Screening for Prostate Cancer by PSA Determination: A Time for Caution

(PC) is becoming increasingly popular in Western countries. The enthusiasm for this procedure might be justified by – the relevance of the disease, which is among the main causes of death among the male population; – the potential of earlier detection (PC is usually detected at an advanced stage); – the expected gain in survival due to earlier treatment, survival being strongly associated with disease stage; – the simplicity and relatively low cost of the PSA assay; – last but not least, the fact that it is “big business”. On the other hand there are several reasons why, even on an ethical basis, the promotion of screening as routine practice should not be encouraged: 1. No screening procedure should be recommended before unequivocal evidence is available for its costeffectiveness: two large controlled studies are currently ongoing in Europe (ERSPC) (1) and the US (PCLO) (2) but according to their statistical power they will most likely produce no evidence for efficacy (if any) until 2010. 2. The currently available evidence comes from pilot studies which may only prove the feasibility of screening and suggest a relevant diagnostic anticipation, but allow no prediction of the efficacy of screening in reducing prostate cancer mortality and/or patient quality of life (3). In fact, a Canadian study recently reported by Labrie et al (4), which claimed a reduction in mortality due to screening, does provide further evidence for early detection but the mortality analysis is flawed by a gross methodological error (imbalance of inclusion of subjects in screening and control arms which tends to introduce severe selection biases). The history of oncologic screening teaches us that even significantly earlier detection of cancer may be entirely useless in improving prognosis, as was clearly shown by the controlled trials on lung cancer detection (2). 3. Data from pilot studies and preliminary data from controlled trials suggest a large diagnostic anticipation which foretells quite a relevant overdiagnosis (5, 6). The ratio of observed (prevalent, screen-detected) to expected (underlying incidence) cancers in screening experiences ranges between 10 and 30:1, suggesting a detection lead time (diagnostic anticipation) of 7 to 15 years (6). Considering the life expectancy of men enrolled in screening programs it is likely that a relevant proportion of subjects with screen-detected cancers would die of other causes before their cancers would become symptomatic. Overdiagnosis can be estimated in a screening scenario when the cancer prevalence, underlying incidence and interval cancer rate are known: estimates which can currently be based on existing data suggest a magnitude ranging from 100% to 300%, depending on the age group considered and the screening frequency and aggressiveness (see Table I). 4. Latent carcinoma (carcinoma according to typical pathological appearance, but with very low progression potential, so that it would never surface as clinical in a lifetime or would never become life threatening) is very common in males over 50 years of age, as shown by well-known autopsy studies (7). The practice of random biopsy of the peripheral prostate routinely performed in screening practice in all subjects with elevated PSA (usually with a cutoff of 4 or even 3 ng/mL) offers a special chance of sampling one of these latent cancers. 5. At present it is not possible to differentiate latent cancer from cancer of clinical significance. Stage is not a criterion as even advanced stage prostate cancer may be slow growing and death from other causes may occur prior to symptomatic onset. Thus overtreatment will be the natural consequence of overdiagnosis, and a large majority of subjects in whom adenocarcinoma of the prostate is detected at biopsy, irrespective of its stage and grade, will undergo some sort of treatment. 6. The optimal treatment for early prostate cancer has not yet been defined. Evidence of prolonged survival of untreated patients suggests that aggressive therapy is not always the best choice, and watchful waiting is currently listed among the possible therapeutic options (8). Nevertheless, in most instances radical – The International Journal of Biological Markers, Vol. 15 no. 4, pp. 285-287 © 2000 Wichtig Editore