Response to Kratz, Plon, and Lupo

We thank Kratz and colleagues¹ for their interest and comments on our study titled “Childhood, adolescent and young adulthood cancer risk in BRCA1 or BRCA2pathogenic variant carriers” in which we did not find evidence that people with a BRCA1 or BRCA2 pathogenic variant have an increased cancer risk before age 30 years aside from breast cancer for women in their 20s.

Kratz and colleagues¹ commented that our study could not refute previously reported associations between BRCA1 or BRCA2 pathogenic variants and certain rare cancer types, including rhabdomyosarcoma,^2,3 medulloblastoma,⁴ and neuroblastoma.⁵ We acknowledge that our data collection methods, data availability, and the rarity of these cancers limited our study’s ability to investigate these cancers specifically; as noted, our study had 80% power to detect a relative risk greater than 2 for childhood cancers combined and was not able to detect low relative risks for rarer diagnoses individually.

We cannot exclude a low increase in relative risk for specific cancers during childhood, but the absolute risk for cancer in general during childhood is low, and we agree that clinical testing of children in these kindreds is not indicated. Kratz and colleagues’ suggestion regarding testing is essentially consistent with what we encouraged in our publication: “further data collection, as well as translational studies of the impact of germline pathogenic variant on tumor behavior and therapeutic choices.” Specifically, we call for research that collects data from children with cancers, allowing us to obtain more precise estimates of frequencies of BRCA1 and BRCA2 pathogenic variants in them, and from a large number of children without cancers, which is critical in providing empirical population frequencies of BRCA1 and BRCA2 pathogenic variants in childhood and investigating whether there is an association using a case-control study design.⁶

Our pedigree-based study provides an alternative to the methodology used in cancer cohort studies in which children with cancer are tested for germline variants in these genes, and the frequency of a variant is compared with the frequency in noncancer cohorts. It should be noted, however, that these frequencies have wide confidence intervals and that data must be considered carefully, with consideration of analytic methods and demographics (age, ethnicity, race) of the comparison group. This is consistent with Kratz and colleagues’ statement that “well-powered burden testing studies using large cohorts of cases with rare tumors, suitable controls, and identical analytic pipelines are required.”⁷

Author contributions

Shuai Li, PhD (Conceptualization; Writing—original draft), Laura Madanat-Harjuoja, MD, PhD (Conceptualization; Writing—review & editing), Timothy Rebbeck, PhD (Conceptualization; Writing—review & editing), and Lisa Diller, MD, PhD (Conceptualization; Writing—original draft).

Acknowledgments

Consortium of Investigators of Modifiers of BRCA1/2 (CIMBA) members contributed to our study include Goska Leslie, Daniel R. Barnes, Manjeet K. Bolla, Joe Dennis, Michael T. Parsons, Paraskevi Apostolou, Norbert Arnold, Kristin Bosse, EMBRACE Collaborators, Jackie Cook, Christoph Engel, D. Gareth Evans, Florentia Fostira, Megan N. Frone, Andrea Gehrig, Mark H. Greene, Karl Hackmann, Eric Hahnen, Nadia Harbeck, Jan Hauke, Julia Hentschel, Judit Horvath, Louise Izatt, Marion Kiechle, Irene Konstantopoulou, Fiona Lalloo, Joanne Ngeow Yuen Yie, Dieter Niederacher, Julia Ritter, Marta Santamariña, Rita K. Schmutzler, Claire Searle, Christian Sutter, Marc Tischkowitz, Vishakha Tripathi, Ana Vega, Hannah Wallaschek, Shan Wang-Gohrke, Barbara Wappenschmidt, Bernhard H. F. Weber, Drakoulis Yannoukakos, Emily Zhao, Douglas F. Easton, Antonis C. Antoniou, and Georgia Chenevix-Trench.

Funding

SL is a National Health and Medical Research Council Emerging Leadership Fellow (grant number GNT2017373).

Conflict of interest

The authors have no conflict of interest to declare.

Data availability

No new data were generated or analyzed for this response.