Diagnosis of prostate cancer in brothers with signs of detection bias and related consequences

Abstract

Background: Swedish nationwide family and cancer data offer an unbiased opportunity to study familial prostate cancer (PC) and allow an assessment of the possible diagnostic bias that might have been introduced in the prostate-specific antigen (PSA) era. We address this bias in a setting between brothers diagnosed with PC.Methods: Brothers were identified from year 1932 onwards until the end of the study 2021. Differences between diagnostic ages, incidence, TNM stage and survival were compared between brothers.Results: A total of 18317 brothers were diagnosed in brotherships of two or more PCs. If there were no bias in diagnostics between the brothers, the null hypothesis was that their diagnostic ages were equal. This was clearly refuted by the data: subsequent brothers were preferentially diagnosed within a year of the first brother. The bias for subsequent brothers was higher when the brothers lived in the same community, and it was highest for large cities, particularly Stockholm. Comparing incidence rates between the first and subsequent brothers, subsequent brothers showed higher rates except at very high age. Proportions of patients detected with PSA (stage T1c) were higher in subsequent brothers compared to the first ones; the proportions were reversed for stages T2-4. Subsequent brothers were diagnosed less frequently with metastases compared to first brothers. Subsequent brothers survived better than first, but the difference depended on age and time since diagnosis.Conclusions: The results between brothers, using objective measures of diagnostic age, incidence and stage of disease, confirmed the existence of detection bias which however appeared to weaken towards high age.