Article - Comment

Prediction of Response to Androgen Deprivation Therapy and Castration Resistance in Primary Metastatic Prostate Cancer

  • Ali F. Şahin

Bull Urooncol 2012;11(2):178-180

Propose:

In one of the largest and relatively homogeneous (primary advanced disease with bone metastasis) retrospective multicenter study, we tried to establish the predictive factors influencing the initial response as well as its duration (the median time to biochemical progression) and time to castration resistance (CR).

Methods:

We evaluated all patients initially receiving androgen deprivation therapy (ADT) for primary advanced prostate cancer (PC) with bone metastasis. A total of 982 patients with complete medical records available for analysis from 18 centers were included in this study. Age, initial PSA, Gleason score (GS) and extent of bone involvement (EBI) were recorded in a database. Patients with solid organ metastasis (lung and/or liver) and inadequate data which were required for this study to predict the end-points were excluded. Definitions · Initial response: decrease in PSA by ≥50% from pretreatment value lasting ≥1 month. · PSA ‘normalization': reaching a PSA level of ≤4 ng/ml during the treatment. · Nadir PSA (nPSA): the lowest value of serum PSA observed during the treatment. · Time to nPSA (TTnPSA): duration between the dates of initiation of ADT and nPSA. · Progression to CR: two consecutive increases in PSA by ≥25% of the nadir value if nPSA ≥4 ng/ml or increase in PSA >4 ng/ml. · The date of CR was chosen to be the date of the first PSA increase. · Duration of response (time to CR): the period from the initiation of ADT to CR.

Results:

Among all the patients, 896 (91.2%) responded to ADT initially. Pretreatment PSA and EBI were significant predictors in the multivariate model. Among the 659 patients who progressed into a CR state, the mean duration of response was 22.4 months. There was a significant correlation between the CR state and nPSA level and time to nPSA. Pretreatment PSA, EBI, GS, highest tumor volume in biopsy cores (%), number of positive biopsy cores, percent positive biopsy cores and time to nPSA were proven to be significant to predict a nPSA. Pretreatment PSA, GS and EBI were statistically significant predictors of PSA normalization in multivariate analysis. The limitation of the study depends on the retrospective design and a model was developed for low standardization as a result of using multicenter data. The patients enrolled in this study were from a relatively long period of time (1989–2008).

Discussion:

There is no consistency in response rates to ADT among various investigations basically due to the definition of the response to ADT, and possibly patient and disease characteristics such as pretreatment PSA, percentage of GS≥8, painless bone metastasis, performance status≤1, burden of the bone involvement, and distribution of age. In our study, only initial PSA and EBI were independent predictors of response to ADT.

PSA and initial Gleason grade were reported as the most important predictors of the time to androgen independence in locally advanced PC [13] . These findings were further corroborated by Kwak et al. [12] who reported that pretreatment PSA, PSA at 6 months after treatment and bone metastasis were significantly associated with progression to castration-resistant PC. In another study, Dijkman et al. [14] reported early normalization of PSA was shown to predict and improved long-term response to hormonal therapy in terms of CR and death. Our data indicate that approximately a quarter of the patients respond favorably to ADT with a longer failure-free period while the rest succumb to CR within a median of 12–18 months.

It is retrospective in design and the model developed a low standardization as a result of using multicenter data. Patients enrolled in this study were from a relatively long period of time (1989–2008). Other factors such as pain score and PSA doubling time, previously reported as predictive factors for survival in metastatic PC, were not assessed in this analysis. Despite these limitations, our cohort was uniform in terms of newly diagnosed primary advanced PC with bone metastasis without any previous treatment with a large size.

Conclusions:

The results of this study indicate that it is possible to predict the initial response to ADT by pretreatment PSA levels and EBI, while the duration of response can be reflected by a multitude of clinical factors including nPSA, TTnPSA, percent positive cores, biopsy GS and EBI.

Keywords: prostate cancer, androgen deprivation therapy, treatment prediction, castration resistance