To the Editor In the study by Shu et al,1 the investigators conducted a multicenter prospective cohort study of 1083 women with a BRCA1 or BRCA2 mutation at 9 academic medical centers in the United States or the United Kingdom. They observed 8 incident uterine cancers, compared with 4.3 expected. In my opinion this is erroneously described as a “prospective cohort study” because all of the incident cancers had been diagnosed at the time when the cohort was designed and assembled. The patients were ascertained from 1995 on, but the cohort was assembled in 2014 by inviting potential participants around the world to join the study group after the all cancers had already been diagnosed. It is my understanding that for a cohort study to be valid, the cohort should be assembled and followed forward in time (prospectively) or constructed through a formal and rigid and evaluable process of an existing data set (historical). The study by Shu et al1 does not qualify in either case, and at Womens’ College Hospital we will not be relying on the results of this study on its own to advise our patients on the benefits of a hysterectomy in addition to salpingo-oophorectomy.