[Skip to Content]
[Skip to Content Landing]
Views 187
Citations 0
Comment & Response
July 11, 2019

Germline Variants in Highly Selected Patients With Prostate Cancer—In Reply

Author Affiliations
  • 1Tulane University School of Medicine, New Orleans, Louisiana
  • 2Invitae Corporation, San Francisco, California
JAMA Oncol. Published online July 11, 2019. doi:10.1001/jamaoncol.2019.2120

In Reply We thank the Editors of JAMA Oncology for the opportunity to respond to the comments by Benusiglio et al and Greenberg et al concerning our article.1 The article was clear in describing the limitation of using a convenience sample derived from patients referred for testing and the possible selection bias by the ordering clinicians. However, the sample was unselected with regard to whether the patients met clinical guidelines for testing and which genes should be tested. The main point of the article1 was that the similar yield of pathogenic mutations among patients who either met or did not meet personal or family history testing guidelines demonstrated that such guidelines reduced the sensitivity of testing. The limitation of using a convenience sample makes a direct comparison of the absolute yield of pathogenic mutations among patients with prostate cancer, such as the comparisons reported by Slavin et al2 or Latham et al3 in their cohorts with prostate cancer, ill-advised. Notably, Latham et al3 supported our proposal for expanded testing in that they recommend that patients with microsatellite instability-high tumors undergo germline genetic assessment for mutations in mismatch-repair genes regardless of cancer type or family cancer history because of the importance of determining the presence of Lynch syndrome. Given that microsatellite instability status may not be routinely available and is not definitive for diagnosis of Lynch syndrome, the availability of low-cost germline assessment for mismatch-repair genes is a reasonable strategy for making such a determination. It is true that many of the genes we listed as having mutations are not known at this time to be associated with prostate cancer; however, the history of hereditary cancer syndrome research has been that mutations in tumor-suppressor genes are often involved in more tumor types than originally seen in the earliest and most obvious case presentations. The suggestion made by Benusiglio et al that research should only be done as part of a clinical trial paid for by a sponsor would prohibit learning from research in the real-life settings of patient care. Finally, 2 authors (E.L. and O.S.) of the article1 were not associated with Invitae, and the Invitae employees disclosed their financial conflicts of interest, just as Benusiglio et al did for their article.