnic Cohort, the ROOT Cohort, and the UK Biobank.
The caIRS outperformed the T-C model alone in predicting breast cancer among all populations tested in both validation cohorts.
In particular, the largest gains were seen among Black and Hispanic women.
The positive predictive value of the caIRS, for example, showed a twofold increase over the T-C model alone in identifying high-risk Black women in cohort 1, in which approximately 30 percent of women identified as African American/Black.
The largest improvement in remaining lifetime risk in cohort 1 was observed in Hispanic women, where the odds ratio per standard deviation — meaning the risk associated with those women in terms of the change in risk per standard deviation — increased from 1.31 to 1.88. Cohort 2 did not include any women who identified as Hispanic, and the greatest gain in odds ratio per standard deviation in this cohort was observed in white women, who comprised nearly 88 percent of the cohort.
The caIRS also recategorized the estimated risk for more women relative to the T-C model alone in both five-year and lifetime risk estimations, suggesting that had it been included in guidelines, more women may have received risk-lowering medication and/or enhanced surveillance.
Specifically, the caIRS reclassified 8.8 percent of cohort 1 and 6.7 percent of cohort 2 differently, compared to the T-C model alone.
MyOme had presented preliminary data from this study at last year’s American Society of Clinical Oncology (ASCO) conference, where it also announced plans to field a commercial product based on its caPRS sometime this year.
A spokesperson for Menlo Park, California-based MyOme said that the company remains on track to meet that goal and that the resulting integrated breast cancer PRS will be marketed as a lab-developed test, although the company is not announcing any commercialization plans at this time.
In addition, although two of the study’s 10 authors are employees of Natera, MyOme declined comment on any possible commercial collaboration with Natera, but pointed out that Natera is an investor in the company. Matthew Rabinowitz, the cofounder and executive chairman of Natera, also cofounded MyOme.
The cohorts used in the present study included few Asian women, and cohort 2 included less clinical information overall than did cohort 1. Although MyOme does not expect this to significantly impact the study’s findings, it does hope to analyze larger datasets with better Asian representation in the future.
MyOme’s breast cancer PRS joins several others in various stages of development.
Earlier today, Allelica announced a joint study with Taiwanese firm SP BioMed to use Allelica’s breast cancer PRS in evaluating different genotyping technologies for future applications in the Taiwanese population.
Late last year, Myriad Genetics presented data from a prospective, longitudinal study of over 130,000 women that also showed improved risk prediction compared to the T-C model alone.”