Prostate-specific antigen flare phenomenon with docetaxel-based chemotherapy in patients with androgen-independent prostate cancer
Avishay Sella, Department of Oncology, Assaf Harofeh Medical Center, Zerifin 70300, Israel.
To evaluate the prostate-specific antigen (PSA) ‘flare’ phenomenon in patients with androgen-independent prostate cancer (AIPC) treated with docetaxel, as flare is a known effect of androgen-deprivation therapy in hormone-dependent prostate cancer.
PATIENTS AND METHODS
The charts of 56 patients who received docetaxel-based chemotherapy in three different centres from August 1999 to August 2007 were reviewed retrospectively. The biochemical response was characterized according to the Bubley criteria. There was an immediate PSA response (PSA decline ≥ 50%) in 23 (41%) patients, PSA stabilization (PSA decline < 50%) in 16 (29%) and PSA progression in nine (16%). There was also a fourth response, i.e. PSA flare, defined as an increase in PSA level with no symptomatic progression, after starting docetaxel-based chemotherapy administered every 3 weeks.
Eight (14%) patients with PSA flare were identified; all had osseous disease and five had additional soft-tissue disease. The PSA flare lasted a median (range) of 21 (21–42) days and it spread over a median of 1 (1–2) cycles. The temporary PSA surge exceeded baseline values by a median (range) of 61.5 (12–404)%. There was a subsequent PSA response in six of the eight patients and PSA stabilized in the remaining two. Patients with flare received a median of 8.5 (5–12) treatment cycles, vs a median of 8 (2–12) in the immediate PSA response group (P = 0.103, Student’s t-test). The Response Evaluation Criteria in Solid Tumors criteria evaluation showed one patient with a partial response and six with stable disease. The median survival of patients with PSA flare was 12.5 months, while that of the immediate PSA responders was 20.1 months (not statistically significant, P = 0.168, log-rank test).
Of patients with AIPC, 14% had an initial PSA flare after starting docetaxel-based chemotherapy. The occurrence of PSA flare had no effect on treatment duration or outcome. With lack of clinical progression, docetaxel-based chemotherapy should be administered for at least two 3-week cycles before further decisions are made about efficacy.
androgen-independent prostate cancer
Response Criteria in Solid Tumors.
The difficulty in assessing the response to treatment in patients with androgen-independent prostate cancer (AIPC) is a complex issue, extending across the biochemical and clinical spectrum of the disease. Investigators have focused on changes of PSA level to aid in this assessment, but changes in PSA level might be dissociated from the clinical outcome [1,2]. However, it is well-recognized that elevations in PSA level antedate the appearance of new lesions, can reflect a worsening of the disease, and thus be a source of anxiety for patients .
An initial PSA surge before a therapeutic response is a known phenomenon with androgen-deprivation therapy in hormone-dependent prostate cancer [3,4]. This ‘flare’ has also been reported in patients with AIPC treated with liposomal doxorubicin and docetaxel [5–8]. As docetaxel-based therapy is the standard treatment for patients with AIPC, it is important to expand the data on PSA flare during docetaxel-based therapy, and further investigate its magnitude and clinical significance [9,10].
PATIENTS AND METHODS
We retrieved data from the charts of 56 patients who received docetaxel-based therapy from August 1999 to December 2005 in three different centres. The treatment protocols included docetaxel 60–70 mg/m2 every 21 days, with either prednisone 5 mg twice daily or estramustine 280 mg three times daily on days 1–5; dexamethasone preceded docetaxel therapy [9,10].
All patients received chemotherapy after documentation of progression by PSA measurements and radiological studies that included CT of the abdomen and pelvis, chest X-ray and nuclear bone scintigraphy .
Treatment efficacy was determined according to the criteria of Bubley et al. and The Response Evaluation Criteria in Solid Tumors ; PSA was measured on day 1 of each cycle and radiological studies assessed every three cycles. Progression was based on RECIST criteria, appearance of two or more new bone lesions, occurrence of a new skeletal-related event, and worsening of pain, as measured by an increase of ≥25% in pain medication consumption [10,12–14]. To qualify for progression the patient had to have one of these criteria, with PSA progression according to Bubley et al.; the latter were modified to include a PSA flare, defined as an increase in PSA level with no evidence of clinical progression.
The survival was assessed using Kaplan-Meier analysis and log-rank tests, and Student’s t-tests used for assessing differences in means or medians.
The 56 patients had a median (range) age of 68.5 (50–81) years and metastatic disease that involved bone only in 15 (27%), soft-tissue disease only in 13 (23%) and both sites in 28 (50%). Treatments included docetaxel/prednisone in 31 patients (55%) and docetaxel/estramustine in 25 (45%) [9,10]. Of these patients, 23 (41%) had a PSA response, 16 (29%) PSA stabilization, nine (16%) PSA progression and eight (14%) a PSA flare.
The median age of the PSA-flare group was 75 (58–81) years. The therapy of these patients included docetaxel/prednisone in five and docetaxel/estramustine in three. All eight patients had osseous disease and five patients had additional soft-tissue disease. The PSA flare lasted a median (range) 21 (21–42) days and spread over a median of 1 (1–2) cycles. The temporary PSA flare exceeded baseline values by a median of 61.5 (12–404)%.
The overall response of the PSA-flare group showed a PSA response in six, PSA stabilization in two, RECIST partial response in one, and six had stable disease. When compared with the 23 patients with an immediate PSA response, both groups had received similar treatment cycles, the PSA-flare group having a median 8.5 (5–12) and the immediate PSA-response group 8 (2–12) cycles (P = 0.103, Student’s t-test). The median survival of the PSA flare group was not significantly shorter than that of the immediate response group, at 12.5 (6–52) vs 20.1 (5–107+) (P = 0.168, log-rank test). However, it was not significantly longer than the survival of those with PSA failure, of 10.2 (1–24) months (P = 0.789).
A rapid increase in tumour marker levels is a known phenomenon in solid tumours, and has been described in testicular germ cell tumours, breast and colon cancers, and involving βhCG, α-fetoprotein, carcinoembryonic antigen and CA15-3 [15–18]. Similarly, a PSA flare is a recognized phenomenon in androgen-dependent prostate cancer, and has been reported with chemotherapy in AIPC [3–8].
In the present study a considerable proportion (14%) of patients with AIPC had an initial PSA flare under docetaxel-based chemotherapy, with a median increase in PSA level of 61.5 (12–404)%. This is similar to previous reports of a PSA flare with docetaxel-based therapy in AIPC, where the rate was 10.4% or 18% and the maximum PSA increase was 3–514%[7,8].
While PSA flare in androgen-dependent prostate cancer treated with LHRH agonist is related to transient increases in testosterone level, and can be associated with clinical symptoms, the PSA flare produced by chemotherapy in AIPC is not associated with symptoms [3,4]. It is attributed to PSA released from a tumour undergoing lysis, or transactivation of mutated androgen-receptors by the estramustine component of the therapy, and the premedication with dexamethasone [7,8].
The increase in PSA level in the PSA-flare group continued for a median of one (one or two) cycles. Therefore, in the absence of clinical signs of progression, an early transient increase in PSA level under docetaxel-based therapy is not a reliable sign of progressive disease [7,8].
Patients with flare had a similar treatment duration and outcome to those who had a PSA response; they benefited from therapy and had biochemical and clinical responses. However, we and Nelius et al. were unable to confirm the results of Olbert et al. of a significant survival difference between patients who had a PSA flare and those who failed to respond to docetaxel-based therapy. This might relate to different definitions of the PSA flare phenomenon. While all studies included any increase in PSA level, Nelius et al. and Olbert et al. selected only those patients whose PSA level decreased to below baseline values (≥50%, or ≥50% and less than the maximum PSA level), but we also included in the flare group two patients whose PSA increase stabilized. This emphasizes the need for a definition of PSA flare and the need for a larger sample to better define its clinical effect. It is also possible that with the additional therapeutic options of intermittent docetaxel, second-line chemotherapy and the emergence of hormonal sensitivity after docetaxel-based chemotherapy, it will be difficult to isolate the relevance of a PSA flare on patient survival [19–21].
A PSA increase, particularly peaking at 400%, can be a source of concern for the patient. Such an increase might raise the question as to whether it should result in early withdrawal from therapy in a palliative setting . Based on the present data we suggest that docetaxel based-chemotherapy should be continued for ≥7 weeks (two cycles). A similar conclusion originated recently from a different perspective, by analysing the relationship between PSA progression-free survival and overall survival . This analysis provided further support for an initial 12-week period of drug exposure independent of PSA levels, because the association between PSA progression-free survival and overall survival increased during this interval. The present data are useful to clinicians to avoid early discontinuation of potentially beneficial therapy. Indeed, the recent Prostate Cancer Clinical Trials Working Group recommendations advocated that PSA measurements obtained during the first 12 weeks not be used as the sole criterion for clinical decision making .
CONFLICT OF INTEREST