Response to K. Hemminki et al.

Our research team acknowledges some of the limitations highlighted by Hemminki and colleagues (specifically the low prevalence of variant histology and the weak impact of smoking), which we previously addressed in our article. The purpose of our study was to better characterize the pattern of familial cancer risk (both concordant and discordant) in relatives of urothelial cancer patients. This is part of a larger goal of defining a phenotype of cancer across anatomic sites that includes bladder cancer. Central to this work is making analytic decisions that reduce phenotypic heterogeneity, such as excluding variant histology. Squamous cancer is remarkably distinct in its biology, epidemiology, and treatment. While the relative frequency of squamous cell carcinoma is low, and therefore does not drive the observed associations, removing it allows us to refine this phenotypic definition. To do this, we excluded 6462 individuals from the analysis: 1869 individuals with a bladder cancer diagnosis that was not the first primary, 368 individuals with nonurothelial bladder cancer, 4217 individuals with kidney and renal cancer, and eight ureteral cases.

Phenotypic definitions of cancer can be further refined by considering the genetic and environmental underpinnings potentially underlying the observed patterns. Families share genetic and environmental factors that contribute to cancer risk, and both of these factors may contribute to differential patterns of familial cancer clustering. By excluding families with an observed familial risk pattern that suggests behavioral mechanisms, such as smoking, we can further refine this phenotypic definition and generate more homogenous clusters of the familial patterns of cancer risk. Identifying more precise definitions of cancer phenotypes that include bladder cancer is essential for improving the understanding of the etiology associated with patterns of familial cancer clustering in a heterogeneous disease such as bladder cancer. We are fortunate to be able to contribute to the evolution of the understanding of patterns of cancer clustering in families, an iterative process that requires the validation of previously reported findings in multiple populations, extending current knowledge by testing for differential effects across subpopulations or disease patterns. Our team is excited to be working on a project that furthers the work presented in our article in the Journal. We are hopeful that this ongoing work will contribute to the scientific understanding of bladder cancer.

Notes

Affiliations of authors: Division of Urology, Department of Surgery (HAH, CM, BON), Population Sciences (HAH, CLL, KRS), Huntsman Cancer Institute, and Department of Family and Consumer Studies (KRS), University of Utah, Salt Lake City, UT.