The present burden of prostate cancer (PCa) in the Western world has continuously increased over the past 20 yr. However, the median age at death from PCa is >80 yr . Consequently, a significant proportion of men with PCa are considered to be overdiagnosed in the sense that a diagnosis may not improve their lifespan or quality of life. The concept of active surveillance has progressively emerged as a realistic alternative treatment for “low-grade” tumours. Ongoing studies seek to identify the optimum schedule for prostate-specific antigen (PSA) testing and repeat biopsies, the appropriate indications for intervention, and the long-term efficacy of surveillance in comparison to immediate radical treatment.
2. Current limitations of active surveillance
The outcome of focal or multifocal microscopic and low-grade prostatic neoplasia with regard to aggressiveness remains difficult to determine with certainty at the time of diagnosis and during the follow-up period. At an early stage of the disease, urologists and pathologists diagnose and classify PCa according to anatomopathologic features. This strategy is particularly true for the empirical definition of indolent/insignificant PCa (T1cN0M0 of low volume, as estimated by two or fewer biopsy cores with ≤20% tumour in each core, a low Gleason score [<7, without grade 4], and a low PSA level [<10 ng/ml] ), which is a candidate for active surveillance. However, this definition reflects only the probability that the disease will be indolent in the sample specimen obtained by biopsy. Biopsy sampling does not reflect the overall aggressiveness of the disease, and the small amount of tumour tissue obtained by biopsy does not routinely allow for the subsequent study of relevant biomarkers that reflect tumour aggressiveness.
The second drawback of active surveillance is related to difficulties in clearly assessing disease progression despite surveillance with PSA levels and repeat biopsies. PSA levels and PSA doubling times (PSA DTs) are difficult to interpret, especially when associated with benign prostate hypertrophy. Loblaw et al.  reported that 70% of approximately 450 patients under active surveillance with a 6.8-yr median follow-up remained untreated. However, the best indicators for improving treatment decisions remain unknown. Loblaw et al observed that in 37% of untreated cases, PSA DT was <2 yr at least once during follow-up, suggesting that more work is needed to identify a better prognostic factor for patients under active surveillance. In addition, conclusions regarding disease outcomes remain unclear when the secondary biopsies demonstrate similar neoplastic lesions with the same tumour volume and a Gleason score <7 (in just >50% of cases). Soloway et al.  reported that positive rebiopsies were found in 46% of 230 low-risk PCa patients on active surveillance, and these rebiopsies had the same features found with the initial biopsy. Moreover, active surveillance, without any form of treatment, potentially allows cells to acquire new molecular changes after mitosis, which can suddenly increase aggressiveness. Lastly, PSA DT is not a clear predictor for the decision to pursue treatment, and an increase in Gleason score or tumour volume is the primary factor associated with radical treatment (Fig. 1).
Because much lethal PCa is associated with outlaw or parallel forms of androgen receptor (AR) activation , the current knowledge of PCa biology and genomics suggests that aggressiveness should be assessed not only in terms of anatomic and histopathologic features but also in terms of functional features based on a response to androgen deprivation therapy (ADT). Thus AR disruption could suppress truly indolent PCa and thereby highlight cancers that require radical therapy. What arguments exist to support the idea that a subtype of asymptomatic PCa is reversible with ADT?
3. Is microscopic prostate cancer reversible under androgen deprivation?
Since the discovery of ADT efficacy for the treatment of PCa by Huggins in 1941, it has become generally accepted that ADT is not a cure for PCa, despite a reduction in tumour volume and a state of remission usually observed after castration. However, this historic dogma has not been established via microscopic PCa staging. Recent clinical observations related to the chemoprevention of PCa or androgen blockage at early stages of the disease suggest that some low-grade forms of the disease can be reversed by ADT.
3.1. A view from the bench
Experimental data in animal models suggest that ADT can be used to cure precancerous lesions or even very small tumours. However, the question regarding androgens resistant to the castration of PCa progenitor cells still remains unclear. Recently, controversies about the luminal or basal cell origin of PCa have been debated. Goldstein et al.  suggested that the cooperative effects of early molecular events, such as v-ets erythroblastosis virus E26 oncogene homolog (ERG) rearrangements and AR activation, in basal or stem cells recapitulate the histologic features of PCa, including the loss of basal or stem cell phenotypes and the expansion of cells with a luminal phenotype. Conversely, stem cell markers can be expressed by luminal cells  that have acquired a stem cell–like dedifferentiation programme under the control of AR/transmembrane protease, serine 2 (TMPRSS2)–ERG.
Based on the potential heterogeneity in the cellular origins of PCa, we propose a model in which the initiation of PCa can occur in two ways, both of which lead to the same luminal phenotype but to different physiologic outcomes under ADT. It is suggested by colon cancer models that segregation of the older DNA strand during the renewal of epithelial glands minimises the effective mutation rate in somatic epithelial cells. If the older DNA strand, which has been exposed to cumulative mutations, is transmitted to the differentiated daughter cells, it is lost when terminally differentiated cells are sloughed from the epithelium. If the same model is applied to the prostate gland, it can be hypothesised that apoptosis of luminal cells following ADT could result in the sloughing of transformed cells if oncogenic events are carried by luminal daughter cells and are not retained in castration-resistant basal or stem daughter cells.
TMPRSS2:ERG fusions are seen in approximately 50% of localised PCa cases treated by radical prostatectomy and in approximately 20% of high-grade prostatic intraepithelial neoplasia cases. These reports strongly suggest that ETS gene rearrangements might be the early events in prostate carcinogenesis and cooperate in a multistep process with AR activity. Consequently, it is legitimate to assume that during the early stages of carcinogenesis, ADT can disrupt androgen-dependent gene fusions and induce apoptosis in luminal cells that have not acquired secondary molecular events that lead to castration resistance. However, carcinogenesis initiated in the basal or stem lineage also leads to a transformed luminal cell phenotype after additional cell divisions. After castration, basal or stem cells carrying silent oncogenic events resist apoptosis and can regenerate into recurrent progressive PCa if AR or androgen activity is restored.
3.2. A view from the bedside
Recent clinical reports have shown that well-differentiated (Gleason 6), low-volume, and localised PCa (1) could be prevented by a 5α-reductase inhibitor (5-ARI) and (2) could progress to complete remission. For the most part, the results of the Reduction by Dutasteride of Prostate Cancer Events (REDUCE) trial , which studied the primary chemoprevention of PCa, are similar to those of the Prostate Cancer Prevention Trial (PCPT; using finasteride) . The 22.8% reduction in PCa diagnosis with dutasteride after 4 yr of treatment was similar to that with finasteride (24.8%) after 7 yr. In these two studies, the results demonstrated a preventive effect on the incidence of well-differentiated (Gleason 6) cancers.
Regarding the natural history of PCa and the time scale of the REDUCE trial, it is clear that 5-ARI therapy suppresses preexisting microscopically well-differentiated cancers that were missed by a first biopsy. Moreover, some studies report a curative effect of ADT at an early stage of the disease (local PCa), but these effects have not been well documented. These studies included heterogeneous stages of tumours, T1c to T3a, including cancers with a Gleason score of ≥7 or an initial PSA >10 ng/ml  and . Interestingly, these reports determined that an undetectable level of PSA during ADT was a strong, adjusted predictive factor for progression-free survival. In the same way, tumour burnout (pT0 stage) was observed in 30% of prostate specimens with Gleason scores <7 at diagnosis when a radical prostatectomy was performed with ADT . Interestingly, if the study demonstrates that primary tumours can disappear with ADT, it fails to demonstrate significant differences in relapse-free survival between the pT0 and non-pT0 groups. This suggests that early metastatic disease exists, without signs of local extraprostatic extension, and can explain the subset of tumours exhibiting a failure of local therapy with or without neoadjuvant ADT. It could also be suggested that radical prostatectomy has not been useful for the pT0 group after neoadjuvant ADT.
4. Conclusion: Secondary chemoprevention is a potential answer to the dilemma faced when choosing between overtreatment and no treatment (active surveillance)
Combining clinical results obtained with ADT at the early stage of PCa with recent biological advances in the understanding of prostate carcinogenesis, we suggest that different functional subtypes of PCa, according to the cell of origin and the multistep nature of cancer, exist under AR modulation at early stages of the disease. Thus a functional classification of PCa based on an initial period of ADT can be used to help identify potentially lethal disease. After this initial and limited period of ADT and according to a secondary prevention concept (defined as the treatment of an existing disease in its early stages before it results in significant morbidity), tumour recurrence in a biopsy would then be a signature for PCa relapse from quiescent transformed progenitor cells or from luminal cells that underwent molecular events leading to castration resistance, which would then justify radical treatment.
Conflicts of interest
The authors have nothing to disclose.
-  M.S. Soloway, C.T. Soloway, A. Eldefrawy, K. Acosta, B. Kava, M. Manoharan. Careful selection and close monitoring of low-risk prostate cancer patients on active surveillance minimizes the need for treatment. Eur Urol. 2010;58:831-835 Abstract, Full-text, PDF, Crossref.
-  P.J. Bastian, B.H. Carter, A. Bjartell, et al. Insignificant prostate cancer and active surveillance: from definition to clinical implications. Eur Urol. 2009;55:1321-1332 Abstract, Full-text, PDF, Crossref.
-  A. Loblaw, L. Zhang, A. Lam, et al. Comparing prostate specific antigen triggers for intervention in men with stable prostate cancer on active surveillance. J Urol. 2010;184:1942-1946 Crossref.
-  W. Liu, S. Laitinen, S. Khan, et al. Copy number analysis indicates monoclonal origin of lethal metastatic prostate cancer. Nat Med. 2009;15:559-565 Crossref.
-  A.S. Goldstein, J. Huang, C. Guo, I.P. Garraway, O.N. Witte. Identification of a cell of origin for human prostate cancer. Science. 2010;329:568-571 Crossref.
-  X. Wang, M. Kruithof-de Julio, K.D. Economides, et al. A luminal epithelial stem cell that is a cell of origin for prostate cancer. Nature. 2009;461:495-500 Crossref.
-  W.M. White, E.D. Kim. Evolving role of 5-alpha reductase inhibitors in chemoprevention. Nat Rev Clin Oncol. 2010;7:487-488 Crossref.
-  F. Labrie, B. Candas, J.L. Gomez, L. Cusan. Can combined androgen blockade provide long-term control or possible cure of localized prostate cancer? Urology. 2002;60:115-119 Crossref.
-  R.L. Leibowitz, S.J. Tucker. Treatment of localized prostate cancer with intermittent triple androgen blockade: preliminary results in 110 consecutive patients. Oncologist. 2001;6:177-182 Crossref.
-  J. Köllermann, W. Hopfenmüller, J. Caprano, et al. Prognosis of stage pT0 after prolonged neoadjuvant endocrine therapy of prostate cancer: a matched-pair analysis. Eur Urol. 2004;45:42-45
a Department of Urology, GHU EST, Hôpital Pitié-Salpétrière and Hôpital Tenon, Assistance-Publique Hôpitaux de Paris; Faculté de Médecine Pierre et Marie Curie, University Paris VI, Paris, France
b Department of Pathology, GHU EST, Hôpital Pitié-Salpétrière and Hôpital Tenon, Assistance-Publique Hôpitaux de Paris; Faculté de Médecine Pierre et Marie Curie, University Paris VI, Paris, France
c CeRePP, Centre d ‘Etudes et de Recherche sur les Pathologies Prostatiques, Paris, France
© 2011 European Association of Urology, Published by Elsevier B.V.