We appreciate the thoughtful comments of Professors Cai and Bartoletti about our article regarding hematuria home screening for bladder cancer in men.1 In answer to whether cytology would have been preferable to hematuria testing, it must be remembered that, although it is highly specific, cytology reportedly has a sensitivity <50% for high-grade cancers.2, 3 False-negative results in these patients would defeat the purpose of screening. Moreover, the deleterious impact of false-negative tests on screenees and their primary physicians cannot be underestimated in terms of ignoring symptoms of bladder cancer (BC), such as visible hematuria, because they take comfort in the negative screening test result.
In addition, unlike most other malignancies, BC almost never is found incidentally at autopsy,4 indicating that even those screenees with low-grade cancers were not harmed by screening; because, at some time in their future, their tumor would have caused symptoms, been diagnosed, and eventually treated.
We lack sufficient information to comment extensively on Professor Cai's and Bartoletti's unpublished reference to their artificial neural network (ANN), except that monitoring patients with previously resected, superficial BC and screening individuals who never had a diagnosis of BC are not analogous scenarios, and the performance of a diagnostic test in both circumstances may be quite different. For example, cystoscopy had previously been performed on all patients with prior tumors and in no screenees. Because instrumentation induces subtle hematuria, hematuria testing would have produced too many false-positive results to be useful in the monitoring scenario (as the ANN revealed).
Professors Cai and Bartoletti appear to have calculated incorrectly the positive predictive value (PPV) (true positives/true positives + false positives) for detecting men at risk for dying from BC with Hemastix strips. The PPV actually was 4.3% (and, for all patients with BC, it was 8.9%), and not 0.5%. However, we agree that restricting the screened population to individuals at higher risk, including those with known industrial or smoking exposures, certainly would increase the prevalence of BC in the screened cohort and, hence, the PPV. This strategy currently is being tested.
Finally, we agree that screening has a variety of spin-off benefits; and, although some reflect self-selection, we do not know how many screenees ceased smoking or exhibited greater health consciousness in lifestyle and medical awareness because they took part in screening. This may have had a role in the reduction of mortality in screenees with BC compared with the reduction of mortality in the general Wisconsin BC population.